스탯 겨울특강 PDF

Mike's AP Statistics
Chapter 1, Section 1
Introduction
Chapter 1 Introduction & Data Analysis: Making Sense of
1.1: Analyzing Categorical Data
Data After this section, you should be able to…
ü DEFINE “Individuals” and “Variables”

ü DISTINGUISH between “Categorical” and “Quantitative”
variables
ü DEFINE “Distribution”
ü DESCRIBE the idea behind “Inference”
What is the Study of Statistics?! Population

• Statistics is the science of data.
• In this course we study four different aspects of statistics:
Sample
– Data Analysis (Chapters 1 to 3)
• The process of organizing, displaying, summarizing, and asking
questions about data.
– Data Collection (Chapter 4)
• The process of conducting and interpreting surveys and experiments. Collect data from a
– Anticipating Patterns/Probability (Chapter 5 to 7) representative
• The process of using probability and chance to explain natural Make an inference about sample...
phenomena. the population.
– Inference (Chapter 8 to 12)
• The process of making predications and evaluations about a Perform Data Analysis,
population from a sample.
keeping probability in
mind…
Variable - any characteristic of an Distribution

individual or object
• Distribution: describes what values a variable takes and
how often it takes those values
• Essentially “distribution” replaces the words “data” or
Categorical Variable Quantitative Variable “graph”.
- Usually an adjective - Always a number • The median of the distribution is 28.
- Rarely a number - Must be able to find the • The distribution is skewed left.
Examples: mean of the numbers
- Gender Examples: Dotplot of MPG Dot Plot
- Race - Weight 2009 F uel Economy Guide
Distribution
- Grade in School - Height
(Sophomore, Jr., Sr.) - GPA
- Zip Code - # of AP Classes taken
- Square footage 14 16 18 20 22 24 26 28 30 32 34
MPG
Organizing a Statistical Problem Section 1.1

The Four-Step Process
Analyzing Categorical Data
After this section, you should be able to…
State: What’s the question that you’re trying to answer?
Plan: How will you go about answering the question? What ü CONSTRUCT and INTERPRET bar graphs and pie charts
statistical techniques does this problem call for?
ü RECOGNIZE “good” and “bad” graphs
Do: Make graphs and carry out needed calculations.
ü CONSTRUCT and INTERPRET two-way tables
Conclude: Give your practical conclusion in the setting of the
ü DESCRIBE relationships between two categorical variables
real-world problem.
ü ORGANIZE statistical problems
***Using this method is NOT required; however, all
complete answers MUST include the “Do” and “Conclude”
steps***
Distribution & Categorical Variables Displaying Categorical Data

The distribution of a categorical variable lists the count or Frequency tables can be difficult to read. Sometimes it is
percent of individuals who fall into each category. easier to analyze a distribution by displaying it with a bar
graph or pie chart.
Favorite Course Count Favorite Course Percentage Count of Stations Percent of Stations
English 8 English 16% 2500 Adult
Contemporary
Adult Standards
Foreign Language 8% 2000
Foreign Language 4 Contemporary hit
11%
1500 11%
Histroy 11 Histroy 22% 5% 9%
Country
1000 6% 4% News/Talk
Math 15 Math 30% Oldies

500 15% 15%
Science 12 Science 24% 0
Religious
8% Rock
16%
Spanish
Other
2014 AP Exam Scores

Graphs: Good and Bad Two-Way Tables

Bar graphs compare several quantities by comparing the Two-Way Tables: describe two categorical variables,
heights of bars that represent those quantities. organizing counts according to a row variable and a
Our eyes react to the area of the bars as well as height. Be column variable.
sure to make your bars equally wide. When a dataset involves two categorical variables, we
Avoid the temptation to replace the bars with pictures for begin by examining the counts or percents in various
greater appeal…this can be misleading! categories for one of the variables.
Member of No Member of Member of 2 or

Clubs One Club More Clubs Total
This ad for DIRECTV has
Rides the School Bus 55 33 20 108
multiple problems. How
many can you point out? Does not Ride Bus 16 44 82 142
Total 71 77 102 250
• What proportion of students that ride the school bus are • What proportion of males have “a good chance” at being
members of two or more clubs? rich?
• What proportion of students that are members of no clubs • What proportion of females have a “50-50 chance” at being
do not ride the school bus? rich?
• What proportion of students that do not ride the school bus • What proportion of young adults that have an “almost
are members of at least one club? certain” chance of being rich are male?
Member of No Member of Member of 2 or

Clubs One Club More Clubs Total
Rides the School Bus 55 33 20 108
Does not Ride Bus 16 44 82 142

Total 71 77 102 250
Comparing Categorical Comparing Categorical

Distributions Distributions
Sophomore Junior Senior Total
One 0 0 4 4
Two 1 3 12 16
Three 4 7 6 17
Four 7 4 8 19
Five 2 0 3 5
Total 14 14 33 61
Comparing Categorical Writing to Compare Categorical

• Cite specific numerical values/proportions.
Member of
• Use comparison words.
Does not Ride Bus
No Clubs – Greater, smaller, less, while only, more, wider,
Member of narrower, etc.
One Club
• Use transition words
Rides the School Bus Member of 2 – However, whereas, similarly, additionally, etc.
or More Clubs
• Discuss at least two points of comparison.
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Comparing Categorical Comparing Categorical

Is there an association between after-school club
Sample Answer:
participation and whether or not the student rides the
school bus? Support your answer with a discussion of the Yes, there is a clear association between after-school club
provided graphs. participation and transportation. Only 11% of students who
don’t ride the bus do not participate in after school clubs,
whereas 51% of students who do ride the bus do not
Does not Ride Bus Member of participate. Similarly, 58% of students who do not ride the
No Clubs bus are involved in 2 or more clubs, while only 19% of
Member of students riding the bus are involved in 2 or more clubs.
One Club
However, the proportion of students who participate in one
Rides the School Bus Member of 2 club is the same for students who ride and students who
or More Clubs
don’t ride the bus.
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Section 1.2
Displaying Quantitative Data with Graphs
1.2: Displaying Quantitative
Data with Graphs After this section, you should be able to…
ü CONSTRUCT and INTERPRET dotplots, stemplots, and

histograms
ü DESCRIBE the shape of a distribution
ü COMPARE distributions
ü USE histograms wisely
Dotplots
– Each data value is shown as a dot above its
How to Make a Dotplot
location on a number line. 1. Draw a horizontal axis (a number line) and
label it with the variable name.
Number of Goals Scored Per Game by the 2004 US Women’s Soccer
Team
3 0 2 7 8 2 4 3 5 1 1 4 5 3 1 1 3 2. Scale the axis from the minimum to the
maximum value.
3 3 2 1 2 2 2 4 3 5 6 1 5 5 1 1 5
3. Mark a dot above the location on the

horizontal axis corresponding to each data
value.
How to Describe Quantitative Describing Shape

Data When you describe a distribution’s shape, concentrate on the
main features. Look for rough symmetry or clear skewness.
In any graph, look for the overall pattern and for striking
departures from that pattern.
Describe the overall pattern of a distribution by its:
•Shape
•Outliers
Don’t forget
•Center
your SOCS!
•Spread
Shape Definitions:
Symmetric: if the right and left sides of the graph are
approximately mirror images of each other.
Skewed to the right (right-skewed) if the right side of the
graph is much longer than the left side.
Skewed to the left (left-skewed) if the left side of the graph
Dot Plot Dot Plot Dot Plot
is much longer than the right side.
Collection 1 Collection 1 Collection 1
0 2 4 6 8 10 12 70 75 80 85 90 95 100 0 1 2 3 4 5 6 7
DiceRolls Score Siblings
Symmetric Skewed-left Skewed-right
Other Ways to Describe Shape:

• Unimodal
• Bimodal
• Multimodal
Outliers Center
Definition: Values that differ from the overall pattern We can describe the center by finding a value that
are outliers. divides the observations so that about half take larger
We will learn specific ways to find outliers in a later values and about half take smaller values.
chapter. For now, we can only identify “potential Ways to describe center:
outliers.” • Calculate median (best when distribution is
skewed)
• Calculate mean (best when distribution is
symmetric)
Describe the shape, center, and spread of the

Spread distribution. Are there any potential outliers?
The spread of a distribution tells us how much Remember to include CONTEXT!!!
variability there is in the data.
Ways to ‘describe’ spread:
• Calculate the range 2009 Fuel Economy Guide Dot Plot
• IQR (coming later)
• Standard Deviation (coming later)
14 16 18 20 22 24 26 28 30 32 34
MPG
Stemplots (Stem-and-Leaf Plots)

Sample Answer: Stemplots give us a quick picture of the distribution while
• Shape: The shape of the distribution is including the actual numerical values.
roughly unimodal and skewed left.
• Center: The mean is 25.9 mpg and the median
is 28 mpg. (only need one measure)
• Spread: The range is 19 mpg.
• Outliers: There are two potential
outliers/influential values: 14 mpg and 18
mpg.
How to Make a Stemplot Stemplots (Stem-and-Leaf Plots)

These data represent the responses of 20 female AP Statistics
1)Separate each observation into a stem (all but the final students to the question, “How many pairs of shoes do you
digit) and a leaf (the final digit). have?”
50 26 26 31 57 19 24 22 23 38
2)Write all possible stems from the smallest to the largest in a 13 50 13 34 23 30 49 13 15 51
vertical column and draw a vertical line to the right of the
column.
3)Write each leaf in the row to the right of its stem.
4)Arrange the leaves in increasing order out from the stem.
5)Provide a key that explains in context what the stems and
leaves represent.
Two Special Types of Stem Plots

– Spilt Stemplots: Best when data values are “bunched
Histograms
up” – Quantitative variables often
• Spilt 0-4 and 5-9 take many values. A graph
– Back-to-Back Stemplot: Compares two distributions of the distribution may be
of the same quantitative variable clearer if nearby values are
grouped together.
Females Males
0
0
0 4 – The most common graph of
0 555677778 Back-to-Back
1 333 1 0000124 the distribution of one
1
2
95
4332
1
2 2
quantitative variable is a
“split stems”
2
3
66 2 histogram.
410 3 Key: 4|9
3 8 3 58 represents a
4 4 student who
4 9 4 reported
5 100 5 having 49
5 7 5 pairs of shoes.
How to Make a Histogram Making a Histogram

1)Divide the range of data into classes of equal width.
2)Find the count (frequency) or percent (relative frequency) Frequency Table
of individuals in each class. Class Count
Number of States
0 to <5 20
3)Label and scale your axes and draw the histogram. The
5 to <10 13
height of the bar equals its frequency. Adjacent bars should
10 to <15 9
touch, unless a class contains no individuals.
15 to <20 5
20 to <25 2
25 to <30 1
Total 50
Percent of foreign-born residents
Caution: Using Histograms Wisely 1.3: Describing

1)Don’t confuse histograms and bar graphs. Quantitative Data with
2)Don’t use counts (in a frequency table) or percents (in a
relative frequency table) as data. Numbers
3)Use percents instead of counts on the vertical axis when
comparing distributions with different numbers of
observations.
4)Just because a graph looks nice, it’s not necessarily a
meaningful display of data.
Section 1.3
Describing Quantitative Data with Measuring Center: The Mean
Numbers
To find the mean x (pronounced “x-bar”) of a set of
After this section, you should be able to… observations, add their values and divide by the number of
observations. If the n observations are x1, x2, x3, …, xn, their
ü MEASURE center with the mean and median mean is:
ü MEASURE spread with standard deviation and sum of observations x1 + x 2 + ...+ x n

interquartile range
x= =
n n
ü IDENTIFY outliers
ü CONSTRUCT a boxplot using the five-number summary
Compact Notation: x=
åx i
ü CALCULATE numerical summaries with technology n

Measuring Center: The Median Comparing the Mean and the

The median M is the midpoint of a distribution, the number
Median
such that half of the observations are smaller and the other The mean and median measure center in different ways, and
half are larger. both are useful.
Mean: “average” value Median: “typical” value
To find the median of a distribution:
1)Arrange all observations from smallest to largest. Relationship between Mean & Median:
• The mean and median of a roughly symmetric distribution
2)If the number of observations n is odd, the median M is the
are close together.
center observation in the ordered list.
3)If the number of observations n is even, the median M is the • If the distribution is exactly symmetric, the mean and median
average of the two center observations in the ordered list. are exactly the same.
• In a skewed distribution, the mean is usually farther out in
the long tail than is the median.
Why is the mean more affected by

the presence of outliers than the
median?
Standard Deviation Standard Deviation

Standard deviation is a number used to tell how
• A relatively low standard deviation value indicates that the
measurements for a group are spread out from the data points tend to be very close to the mean.
mean. • A relatively high standard deviation value indicates that the
data points are spread out over a large range of values.
Standard Deviation Formula FYI: Why n-1?!

The standard deviation sx measures the average distance of the
observations from their mean. It is calculated by finding an
Applet:
average of the squared distances and then taking the square http://www.uvm.edu/~dhowell/SeeingStati
root. This average squared distance is called the variance. sticsApplets/N-1.html
( x1 - x ) 2 + ( x2 - x ) 2 + ... + ( xn - x ) 2 1
variance = sx2 =
n -1
=
n -1
å ( xi - x )2 Proof
1
standard deviation = sx =
n -1
å (x i - x ) 2
How to Calculate Standard Calculate the Standard Deviation

Deviation by Hand Calculate the standard deviation.
1. Calculate mean.
2. Calculate each deviation. Subtract your mean score
from every actual (observed) score.
3. Square each deviation.
4. Find the “average” squared deviation by calculating Collection 5 Dot Plot
the sum of the squared deviations divided by (n-1).
4. Divide that sum by the number of cases in your data
5. Finally, calculate the square root of the number
calculate in step #4
0 2 4 6 8 10
NumberOfPets
Calculate the Standard Deviation

Calculate the Standard Deviation
xi (xi-mean) 3) Square each deviation. xi (xi-mean) (xi-mean)2
1) Calculate the mean. Step 3: See Table 1 1 - 5 = -4
1
Step 1: 5 3 3 - 5 = -2
3 4) Find the “average”
4 squared deviation by 4 4 - 5 = -1
2) Calculate each deviation.
4 calculating the sum of the 4 4 - 5 = -1
deviation = observation – mean
squared deviations 4 4 - 5 = -1
4
Collection 5 Dot Plot divided by (n-1).
deviation: 8 - 5 = 3 5 5 5 - 5 = 0
Step 4: “Average”
deviation: 1 - 5 = -4
7 squared deviation = 7 7 - 5 = 2
8 52/(9-1) = 6.5 8 8 - 5 = 3
0 2 4 6 8 10
NumberOfPets 9 Variance = 6.5 9 9 - 5 = 4
x=5 Sum= Sum= Sum=

Calculate the Standard Deviation Two Extreme Examples:

xi (xi-mean) (xi-mean)2
5) Calculate the square root 1 1 - 5 = -4 (-4)2 = 16 • In dataset #1, we have five people that report
of the variance…this is eating 4 pieces of cake and five people that
3 3 - 5 = -2 (-2)2 = 4
the standard deviation. report eating 6 pieces of cake, for a mean of 5
4 4 - 5 = -1 (-1)2 = 1
4 4 - 5 = -1 (-1)2 = 1
pieces of cake ([4+4+4+4+4+6+6+6+6+6]/10=5).
Step 5: Square root
4 4 - 5 = -1 (-1)2 = 1 – Mean =5; Variance = 1
of variance
5 5 - 5 = 0 (0)2 = 0 • In dataset #2, we have five people that report
6.5 = 2.55 7 7 - 5 = 2 (2)2 = 4 eating 0 piece of cake and five people that report
Standard Deviation =
2.55
8 8 - 5 = 3 (3)2 = 9 eating 10 pieces of cake, for a mean of 5 pieces of
9 9 - 5 = 4 (4)2 = 16 cake ([0+0+0+0+0+10+10+10+10+10]/10=5).
Sum=? Sum=? – Mean = 5; Variance = 5
Below are dotplots of three different TI-NSpire: Calculate standard deviation and
distributions, A, B, and C. Which one has the mean.
largest standard deviation? Justify your answer.
1. Select “Lists & Spreadsheet” (blue/green
button at bottom of home screen)
2. Type the values into list1.
3. With your cursor on the values, press menu

4. Select 4: Statistics, then 1: Stat Calculations,
press enter.
5. Select 1: One-Variable Stats
TI-NSpire: Calculate standard deviation and

mean.
6. Set screen to:

Mean
and then press enter.
Standard
Deviation
Interquartile Range (IQR) Interquartile Range (IQR)

To calculate:
1)Arrange the observations in increasing order and locate the
median M.
2)The first quartile Q1 is the median of the observations
located to the left of the median in the ordered list.
3)The third quartile Q3 is the median of the observations
located to the right of the median in the ordered list.
The interquartile range (IQR) is defined as:
IQR = Q3 – Q1
Find and Interpret the IQR… Find and Interpret the IQR…
Travel times to work for 20 randomly selected New Yorkers Travel times to work for 20 randomly selected New Yorkers
10 30 5 25 40 20 10 15 30 20 15 20 85 15 65 15 60 60 40 45 10 30 5 25 40 20 10 15 30 20 15 20 85 15 65 15 60 60 40 45
5 10 10 15 15 15 15 20 20 20 25 30 30 40 40 45 60 60 65 85
Q1 = 15 M = 22.5 Q3= 42.5
IQR = Q3 – Q1
= 42.5 – 15
= 27.5 minutes
Interpretation: The range of the middle half of travel
times for the New Yorkers in the sample is 27.5 minutes.
Identifying Outliers In the New York travel time data, we found Q1=15
minutes, Q3=42.5 minutes, and IQR=27.5 minutes.
Calculate the outlier cutoffs using the IQR rule.
In addition to serving as a measure of spread, the
interquartile range (IQR) is used as part of a rule of thumb for
identifying outliers.
1.5 x IQR Rule for Outliers

Call an observation an outlier if it falls more than 1.5 x
IQR above the third quartile or below the first quartile.
In the New York travel time data, we found Q1=15 The Five-Number Summary
minutes, Q3=42.5 minutes, and IQR=27.5 minutes.
The five-number summary of a distribution consists of the
Calculate the outlier cutoffs using the IQR rule. smallest observation, the first quartile, the median, the third
quartile, and the largest observation, written in order from
smallest to largest.
For these data, 1.5 x IQR = 1.5(27.5) = 41.25
Q1 - 1.5 x IQR = 15 – 41.25 = -26.25 Minimum Q1 M Q3 Maximum
Q3+ 1.5 x IQR = 42.5 + 41.25 = 83.75
Any travel time shorter than -26.25 minutes or longer than

83.75 minutes is considered an outlier.
TI- Nspire: 5 Number Summary TI- Nspire: 5 Number Summary

1. Select “Lists & Spreadsheet” (bottom of home 6. Set screen to:
screen)
and then press enter.
2. Type the values into list1.
3. With your cursor on the values, press menu

4. Select 4: Statistics, then 1: Stat Calculations,
press enter. 7. Scroll down to see the 5 number summary.
5. Select 1: One-Variable Stats
Boxplots (Box-and-Whisker Plots)

•Draw and label a number line
that includes the range of the
distribution.
•Draw a central box from Q1 to
Q3.
•Note the median M inside the
box.
•Extend lines (whiskers) from the
box out to the minimum and
maximum values that are not
outliers.
Construct a Boxplot Construct a Boxplot

Using our NY travel times data. Construct a boxplot. Using our NY travel times data. Construct a boxplot.
10 30 5 25 40 20 10 15 30 20 15 20 85 15 65 15 60 60 40 45 10 30 5 25 40 20 10 15 30 20 15 20 85 15 65 15 60 60 40 45
5 10 10 15 15 15 15 20 20 20 25 30 30 40 40 45 60 60 65 85
Min=5 Q1 = 15 M = 22.5 Q3= 42.5 Max=85

Recall, this is an
outlier by the
1.5 x IQR rule
Collect ion 5 Box Plot Collect ion 5 Box Plot
0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90
TravelTime TravelTime
Choosing Best Measures of Center

& Spread
Symmetric Skewed
Distribution Distribution
Best Measure of
Center
Best Measure of
Spread
Section 2.1
2.1: Describing Location Describing Location in a Distribution
in a Distribution After this section, you should be able to…
ü MEASURE position using percentiles

ü INTERPRET cumulative relative frequency graphs
ü TRANSFORM data
ü DEFINE and DESCRIBE density curves
Measuring Position: Percentiles Percentiles

– One way to describe the location of a value in a A percentile measure the position of a single data item based
distribution is to tell what percent of observations are on the percentage of data items below that single data item.
less than it.
– The pth percentile of a distribution is the value with p Standardized tests taken by larger numbers of students,
percent of the observations less than it. convert raw scores to a percentile score.
If approximately n percent of the items in a distribution are
less than the number x, then x is the nth percentile of the
distribution, denoted Pn.
Percentiles Other Percentiles: Deciles and Quartiles

Example: Deciles are the nine values (denoted D1, D2,…, D9) along the
The following are test scores (out of 100) for a particular math scale that divide a data set into ten (approximately) equal parts.
class.
44 56 58 62 64 64 70 72 72 72 74 74 75 78 78 79 80 82 82 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, and 90%
84 86 87 88 90 92 95 96 96 98 100
Quartiles are the three values (Q1, Q2, Q3) that divide the data
Find the fortieth percentile. set into four (approximately) equal parts.
40% = 0.4 25%, 50%, and 75%
The average of the 12th and 13th items
0.4(30) represents the 40th percentile (P40).
12
40% of the scores were below 74.5.
Other Percentiles: Deciles and Quartiles Other Percentiles: Deciles and Quartiles
Example: Deciles Quartiles
The following are test scores (out of 100) for a particular math For any set of data (ranked in order from least to greatest):
class.
The second quartile, Q2 (50%) is the median.
44 56 58 62 64 64 70 72 72 72 74 74 75 78 78 79 80 82 82
84 86 87 88 90 92 95 96 96 98 100 The first quartile, Q1 (25%) is the median of items below Q2.
The third quartile, Q3 (75%) is the median of items above Q2.
Find the sixth decile.
Sixth decile = 60% The average of the 18th and 19th items
60% = 0.6 represents the 6th decile (D6).
0.6(30) 60% of the scores were at or below 82.
18
Other Percentiles: Deciles and Quartiles Other Percentiles: Deciles and Quartiles
Example: Quartiles Example: Quartiles
The following are test scores (out of 100) for a particular math The following are test scores (out of 100) for a particular math
class. class.
44 56 58 62 64 64 70 72 72 72 74 74 75 78 78 79 80 82 82 44 56 58 62 64 64 70 72 72 72 74 74 75 78 78 79 80 82 82
84 86 87 88 90 92 95 96 96 98 100 84 86 87 88 90 92 95 96 96 98 100
Find the three quartiles. Find the three quartiles.

Q1= 25% The 8th item represents the 1st quartile Q2= 50% = median The average of the 15th and 16th items
25% = 0.25 (Q1) 50% = 0.5 represents the 2nd quartile (Q2) or the
0.25(30) 0.5(30) median
25% of the scores were below 72.
7.5 15 50% of the scores were below 78.5.
Other Percentiles: Deciles and Quartiles Cumulative Relative Frequency

Example: Quartiles Graphs
The following are test scores (out of 100) for a particular math A cumulative
class. relative frequency
44 56 58 62 64 64 70 72 72 72 74 74 75 78 78 79 80 82 82 graph displays the
84 86 87 88 90 92 95 96 96 98 100 cumulative relative
frequency of each
Find the three quartiles.
class of a frequency
Q3= 75% The 23rd item represents the 3rd quartile distribution.
75% = 0.75 (Q3)
0.75(30) 75% of the scores were below 88.
22.5
Age of First 44 Presidents When They Were Inaugurated Age of First 44 Presidents When They Were Inaugurated
Age Frequency Relative Cumulative Cumulative Age Frequency Relative Cumulative Cumulative
frequency frequency relative frequency frequency relative
frequency frequency
40- 2 2/44 = 2 2/44 = 40- 2 2/44 = 2 2/44 =

44 4.5% 4.5% 44 4.5% 4.5%
45- 7 7/44 = 9 9/44 =
Age of Presidents 45- 7 7/44 = 9 9/44 =
Age of Presidents
49 15.9% 20.5% When Inaugurated 49 15.9% 20.5% When Inaugurated
50- 13 13/44 = 22 22/44 = 50- 13 13/44 = 22 22/44 =
100 100
54 29.5% 50.0% 54 29.5% 50.0%
Cumulative relative frequency (%)
Cumulative relative frequency (%)

55- 12 12/44 = 34 34/44 = 55- 12 12/44 = 34 34/44 =
80 80
59 27.3% 77.3% 59 27.3% 77.3%
60- 7 7/44 = 41 41/44 = 60- 7 7/44 = 41 41/44 =
60 60
64 15.9% 93.2% 64 15.9% 93.2%
65- 3 3/44 = 44 44/44 = 65- 3 3/44 = 44 44/44 =
40 40
69 6.8% 100% 69 6.8% 100%
20 20
0 0
40 45 50 55 60 65 70 40 45 50 55 60 65 70
Age at inauguration Age at inauguration
Transforming Data Transforming Data: Add/Sub a

Constant
• Transforming converts the original observations from the
original units of measurements to another scale. Adding the same number a (either positive, zero, or
negative) to each observation:
• Some transformations can affect the shape, center, and
spread of a distribution. •adds a to measures of center and location (mean,
median, quartiles, percentiles), but
•Does not change the shape of the distribution or
measures of spread (range, IQR, standard deviation).
Transforming Data: Add/Sub a

Constant Transforming Data:
n Mean sx Min Q1 M Q3 Max IQR Range
Multiplying/Dividing
Guess(m) 44 16.02 7.14 8 11 15 17 40 6 32
Multiplying (or dividing) each observation by the same
number b (positive, negative, or zero):
Error 44 3.02 7.14 -5 -2 2 4 27 6 32
(m -13)
•multiplies (divides) measures of center and location by b
•multiplies (divides) measures of spread by |b|
•does not change the shape of the distribution
Transforming Data Density Curve

Change data from feet to meters
A density curve:
n Mean sx Min Q1 M Q3 Max IQR Range • is ralways on or above the
Error 44 9.91 23.43 -16.4 -6.56 6.56 13.12 88.56 19.68 104.96 horizontal axis, and
(feet) • has area exactly 1 underneath
Error 44 3.02 7.14 -5 -2 2 4 27 6 32
it.
(meters) • A density cuve describes the
overall pattern of a
distribution.
• The area under the curve and
above any interval of values on
the horizontal axis is the
The overall pattern of this histogram of the scores
proportion of all observations of all 947 seventh-grade students in Gary, Indiana,
that fall in that interval. on the vocabulary part of the Iowa Test of Basic
Skills (ITBS) can be described by a smooth curve
drawn through the tops of the bars.
Density Curves Continued… Describing Density Curves

• Areas under a density curve represent
proportions of observations The median and the mean are the same for a symmetric
density curve. They both lie at the center of the curve. The
• The scale of a density curve is adjusted in mean of a skewed curve is pulled away from the median in the
such a way that the total area under the direction of the long tail.
curve is always equal to 1
Section 2.2
2.2: Normal Distributions
Normal Distributions After this section, you should be able to…
ü DESCRIBE and APPLY the 68-95-99.7 Rule

ü DESCRIBE the standard Normal Distribution
ü PERFORM Normal distribution calculations
ü ASSESS Normality
Normal Distributions Normal Distributions

• All Normal curves are symmetric, single-peaked, and bell- • We abbreviate the Normal distribution with mean µ and
shaped standard deviation σ as N(µ,σ).
• A Specific Normal curve is described by giving its mean µ • Any particular Normal distribution is completely specified by
and standard deviation σ. two numbers: its mean µ and standard deviation σ.
• The mean of a Normal distribution is the center of the
symmetric Normal curve.
• The standard deviation is the distance from the center to the
change-of-curvature points on either side.
Two Normal curves, showing the mean µ and

standard deviation σ.
Normal Distributions are Useful… The 68-95-99.7 Rule

• Normal distributions are good descriptions for some Although there are many different sizes and shapes of Normal
distributions of real data. curves, they all have properties in common.
• Normal distributions are good approximations of the results
of many kinds of chance outcomes. The 68-95-99.7 Rule (“The Empirical Rule”)
• Many statistical inference procedures are based on Normal In the Normal distribution with mean µ and standard
distributions. deviation σ:
•Approximately 68% of the observations fall
within σ of µ.
•Approximately 95% of the observations fall
within 2σ of µ.
•Approximately 99.7% of the observations fall
within 3σ of µ.
The distribution of Iowa Test of Basic Skills (ITBS) vocabulary

scores for 7th grade students in Gary, Indiana, is close to
Normal. Suppose the distribution is N(6.84, 1.55) and the
range is between 0 and 12.
a) Sketch the Normal density curve for this distribution.
The distribution of Iowa Test of Basic Skills (ITBS) vocabulary The distribution of Iowa Test of Basic Skills (ITBS) vocabulary
scores for 7th grade students in Gary, Indiana, is close to scores for 7th grade students in Gary, Indiana, is close to
Normal. Suppose the distribution is N(6.84, 1.55) and the Normal. Suppose the distribution is N(6.84, 1.55).
range is between 0 and 12. b) Using the Empirical Rule, what percent of ITBS vocabulary
a) Sketch the Normal density curve for this distribution. scores are less than 3.74?
The distribution of Iowa Test of Basic Skills (ITBS) vocabulary The distribution of Iowa Test of Basic Skills (ITBS) vocabulary
scores for 7th grade students in Gary, Indiana, is close to scores for 7th grade students in Gary, Indiana, is close to
Normal. Suppose the distribution is N(6.84, 1.55). Normal. Suppose the distribution is N(6.84, 1.55).?
b) Using the Empirical Rule, what percent of ITBS vocabulary c) Using the Empirical Rule, what percent of the scores are
scores are less than 3.74? between 5.29 and 9.94?
The distribution of Iowa Test of Basic Skills (ITBS) vocabulary

scores for 7th grade students in Gary, Indiana, is close to Importance of Standardizing
Normal. Suppose the distribution is N(6.84, 1.55).?
c) Using the Empirical Rule, What percent of the scores are • There are infinitely many different Normal
between 5.29 and 9.94? distributions; all with unique standard deviations
and means.
• In order to more effectively compare different
Normal distributions we “standardize”.
• Standardizing allows us to compare apples to
apples.
• We can compare SAT and ACT scores by
standardizing.
The Standardized Normal Distribution How to Standardize a Variable:

All Normal distributions are the same if we measure in
1. Draw and label an Normal curve with the mean and
units of size σ from the mean µ as center.
standard deviation.
The standardized Normal distribution 2. Calculate the z- score
is the Normal distribution with mean 0
x= variable
and standard deviation 1.
µ= mean
σ= standard deviation
3. Determine the p-value by looking up the z-score in

the Standard Normal table.
4. Conclude in context.
The Standard Normal Table

Because all Normal distributions are the same when we
standardize, we can find areas under any Normal curve
from a single table.
The Standard Normal table is a table of the areas under

the standard normal curve. The table entry for each value
“z” is area under the curve to the LEFT of z.
The area to left is called the “p-value”

• Probability
• Percent
Using the Standard Normal Table

Using the Standard Normal Table
Row: Ones and tenths digits
Column: Hundredths digit Using the Standard Normal Table, find the
Practice: What is the p-value for a z-score of -2.33? following:
Z-Score P-value
-2.23
1.65
.52
.79
.23
1. Draw and label an Normal curve with the mean and

Let’s Practice standard deviation.
In the 2008 Wimbledon tennis tournament,

Rafael Nadal averaged 115 miles per hour (mph) 2. Calculate the z- score
on his first serves. Assume that the distribution x= variable
of his first serve speeds is Normal with a mean of µ= mean
115 mph and a standard deviation of 6.2 mph. σ= standard deviation
About what proportion of his first serves would
you expect to be less than 120 mph? Greater
than?
4. Conclude in context.
3. Determine the p-value by looking up the z-score in We expect that 79.1% of Nadal’s first serves will be less
the Standard Normal table. than 120 mph.
P(z < 0.81) = .7910
We expect that 20.9% of Nadal’s first serves will be greater
Z .00 .01 .02 than 120 mps.
0.7 .7580 .7611 .7642
0.8 .7881 .7910 .7939
0.9 .8159 .8186 .8212
Let’s Practice When Tiger Woods hits his driver, the distance the ball
travels can be described by N(304, 8). What percent of
When Tiger Woods hits his driver, the Tiger’s drives travel between 305 and 325 yards?
distance the ball travels can be Step 1: Draw Distribution
described by N(304, 8). What percent of
Tiger’s drives travel between 305 and
325 yards?
Step 2: Z- Scores
325 - 304
When x = 325, z = = 2.63
8
305 - 304
When x = 305, z = = 0.13
8
Step 3: P-values
Normal Calculations on Calculator
Calculates Example
NormalCDF Probability of What percent of students
obtaining a value scored between 70 and
BETWEEN two values 95 on the test?
NormalPDF Probability of What is the probability
Using Table A, we can find the area to the left of z=2.63 and
obtaining PRECISELY that Suzy scored a 75 on
the area to the left of z=0.13.
or EXACTLY a specific the test?
0.9957 – 0.5517 = 0.4440.
x-value
InvNorm X-value given Tommy scored a 92 on
Step 4: Conclude In Context probability or the test; what proportion
percentile of students did he score
About 44% of Tiger’s drives travel between 305 and 325 yards. better than?
TI-Nspire: NormalCDF TI-Nspire: NormalPDF

Normalcdf- “Area under the curve between two points” Normalpdf- “Exact percentile/probability of a specific event occurring”
1. Select Calculator (on home screen), press center button.

1. Select Calculator (on home screen), press center button.
2. Press menu, press enter.
3. Select 6: Statistics, press enter.
4. Select 5: Distributions, press enter.
5. Select 1: Normal Pdf press enter.
5. Select 2: Normal Cdf, press enter.
6. Enter the following information:
1. Xvalue (not a percent)
1. Lower: (the lower bound of the region OR 1^-99)
2. µ: (mean)
2. Upper: (the upper band of the region OR 1,000,000)
3. Ơ: (standard deviation)
3. µ: (mean)
7. Press enter, number that appears is the p-value
When Tiger Woods hits his driver, the distance

TI-Nspire: InvNorm the ball travels can be described by N(304, 8).
invNorm- “Exact x-value at which something occurred” What percent of Tiger’s drives travel between
1. Select Calculator (on home screen), press center button. 305 and 325 yards?
5. Select 3: Inverse Norm press enter.
1. Area (enter as a decimal)
2. µ: (mean)
When Tiger Woods hits his driver, the distance Suzy bombed her recent AP Stats exam; she
the ball travels can be described by N(304, 8). scored at the 25th percentile. The class average
What percent of Tiger’s drives travel between was a 170 with a standard deviation of 30.
305 and 325 yards? Assuming the scores are normally distributed,
what score did Suzy earn of the exam?
Suzy bombed her recent AP Stats exam; she When Can I Use Normal
scored at the 25th percentile. The class average Calculations?!
was a 170 with a standard deviation of 30.
Assuming the scores are normally distributed, • Whenever the distribution is Normal.
what score did Suzy earn of the exam? • Ways to Assess Normality:
– Plot the data.
• Make a dotplot, stemplot, or histogram and see if the
graph is approximately symmetric and bell-shaped.
– Check whether the data follow the 68-95-99.7
rule.
– Construct a Normal probability plot.
Normal Probability Plot Interpreting Normal Probability Plot

• If the points on a Normal probability plot lie close to a
• These plots are constructed by plotting each observation straight line, the plot indicates that the data are Normal.
in a data set against its corresponding percentile’s z-
score. • Systematic deviations from a straight line indicate a non-
Normal distribution.
• Outliers appear as points that are far away from the
overall pattern of the plot.
Summary: Normal Distributions

• The Normal Distributions are described by a special family of bell-
shaped, symmetric density curves called Normal curves. The mean µ and
standard deviation σ completely specify a Normal distribution N(µ,σ). The
mean is the center of the curve, and σ is the distance from µ to the
change-of-curvature points on either side.
• All Normal distributions obey the 68-95-99.7 Rule, which describes what
percent of observations lie within one, two, and three standard
deviations of the mean.
• All Normal distributions are the same when measurements are
standardized. The standard Normal distribution has mean µ=0 and Additional Help
standard deviation σ=1.
• Table A gives percentiles for the standard Normal curve. By standardizing,
we can use Table A to determine the percentile for a given z-score or the
z-score corresponding to a given percentile in any Normal distribution.
• To assess Normality for a given set of data, we first observe its shape. We
then check how well the data fits the 68-95-99.7 rule.
Finding Areas Under the Standard Normal Finding Areas Under the Standard
Curve Normal Curve
Find the proportion of observations from the standard Find the proportion of observations from the standard
Normal distribution that are between -1.25 and 0.81. Normal distribution that are between -1.25 and 0.81.
Step 1: Look up area to
Step 3: Subtract.
the left of 0.81 using
table A.
Step 2: Find the area to

the left of -1.25
Section 3.1
3.1: Scatterplots & Scatterplots and Correlation
Correlation After this section, you should be able to…
ü IDENTIFY explanatory and response variables

ü CONSTRUCT scatterplots to display relationships
ü INTERPRET scatterplots
ü MEASURE linear association using correlation
ü INTERPRET correlation
Explanatory & Response Scatterplots

Variables • A scatterplot shows the relationship between two
quantitative variables measured on the same individuals.
Explanatory Variables Response Variables • The values of one variable appear on the horizontal axis,
(Independent Variables ) (Dependent Variables) and the values of the other variable appear on the
vertical axis.
Car weight Accident death rate • Each individual in the data appears as a point on the
graph.
Number of cigarettes Life expectancy
smoked
Number of hours studied SAT scores

Scatterplots
Scatterplots Make a scatterplot of the relationship between body weight and
pack weight. Body weight is our eXplanatory variable.
1. Decide which variable should go on each axis. Body weight (lb) 120 187 109 103 131 165 158 116
Backpack weight 26 30 26 24 29 35 31 28
• Remember, the eXplanatory variable goes on the X- (lb)
axis!
2. Label and scale your axes.
3. Plot individual data values.
Constructing a Scatterplot: Constructing a Scatterplot:

TI-Nspire TI-Nspire
1. Enter x values into list 4. Move the cursor to
1 and enter y values the bottom of the
into list 2. screen and “click to
2. Label each column. add variable”. Select
Label column x : weight.
weight and column y: 5. Move the cursor to
bpack. the left of the screen
3. Press HOME/On, click and “click to add
Add Data & Statistics variable”. Select
bpack.
Constructing a Scatterplot Describing Scatterplots

As in any graph of data, look for the overall pattern and for
striking departures from that pattern.
• You can describe the overall pattern of a scatterplot by
the direction, form, and strength of the relationship.
• An important kind of departure is an outlier, an
individual value that falls outside the overall pattern of
the relationship.
• Also, clustering.
Words That Describe… More on Strength…

• Direction (slope) • Strength refers to how tightly grouped the
– Positive or Negative points are in a particular pattern.
• Form • Later on we use describe strength as
– Linear, quadratic, cubic, exponential, curved, non- “correlation”
linear, etc.
• Strength
– Strong, weak, somewhat strong, very weak,
moderately strong, etc.
Describe this Scatterplot Describe this Scatterplot
Describe this Scatterplot Interpreting a Scatterplot

• Interpret….tell what the data suggests in real
world terms.
• Example: The data suggests that the more
hours a student studied for Mrs. Daniel’s AP
Stats test the higher grade the student earned.
There is a positive relationship between hours
studied and grade earned.
Describe and interpret the scatterplot below. The y-axis Describe and interpret the scatterplot below. The y-axis
refers to backpack weight in pounds and the x-axis refers to refers to backpack weight in pounds and the x-axis refers to
body weight in pounds. body weight in pounds.
Sample Answer:
There is a moderately strong,
positive, linear relationship
between body weight and pack
weight. There is one possible
outlier, the hiker with the body
weight of 187 pounds seems to
be carrying relatively less weight
than are the other group
members. It appears that lighter
students are carrying lighter
backpacks
Describe and interpret the scatterplot below. The y-axis refer Describe and interpret the scatterplot below. The y-axis refer
to a school’s mean SAT math score. The x-axis refers to the to a school’s mean SAT math score. The x-axis refers to the
percentage of students at a school taking the SAT. percentage of students at a school taking the SAT.
Sample Answer:
There is a moderately strong,
negative, curved relationship
between the percent of
students in a state who take
the SAT and the mean SAT
math score.
Further, there are two distinct
clusters of states and at least
one possible outliers that falls
outside the overall pattern.
What is Correlation? What does “r” tell us?!

• A mathematical value that describes the • Correlation describes what percent of
strength of a linear relationship between variation in y is ‘explained’ by x.
two quantitative variables. • Notice that the formula is the sum of the z-
• Correlation values are between -1 and 1. scores of x multiplied by the z-scores of y.
• Correlation is abbreviated: r
• The strength of the linear relationship
increases as r moves away from 0 towards -1
or 1.
What does “r” mean?

R Value Strength
Scatterplots and Correlation
-1 Perfectly linear; negative

-0.75 Strong negative relationship
-0.50 Moderately strong negative relationship
-0.25 Weak negative relationship
0 nonexistent
0.25 Weak positive relationship
0.50 Moderately strong positive relationship
0.75 Strong positive relationship
1 Perfectly linear; positive
Describe and interpret the scatterplot below. Be

How strong is the correlation? Is it sure to estimate the correlation.
positive or negative?
• 0.235
• -0.456
• 0.975
• -0.784
Sample Answer: Describe and interpret the scatterplot below. Be

As the number of boats registered in Florida sure to estimate the correlation.
increases so does the number of manatees killed
by boats. This relationship is evidenced in the
scatterplot by a strong, positive linear
relationship. The estimated correlation is
approximately r =0.85.
**Answers between 0.75-0.95 would be

acceptable.
Sample Answer: Estimate the Correlation Coefficient

As the number of predicted storms increases,
so does the number of observed storms, but
the relationship is weak. The relationship
evidenced in the scatterplot is a fairly weak
positive linear relationship. The estimated
correlation is approximately r = 0.25.
**Answers between 0.15 and 0.45 would be

acceptable.
Estimate the Correlation Coefficient Calculate Correlation: TI-Nspire

1. Enter x values in list 1 and y values in list 2.
2. Press MENU, then 4: Statistics
3. Option 1: Stat Calculations
4. Option 3: Linear Regression mx + b
5. X: a[] , Y: b[] , ENTER
6. Correlation = r
Correlation should be 0.79

Find the Correlation Facts about Correlation

1. Correlation requires that both variables be quantitative.
Height in 5.5 6.0 5.25 6.25 5.75 6.0 5.75 5.5 5.75 2. Correlation does not describe curved relationships between
Feet variables, no matter how strong the relationship is.
Weight in 150 180 138 191 172 181 168 148 172 3. Correlation is not resistant. r is strongly affected by a few
pounds outlying observations.
4. Correlation makes no distinction between explanatory and
response variables.
R = 0.97 5. r does not change when we change the units of measurement

of x, y, or both.
6. r does not change when we add or subtract a constant to
either x, y or both.
7. The correlation r itself has no unit of measurement.
R: Ignores distinctions R: Highly Effected By

between X & Y Outliers
Why?! Correlation Formula:

Suppose that we have data on variables x and y for n
• Since r is calculated using standardized values individuals.
(z-scores), the correlation value will not The values for the first individual are x1 and y1, the values
change if the units of measure are changed for the second individual are x2 and y2, and so on.
(feet to inches, etc.) The means and standard deviations of the two variables are
x-bar and sx for the x-values and y-bar and sy for the y-
• Adding a constant to either x or y or both will values.
not change the correlation because neither The correlation r between x and y is:
the standard deviation nor distance from the 1 éæ x1 - x öæ y1 - y ö æ x 2 - x öæ y 2 - y ö æ x - x öæ y n - y öù
r= êç ÷ç ÷ +ç ÷ç ÷ + ...+ ç n ÷çç ÷÷ú
mean will be impacted. n -1 êëè sx øçè sy ÷ø è sx øçè sy ÷ø è sx øè sy øúû
1 æ x - x öæ y - y ö
r= åç i ÷ç i ÷
n -1 è sx øçè sy ÷ø
Section 3.2
3.2: Least Squares Least-Squares Regression
Regressions After this section, you should be able to…
ü INTERPRET a regression line

ü CALCULATE the equation of the least-squares regression
line
ü CALCULATE residuals
ü CONSTRUCT and INTERPRET residual plots
ü DETERMINE how well a line fits observed data
ü INTERPRET computer regression output
Regression Lines Regression Lines

A regression line summarizes the relationship
between two variables, but only in settings where one Regression lines are used to conduct analysis.
of the variables helps explain or predict the other. • Colleges use student’s SAT and GPAs to predict
college success
A regression line is a line • Professional sports teams use player’s vital stats
that describes how a (40 yard dash, height, weight) to predict success
response variable y changes • The Federal Reserve uses economic data (GDP,
as an explanatory variable x unemployment, etc.) to predict future economic
changes. trends.
We often use a regression
line to predict the value of y
• Macy’s uses shipping, sales and inventory data
predict future sales.
for a given value of x.
Regression Line Equation Regression Line Equation

Suppose that y is a response variable (plotted on the
vertical axis) and x is an explanatory variable (plotted on
the horizontal axis).
A regression line relating y to x has an equation of the
form:
ŷ = ax + b
In this equation,
•ŷ (read “y hat”) is the predicted value of the response
variable y for a given value of the explanatory variable x.
•a is the slope, the amount by which y is predicted to
change when x increases by one unit.
•b is the y intercept, the predicted value of y when x = 0.
Format of Regression Lines Interpreting Linear Regression

• Y-intercept: A student weighing zero pounds is predicted
Format 1: to have a backpack weight of 16.3 pounds (no practical
interpretation).
= 0.0908x + 16.3 • Slope: For each additional pound that the student
= predicted back pack weight weighs, it is predicted that their backpack will weigh an
additional 0.0908 pounds more, on average.
x= student’s weight
Format 2:
Predicted back pack weight= 16.3 +
0.0908(student’s weight)
Interpreting Linear Regression Interpreting Linear Regression

Interpret the y-intercept and slope values in = 37x + 270
context. Is there any practical interpretation?
Y-intercept: If a student studies for zero hours,
then the student’s predicted SAT score is 270
= 37x + 270 points. This makes sense.
x= Hours Studied for the SAT
Slope: For each additional hour the student
Predicted SAT Math Score
studies, his/her score is predicted to increase
37 points, on average. This makes sense.
Predicted Value Predicted Value

What is the predicted SAT Math score for a student who What is the predicted SAT Math score for a student who
studies 12 hours? studies 12 hours?
= 37x + 270 = 37x + 270

Hours Studied for the SAT (x) Hours Studied for the SAT (x)
Predicted SAT Math Score (y) Predicted SAT Math Score (y)
= 37(12) + 270
Predicted Score: 714 points
Self Check Quiz: Calculate the Regression

Equation
Self Check Quiz! A crazy professor believes that a child with IQ 100 should
have a reading test score of 50, and that reading score should
increase by 1 point for every additional point of IQ. What is
the equation of the professor’s regression line for predicting
reading score from IQ? Be sure to identify all variables used.
Self Check Quiz: Calculate the Regression Self Check Quiz: Interpreting Regression Lines &
Equation Predicted Value
Data on the IQ test scores and reading test scores for a
A crazy professor believes that a child with IQ 100 should group of fifth-grade children resulted in the following
have a reading test score of 50, and that reading score should regression line:
increase by 1 point for every additional point of IQ. What is predicted reading score = −33.4 + 0.882(IQ score)
the equation of the professor’s regression line for predicting
reading score from IQ? Be sure to identify all variables used. (a) What’s the slope of this line? Interpret this value in
context.
Answer: (b) What’s the y-intercept? Explain why the value of the
= 50 + x intercept is not statistically meaningful.
= predicted reading score (c) Find the predicted reading scores for two children
x = number of IQ points above 100 with IQ scores of 90 and 130, respectively.
predicted reading score = −33.4 + 0.882(IQ score) Least-Squares Regression Line

Different regression lines produce different residuals. The
(a) Slope = 0.882. For each 1 point increase of IQ regression line we use in AP Stats is Least-Squares
score, the reading score is predicted to increase Regression.
0.882 points, on average. The least-squares regression line of y on x is the line that
makes the sum of the squared residuals as small as possible.
(b) Y-intercept= -33.4. If the student has an IQ of
zero, which is essential impossible (would not be
able to hold a pencil to take the exam), the score
would be -33.4. This has no practical interpretation.
(c) Predicted Value: 90: -33.4 + 0.882(90) = 45.98

130: -33.4 + 0.882(130) = 81.26 points.
TI-NSpire: LSRL
1. Enter x data into list 1 and y data into list 2.
2. Press MENU, 4: Statistics, 1: Stat
Calculations
3. Select Option4: Linear Regression.
4. Insert either name of list or a[] for x and
name of list or b[] of y. Press ENTER.
TI-NSPIRE: LSRL to View

Graph
1. Enter x data into list 1 and y data into list 2.
Be sure to name lists
2. Press HOME/ON, Add Data & Statistics
3. Enter variables to x and y axis.
4. Click MENU, 4: Analyze
5. Option 6: Regression
6. Option 2: Show Linear (a + bx), ENTER
Residuals How to Calculate the

A residual is the difference between an observed value Residual
of the response variable and the value predicted by the
regression line. That is, 1. Calculate the predicted value, by
residual = observed y – predicted y
plugging in x to the LSRE.
2. Determine the observed/actual value.
residual = y - ŷ Positive residuals
(above line) 3. Subtract.
residual
Negative residuals
(below line)
Calculate the Residual Calculate the Residual

1. If a student weighs 170 pounds and their backpack weighs 1. If a student weighs 170 pounds and their backpack
35 pounds, what is the value of the residual? weighs 35 pounds, what is the value of the residual?
2. If a student weighs 105 pounds and their backpack weighs Predicted: ŷ = 16.3 + 0.0908 (170) = 31.736
24 pounds, what is the value of the residual?
Observed: 35
Residual: 35 - 31.736 = 3.264 pounds
The student’s backpack weighs 3.264 pounds more
than predicted.
Calculate the Residual Residual Plots

2. If a student weighs 105 pounds and their backpack A residual plot is a scatterplot of the residuals against the
explanatory variable. Residual plots help us assess how well
weighs 24 pounds, what is the value of the residual?
a regression line fits the data.
Predicted: ŷ = 16.3 + 0.0908 (105) = 25.834
Observed: 24
Residual: 24 – 25.834= -1.834
The student’s backpack weighs 1.834 pounds less
than predicted
TI-NSpire: Residual Plots Interpreting Residual Plots

A residual plot magnifies the deviations of the points from
1. Press MENU, 4: Analyze the line, making it easier to see unusual observations and
2. Option 6: Residual, Option 2: Show Residual Plot patterns.
1) The residual plot should show no obvious patterns
2) The residuals should be relatively small in size.
A valid residual plot should look like the “night sky” with
approximately equal amounts of positive and negative
residuals.
Pattern in residuals
Linear model not
appropriate
Should You Use LSRL? Interpreting Computer Regression

1. Output
Be sure you can locate: the slope, the y intercept and
determine the equation of the LSRL.
2.
" = -0.0034415x + 3.5051

𝒚
" = predicted....
𝒚
x = explanatory variable
r2: Coefficient of Determination Interpret r2

r 2 tells us how much better the LSRL does at predicting values of y
than simply guessing the mean y for each value in the dataset.
Interpret in a sentence (how much variation is
In this example, r2 equals 60.6%.
accounted for?)
60.6% of the variation in pack
weight is explained by the linear 1. r2 = 0.875, x= hours studied, y= SAT score
relationship with bodyweight. 2. r2 = 0.523, x= hours slept, y= alertness score
(Insert r2)% of the variation in y is
explained by the linear
relationship with x.
Interpret r2 S: Standard Deviation of the

Residuals
Answers:
1. 87.5% of the variation in SAT score is
explained by the linear relationship with the
number of hours studied.
2. 52.3% of the variation in alertness score is

explained by the linear relationship with the
number of hours slept. 1. Identify and interpret the standard deviation of the
residual.
S: Standard Deviation of the S: Standard Deviation of the

Residuals Residuals
If we use a least-squares regression line to predict the
values of a response variable y from an explanatory variable
x, the standard deviation of the residuals (s) is given by
s=
å residuals 2
=
å(y i - yÙ) 2
Answer: n -2 n -2
S= 0.740
S represents the typical or average error (residual).
Interpretation: On average, the model under predicts fat
gain by 0.740 kilograms using the least-squares regression Positive = UNDER predicts
line. Negative = OVER predicts
1. ŷ = 2.1495x+ 10.667
Self Check Quiz!
ŷ = predicted fuel consumption in pounds of coal
The data is a random sample of 10 trains comparing number of
cars on the train and fuel consumption in pounds of coal. x = number of rail cars
• What is the regression equation? Be sure to define all variables. 2. 96.7 % of the varation is fuel consumption is explained by the
• What is r2 telling you? linear realtionship with the number of rail cars.
• Define and interpret the slope in context. Does it have a 3. Slope = 2.1495. With each additional car, the fuel consuption
practical interpretation? increased by 2.1495 pounds of coal, on average. This makes
• Define and interpret the y-intercept in context. practical sense.
• What is s telling you? 4. Y-interpect = 10.667. When there are no cars attached to the
train the fuel consuption is 10.667 pounds of coal. This has no
practical intrepretation beacuse there is always at least one car,
the engine.
5. S= 4.361. On average, the model over predicts fuel
consumption by 4.361 pounds of coal using the least-squares
regression line.
Extrapolation Outliers and Influential

We can use a regression line to predict the response ŷ for a Points
specific value of the explanatory variable x. The accuracy of
the prediction depends on how much the data scatter • An outlier is an observation that lies outside the overall
about the line. Exercise caution in making predictions pattern of the other observations.
outside the observed values of x. • An observation is influential for a statistical calculation if
removing it would markedly change the result of the
Extrapolation is the use of a regression line for prediction calculation.
far outside the interval of values of the explanatory • Points that are outliers in the x direction of a scatterplot
variable x used to obtain the line. Such predictions are are often influential for the least-squares regression line.
often not accurate. • Note: Not all influential points are outliers, nor are all
outliers influential points.
Outliers and Influential Correlation and

Points Regression Limitations
The distinction between explanatory and
response variables is important in regression.
The left graph is perfectly linear. In the right graph, the last value was
changed from (5, 5) to (5, 8)…clearly influential, because it changed
the graph significantly. However, the residual is very small.
Correlation and Correlation and Regression

Regression Limitations Limitations
Correlation and least-squares regression
Correlation and regression lines describe lines are not resistant.
only linear relationships.
NO!!!
Correlation and Regression

Wisdom
Association Does Not Imply Causation
An association between an explanatory variable x and a
response variable y, even if it is very strong, is not by itself
good evidence that changes in x actually cause changes in y.
A serious study once found that people

with two cars live longer than people
Additional Calculations
who only own one car. Owning three
cars is even better, and so on. There is a
& Proofs
substantial positive correlation
between number of cars x and length of
life y. Why?
Least-Squares Regression Line Calculate the Least Squares Regression Line

We can use technology to find the equation of the least-
squares regression line. We can also write it in terms of the Some people think that the behavior of the stock market in
means and standard deviations of the two variables and their January predicts its behavior for the rest of the year. Take the
correlation. explanatory variable x to be the percent change in a stock
market index in January and the response variable y to be the
Equation of the least-squares regression line
change in the index for the entire year. We expect a positive
We have data on an explanatory variable x and a response correlation between x and y because the change during
variable y for n individuals. From the data, calculate the January contributes to the full year’s change. Calculation from
means and standard deviations of the two variables and data for an 18-year period gives
their correlation. The least squares regression line is the line
Mean x =1.75 % Sx= 5.36% Mean y = 9.07%
ŷ = a + bx with
Sy = 15.35% r = 0.596
sy
slope b=r and y intercept a = y - bx Find the equation of the least-squares line for predicting
sx full-year change from January change. Show your work.
The Role of r2 in Regression Accounted for Error

The standard deviation of the residuals gives us a numerical
estimate of the average size of our prediction errors. 1 – SSE/SST = 1 –
30.97/83.87
The coefficient of determination r2 is the fraction of the r2 = 0.632
variation in the values of y that is accounted for by the least- 63.2 % of the
squares regression line of y on x. We can calculate r2 using the variation in
following formula: backpack weight is
SSE
r 2 = 1- accounted for by
SST the linear model
relating pack
SSE = å residual2 SST = å ( yi - y ) 2
weight to body
If we use the LSRL to make our predictions,
In practicality, just square the correlation r. the sum of the squared residuals is 30.90.
weight.
SSE = 30.90
Interpreting a Regression Line

Unaccounted for Error Consider the regression line from the example (pg. 164)
“Does Fidgeting Keep You Slim?” Identify the slope and
y-intercept and interpret each value in context.
SSE/SST =
30.97/83.87 fatgain = 3.505 - 0.00344(NEA change)
SSE/SST = 0.368
The slope b = -0.00344 tells
us that the amount of fat
Therefore, 36.8% of gained is predicted to go down
the variation in pack by 0.00344 kg for each added
calorie of NEA.
weight is
unaccounted for by
the least-squares The y-intercept a = 3.505 kg is
regression line. the fat gain estimated by this
If we use the mean backpack weight as model if NEA does not change
when a person overeats.
our prediction, the sum of the squared
residuals is 83.87.
SST = 83.87
Section 4.1
4.1: Samples & Surveys
Samples and Surveys
ü IDENTIFY the population and sample in a sample survey

ü IDENTIFY voluntary response samples and convenience
samples
ü DESCRIBE how to use a table of random digits to select a
simple random sample (SRS)
ü DESCRIBE simple random samples, stratified random
samples, and cluster samples
ü EXPLAIN how undercoverage, nonresponse, and question
wording can lead to bias in a sample survey
Populations and Samples How do we gather data?

The population in a statistical study is the entire group
of individuals about which we want information. • Surveys
A sample is the part of the population from which we • Opinion polls
actually collect information. We use information from a • Interviews
sample to draw conclusions about the entire
population. • Studies
– Observational
Collect data from a – Retrospective (past)
Population representative Sample... • Experiments
Sample Make an Inference about
the Population.
The Idea of a Sample Survey Sampling Design

Step 1: Define the population we want to describe. • Sampling Design: method used to choose the
Step 2: Say exactly what we want to measure. sample from the population
A “sample survey” is a study that uses an organized • Types of Samples:
plan to choose a sample that represents some – Simple Random Sample
specific population. – Stratified Random Sample
– Systematic Random Sample
Step 3: Decide how to choose a sample from the
– Cluster Sample
population.
– Multistage Sample
Simple Random Sample SRS

(SRS)
• Advantages • Disadvantages
• Consist of n individuals from the population chosen – Unbiased – Large variance/high
in such a way that – Easy variability
– every individual has an equal chance of being – May not be
selected representative
– Must be able to identify
– every set of n individuals has an equal chance of
entire population
being selected
Methods of Selecting an
SRS
• Draw names from a hat
• Assign each person in the group and randomly How to Make a Simple Random
generate chosen numbers Sample
– Ways to randomly generate numbers
• Computer
• Random Table of Digits
Use of a Random Digits Table
• Calculator
The students are:
If we want to select a simple random sample of five students from a

As you can see, they are
class of 22 students, we can use a table of random digits as a tool.
already numbered.
The numbers 1 through 9

First we must number each student, in a process also called labeling. are single digits, however, so
The class list is the sampling frame. we renumber them 01
through 09 so that each
label has the same number
of digits.
RANDOM DIGITS TABLE
The labels are on the right-hand

side.
Now we go to a random digits

table, similar, but not identical, to
the one in your text.
This is just a portion of an entire random digits table. Ordinarily, the starting Comparing our selection
point is randomly selected. For this example we start at the left end of row 2. {21,09,08,20,18} and our student
list (sampling frame) we determine
our SRS.
We read digits in groups of two since that is the number of digits in our labels.
Our SRS is Shae Wilson, Tommy

McDowell, Jenna Kanoy, Dania
Warren, and Adam Tesh.
(If we had one digit labels we would read one digit at a time; for three digit labels,
read three digits at a time, etc.)
(Spaces between digits are there simply to make the chart easier to read.)
Our selection is {21,09,08,20,18}.
Table of Random Digits Use Table D at line 130 to choose an SRS of 4

A table of random digits is a long string of the digits 0, 1, 2, 3, hotels.
4, 5, 6, 7, 8, 9 with these properties:
01 Aloha Kai 08 Captiva 15 Palm Tree 22 Sea Shell
• Each entry in the table is equally likely to be any of the 10 02 Anchor Down 09 Casa del Mar 16 Radisson 23 Silver Beach
digits 0 - 9. 03 Banana Bay 10 Coconuts 17 Ramada 24 Sunset Beach
04 Banyan Tree 11 Diplomat 18 Sandpiper 25 Tradewinds
• The entries are independent of each other. That is, 05 Beach Castle 12 Holiday Inn 19 Sea Castle 26 Tropical Breeze
knowledge of one part of the table gives no information 06 Best Western 13 Lime Tree 20 Sea Club 27 Tropical Shores
07 Cabana 14 Outrigger 21 Sea Grape 28 Veranda
about any other part.
69051 64817 87174 09517 84534 06489 87201 97245
How to Choose an SRS Using Table D
Step 1: Label. Give each member of the population a numerical
label of the same length.
69 05 16 48 17 87 17 40 95 17 84 53 40 64 89 87 20
Step 2: Table. Read consecutive groups of digits of the
appropriate length from Table D. Your sample contains the Our SRS of 4 hotels for the editors to contact is: 05 Beach Castle,
individuals whose labels you find. 16 Radisson, 17 Ramada, and 20 Sea Club.
Stratified Random Sample Stratified Random Sample

• Population is divided into homogeneous (alike)
groups called strata
– Strata 1: Seniors
– Strata 2: Juniors
• SRS’s are pulled from each strata
• Helps control for lurking variables
Stratified Random Sample Systematic Random Sample

• Advantages • Disadvantages • Pick a method of identifying subjects randomly before
starting
– More precise – Difficult to do if you must
unbiased estimator divide stratum • Requires strict adherence
than SRS – Formulas for SD & • Example: Suppose a supermarket wants to study buying
– Less variability confidence intervals are habits of their customers, then using systematic
more complicated sampling they can choose every 10th or 15th customer
– Cost reduced if strata entering the supermarket and conduct the study on this
already exists sample.
Cluster Sample Cluster Samples

• Based upon location • Advantages • Disadvantages
• Randomly pick a location & sample all there – Unbiased – Clusters may not be
• Examples: representative of
– Cost is reduced
population
– All houses on a certain block
– Formulas are complicated
– All houses in a specific zip code
– All students at specific schools in MDCPS
– All students in specific homeroom classes
Multistage Sample Inference for Sampling

• The purpose of a sample is to give us information about
• At least two separate levels/stages of SRS. a larger population.
• The process of drawing conclusions about a population
• Example: on the basis of sample data is called inference.
– Stage 1: Juniors vs. Seniors
– Stage 2: Divide the above groups (Juniors and Seniors) by Why should we rely on random sampling?
AP, Regular and Honors….select 10 for each of the groups 1)To eliminate bias in selecting samples from the list of
for a total of 60. available individuals.
2)The laws of probability allow trustworthy inference about the
population
• Results from random samples come with a margin of
error that sets bounds on the size of the likely error.
• Larger random samples give better information about
the population than smaller samples.
Errors in Surveys Sources of Error in Sample Surveys

Undercoverage occurs when some groups in the population are
left out of the process of choosing the sample.
Nonresponse occurs when an individual chosen for the sample
can’t be contacted or refuses to participate.
A systematic pattern of incorrect responses in a sample survey
leads to response bias (wanting to look cool, not wanting to be
a prude, etc.).
The wording of questions is the most important influence on
the answers given to a sample survey.
Voluntary response bias occurs when participation is optional.
Usually only people with strong opinions respond.
Errors?!
• How much do you weigh?
4.2: Experiments
• Will you not vote for President Obama’s
reelection?
• Why should guns be outlawed?
• How often do you exercise?
• How many cigarettes do you smoke each
week?
• How often should Mrs. Daniel give quizzes?
Section 4.2 Observational Study vs.

Experiments Experiment
After this section, you should be able to… An observational study observes individuals and
measures variables of interest but does not
ü DISTINGUISH observational studies from experiments attempt to influence the responses.
ü DESCRIBE the language of experiments An experiment deliberately imposes some
ü APPLY the three principles of experimental design treatment on individuals to measure their
responses.
ü DESIGN comparative experiments utilizing completely
randomized designs and randomized block designs,
including matched pairs design
***When our goal is to understand cause and
effect, experiments are the only source of fully
convincing data.***
SAT Survey vs. SAT • Survey: Asking students about how many
hours they studied for the SAT and their
Experiment resulting scores.
• Experiment: Selecting a group of same IQ
Describe a survey and an experiment that students and assigning each student a
can be used to determine the relationship different random number of hours to
between SAT scores and hours studied? studying for the SAT. The student is ONLY
allowed to study the mandated amount of
hours. Then, compare their result scores
Lurking & Confounding Variables Confounding Variables

A lurking variable is a variable that is not among • Confounding refers to a problem that can arise in an
the explanatory or response variables in a study experiment, when there is another variable that may
effect the response and is in some way tied together with
but that may influence the response variable.
Experiments
the factor under investigation, leaving us unable to tell

Lurking = not included. which of the two variables (or perhaps some interaction)
A confounding variable is one whose effects on caused the observed response.
the response variable cannot be distinguished
from one or more of the explanatory variables in
the study. Confounding = included.
Confounding Variables Examples

• For example, we plant tomatoes in a garden that's half-
shaded. We test a fertilizer by putting it on the plants in What’s Lurking?!
the sun and apply none to the shaded plants. Months later
the fertilized plants bear more and better tomatoes.
• Why?
1. As shoe size increases…so does reading ability.
• Well, maybe it's the fertilizer, maybe it's the sun, maybe 2. An increase in ice cream consumption equals
we need both. We're unable to conclude that the fertilizer an increase in the number of drowning deaths
works because any effect of fertilizer is confounded with for a given period.
any effect of the extra sunshine.
A high school regularly offers a review course to The Randomized Comparative

prepare students for the SAT. This year, budget cuts
will allow the school to offer only an online version Experiment
of the course. Over the past 10 years, the average • The remedy for confounding is to perform a comparative
SAT score of students in the classroom course was experiment in which some units receive one treatment
and similar units receive another. Most well designed
1620. The online group gets an average score of
experiments compare two or more treatments.
1780. That’s roughly 10% higher than the long- time
• Comparison alone isn’t enough, if the treatments are
average for those who took the classroom review given to groups that differ greatly, bias will result. The
course. solution to the problem of bias is random assignment.
Is the online course more effective? Is there a
In an experiment, random assignment means that
lurking variable? Is there a confounding variable?
experimental units are assigned to treatments at
random, that is, using some sort of chance process.
The Randomized Comparative Experiment

The Language of Experiments
In a completely randomized design, the treatments are
assigned to all the experimental units completely by Experimental Units: smallest collection of individuals to
chance. which treatments are applied. When the units are human
Some experiments may include a control group that beings, they often are called subjects.
receives an inactive treatment or an existing baseline
treatment. Factors: General name for explanatory variables in an
Treatment
Group
1
experiment (multi-vitamin regime).
1
Treatment: a specific condition (given vitamin A vs. vitamin

Experimental Random Compare
Assignment Results
B; time frame vitamin taken) applied to the individuals in an
Units
experiment.
Group Treatment
2 2
Factor v. Treatment A cookie manufacturer is trying to determine

how long cookies stay fresh on store shelves,
A factor is a specific type or
category of treatments.
and the extent to which the type of packaging
Whereas the specific different and the store’s temperature influences how
treatments constitute levels of a long the cookies stay fresh. He designs a
factor. completely randomized experiment involving
For example, three different
low (64 Fº) and high (75 Fº) temperatures and
groups of runners are subjected
to different training methods. two types of packaging—plastic and waxed
Experimental units- runners Factor = General Group
cardboard.
Factor- Training methods List the experimental units, factors, and
Treatments- Specific type of Treatment = Specific treatments in this experiment.
workout: Speed, strength Implementation
training and distance workouts
Three Principles of Experimental

Experimental units: packages of cookies.
Design
Factors: Temperature and packaging. 1. Control for lurking variables that might affect the response:
Treatments: Low temp and plastic, high Use a comparative design and ensure that the only systematic
temp and plastic, low temp and waxed difference between the groups is the treatment administered.
cardboard, high temp and waxed cardboard. 2. Random assignment: Use impersonal chance to assign
experimental units to treatments. This helps create roughly
equivalent groups of experimental units by balancing the
effects of lurking variables that aren’t controlled on the
treatment groups.
3. Replication: Use enough experimental units in each group so
that any differences in the effects of the treatments can be
distinguished from chance differences between the groups.
Specific Types of Double-Blind

Experimental Design • In a double-blind experiment, neither the
subjects nor the experimenters know which
• Double-Blind treatment a subject received.
• Single-Blind
• Matched Pairs
• Block Design
Matched Pair Design Matched Pair Design

• In a matched pair design, subjects are paired • Example:
by matching common important attributes. – Tire wear and tear.
• Some times the results are a pre-test and – Put one set of tires on the left
post-test with the unit being “matched” to side of the car and a different set
on the right side of the car.
itself.
– This would help control the
lurking variable of different
driving styles (between teenage
boys vs. teachers) and mileage
driven.
Block Design
• A block is a group of experimental units or
subjects that are known before the
experiment to be similar in some way that is
expected to affect the response to the
treatments.
• In a block design, the random assignment of
units to treatments is carried out separately
within each block.
• Helps control for lurking variables.
Block Design
• Experiments are often blocked by
– Age
– Gender
– Race
– Achievement Level (Regular, Honors, AP, IQ level,
etc.)
Inference for Experiments 4.3: Using Studies Wisely

An observed effect so large that it would rarely
occur by chance is called statistically significant.
A statistically significant association in data from a

well-designed experiment does imply causation.
Section 4.3 Scope of Inference

Using Studies Wisely What type of inference can be made from a
particular study?
Well-designed experiments randomly assign individuals
to treatment groups. However, most experiments don’t
ü DESCRIBE the challenges of establishing causation select experimental units at random from the larger
ü DEFINE the scope of inference population. That limits such experiments to inference
about cause and effect.
ü DESCRIBE data ethics in designing studies
Most observational studies don’t randomly assign
individuals to groups, which rules out inference about
cause and effect. However, observational studies that
use random sampling can make inferences about the
population.
The Challenges of Establishing

Causation
When we can’t do an experiment, we can use the following
criteria for establishing causation.
§ The association is strong.
§ The association is consistent.
§ Larger values of the explanatory variable are
associated with stronger responses.
§ The alleged cause precedes the effect in time.
§ The alleged cause is plausible.
Discuss how each of these criteria apply to the observational

studies of the relationship between smoking and lung cancer.
Data Ethics
• All planned studies must be reviewed in advance by an
institutional review board charged with protecting the
safety and well-being of the subjects.
• All individuals who are subjects in a study must give their
informed consent before data are collected.
• All individual data must be kept confidential. Only
statistical summaries for groups of subjects may be made
public.
Section 5.1
5.1: Randomness, Randomness, Probability and Simulation
Probability and Simulation After this section, you should be able to…
ü DESCRIBE the idea of probability

ü DESCRIBE myths about randomness
ü DESIGN and PERFORM simulations
The Idea of Probability The Law of Large Numbers

Chance behavior is unpredictable in the short run, but has a The law of large numbers says that if we observe more and
regular and predictable pattern in the long run. more repetitions of any chance process, the proportion of
The probability of any times that a specific outcome occurs approaches a single
outcome of a chance value.
process is a number
between 0 (never
occurs) and 1 (always
occurs) that describes
the proportion of times
the outcome would
occur in a very long
series of repetitions.
1
Performing a Simulation
Myths about Randomness
The imitation of chance behavior, based on a model that accurately
The myth of short-run regularity: reflects the situation, is called a simulation. Simulations are usually
done with a table of random digits, calculator random number
The idea of probability is that randomness is predictable in generator (RandInt) or computer software.
the long run (1 million plus occurrences). Probability does
not allow us to make short-run predictions. State: Identify the probability calculation at interest.
Plan: Describe how to use a chance device/tool to implement
The myth of the “law of averages”: one repetition of the process. Explain clearly how to identify
Probability tells us random behavior evens out in the long the outcomes of the chance process.
run. Future outcomes are not affected by past behavior. Do: Perform many (at least 20) repetitions of the simulation.
Conclude: Use the results of your simulation to answer the
Women have a 50% of having a boy with each pregnancy;
question of interest, in context.
the gender of any previous children do not matter!
Performing a Simulation Required Elements:

– State must include:
For Example: What is the probability that a student earns an
• Identify variable
80% on a true/false quiz written in Chinese? (Assume the
exam taker does not know any Chinese). Should the instructor • Statement of probability in symbols or words.
be concerned about cheating? – Plan must include:
• What tool?
How can we simulate the probability of guessing 80% correct • What values are you assigning?
on a True/False quiz? • How many values are you picking each time?
• How many times are you conducting the simulation?
• What about repeat digits or ignored digits?
• What are you recording?
2
Required Elements: The Golden Ticket

At a local high school, 95 students have permission to park
– Do must include:
on campus. Each month, the student council holds a “golden
• Simulation data, if number of trials is 20 or less ticket parking lottery” at a school assembly. The two lucky
• Summary of data for larger trials winners are given reserved parking spots next to the school’s
main entrance. Last month, the winning tickets were drawn
– Conclude must include: by a student council member from the AP Statistics class.
• Statement of probability When both golden tickets went to members of that same
• Answer to question class, some people thought the lottery had been rigged.
– Usually about being surprised/reasonable/expected, etc. There are 28 students in the AP Statistics class, all of whom
are eligible to park on campus. Design and carry out a
simulation to decide whether it’s plausible that the lottery
was carried out fairly.
**See 5.1 WS
STATE: PLAN:
• What is the probability that the lottery would Using the table of random digits, we will randomly
assign each student a two digit number from 01 to 95.
result in two winners from the AP Stats class? We’ll label the students in the AP Statistics class from
• P (X=2), where x is the number of winners from 01 to 28, and the remaining students from 29 to 95.
AP Stats (Numbers from 96 to 00 will be skipped.) Starting at
the randomly selected row 139 and moving left to
right across the row, we’ll look at pairs of digits until
we come across two different values from 01 to 95.
These two values will represent the two students with
these labels will win the prime parking spaces. We will
record whether both winners are members of the AP
Statistics class (Yes or no). We will conduct the
simulation 18 times.
3
Required Elements: DO:

– Plan must include: Students Labels Reading across row 139 in Table
• What tool? D, look at pairs of digits until you
AP Statistics Class 01-28 see two different labels from 01-
– Table of Random of Digits, Calculator Random Number
Generator (RandInt), etc. Other 29-95 95. Record whether or not both
winners are members of the AP
• What values are you assigning? Skip numbers from 96-00 Statistics Class.
– 01 to 95
• How many values are you picking each time? 55 | 58 89 | 94 04 | 70 70 | 84 10|98|43 56 | 35 69 | 34 48 | 39 45 | 17
– 2 values X|X X|X ✓|X X|X ✓|Sk|X X|X X|X X|X X|✓
• How many times are you conducting the simulation? No No No No No No No No No
– 18 times 19 | 12 97|51|32 58 | 13 04 | 84 51 | 44 72 | 32 18 | 19 40|00|36 00|24|28
• What about repeat digits or ignored digits? ✓|✓ Sk|X|X X|✓ ✓|X X|X X|X ✓|✓ X|Sk|X Sk|✓|✓
– Ignore repeat digits within a single draw Yes No No No No No Yes No Yes
• What are you recording?

– Yes for both AP Stats.
CONCLUDE: NASCAR
Based on 18 repetitions of our simulation, both In an attempt to increase sales, a breakfast cereal company
decides to offer a NASCAR promotion. Each box of cereal will
winners came from the AP Statistics class 3 contain a collectible card featuring one of these NASCAR
times, so the probability is estimated as drivers: Jeff Gordon, Dale Earnhardt, Jr., Tony Stewart, Danica
16.67%. Therefore is definitely possible for two Patrick, or Jimmie Johnson. The company says that each of the
5 cards is equally likely to appear in any box of cereal. A
AP Stats students to be selected in a “fair”
NASCAR fan decides to keep buying boxes of the cereal until she
drawing. has all 5 drivers’ cards. She is surprised when it takes her 23
boxes to get the full set of cards. Should she be surprised?
Design and carry out a simulation to help answer this question.
4
STATE: PLAN:
What is the probability of needing to buy 23 or Using the calculator's random number generator
more cereal boxes to obtain one card from each (RandInt) we are going to simulate 50 trials. We
driver? will assign each driver a unique number 1 through
5. We will record how many trials it takes to get all
five values (drivers). We will record the total
number of digits required each time.
Driver Label
Jeff Gordon 1
Dale Earnhardt, Jr. 2
Tony Stewart 3
Danica Patrick 4
Jimmie Johnson 5
DO: CONCLUDE:
Dotplot of 50 Trials We never had to buy more than 22 boxes to get
the full set of cards in 50 repetitions of our
simulation. Our estimate of the probability that it
takes 23 or more boxes to get a full set of driver is
roughly 0. Therefore, she should be surprised that
it took 23 cereal box purchases to find all 5 driver
cards.
5
Section 5.2
5.2: Probability Rules Probability Rules
ü DESCRIBE chance behavior with a probability model

ü DEFINE and APPLY basic rules of probability
ü DETERMINE probabilities from two-way tables
ü CONSTRUCT Venn diagrams and DETERMINE probabilities
Basic Rules of Probability Probability Models

The sample space S of a chance
process is the set of all possible
outcomes.
A probability model is a description of
some chance process that consists of
two parts: a sample space S and a
probability for each outcome.
Example of Coin Toss:
Sample Space: Either heads or tails.
Probability: Heads (0.5) and Tails (0.5)
6
Probability Models Probability Models

Venn Diagram, Tree Diagram, List, Chart, etc.
Probability models allow us to find the probability of any
collection of outcomes.
An event is any collection of outcomes from some chance
process. That is, an event is a subset of the sample space.
Events are usually designated by capital letters, like A, B, C,
and so on.
Specific “event” examples:

- Flipping 3 heads in a row
- Rolling two dice that sum to 5
Sample Space: Rolling Two Dice

The probability model for the chance process of rolling two
Probability Models
fair, six-sided dice – one that’s red and one that’s green. Event:
Rolling a sum of “5” with 2 dice or P(A)= sum of 5
Event Space:
There are 4 different combination of dice rolls that sum
to 5.
Solution:
Since the dice are fair, each Since each outcome has probability 1/36:
Sample
outcome is equally likely.
Space
Each outcome has P(A) = 4/36 or 1/9.
36 Outcomes
probability 1/36.
7
What type of Pizza do you

like?
Meat &
- Meat
Veggies
-
-
Veggies
Veggies and meat
Meat Veggies
- Neither (cheese)
***There are NO other choices at Mrs.

Daniel's pizzeria***
Neither (Cheese)
What is the probability that a randomly

selected student…
- Likes meat
- Likes veggies
- Likes veggies and meat
- Likes neither (cheese)
- Like either veggies or meat
8
Complement Mutually Exclusive

• The event that did not occur….not A • Two events that cannot occur at the same
time. There are no common outcomes.
• A= airplane takes off on time
• Ac= airplane does not take off on time • Student is EITHER a Junior or Senior
Intersection Calculating Probabilities

• Probability of both events occurring. § Complement of A
• For example: A = likes salad, B = likes meat, § Complement = 1 – P(A)
therefore P(A and B) = likes both salad and § Mutually Inclusive (A or B)
meat § P(A or B) = P(A) + P(B) – P(A and B)
§ Intersection (A and B)
§ P(A and B)
9
2014 AP Statistics Exam Scores Probabilities: 2014 AP Statistics Exam Scores Probabilities:
Score 1 2 3 4 5 Score 1 2 3 4 5
Probability 0.223 0.183 0.235 0.224 0.125 Probability 0.223 0.193

0.183 0.235 0.224 0.125
(a) Is this a legitimate probability model? Justify.

(a) Is this a legitimate probability model? Justify. Each probability is between 0 and 1 and the sum
of the probabilities : 0.223 + 0.193 + 0.235 + 0.224
+ 0.125 = 1.
(b) Find the probability that the chosen student
(b) Find the probability that the chosen student
scored 3 or better. scored 3 or better.
The probability of scoring a 3 or better: 0.235 +
0.224 + 0.125 = 0.584
Online-learning courses are rapidly gaining popularity Distance-learning courses are rapidly gaining popularity
among college students. Randomly select an among college students. Randomly select an
undergraduate student who is taking online-learning undergraduate student who is taking distance-learning
courses for credit and record the student’s age. Here is the courses for credit and record the student’s age. Here is the
probability model: probability model:
Age group (yr): 18 to 23 24 to 29 30 to 39 40 or over Age group (yr): 18 to 23 24 to 29 30 to 39 40 or over
Probability: 0.57 0.17 0.14 0.12 Probability: 0.57 0.17 0.14 0.12
(a) Is this a legitimate probability model? Justify. (a)Is this a legitimate probability model? Justify.
Each probability is between 0 and 1 and
0.57 + 0.17 + 0.14 + 0.12 = 1
(b)Find the probability that the chosen student is not in (b)Find the probability that the chosen student is not in
the traditional college age group (18 to 23 years). the traditional college age group (18 to 23 years).
P(not 18 to 23 years) = 1 – P(18 to 23 years)
= 1 – 0.57 = 0.43
10
What is the relationship between educational achievement What is the relationship between educational achievement
and home ownership? A random sample of 500 people and and home ownership? A random sample of 500 people and
each member of the sample was identified as a high school each member of the sample was identified as a high school
graduate (or not) and as a home owner (or not). The two-way graduate (or not) and as a home owner (or not). The two-way
table displays the data. table displays the data.
High Not a High Not a
School High School Total School High School Total
Graduate Graduate Graduate Graduate
Homeowner 221 119 340 Homeowner 221 119 340
Not a Homeowner 89 71 160 Not a Homeowner 89 71 160
Total 310 190 500 Total 310 190 500
What is the probability that a randomly selected person… What is the probability that a randomly selected person…
• (a) is a high school graduate (a) is a high school graduate = 310/500
• (b) is a high school graduate and owns a home (b) is a high school graduate and owns a home = 221/500
• (c) is a high school graduate or owns a home (c) is a high school graduate or owns a home = 310 + 119 =
429/500
5.3: Conditional Probability After this section, you should be able to…
and Independence ü DEFINE conditional probability

ü COMPUTE conditional probabilities
ü DESCRIBE chance behavior with a tree diagram
ü DEFINE independent events
ü DETERMINE whether two events are independent
ü APPLY the general multiplication rule to solve probability
questions
11
Basic Probability Basic Probability

Assume a spinner has 8 equal sized sections; Assume a spinner has 8 equal sized sections;
each section is numbered a unique number each section is numbered a unique number
from 1 to 8. from 1 to 8.
A. What is the probability of getting an even A. What is the probability of getting an even
number? number? 4/8 or 1/2
B. What is the probability of getting a prime B. What is the probability of getting a prime
number? number? 5/8
C. What is the probability of getting a multiple C. What is the probability of getting a multiple
of 3? of 3? 2/8 or 1/4
Mixed Probability Mixed Probability

Assume a spinner has 8 equal sized sections; Assume a spinner has 8 equal sized sections;
each section is numbered a unique number each section is numbered a unique number
from 1 to 8. from 1 to 8.
A. What is the probability of getting 2 even A. What is the probability of getting 2 even
spins in a row? spins in a row? 1/4
B. What is the probability of getting a prime B. What is the probability of getting a prime
number or an odd number? number or an odd number? 5/8
C. What is the probability of getting a multiple C. What is the probability of getting a multiple
of 3 or an even spin? of 3 or an even spin? 5/8
12
What is Conditional Probability? Example: Grade Distributions

E: the grade comes from an EPS course, and
L: the grade is lower than a B.
When we are trying to find the probability that one event
will happen under the condition that some other event is
already known to have occurred, we are trying to determine Total
a conditional probability. 6300
1600
The probability that one event happens given that another 2100
event is already known to have happened is called a Total 3392 2952 3656 10000
conditional probability. Suppose we know that event A has
Find P(L)
happened. Then the probability that event B happens given P(L) = 3656 / 10000 = 0.3656
that event A has happened is denoted by P(B | A).
Find P(E | L)
P(E | L) = 800 / 3656 = 0.2188
Read | as “given that”
or “under the Find P(L | E)
condition that” P(L| E) = 800 / 1600 = 0.5000
13
Who Reads the Newspaper? Who Reads the Newspaper?

Residents of a large apartment complex can be classified Residents of a large apartment complex can be classified
based on the events A: reads USA Today and B: reads the based on the events A: reads USA Today and B: reads the
New York Times. What is the probability that a randomly New York Times. What is the probability that a randomly
selected resident who reads USA Today also reads the New selected resident who reads USA Today also reads the
York Times? New York Times?
0.05
P(B | A) = = 0.125
0.40
There is a 12.5% chance that a randomly selected resident
who reads USA Today also reads the New York Times.
Conditional Probability and Conditional Probability and Independence

Independence P(A | B) = P(A)
OR
When knowledge that one event has happened does not P(B | A) = P(B).
change the likelihood that another event will happen, we say
the two events are independent. Are the events “male” and
“left-handed” independent?
Two events A and B are independent if the occurrence of A: left-handed
one event has no effect on the chance that the other event B: male
will happen. In other words, events A and B are
independent if:
P(A | B) = P(A)
OR
P(B | A) = P(B).
14
Conditional Probability and Independence Are these events independent?

Earns A in AP Earns A in AP Total
P(A | B) = P(A) Stats Calc
OR
P(B | A) = P(B). Junior 5 7 12
Senior 12 9 21
Are the events “male” and “left-
handed” independent?. Total 17 16 33
A: left-handed 1. Junior and AP Calc?

B: male
P(left-handed | male) = 3/23 = 0.13

2. Senior and AP Stats?
P(left-handed) = 7/50 = 0.14
General Multiplication Rule

Earns A in AP Earns A in AP Total
Stats Calc The probability that events A and B both occur can be found
using the general multiplication rule
Junior 5 7 12
P(A ∩ B) = P(A) • P(B | A)
Senior 12 9 21
where P(B | A) is the conditional probability that event B
Total 17 16 33
occurs given that event A has already occurred.
15
Tree Diagrams Tree Diagrams

Tree Diagrams are best for events that follow each other,
events that happen multiple times or events that are Consider flipping a
logically related (example: graduate high school first, then coin twice.
attend college OR having cancer, then testing positive).
What is the
probability of getting
two heads?
Sample Space:
HH HT TH TT
So, P(two heads) = P(HH) = 1/4
Example: Teens with Online Profiles Example: Teens with Online Profiles
The Pew Internet and American Life Project finds that 93% of The Pew Internet and American Life Project finds that 93% of
teenagers (ages 12 to 17) use the Internet, and that 55% of teenagers (ages 12 to 17) use the Internet, and that 55% of
online teens have a Facebook profile. online teens have a Facebook profile.
What percent of teens are online and have a Facebook profile? What percent of teens are online and have a Facebook profile?
P(online) = 0.93
P(profile | online) = 0.55
P(online and have profile) = P(online) × P(profile | online)
= (0.93)(0.55)
= 0.5115
51.15% of teens are online

and have posted a profile.
16
Consecutive Probability
Assume a spinner has 8 equal sized sections;
each section is numbered a unique number
from 1 to 8. You spin the spinner three times.
A. What is the probability of getting at least two

even spins?
B. What is the probability of getting a prime
number exactly twice?
C. What is the probability of getting a multiple
of 3 or an even spin only once?
Internet & YouTube Usage

About 27% of adult Internet users are 18 to 29
years old, another 45% are 30 to 49 years old,
and the remaining 28% are 50 and over.
The Pew Internet and American Life Project finds
that 70% of Internet users aged 18 to 29 have
visited a video-sharing site, along with 51% of
those aged 30 to 49 and 26% of those 50 or
older.
Make a Tree Diagram of the probabilities.
Questions on next slide.
17
B. What proportion of adults are 18 to 29 year old B. What proportion of adults are 18 to 29 year old
Internet users that visit video-sharing sites? Internet users that visit video-sharing sites?
.27 x .7 = .189
C. What proportion of adults are 30 to 49 year old
Internet users that visit video-sharing sites? C. What proportion of adults are 30 to 49 year old
Internet users that visit video-sharing sites?
D. What proportion of adults are 50 and over year old .45 x .51 = .2295
Internet users that visit video-sharing sites?
D. What proportion of adults are 50 and over year old
E. What proportion of all adult Internet users visit video- Internet users that visit video-sharing sites?
sharing sites? Do most Internet users visit YouTube .28 x 26 = .0728
and/or similar sites? Justify your answer.
E. P(video yes ∩ 18
to 29) = 0.27 • P(video yes) = 0.1890 + 0.2295 + 0.0728 = 0.4913
0.7
=0.1890
49.13% of all adult Americans that use the Internet watch
videos online. While 49.13% represents a large proportion of
the population, it is not a majority, so it is not fair to say “most”
P(video yes ∩ 30
adult American Internet users visit video-sharing sites.
to 49) = 0.45 •
0.51
=0.2295
P(video yes ∩ 50
+) = 0.28 • 0.26
=0.0728
18
Independence: A Special Multiplication Rule

When events A and B are independent, we can simplify the
general multiplication rule since P(B| A) = P(B).
Multiplication rule for independent events
If A and B are independent events, then the probability that A
and B both occur is
P(A ∩ B) = P(A) • P(B)
Special Probability Rules Example:

Following the Space Shuttle Challenger disaster, it was determined that the failure
of O-ring joints in the shuttle’s booster rockets was to blame. Under cold
conditions, it was estimated that the probability that an individual O-ring joint would
function properly was 0.977. Assuming O-ring joints succeed or fail independently,
what is the probability all six would function properly?
P(joint1 OK and joint 2 OK and joint 3 OK and joint 4 OK and joint 5 OK and joint 6 OK)
=P(joint 1 OK) • P(joint 2 OK) • … • P(joint 6 OK)
=(0.977)(0.977)(0.977)(0.977)(0.977)(0.977) = 0.87
19
6.1: Section 6.1

Discrete & Continuous Random Variables
Discrete and Continuous After this section, you should be able to…
Random Variables ü APPLY the concept of discrete random variables to a
variety of statistical settings
ü CALCULATE and INTERPRET the mean (expected value) of
a discrete random variable
ü CALCULATE and INTERPRET the standard deviation (and
variance) of a discrete random variable
ü DESCRIBE continuous random variables
Random Discrete Random Variables

Variables • A discrete random variable is one which may take on
only a countable number of distinct values such as 0,
1, 2, 3, 4,....
• A random variable, usually • Discrete random variables are usually (but not
written as X, is a variable
necessarily) counts.
whose possible values are
numerical outcomes of a • Examples:
random phenomenon. • number of children in a family
• There are two types of • the Friday night attendance at a cinema
random variables, discrete • the number of patients a doctor sees in one day
random variables and
• the number of defective light bulbs in a box of ten
continuous random
variables. • the number of “heads” flipped in 3 trials
1
Probability Distribution Discrete Random Variables

The probability distribution of a discrete random variable is A discrete random variable X takes a fixed set of possible
a list of probabilities associated with each of its possible values with gaps between. The probability distribution of
values a discrete random variable X lists the values xi and their
probabilities pi:
Consider tossing a fair coin 3 times. Value: x1 x2 x3 …
Define X = the number of heads obtained Probability: p1 p2 p3 …
The probabilities pi must satisfy two requirements:
X = 0: TTT
X = 1: HTT THT TTH 1. Every probability pi is a number between 0 and 1.
X = 2: HHT HTH THH 2. The sum of the probabilities is 1.
X = 3: HHH
To find the probability of any event, add the probabilities pi
Value 0 1 2 3 of the particular values xi that make up the event.
Probability 1/8 3/8 3/8 1/8
Example: Babies’ Health at Birth

Describing the (Probability) Distribution Background details are on page 343.
When analyzing discrete random variables, we’ll follow the (a)Show that the probability distribution for X is legitimate.
same strategy we used with quantitative data – describe the (b)Make a histogram of the probability distribution. Describe
shape, center (mean), and spread (standard deviation), and what you see.
identify any outliers.
(c)Apgar scores of 7 or higher indicate a healthy baby. What is
P(X ≥ 7)?
Value: 0 1 2 3 4 5 6 7 8 9 10
Probability: 0.001 0.006 0.007 0.008 0.012 0.020 0.038 0.099 0.319 0.437 0.053
2
Example: Babies’ Health at Birth Example: Babies’ Health at Birth

Background details are on page 343. b. Make a histogram of the probability distribution. Describe what you see.
(a)Show that the probability distribution for X is legitimate. c. Apgar scores of 7 or higher indicate a healthy baby. What is P(X ≥ 7)?
(b)Make a histogram of the probability distribution. Describe
Value: 0 1 2 3 4 5 6 7 8 9 10
the distribution.
Probability: 0.001 0.006 0.007 0.008 0.012 0.020 0.038 0.099 0.319 0.437 0.053
(c)Apgar scores of 7 or higher indicate a healthy baby. What is
P(X ≥ 7)?
(c) P(X ≥ 7) = .908
Value: 0 1 2 3 4 5 6 7 8 9 10
We’d have a 91 % chance
Probability: 0.001 0.006 0.007 0.008 0.012 0.020 0.038 0.099 0.319 0.437 0.053 of randomly choosing a
healthy baby.
(a) All probabilities are between 0 and 1 and the probabilities
sum to 1. This is a legitimate probability distribution. (b) The left-skewed shape of the distribution suggests a randomly
selected newborn will have an Apgar score at the high end of the scale.
While the range is from 0 to 10, there is a VERY small chance of getting
a baby with a score of 5 or lower. There are no obvious outliers. The
center of the distribution is approximately 8.
Example: Apgar Scores – What’s Typical?

Mean of a Discrete Random Variable Consider the random variable X = Apgar Score
The mean of any discrete random variable is an average of Compute the mean of the random variable X and interpret it in context.
the possible outcomes, with each outcome weighted by its
Value: 0 1 2 3 4 5 6 7 8 9 10
probability.
Probability: 0.001 0.006 0.007 0.008 0.012 0.020 0.038 0.099 0.319 0.437 0.053
Suppose that X is a discrete random variable whose

probability distribution is µx = E(X) = å x i pi
Value: x1 x2 x3 …
Probability: p1 p2 p3 …
To find the mean (expected value) of X, multiply each
possible value by its probability, then add all the products:
µ x = E(X) = x1 p1 + x 2 p2 + x 3 p3 + ...
= å x i pi
3
Example: Apgar Scores – What’s Typical?

Consider the random variable X = Apgar Score Analyzing Discrete Random
Compute the mean of the random variable X and interpret it in context.
Variables on the Calculator
Value: 0 1 2 3 4 5 6 7 8 9 10
Probability: 0.001 0.006 0.007 0.008 0.012 0.020 0.038 0.099 0.319 0.437 0.053
1. Using one-variable statistics to calculate:
µx = E(X) = å x i pi 2. Enter “ascre” for X1 and “freqas” for
= (0)(0.001) + (1)(0.006) + (2)(0.007) + ...+ (10)(0.053) frequency list.
= 8.128
The mean Apgar score of a randomly selected newborn is 8.128.
This is the long-term average Agar score of many, many
randomly chosen babies.
Note: The expected value does not need to be a possible value of

X or an integer! It is a long-term average over many repetitions.
Standard Deviation of a Discrete

Analyzing Discrete Random Random Variable
Variables on the Calculator The definition of the variance of a random variable is similar to
the definition of the variance for a set of quantitative data. To
get the standard deviation of a random variable, take the square
root of the variance.
Suppose that X is a discrete random variable whose

probability distribution is
Value: x1 x2 x3 …
Probability: p1 p2 p3 …
and that µX is the mean of X. The variance of X is
Var(X) = s X2 = (x1 -µX ) 2 p1 + (x 2 -µ X )2 p2 + (x 3 -µ X )2 p3 + ...
= å(x i -µ X )2 pi
4
Example: Apgar Scores – How Variable Are They?

Consider the random variable X = Apgar Score Continuous Random Variable
Compute the standard deviation of the random variable X
and interpret it in context. • A continuous random variable is one which
Value: 0 1 2 3 4 5 6 7 8 9 10 takes an infinite number of possible values.
Probability: 0.001 0.006 0.007 0.008 0.012 0.020 0.038 0.099 0.319 0.437 0.053
• Continuous random variables are usually
s X2 = å(x i -µX )2 pi measurements.
= (0 - 8.128)2 (0.001) + (1- 8.128)2 (0.006) + ...+ (10 - 8.128)2 (0.053) • Examples:
= 2.066 Variance
– height
s X = 2.066 =1.437 – weight
The standard deviation of X is 1.437. On average, a randomly
– the amount of sugar in an orange
selected baby’s Apgar score will differ from the mean 8.128 by – the time required to run a mile.
about 1.4 units.
Continuous Random Variables Continuous Random Variables

A continuous random variable X takes on all values in an • A continuous random variable is
interval of numbers. The probability distribution of X is not defined at specific values.
described by a density curve. The probability of any • Instead, it is defined over an
event is the area under the density curve and above the interval of value; however, you
values of X that make up the event. can calculate the probability of a
range of values.
• It is very similar to z-scores and
normal distribution calculations.
5
Example: Young Women’s Heights Example: Young Women’s Heights

The height of young women can be defined as a continuous random variable (Y) with a
The height of young women can be defined as a continuous probability distribution is N(64, 2.7).
random variable (Y) with a probability distribution is N(64, A. What is the probability that a randomly chosen young woman has height between
2.7). 68 and 70 inches?
A. What is the probability that a randomly chosen young
woman has height between 68 and 70 inches? P(68 ≤ Y ≤ 70) = ???
P(68 ≤ Y ≤ 70) = ??? 68 - 64 70 - 64

z= z=
2.7 2.7
= 1.48 = 2.22
P(1.48 ≤ Z ≤ 2.22) = P(Z ≤ 2.22) – P(Z ≤ 1.48)
= 0.9868 – 0.9306
= 0.0562
There is about a 5.6% chance that a randomly chosen young

woman has a height between 68 and 70 inches.
Example: Young Women’s Heights Example: Young Women’s Heights

The height of young women can be defined as a continuous random variable (Y) with a
The height of young women can be defined as a continuous probability distribution is N(64, 2.7).
random variable (Y) with a probability distribution is N(64, B. At 70 inches tall, is Mrs. Daniel unusually tall?
2.7).
B. At 70 inches tall, is Mrs. Daniel unusually tall? P(Y ≤ 70) = ???
70 - 64
z= P value: 0.9868
2.7
= 2.22
Yes, Mrs. Daniel is unusually tall because 98.68% of the

population is shorter than her.
6
üDESCRIBE the effect of performing a linear

6.2: Transforming and transformation on a random variable
Combining Random üCOMBINE random variables and CALCULATE the
resulting mean and standard deviation
Variables üCALCULATE and INTERPRET probabilities
involving combinations of Normal random
variables
Linear Review: Linear Transformations

In Chapter 2, we studied the effects of linear transformations
Transformations
on the shape, center, and spread of a distribution of data.
Remember:
1. Adding (or subtracting) a constant, a, to each observation:
on Random
• Adds a to measures of center and location.
• Does not change the shape or measures of spread.
2. Multiplying (or dividing) each observation by a constant,
Variables b:
• Multiplies (divides) measures of center and location by
b.
• Multiplies (divides) measures of spread by |b|.
• Does not change the shape of the distribution.
7
Linear Transformations on Linear Transformations

Pete’s Jeep Tours offers a popular half-day trip in a tourist area.
Random Variables There must be at least 2 passengers for the trip to run, and the
vehicle will hold up to 6 passengers. Define X as the number of
passengers on a randomly selected day.
Multiplying (or dividing) each value of a random variable
by a number b: Passengers xi 2 3 4 5 6
Probability pi 0.15 0.25 0.35 0.20 0.05
• Multiplies (divides) measures of center and location
(mean, median, quartiles, percentiles) by b.
• Multiplies (divides) measures of spread (range, IQR, The mean of X is 3.75 and
standard deviation) by |b|. the standard deviation is
• Does not change the shape of the distribution. 1.090.
Note: Multiplying a random variable by a constant b

multiplies the variance by b2.
Linear Transformations Compare the shape, center and spread of each

Pete charges $150 per passenger. The random distribution.
variable C describes the amount Pete collects on a
randomly selected day.
Collected ci 300 450 600 750 900
Probability pi 0.15 0.25 0.35 0.20 0.05
The mean of C is
$562.50 and the
standard deviation is
$163.50.
8
Linear Transformations on Linear Transformations

Consider Pete’s Jeep Tours again. We defined C as the amount of money Pete
Random Variables collects on a randomly selected day.
Collected ci 300 450 600 750 900
Adding the same number a (which could be negative) to Probability pi 0.15 0.25 0.35 0.20 0.05
each value of a random variable: The mean of C is $562.50 and the

standard deviation is $163.50.
• Adds a to measures of center and location (mean,
median, quartiles, percentiles). It costs Pete $100 per trip to buy permits, gas, and a ferry pass. The random
variable V describes the profit Pete makes on a randomly selected day.
• Does not change measures of spread (range, IQR,
Profit vi 200 350 500 650 800
standard deviation).
Probability pi 0.15 0.25 0.35 0.20 0.05
• Does not change the shape of the distribution.
The mean of V is $462.50 and the
standard deviation is $163.50.
Compare the shape, center, and spread of the two probability distributions.
Bottom Line:
Whether we are dealing with data or random
Combining
Random
variables, the effects of a linear transformation
are the same!!!
Variables
9
Combining Random Variables Combining Random Variables

Let D = the number of passengers on a randomly selected
Delta flight to Atlanta
Before we can combine random variables, a determination
about the independence of each variable from the other Let A = the number of passengers on a randomly selected trip
must be made. Probability models often assume American Airlines flight to Atlanta
independence when the random variables describe Define T = X + Y. Calculate the mean and standard deviation
outcomes that appear unrelated to each other. of T.
Passengers xi 75 76 77 78 79
Probability pi 0.15 0.25 0.35 0.20 0.05
You should always ask yourself whether the assumption of

independence seems reasonable.
Passengers yi 75 76 77 78
Probability pi 0.3 0.4 0.2 0.1
Combining Random Variables Combining Random

Let D = the number of passengers on a randomly selected
Delta flight to Atlanta
Variables: Mean
How many total passengers fly to Atlanta on a randomly
Let A = the number of passengers on a randomly selected trip selected day?
American Airlines flight to Atlanta
Delta: µD = 76.75
Define T = X + Y. Calculate the mean and standard deviation
American: µA = 76.10
of T.
Total: 76.75+ 76.10 = 142.85 passengers to Atlanta daily
Passengers xi 75 76 77 78 79
Probability pi 0.15 0.25 0.35 0.20 0.05
For any two random variables X and Y, if T = X + Y, then the
Mean µD = 76.75 Standard Deviation σD= 1.0897 expected value of T is
Passengers yi 75 76 77 78
E(T) = µT = µX + µY
Probability pi 0.3 0.4 0.2 0.1 In general, the mean of the sum of several random
variables is the sum of their means.
Mean µA = 76.1 Standard Deviation σA = 0.943
10
Combining Random Variables: Combining Random Variables:

Variance Variance
How much variability is there in the total number of Delta = (1.090)2
passengers who fly to Atlanta on a randomly selected day? American = (0.943)2
(Hint: find the combined variance) Total Variance = (1.090)2 + (0.943)2 = 2.077
Delta: Mean µD = 76.75 Standard Deviation σD= 1.0897

For any two independent random variables X and Y, if T = X +
American: Mean µ = 76.1 Standard Deviation σ = 0.943 Y, then the variance of T is
A A
sT2 = sX2 + sY2

REMEMBER: Standard Deviations do not add!!!
In general, the variance of the sum of several independent
random variables is the sum of their variances.
Subtracting Random Variables: Subtracting Random Variables:

Mean Variance
Mean of the Difference of Random Variables
For any two random variables X and Y, if D = X - Y, then the Variance of the Difference of Random Variables
expected value of D is For any two independent random variables X and Y, if D = X -
Y, then the variance of D is
E(D) = µD = µX - µY
In general, the mean of the difference of several random sD2 = sX2 + sY2
variables is the difference of their means. The order of In general, the variance of the difference of two independent
Variance of the Difference of Random Variables
subtraction is important! random variables is the sum of their variances.
**This was an FRQ on the 2013 exam**
11
Combining Normal Random Combining Normal Random Variables:

Calculating Probabilities
Variables:
Mrs. Daniel likes between 8.5 and 9 grams of sugar in her iced
Calculating Probabilities coffee. Suppose the amount of sugar in a randomly selected
packet follows a Normal distribution with mean 2.17 g and
If a random variable is Normally distributed, we can use standard deviation 0.08 g. If Mrs. Daniel selects 4 packets at
its mean and standard deviation to compute random, what is the probability her iced coffee will taste right?
probabilities.
Important Fact: Any sum or difference of independent

Normal random variables is also Normally distributed.
Combining Normal Random Variables: Combining Normal Random Variables:

Calculating Probabilities Calculating Probabilities
DO:
STATE & PLAN: Let X = the amount of sugar in a randomly
1. Calculate combined mean
selected packet. Then, T = X1 + X2 + X3 + X4. We want to
find P(8.5 ≤ T ≤ 9). µT = µX1 + µX2 + µX3 + µX4 = 2.17 + 2.17 + 2.17 +2.17 = 8.68
2. Calculate combined variance

sT2 = sX2 + sX2 + sX2 + sX2 = (0.08)2 + (0.08)2 + (0.08)2 + (0.08)2 = 0.0256
1 2 3 4
3. Calculate combined standard deviation.

sT = 0.0256 = 0.16
12

4. Calculate z-scores.
8.5 - 8.68 9 - 8.68
z= = -1.13 z= = 2.00
0.16
and
0.16 YES, you may use your calculator!
5. Find p-values
p-values (z = -1.13) = 0.1292 and (z = 2) 0.9772
Just remember to recalculate the combined
6. Final Calculations mean and standard deviation, before using
0.9772 - 0.1292 = 0.8480 the calculator!!!!
OR normcdf (8.5, 9, 8.68, 0.16) =
CONCLUDE:
There is an 84.8% percent chance that Mrs. Daniel’s iced coffee will taste right.

The diameter C of a randomly selected large drink cup at a fast-
STATE & PLAN: We’ll define the random variable D = L − C to
food restaurant follows a Normal distribution with a mean of 3.96
inches and a standard deviation of 0.01 inches. The diameter L of represent the difference between the lid’s diameter and the cup’s
diameter. Our goal is to find P(0.00 < D ≤ 0.06).
a randomly selected large lid at this restaurant follows a Normal
distribution with mean 3.98 inches and standard deviation 0.02
inches. DO:
1. Calculate combined mean.
For a lid to fit on a cup, the value of L has to be bigger than the
value of C, but not by more than 0.06 inches. μD = μL − μC = 3.98−3.96 = 0.02
2. Calculate combined variance
What’s the probability that a randomly selected large lid will fit (0.02)2 + (0.01)2 = 0.0005
on a randomly chosen large drink cup? 3. Calculate combined standard deviation.
0.0005 = 0.0224
13

DO, cont.:
CONCLUDE:
4. Calculate z-scores:
% '%.%( %.%*'%.%( We predict that the lids will fit properly 77.66% of the time. This
z = = -0.89 and = 1.79 means the lids will not fit properly more than 22% of the time.
%.%(() %.%(()
That is annoying!
5. Find p-values:
pvalues (z = -0.89) = 0.1867 and (z = 1.79) = 0.9633
6. Final calculations:
0.9633 − 0.1867 = 0.7766.
OR normcdf(0, 0.06, 0.02, 0.0224)=

Mrs. Daniel and Mrs. Cooper bowl every Tuesday night. Over
the past few years, Mrs. Daniel’s scores have been
approximately Normally distributed with a mean of 212 and a
standard deviation of 31. During the same period, Mrs.
Cooper’s scores have also been approximately Normally
distributed with a mean of 230 and a standard deviation of 40.
Assuming their scores are independent, what is the probability
that Mrs. Daniel scores higher than Mrs. Cooper on a
randomly-selected Tuesday night?
14
Combining Normal Random Variables:

Calculating Probabilities 6.3: Binomial and
Geometric Random
Variables
CONCLUDE:
There is a 35.94% chance that Mrs. Daniel will score higher than
Mrs. Cooper on any given night.
ü DETERMINE whether the conditions for a binomial setting M & M Lab!

are met
ü COMPUTE and INTERPRET probabilities involving binomial Binomial v. Geometric
random variables
ü CALCULATE the mean and standard deviation of a
Settings
binomial random variable and INTERPRET these values in
context
ü CALCULATE probabilities involving geometric random
variables
15
Binomial Settings Binomial Random Variable

Consider tossing a coin n times. Each toss gives either heads
A binomial setting arises when we perform several or tails. Knowing the outcome of one toss does not
independent trials of the same chance process and record change the probability of an outcome on any other toss.
the number of times that a particular outcome occurs. The If we define heads as a success, then p is the probability
four conditions for a binomial setting are of a head and is 0.5 on any toss.
Binary? The possible outcomes of each trial can be The number of heads in n tosses is a binomial random
B classified as “success” or “failure.”
variable X. The probability distribution of X is called a
Independent? Trials must be independent; that is, binomial distribution.
I knowing the result of one trial must not have any effect
on the result of any other trial. Count the number of successes in a
Number? The number of trials n of the chance process predetermined number of trials!
N must be fixed in advance.
Success? On each trial, the probability p of success must
S be the same.
Binomial Distribution: Mean and Find the mean and standard deviation of
Standard Deviation X.
If a count X has the binomial distribution with number of trials X is a binomial random variable with parameters n = 21
n and probability of success p, the mean and standard and p = 1/3.
deviation of X are
µ X = np
s X = np(1- p)
Note: These formulas work ONLY for binomial distributions.

They can’t be used for other distributions!
16
Find the mean and standard deviation of Binomial Distribution: Describe

We describe the probability distribution of a binomial random
X. variable just like any other distribution – shape, center, and
spread.
X is a binomial random variable with parameters n = 21 Consider the probability distribution of X = number of
and p = 1/3. children with type O blood in a family with 5 children.
xi 0 1 2 3 4 5
µ X = np s X = np(1- p)
pi 0.2373 0.3955 0.2637 0.0879 0.0147 0.00098
= 21(1/3) = 7
= 21(1/3)(2 /3) = 2.16
Binomial Distribution: Describe

xi 0 1 2 3 4 5
Calculator: Binomial Probability
pi 0.2373 0.3955 0.2637 0.0879 0.0147 0.00098
• MENU, 6: Statistics, 5: Distributions
Shape: The probability distribution of X is skewed to the – D: binompdf Same idea as normpdf
right. It is more likely to have 0, 1, or 2 children with type O – E: binomcdf and normcdf
blood than a larger value.
• Binompdf calculates equal to value
Center: The median number of children with type O blood
– For “PERCISE” numbers
is 1. The mean is 1.25.
Spread: The variance of X is 0.9375 and the standard

deviation is 0.96.
17
Binomial Probabilities Binomial Probabilities

Each child of a particular pair of parents CHECK CONDITIONS:
has probability 0.25 of having type O
blood. Genetics says that children receive Binary: Yes. Type O blood = yes and not type O blood = no.
genes from each of their parents There are only two options.
independently. If these parents have 5
children, the count X of children with Independent: Stated.
type O blood is a binomial random
variable with n = 5 trials and probability p Number: Yes. The number of trials is stated as 5.
= 0.25 of a success on each trial. In this
setting, a child with type O blood is a Success: Yes. The probability of success is the same on each
“success” (S) and a child with another attempt, p = 0.25.
blood type is a “failure” (F).
What’s P(X = 2)?
Binomial Probabilities Inheriting Blood Type

Using your calculator:
Each child of a particular pair of parents has probability 0.25 of
Binompdf, enter the following information: having blood type O. Suppose the parents have 5 children.
Trials: 5
P: .25
X value: 2 (a) Find the probability that exactly 3 of the children have type
Answer: 0.263671875 O blood.
We are using binompdf in this example because we want

the “precise” probability of 2. (b) Should the parents be surprised if more than 3 of their
children have type O blood?
CONCLUDE:
We have already checked the conditions, so just do the
There is a 26.37% chance that the family will have two
calculations.
children with type O blood.
18
Inheriting Blood Type

Each child of a particular pair of parents has probability 0.25 of having blood type O.
Binomial Distributions: Statistical
Suppose the parents have 5 children
(a) Find the probability that exactly 3 of the children have type
Sampling
O blood. The binomial distributions are important in statistics when
we want to make inferences about the proportion p of
Binompdf :(5, .25, 3) = 0.08789 successes in a population.
There is an 8.79% percent chance that the family will have
three children with type O blood.
Sampling Without Replacement Condition
(b) Should the parents be surprised if more than 3 of their When taking an SRS of size n from a population of size N, we
children have type O blood? can use a binomial distribution to model the count of successes
Binomcdf: (5, .25, 4, 5) = 0.015625 in the sample as long as
There is a 1.5% percent chance that more than 3 of the children 1

n£ N
(aka at least 4 children) will have type O blood. This is surprising! 10
Example: CDs Example: CDs

Suppose 10% of CDs have defective copy-protection schemes CHECK CONDITIONS:
that can harm computers. A music distributor inspects an SRS Binary: Yes. Defective or not defective, only two options.
of 10 CDs from a shipment of 10,000. Let X = number of Independent: We can safely assume independence in this
defective CDs. What is P (X = 0)? case because we are sampling less than 10% of the
population.
Number: Yes. The number of trials is stated as 10.
Success: Yes. The probability of success is the same on each
attempt, p = 0.10.
DO & CONCLUDE:
Binompdf (10, .1, 0) = 0.3486784401
There is a 34.87% that there will be no defective CDs in the

sample.
19
Binomial Distributions: Normal Binomial Distributions: Normal

Approximation Approximation
Suppose that X has the binomial distribution with n trials and
As n gets larger, something interesting happens to the success probability p. When n is large, the distribution of X is
shape of a binomial distribution. approximately Normal with mean and standard deviation
µ X = np sX = np(1- p)
As a rule of thumb, we will use the Normal approximation
when n is so large that np ≥ 10 and n(1 – p) ≥ 10. That is, the
expected number of successes and failures are both at least
10. We use the normal approximation more in Chapters 8-10.
Example: Attitudes Toward Shopping Binomial:

Binary: There are only 2 options. Success = agree, Failure =
Sample surveys show that fewer people enjoy shopping than don’t agree
in the past. A survey asked a nationwide random sample of Independent: Because the population of U.S. adults is greater
2500 adults if they agreed or disagreed that “I like buying than 25,000, it is reasonable to assume the sampling without
new clothes, but shopping is often frustrating and time- replacement condition is met; we are sampling less than 10%
consuming.” Suppose that exactly 60% of all adult US of the population.
residents would say “Agree” if asked the same question. Let X Number of Trials: n = 2500 trials of the chance process
= the number in the sample who agree. Estimate the Success: The probability of selecting an adult who agrees is p =
probability that 1520 or more of the sample agree. 0.60
Consider the normal approximation for this setting. Binomcdf(2500, 0.6, 1520, 2500) = 0.213139
20
OR:
Normal: Since np = 2500(0.60) = 1500 and n(1 – p) = 2500(0.40) Geometric Settings
= 1000 are both at least 10, we may use the Normal A geometric setting arises when we perform independent
approximation. trials of the same chance process and record the number of
trials until a particular outcome occurs. The four conditions
Calculate the mean. µ = np = 2500(0.60) = 1500 for a geometric setting are
Binary? The possible outcomes of each trial can be
Calculate standard deviation. B classified as “success” or “failure.”
s = np(1 - p) = 2500(0.60)(0.40) = 24.49 Independent? Trials must be independent; that is,
Use Calculator I knowing the result of one trial must not have any effect
Normalcdf (1520, 2500, 1500, 24.49) = 0.207061 on the result of any other trial.
Trials? The goal is to count the number of trials until the
T first success occurs.
CONCLUDE:
There is a 20.61% that 1520 or more of the people in the S Success? On each trial, the probability p of success must
sample agree. be the same.
Geometric Random Variable Calculator: Geometric Probability

Geometric random variable: the number of trials needed to
get the first success.
• MENU, 6: Statistics, 5: Distributions
– F: Geometpdf Same idea as normpdf
– G: Geometcdf and normcdf
Examples:
• How many M&Ms are drawn until a blue one is selected? • Geometpdf calculates equal to value
• How many students will I draw from a hat until a pick a – For “PERCISE” numbers
senior?
• How many households can a surveyor call until someone • Geometcdf calculates the probability of
answers? getting at least one success within a specific
range of number of trials
21
Example: The Birthday Game Example: The Birthday Game

I am going to think of the day of the week of one of my friend’s
birthdays. If the first guesser gets it right you all will receive 1 CHECK CONDITIONS:
homework question. If the second guesser gets the day right
you will receive 2 homework questions, etc. Before playing Binary: There are only 2 options: Success = correct guess,
the game, my plan was to give you all 10 homework questions. Failure = incorrect guess
Independent: The result of one student’s guess has no
The random variable of interest in this game is Y = the number
effect on the result of any other guess.
of guesses it takes to correctly identify the birth day of one of
Trials: We’re counting the number of guesses up to and
your teacher’s friends. What is the probability the first student
including the first correct guess.
guesses correctly? The second? Third? What is the probability
Success: On each trial, the probability of a correct guess is
one of the first three students will be correct?
1/7, which is the same.
Example: The Birthday Game Geometric Distribution: Mean

DO:
Probability First Student: 1/7 = 0.142857 If Y is a geometric random variable with probability p of
success on each trial, then its mean (expected value) is
Probability Second Student: geometPDF(1/7, 2) = 0.1224 E(Y) = µY = 1/p.
Meaning: Expected number of n trials to achieve first success
Probability Third Student: geometPDF (1/7, 3) = 0.10496 (average)
Example: Suppose you are a 80% free throw shooter. You are
What is the probability one of the first three students will
going to shoot until you make.
be correct? GeometCDF(1/7, 1, 3) = 0.37026
For p = .8, the mean is 1/.8 = 1.25. This means we expect the
CONCLUDE: There is a 37.03% percent change that one of shooter to take 1.25 shots, on average, to make first.
the first three students will guess correctly.
22
Binomial vs. Geometric

The Binomial Setting The Geometric Setting Binomial or Geometric??
1. Each observation falls into 1. Each observation falls into • First defective tire
one of two categories. one of two categories.
• Baskets made until first miss
2. The probability of success 2. The probability of success
is the same for each is the same for each • Questions guessed correctly on SAT Math
observation. observation. • Light blubs purchased until third failure
3. The observations are all 3. The observations are all • Jurors selected for trial until first
independent. independent.
disqualification
4. There is a fixed number n 4. The variable of interest is
of observations. the number of trials • Number of students that interrupt class until
required to obtain the 1st Mrs. Daniel gets mad/mean
success.
FRQ Answers Must Include:

1. Name of distribution
Geometric, Binomial
2. Parameters
Binomial: X (define variable), n & p
Geometric: X (define variable), p
3. Probability Statement Calculating Binomial &
Ex: P (X = 7) or P (X ≥ 3)
Geometric Distributions by
4. Calculation and p-value
Calculator notation is okay, but needs to be Hand
labeled.
5. Solution interpreted in context.
23
Binomial Coefficient Binomial Probability

The binomial coefficient counts the number of different
ways in which k successes can be arranged among n
How to Calculate Number of Arrangements: trials. The binomial probability P(X = k) is this count
The number of ways of arranging k successes among n multiplied by the probability of any one specific
observations is given by the binomial coefficient arrangement of the k successes.
Binomial Probability
æ nö n!
ç ÷= If X has the binomial distribution with n trials and probability p of
è k ø k!(n - k)!
success on each trial, the possible values of X are 0, 1, 2, …, n. If k is
any one of these values, æ ön
P(X = k) = ç ÷ p k (1- p) n-k
è kø
Number of Probability of n-
arrangements k failures
Probability of k
of k successes
successes
Binomial Probabilities (Alternative Solution) Inheriting Blood Type (Alternative Solution)

Each child of a particular pair of parents has probability 0.25 of having blood type O.
Each child of a particular pair of parents has probability 0.25 of
having type O blood. Genetics says that children receive genes from Suppose the parents have 5 children
each of their parents independently. If these parents have 5 children, (a) Find the probability that exactly 3 of the children have type O blood.
the count X of children with type O blood is a binomial random
variable with n = 5 trials and probability p = 0.25 of a success on
each trial. In this setting, a child with type O blood is a “success” (S) Let X = the number of children with type O blood. We know X has a binomial
and a child with another blood type is a “failure” (F).
distribution with n = 5 and p = 0.25.
What’s P(X = 2)?
æ5ö
P(SSFFF) = (0.25)(0.25)(0.75)(0.75)(0.75) = (0.25)2(0.75)3 = 0.02637 P(X = 3) = ç ÷(0.25) 3 (0.75) 2 = 10(0.25) 3 (0.75) 2 = 0.08789
è 3ø
(b) Should the parents be surprised if more than 3 of their children have
However, there are a number of different arrangements in which 2 out of type O blood?
the 5 children have type O blood: To answer this, we need to find P(X > 3).
P(X > 3) = P(X = 4) + P(X = 5)

SSFFF SFSFF SFFSF SFFFS FSSFF æ 5ö æ5ö Since there is only a
FSFSF FSFFS FFSSF FFSFS FFFSS = ç ÷(0.25) 4 (0.75)1 + ç ÷(0.25) 5 (0.75) 0 1.5% chance that more
è 4ø è5ø than 3 children out of 5
Verify that in each arrangement, P(X = 2) = (0.25)2(0.75)3 = 0.02637 would have Type O
= 5(0.25) 4 (0.75)1 + 1(0.25) 5 (0.75) 0 blood, the parents
Therefore, P(X = 2) = 10(0.25)2(0.75)3 = 0.2637
= 0.01465 + 0.00098 = 0.01563 should be surprised!
24
Section 7.1
What Is a Sampling Distribution?
7.1: What is a Sampling
Distribution?!?!
ü DISTINGUISH between a parameter and a statistic
ü DEFINE sampling distribution
ü DISTINGUISH between population distribution,
sampling distribution, and the distribution of sample
data
ü DETERMINE whether a statistic is an unbiased
estimator of a population parameter
ü DESCRIBE the relationship between sample size and the
variability of an estimator
The process of statistical inference involves using information

from a sample to draw conclusions about a wider population.
Different random samples yield different statistics.
We need to be able to describe the sampling distribution of
possible statistic values in order to perform statistical
inference. We can think of a statistic as a random variable
because it takes numerical values that describe the outcomes
of the random sampling process.
Population
Sample Collect data from a
representative Sample...
Make an Inference
about the Population.
1
Parameters and Statistics Symbols: Parameters and Statistics
A parameter is a number that describes some characteristic of Proportions Means Standard

the population. In statistical practice, the value of a parameter Deviation
is usually not known because we cannot examine the entire
population. Statistic s
A statistic is a number that describes some characteristic of a Parameter p µ
sample. The value of a statistic can be computed directly from
the sample data. We use a statistic to estimate an unknown
parameter.
Parameter v. Statistic Parameter v. Statistic

Identify the population, the parameter (of interest), the
sample, and the statistic in each of the following settings.
Population: all 10-year-old boys
Parameter: 75th percentile, or Q3
A pediatrician wants to know the 75th Sample: 50 10-year-old boys included in the
percentile for the distribution of heights of 10- sample
year-old boys so she takes a sample of 50 Statistic: Q3 = 56 inches.
patients and calculates Q3 = 56 inches.
2
Parameter v. Statistic Parameter v. Statistic

Identify the population, the parameter, the sample, and Population: All 12-17 year olds in the US
the statistic in each of the following settings.
Parameter: Proportion with cell phones
A Pew Research Center poll asked 1102 12
to 17-year-olds in the United States if they Sample: 1102 12-17 year olds with cell phones
have a cell phone. Of the respondents, 71%
Statistic: 𝑝̂ = 71
said yes.
Sampling Distribution
The sampling distribution of a statistic is the distribution of
values taken by the statistic in all possible samples of the
same size from the same population.
In practice, it’s difficult to take all possible samples of size n

to obtain the actual sampling distribution of a statistic.
Instead, we can use simulation to imitate the process of
taking many, many samples.
One of the uses of probability theory in statistics is to

obtain sampling distributions without simulation. We’ll get
to the theory later.
3
Population Distributions vs.

Sampling Distributions
There are actually three distinct distributions involved when
we sample repeatedly and measure a variable of interest.
1) The population distribution gives the values of the variable

for all the individuals in the population.
2) The distribution of sample data shows the values of the
variable for all the individuals in the sample.
3) The sampling distribution shows the statistic values from
all the possible samples of the same size from the
population.
Bias & Variability Describing Sampling Distributions:

Center
Bias means that our aim is off and
we consistently miss the bull’s-eye A statistic used to estimate a parameter is an
in the same direction. Our sample unbiased estimator (most accurate) if the mean of
values do not center on the its sampling distribution is equal to the true value of
population value. the parameter being estimated.
High variability means that

repeated shots are widely
scattered on the target. Repeated
samples do not give very similar
results.
4
Describing Sampling Distributions: Describing Sampling Distributions:

Spread Shape
The variability of a statistic is described by the spread of its Sampling distributions can take on many shapes. The same
sampling distribution. This spread is determined primarily statistic can have sampling distributions with different
by the size of the random sample. Larger samples give shapes depending on the population distribution and the
smaller spread. The spread of the sampling distribution sample size.
does not depend on the size of the population, as long as
the population is at least 10 times larger than the sample. Sampling distributions for
different statistics used to
n=100 n=1000 estimate the number of tanks
in German during World War 2.
The blue line represents the
true number of tanks.
Section 7.2
Sample Proportions
7.2: Sample Proportions After this section, you should be able to…
ü FIND the mean and standard deviation of the sampling

distribution of a sample proportion
ü DETERMINE whether or not it is appropriate to use the
Normal approximation to calculate probabilities
involving the sample proportion
ü CALCULATE probabilities involving the sample
proportion
ü EVALUATE a claim about a population proportion using
the sampling distribution of the sample proportion
5
http://www.rossmanchance.com/a The Sampling Distribution of pÙ

pplets/Reeses/ReesesPieces.html What do you notice about the shape, center, and spread of each?
n =100 n =400
Normal Approximation & Sample

Sample Proportion Formulas
Proportions
As the sample size increase, sample proportion approach
count of successes in sample X
pÙ= = the normal distribution; therefore, we can use Normal
calculations.
size of sample n
Before using Normal calculation, check Normal
p (1 - p ) conditions:
s pˆ = – (sample size)(proportion) must be greater than 10.
n – (sample size)(1 – proportion) must be greater than

10.
The sample size MUST be less than 10% of the total – Both must be greater than 10
population.
6
Normal Approximation & Sample Normal Approximation & Sample

Proportions Proportions
In the game of Scrabble, each player starts by (a) Yes. Seven tiles is less than 10% of the population
drawing 7 tiles from a bag of 100 tiles. There are 42 of 100 tiles.
vowels, 56 constants and 2 blank tiles. Cait choses an
SRS of 7 tiles. Let be the proportion of vowels in (b) No. Since the total sample size was 7, both np
her sample. and n(1-p) must be less than 10. The Normal
condition is not satisfied.
a) Is the 10% condition met? Justify your answer.
b) Is the Normal condition met? Justify your answer.
Normal Approximation & Sample We have an SRS of size n = 1500 drawn from a population
Proportions in which the proportion p = 0.35 attend college within 50
miles of home.
A polling organization asks an SRS of 1500 first-year college (0.35)(0.65)
students how far away their home is. Suppose that 35% of
µ pÙ = 0.35 s pÙ = = 0.0123
1500
all first-year students actually attend college within 50
miles of home. Conditions:
Independence: It is reasonable to assume that there are
What is the probability that the random sample of 1500 more than 15,000 college freshmen and therefore the
students will give a result within 2 percentage points of sample represents less than 10% of the population.
this true value? Normality: Additionally, np = 1500(0.35) = 525 and n(1 –
p) = 1500(0.65)=975 are both greater than 10, so it is
reasonable to assume normality.
7
The Harvard College Alcohol Study finds that 67% of college

students support efforts to “crack down on underage drinking.”
The study took a random sample of almost 15,000 students, so
the population proportion who support a crackdown is close to p
= 0.67. The administration of a local college surveys an SRS of
100 students and finds that 62 support a crackdown on underage
drinking. Suppose that the proportion of all students attending
this college who support a crackdown is 67%, the same as the
national proportion.
Normalcdf (0.33, 0.37, 0.35, 0.0123) = 0.896054 What is the probability that the proportion in an SRS of
100 students is as small as or smaller than the result of
CONCLUDE: There is an 89.61% chance that the sample will yield the administration’s sample?
results within 2 percentage points of the true value.
(0.67)(0.33)
µ pˆ = 0.67 s pˆ = = 0.04702 Normalcdf (0, 0.62, 0.67, 0.04702) = 0.143805
100
Be sure to include labels!
Conditions: CONCLUDE: There is an 14.38% chance that the sample will yield
Independence: It is reasonable to assume that there are results at or below 62% given that the true population
more than 1000 college freshmen and therefore the proportions is 67%
sample represents less than 10% of the population.
Normality: Additionally, np = 100(0.67) = 67and n(1 – p) =
100(0.33)= 33 are both greater than 10, so it is reasonable
to assume normality.
8
FYI: Derivation of Formulas pÙ

In Chapter 6, we learned that the mean and standard deviation of a binomial
random variable X are
7.3: Sample Means
µ X = np sX = np(1 - p)
Since pÙ= X /n = (1/n) × X, we are just multiplying the random variable X
by a constant (1/n) to get the random variable pÙ. Therefore,
1
µ pÙ = (np) = p pÙis an unbiased estimator or p
n
1 np(1 - p) p(1 - p)
s pÙ = np(1 - p) = =
n n2 n
As sample size increases, the spread decreases.
Section 7.3 Sample Means

Consider the mean household earnings for samples of
Sample Means size 100. Compare the population distribution on the
After this section, you should be able to… left with the sampling distribution on the right. What do
you notice about the shape, center, and spread of
ü FIND the mean and standard deviation of the sampling each?
distribution of a sample mean
ü CALCULATE probabilities involving a sample mean when
the population distribution is Normal
ü EXPLAIN how the shape of the sampling distribution of
sample means is related to the shape of the population
distribution
ü APPLY the central limit theorem to help find probabilities
involving a sample mean
9
Sample Means Formulas REVIEW: Young Women’s Heights

The height of young women follows a Normal distribution
The mean of the sample distribution:
with mean µ = 64.5 inches and standard deviation σ = 2.5
µx = µ inches.
Find the probability that a randomly selected young
The Standard Deviation of the sampling distribution: woman is taller than 66.5 inches.
s
sx =
n
Notes: The sample size must be less than 10% of the
population. The mean and standard deviation of the
sample mean are true no matter the same of the
population distribution.
REVIEW: Young Women’s Heights Example: Young Women’s Heights

STATE: Let X = the height of a randomly selected young
woman. X is N(64.5, 2.5). The height of young women follows a Normal distribution
with mean µ = 64.5 inches and standard deviation σ = 2.5
PLAN: Since the “sample” in this case is only one person, the inches.
sample size is clearly smaller than the 10% of the population. Find the probability that the mean height of an SRS of 10
young women exceeds 66.5 inches.
DO:
66.5 - 64.5
z= = 0.80 P(X > 66.5) = P(Z > 0.80) = 1- 0.7881 = 0.2119
2.5
OR Normalcdf (66.5, 10000, 64.5, 2.5) = 0.2118
CONCLUDE: The probability of choosing a young woman at

random whose height exceeds 66.5 inches is about 0.21.
10
Example: Young Women’s Heights Sample Distributions & Normality

If the population is Normal, then the sample distribution is
Normal. No further checks are need!
If the population is NOT Normal, then…. If the sample is

large enough, the distribution of sample means is
“approximately” Normal, no matter what shape the
66.5 - 64.5 P(x > 66.5) = P(Z > 2.53)
z= = 2.53 population distribution has, as long as the population has
0.79 = 1- 0.9943 = 0.0057 a finite standard deviation.
OR normalcdf(66.5, 10000, 64.5, 0.7905) = 0.0057
CONCLUDE: There is a 0.57% percent chance of getting a sample of

10 women with a mean height of 66.5 It is very unlikely (less than a
1% chance) that we would choose an SRS of 10 young women whose
average height exceeds 66.5 inches.
Sample Distributions & Normality: Sample Distributions & Normality:

HOW LARGE IS LARGE ENOUGH? HOW LARGE IS LARGE ENOUGH?
If the Population is…. Minimum Sample Size to

assume Normal
Normal 0
Slightly Skewed 15
Heavily Skewed 30
11
Example: Servicing Air Conditioners Example: Servicing Air Conditioners

Based on many service records from the past year, PLAN/STATE:
the time (in hours) that a technician requires to Independence: It is reasonable to assume that the company
complete preventative maintenance on an air has serviced more than 700 unit, therefore the 70 units in the
conditioner follows the distribution that is sample represent less than 10% of the population.
strongly right-skewed, and whose most likely Normal: Even though the population has a strong right skew, a
outcomes are close to 0. The mean time is µ = 1 sample size of 70 is large enough to assume normality.
hour and the standard deviation is σ = 1
s 1
Your company will service an SRS of 70 air conditioners. µx = µ = 1 sx =
n
=
70
= 0.12
You have budgeted 1.1 hours per unit. Will this be
enough? What is the chance that the technicians will not
finish within the time frame?
Example: Servicing Air Conditioners

DO:
Important Symbols!!!
1.1 -1 P(x > 1.1) = P(Z > 0.83)
z= = 0.83
0.12 = 1- 0.7967 = 0.2033 You need to know these for your quizzes and
test!
OR Normalcdf (1.1000001, 10000,
1, 0.1195) = 0.2013
CONCLUDE: If you budget 1.1 hours per unit, there is a 20.13%

chance the technicians will not complete the work within the
budgeted time.
12
2017. 12. 10.
Section 8.1 Section 8.1

Confidence Intervals: The Basics
Confidence Intervals: The After this section, you should be able to…
Basics ü INTERPRET a confidence level
ü INTERPRET a confidence interval in context
ü DESCRIBE how a confidence interval gives a range of
plausible values for the parameter
ü DESCRIBE the inference conditions necessary to construct
confidence intervals
ü EXPLAIN practical issues that can affect the interpretation
of a confidence interval
Confidence Intervals The Idea of a Confidence Interval

For example: Based on our SRS we find the mean height of 16
In this chapter, we’ll learn one method of statistical inference – year girls to be 65.28 inches. Is the value of the population
confidence intervals – so we may mean µ exactly 65.28 inches? Probably not.
• Estimate the value of a parameter from a sample statistic However, since the sample mean is 65.28, we could guess that
• Calculate probabilities that would describe what would µ is “somewhere” around 65.28. How close to 240.79 is µ
happen if we used the inference method many times. likely to be?
In Chapter 7 we assumed we knew the population parameter;

however, in many real life situations, it is impossible to know
the population parameter. Can we really weigh all the
uncooked burgers in the US? Can we really measure the weights
of all US citizens?
1
2017. 12. 10.
Confidence Intervals: The Basics Confidence Intervals: Point

Estimator
A point estimator is a statistic that provides an estimate of a
population parameter. The value of that statistic from a
sample is called a point estimate. Ideally, a point estimate is
our “best guess” at the value of an unknown parameter.
The point estimator can be a potential mean, standard
deviation, IQR, median, etc.
Identify the Point Estimators Identify the Point Estimators

(a) The golf ball manufacturer would also like to a. Use the sample IQR as the point estimate for the true IQR.
investigate the variability of the distance
travelled by the golf balls by estimating the b. Use the sample proportion 𝑝" as a point estimator for the
interquartile range. true proportion p. The sample proportion is 𝑝" = 0.28.
(b) The math department wants to know what

proportion of its students own a graphing
calculator, so they take a random sample of 100
students and find that 28 own a graphing
calculator.
2
2017. 12. 10.
Confidence Intervals: The Basics Shape, Center & Spread

How would you describe the shape of this distribution?
Shape : Since the population is Normal, so is the sampling distribution of x .

Center : The mean of the sampling distribution of x is the same as the mean
of the population distribution, µ.
Spread : The standard deviation of x for an SRS of 16 observations is
s 20
sx = = =5
n 16
Confidence Intervals: Conditions Confidence Intervals: Formulas

1) Random: The data should come from a well-designed random sample
or randomized experiment. General Formula:
point estimator z-score standard deviation of the statistic
2) Normal: The sampling distribution of the statistic is approximately
Normal.
For proportions: We can use the Normal approximation to the
CI Proportions:
sampling distribution as long as np ≥ 10 and n(1 – p) ≥ 10.
For means: The sampling distribution is exactly Normal if the
population distribution is Normal. When the population distribution is
not Normal, then the central limit theorem tells us the sampling CI Means:
distribution will be approximately Normal if n is sufficiently large (n ≥
30).
3) Independent: Individual observations are independent. When

z ( )
sampling without replacement, the sample size n should be no more
than 10% of the population size N (the 10% condition).
3
2017. 12. 10.
REVIEW: Finding a Critical Value Common Confidence Intervals &

Use Table A to find the critical value z* for an 80%
confidence interval. Assume that the Normal condition is
met.
z-scores
Since we want to capture the central
80% of the standard Normal
Confidence Level Z-Score
distribution, we leave out 20%, or 10%
in each tail. 99%
Search Table A to find the point z* with
area 0.1 to its left.
95% 1.96
z .07 .08 .09
– 1.3 .0853 .0838 .0823 90%
– 1.2 .1020 .1003 .0985
– 1.1 .1210 .1190 .1170
The closest entry is z = – 1.28.
We usually choose a confidence level of 90% or higher
because we want to be quite sure of our conclusions. The
So, the critical value z* for an 80% confidence interval is z* = 1.28. most common confidence level is 95%.
Interpreting Confidence Levels Interpreting Confidence Intervals

The confidence level tells us how likely it is that the method
we are using will produce an interval that captures the Interpret: “We are 95% confident that the interval from
population parameter if we use it many times. ______ to _______ captures the actual value of the (insert
population parameter details…)”
For Example: “In 95% of all possible samples of the
same size, the resulting confidence interval would NOT There is a 95% percent chance….
capture the true (insert details in context).” For example: “We are 95% confident that the interval from
3.03 inches to 3.35 inches capture the actual mean amount
The confidence level does NOT tell us the chance that a of rain in the month of April in Miami.”
particular confidence interval captures the population
parameter.
NO: There is a 95% chance that the mean is between….
4
2017. 12. 10.
Interpret the Following… Interpret the Following…

According to www.gallup.com, on August 13, Interval: We are 95% confident that the interval from 0.41
to 0.47 captures the true proportion of Americans who
2015, the 95% confidence interval for the true approve of the job Barack Obama was doing as president
proportion of Americans who approved of the at the time of the poll.
job Barack Obama was doing as president was
0.44 +/- 0.03. Level: In 95% of all possible samples of the same size, the
resulting confidence interval would capture the true
Interpret the confidence interval and level. proportion of Americans who approve of the job Barack
Obama was doing as president.
Confidence Intervals: Properties Confidence Intervals: Properties

• The critical value (z-score) depends on the confidence level
and the sampling distribution of the statistic.
The margin of error:
• Greater confidence requires a larger critical value
(critical value) • (standard deviation)
• The standard deviation of the statistic depends on the
sample size n
The margin of error
gets smaller when:
• The confidence
level decreases
• The sample size
n increases
5
2017. 12. 10.
Section 8.2
Section 8.2 Estimating a Population
Estimating a Population Proportion
Proportion ü CONSTRUCT and INTERPRET a confidence interval for a
population proportion
ü DETERMINE the sample size required to obtain a level C
confidence interval for a population proportion with a
specified margin of error
ü DESCRIBE how the margin of error of a confidence
interval changes with the sample size and the level of
confidence C
Estimating p & Constructing a CI:

Conditions
1) Random: The data should come from a well-designed random
sample or randomized experiment.
2) Normal: The sampling distribution of the statistic is
approximately Normal. We can use the Normal approximation
to the sampling distribution as long as np ≥ 10 and n(1 – p) ≥ 10.
However, if we don’t know p, we can replace p with 𝑝" to check
normality.
3) Independent: Individual observations are independent.
When sampling without replacement, the sample size n should
be no more than 10% of the population size N (the 10%
condition).
6
2017. 12. 10.
Formula: Confidence Interval for p The Process: Confidence Intervals

Parameter: p = true proportion….
statistic ± (critical value) × (standard deviation of statistic) Assess Conditions:
Random: SRS or random assignment
CI Proportions: Sample Size: be sure to show both np ≥ 10 and n(1 –
p) ≥ 10, the formula plugged and a sentence stating
$ (𝟏 − 𝒑
𝒑 $)
$ ± 𝒛
𝒑 the values are greater than 10, therefore normal.
𝒏 Independent: 10% rule
Name Interval: 1-proportion z-Interval
Interval: Use your calculator. “We are ____% confident
that the interval from ____ to ____ will capture the true
proportion of (context)…”
Conclude in Context: Answer the question being asked.
Example: Red Beads Example: Red Beads

Calculate and interpret a 90% confidence interval for the
proportion of red beads in the container of 3000 beads. Mrs. Parameter: p = true proportion of red beads
Daniel claims 50% of the beads are red. Use your interval to
comment on this claim. The sample proportion you found Assess Conditions:
was 0.426 with a sample of 251 beads. Random: It is reasonable to assume that the sample was
randomly collected.
z .03 .04 .05 Sample Size: Since both n ≥ 10 (251x 0.426 = 106.9) and n(1
ü For a 90% confidence
– 1.7 .0418 .0409 .0401
level, z* = 1.645 – ) ≥ 10 (251 x 1-.426 = 144.1) are both greater than 10, our
– 1.6 .0516 .0505 .0495
sample size is large enough.
– 1.5 .0630 .0618 .0606
Independent: Since the sample of 251 is less than 10% of the
population (3000 beads), it is reasonable to assume
independence when sampling without replacement.
7
2017. 12. 10.
Example: Red Beads Example: Red Beads

Name Interval: 1-proportion z-Interval Name Interval: 1-proportion z-Interval
Interval: Interval:
We are 90% confident that the interval from 0.375 to 0.477 We are 90% confident that the interval from 0.375 to 0.477
will capture the true proportion of red beads. will capture the true proportion of red beads.
statistic ± (critical value) • (standard deviation of the statistic)
pÙ(1 - pÙ)
0.426 +/- 1.645 * pÙ± z *
n
(0.426)(1 - 0.426) = 0.426 ± 0.051

= 0.426 ± 1.645
251 = (0.375, 0.477)
Example: Red Beads 2011B #5

Conclude in Context:
During a flu vaccine shortage in the United States, it was
It is pretty doubtful that Mrs. Daniel’s claim in true that 50%
believed that 45 percent of vaccine-eligible people received
of the beads are red because 0.50 is not included within the
flu vaccine. The results of a survey given to a random sample
interval.
of 2,350 vaccine-eligible people indicated that 978 of the
2,350 people had received flu vaccine.
(a) Construct a 99 percent confidence interval for the
proportion of vaccine-eligible people who had received flu
vaccine. Use your confidence interval to comment on the
belief that 45 percent of the vaccine-eligible people had
received flu vaccine.
8
2017. 12. 10.
Parameter: p = true proportion of vaccine eligible people Name Interval: 1-proportion z-Interval
receiving flu shot
Interval:
Assess Conditions: We are 99% confident that the interval from 0.38998 to
Random: Random sample, stated. 0.44236 will capture the true proportion of vaccine
eligible adults receiving the flu shot.
Sample Size: Since both n𝑝̂ ≥ 10 (978) and n(1 – 𝑝̂ ) ≥ 10
(1372) are both greater than 10, our sample size is Conclude in Context:
considered large enough. Since 0.45 (45%) is not
Independent: Since the sample of 2350 is less than 10% contained within the
of the population (23,500 adults), it is reasonable to interval, we have
assume independence when sampling without reasonable evidence
replacement. that 45% of eligible
adults did not get
vaccinated.
Scoring Calculating the Minimum Sample

Part a:
Size
In planning a study, we may want to choose a sample size that
1 point:
allows us to estimate a population proportion within a given
• Name interval margin of error by plugging in the desired margin of error
• Check random condition (ME), the desired confidence level (z-score) and the known
• Check sample size condition to the margin of error formula. We then solve for n.
1 point:
pÙ(1 - pÙ)
• Correct interval with sentence ME = z *
1 point: n
• Correct conclusion about 45% vaccination claim If the sample proportion is unknown we use = 0.5 as our
best guess, because ME will be largest when = 0.5.
9
2017. 12. 10.
Example: Customer Satisfaction Example: Customer Satisfaction

MacDonald's wants to determine customer satisfaction with its
McDonald's wants to determine customer satisfaction with its new BBQ Chicken Burger. The VP of New Products has hired you
new BBQ Chicken Burger. The VP of New Products has hired to conduct a survey. At a minimum how many people do you need
you to conduct a survey. At a minimum how many people do to survey, if the company is requiring a margin of error of 0.03 at
you need to survey, if the company is requiring a margin of 95% confidence?
error of 0.03 at 95% confidence? pÙ(1 - pÙ)
1.96 £ 0.03
n
Multiply both sides by 1.96
square root n and divide pÙ(1 - pÙ) £ n
both sides by 0.03. 0.03
æ 1.96 ö2
Square both sides. ç ÷ pÙ(1 - pÙ) £ n We round up to 1068
è 0.03 ø
respondents to ensure
æ 1.96 ö2
Substitute 0.5 for the
ç ÷ (0.5)(1 - 0.5) £ n the margin of error is
sample proportion to
è 0.03 ø
find the largest ME no more than 0.03 at
possible.
1067.111 £ n 95% confidence.
2011B #5b Solution

(b) Suppose a similar survey will be given to The sample-size calculation uses 0.5 as the value of the
proportion in order to provide the minimum required sample
vaccine-eligible people in Canada by Canadian size to guarantee that the resulting interval will have a margin
health officials. A 99 percent confidence interval of error no larger than 0.02.
for the proportion of people who will have
received flu vaccine is to be constructed. What is
the smallest sample size that can be used to
guarantee that the margin of error will be less
than or equal to 0.02 ? A sample of at least 4,148 vaccine-eligible people should be
taken in Canada.
10
2017. 12. 10.
Section 8.3 • The One-Sample z Interval for a

Estimating a Population Mean Population Mean
In Section 8.1, we estimated the “mystery mean” µ (see page 468) by
constructing a confidence interval using the sample mean = 240.79.
Learning Objectives To calculate a 95% confidence interval for µ , we use the familiar formula:
estimate ± (critical value) • (standard deviation of statistic)
Estimating a Population Mean

After this section, you should be able to… s 20
x ± z *× = 240.79 ± 1.96×
n 16
ü CONSTRUCT and INTERPRET a confidence interval for a population mean
= 240.79 ± 9.8
ü DETERMINE the sample size required to obtain a level C confidence = (230.99,250.59)
interval for a population mean with a specified margin of error
ü DESCRIBE how the margin of error of a confidence interval changes with One-Sample z Interval for a Population Mean
the sample size and the level of confidence C
Choose an SRS of size n from a population having unknown mean µ and
ü DETERMINE sample statistics from a confidence interval known standard deviation σ. As long as the Normal and Independent
conditions are met, a level C confidence interval for µ is
s
x ± z*
n
The critical value z* is found from the standard Normal distribution.
• Choosing the Sample Size • Example: How Many Monkeys?

The margin of error ME of the confidence interval for the population mean µ is Researchers would like to estimate the mean cholesterol level µ of a particular variety of
monkey that is often used in laboratory experiments. They would like their estimate
s to be within 1 milligram per deciliter (mg/dl) of the true value of µ at a 95%
z *× confidence level. A previous study involving this variety of monkey suggests that the
n standard deviation of cholesterol level is about 5 mg/dl.
We determine a sample size for a desired margin of error when

estimating a mean in much the same way we did when estimating a ü The critical value for 95% confidence is z* = 1.96.
proportion. ü We will use σ = 5 as our best guess for the standard deviation.
Choosing Sample Size for a Desired Margin of Error When Estimating µ 5

1.96 £1
To determine the sample size n that will yield a level C confidence interval n
for a population mean with a specified margin of error ME: Multiply both sides by
(1.96)(5)
• Get a reasonable value for the population standard deviation σ from an square root n and divide £ n
earlier or pilot study. both sides by 1. 1
• Find the critical value z* from a standard Normal curve for confidence We round up to 97
level C. Square both sides. (1.96× 5)2 £ n monkeys to ensure the
• Set the expression for the margin of error to be less than or equal to ME margin of error is no more
and solve for n: s than 1 mg/dl at 95%
z* £ ME 96.04 £ n confidence.
n
11
2017. 12. 10.
s is Unknown: The t Distributions

• When s
• When is Unknown: The t Distributions
When the sampling distribution of x is close to Normal, we can
find probabilities involving x by standardizing : When we standardize based on the sample
z=
x -µ standard deviation sx, our statistic has a
s n
new distribution called a t distribution.

It has a different shape than the standard
Normal curve:
üIt is symmetric with a single peak at 0,
üHowever, it has much more area in the tails.
When we don’t know σ, we can estimate it using the sample standard
deviation sx. What happens when we standardize?
x -µ Like any standardized statistic, t tells us how far x is from its mean µ
?? =
sx n in standard deviation units.
This new statistic does not have a Normal distribution! However, there is a different t distribution for each sample size, specified by its
degrees of freedom (df).
• The t Distributions; Degrees of Freedom • The t Distributions; Degrees of Freedom

When we perform inference about a When comparing the density curves of the standard Normal
distribution and t distributions, several facts are apparent:
population mean µ using a t distribution,
the appropriate degrees of freedom are üThe density curves of the t distributions
are similar in shape to the standard Normal
found by subtracting 1 from the sample size

curve.
The t Distributions; Degrees of Freedom
Drawn, making df = n - 1. We will write the t üThe spread of the t distributions is a bit
an SRS of size n from a large population that has a Normal
greater than that of the standard Normal
distribution with mean µ and standard deviation σ. The statistic
distribution with n -x 1 degrees of freedom as
-µ
distribution.
t=
tn-1. s
x n üThe t distributions have more probability
in the tails and less in the center than does
has the t distribution with degrees of freedom df = n – 1. The statistic will the standard Normal.
have approximately a tn – 1 distribution as long as the sampling
distribution is close to Normal. üAs the degrees of freedom increase, the t
density curve approaches the standard
Normal curve ever more closely.
We can use Table B in the back of the book to determine critical values t* for t
distributions with different degrees of freedom.
12
2017. 12. 10.
• Using Table B to Find Critical t* Values • Constructing a Confidence Interval for µ

Suppose you want to construct a 95% confidence interval When the conditions for inference are satisfied, the sampling
for the mean µ of a Normal population based on an SRS distribution for x has roughly a Normal distribution. Because we
of size n = 12. What critical t* should you use? donÕ t know s , we estimate it by the sample standard deviation sx .
Upper-tail probability p In Table B, we consult the row

Definition:

corresponding to df = n – 1 = 11. sx
df .05 .025 .02 .01 The standard error of the sample mean x is , where sx is the
n
10 1.812 2.228 2.359 2.764 sample standard deviation. It describes how far x will be from µ, on
11 1.796 2.201 2.328 2.718 average, in repeated SRSs of size n.
12 1.782 2.179 2.303 2.681 To construct a confidence interval for µ,

z* 1.645 1.960 2.054 2.326 We move across that row to the üReplace the standard deviation of x by its standard error in the
90% 95% 96% 98% entry that is directly above 95% formula for the one - sample z interval for a population mean.
confidence level.
Confidence level C
üUse critical values from the t distribution with n - 1 degrees of
freedom in place of the z critical values. That is,
The desired critical value is t * = 2.201. statistic ± (critical value)× (standard deviation of statistic)
s
= x ± t* x
n
• One-Sample t Interval for a Population Mean

The one-sample t interval for a population mean is similar in both reasoning and
computational detail to the one-sample z interval for a population proportion.
As before, we have to verify three important conditions before we estimate a
population mean.
Conditions for Inference about a Population Mean

The One-Sample t Interval for a Population Mean
•Random:
Choose an The
SRSdataof size
comen from
fromaapopulation
random sample
havingofunknown
size n from
mean
theµ.
population
A level C
confidence
of interest orinterval
a randomized
for µ is experiment.s
x ± t*
• Normal: The population has a Normal distribution
x
or the sample size is large
n
(n ≥ 30).
where t* is the critical value for the tn – 1 distribution.
• Independent: The method for calculating a confidence interval assumes that
Use this interval only when:
individual observations are independent. To keep the calculations
(1) reasonably accurate
the population wheniswe
distribution sample
Normal orwithout replacement
the sample from(na ≥finite
size is large 30),
population, we should check the 10% condition: verify that the sample size
(2) the population is at least 10 times as large
is no more than 1/10 of the population size. as the sample.
13
2017. 12. 10.
• Example: Video Screen Tension • Example: Video Screen Tension

Read the Example on page 508. STATE: We want to Read the Example on page 508. We want to estimate the
estimate the true mean tension µ of all the video true mean tension µ of all the video terminals
terminals produced this day at a 90% confidence level. produced this day at a 90% confidence level.
PLAN: If the conditions are met, we can use a one-sample t interval to DO: Using our calculator, we find that the mean and standard deviation of

estimate µ. the 20 screens in the sample are:
Random: We are told that the data come from a random sample of 20
x = 306.32 mV and sx = 36.21 mV
screens from the population of all screens produced that day.
Normal: Since the sample size is small (n < 30), we must check whether it’s Upper-tail probability p Since n = 20, we use the t distribution with df = 19
reasonable to believe that the population distribution is Normal. Examine the df .10 .05 .025 to find the critical value.
distribution of the sample data. 18 1.130 1.734 2.101 From Table B, we find t* = 1.729.
19 1.328 1.729 2.093
Therefore, the 90% confidence interval for µ is:
20 1.325 1.725 2.086
sx 36.21
90% 95% 96%
x ± t* = 306.32 ±1.729
Confidence level C n 20
These graphs give no reason to doubt the Normality of the population = 306.32 ± 14
Independent: Because we are sampling without replacement, we must = (292.32, 320.32)
check the 10% condition: we must assume that at least 10(20) = 200 video CONCLUDE: We are 90% confident that the interval from 292.32 to 320.32 mV captures the
terminals were produced this day. true mean tension in the entire batch of video terminals produced that day.
• Using t Procedures Wisely • Using t Procedures Wisely

The stated confidence level of a one-sample t interval for µ is exactly Except in the case of small samples, the condition that the
correct when the population distribution is exactly Normal. No
population of real data is exactly Normal. The usefulness of the t
data come from a random sample or randomized
procedures in practice therefore depends on how strongly they are experiment is more important than the condition that the
affected by lack of Normality. population distribution is Normal. Here are practical
guidelines for the Normal condition when performing

inference about a population mean.
Definition:
Using One-Sample t Procedures: The Normal Condition
An inference procedure is called robust if the probability calculations
involved in the procedure remain fairly accurate when a condition for • Sample size less than 15: Use t procedures if the data appear close to
using the procedures is violated. Normal (roughly symmetric, single peak, no outliers). If the data are clearly
skewed or if outliers are present, do not use t.
• Sample size at least 15: The t procedures can be used except in the
presence of outliers or strong skewness.
Fortunately, the t procedures are quite robust against non-Normality of
the population except when outliers or strong skewness are present. • Large samples: The t procedures can be used even for clearly skewed
Larger samples improve the accuracy of critical values from the t distributions when the sample is large, roughly n ≥ 30.
distributions when the population is not Normal.
14
2017. 12. 10.
Section 8.3 Section 8.3

Estimating a Population Mean Estimating a Population Mean
Summary
Summary
ü There is a t distribution for every positive degrees of freedom. All are
In this section, we learned that… symmetric distributions similar in shape to the standard Normal distribution.
The t distribution approaches the standard Normal distribution as the
ü Confidence intervals for the mean µ of a Normal population are based number of degrees of freedom increases.
on the sample mean of an SRS. ü A level C confidence interval for the mean µ is given by the one-sample t
ü If we somehow know σ, we use the z critical value and the standard interval s
x ± t* x
Normal distribution to help calculate confidence intervals. n
ü This inference procedure is approximately correct when these conditions
ü The sample size needed to obtain a confidence interval with are met: Random, Normal, Independent.
approximate margin of error ME for a population mean involves solving
s ü The t procedures are relatively robust when the population is non-Normal,
z* £ ME especially for larger sample sizes. The t procedures are not robust against
n
outliers, however.
ü In practice, we usually don’t know σ. Replace the standard deviation of
the sampling distribution with the standard error and use the t
distribution with n – 1 degrees of freedom (df).
15
2017. 12. 10.
Section 9.1
Section 9.1 Significance Tests: The Basics
Significance Tests: The Basics After this section, you should be able to…
ü STATE correct hypotheses for a significance test about a

population proportion or mean.
ü INTERPRET P-values in context.
ü INTERPRET a Type I error and a Type II error in context,
and give the consequences of each.
ü DESCRIBE the relationship between the significance level
of a test, P(Type II error), and power.
Statistical Inference Significance Test

A significance test is a formal procedure for comparing
observed data with a claim (also called a hypothesis) whose
truth we want to assess.
Significance Tests- Confidence Interval -
• ASSESS the evidence • ESTIMATE a population
The claim is a statement about a parameter, like the
provided by data about some parameter. population proportion p or the population mean µ.
claim concerning a • Give range of possible values
population
• Reject or fail to reject (Yes vs.
We express the results of a significance test in terms of a
No) probability (p-value) that measures how well the data and
the claim agree.
1
2017. 12. 10.
The Reasoning of Significance The Reasoning of Significance Tests

Tests We can use software to simulate 400 sets of 50 shots assuming
that the player is really an 80% shooter.
Statistical tests deal with claims about a population. Tests ask
if sample data give good evidence against a claim. A test might
say, “If we took many random samples and the claim were
true, what is the probability we will get a result like this.”
For example: Suppose a basketball player claimed to be an

80% free-throw shooter. To test this claim, we have him
attempt 50 free-throws. He makes 32 of them. His sample
proportion of made shots is 32/50 = 0.64.
What can we conclude about the claim based on this sample
data? What is the probability the player is telling the
truth?!?!
The Reasoning of Significance Tests

The Reasoning of Significance Based on the evidence, we might conclude the player’s claim is
Tests incorrect. In reality, there are two possible explanations for
the fact that he made only 64% of his free throws.
You can say how strong the evidence against the player’s
claim is by giving the probability that he would make as few 1) The player’s claim is correct (p = 0.8), and by horrible luck, a
as 32 out of 50 free throws if he really makes 80% in the very unlikely outcome occurred.
long run. 2) The population proportion is actually less than 0.8, so the
sample result is not an unlikely outcome.
The observed statistic is so
unlikely if the actual
parameter value is p = 0.80 Basic Idea
that it gives convincing
An outcome that would rarely happen if a claim were true
evidence that the player’s
is good evidence that the claim is not true.
claim is not true.
2
2017. 12. 10.
Stating Hypotheses Stating Hypotheses

The claim tested by a statistical test is called the null
hypothesis (H0). The test is designed to assess the strength In any significance test, the null hypothesis has the form
of the evidence against the null hypothesis. Often the null H0 : parameter = value
hypothesis is a statement of “no difference” or that the
claim is true. The alternative hypothesis has one of the forms
The claim about the population that we are trying to find Ha : parameter < value
evidence for is the alternative hypothesis (Ha). Ha : parameter > value
Ha : parameter ≠ value
In the free-throw shooter example, our hypotheses are To determine the correct form of Ha, read the problem
H0 : p = 0.80 carefully.
Ha : p < 0.80
Parameter: p = the long-run proportion of made free
throws.
Stating Hypotheses State the Hypothesis & Parameter:

• The alternative hypothesis is one-sided if it states that a
A high school junior running for student body president,
parameter is larger than the null hypothesis value or if it
Sally, claims that 80% of the student body favors her in the
states that the parameter is smaller than the null value.
school election. Her opponent believes this percentage to be
• It is two-sided if it states that the parameter is different from lower, write the appropriate null and alternative hypotheses.
the null hypothesis value (it could be either larger or smaller).
Use Ha : parameter ≠ value for two sided.
• Hypotheses always refer to a population, not to a sample. Be

sure to state H0 and Ha in terms of population parameters (p
or 𝜇).
• It is never correct to write a hypothesis about a sample

statistic, such as pÙ= 0.64 or x = 85.
3
2017. 12. 10.
State the Hypothesis & Parameter: Interpreting P-Values

• The null hypothesis H0 states the claim that we are
seeking evidence against.
• The probability that measures the strength of the
p= true proportion of students that favor Sally for president. evidence against a null hypothesis is called a P-value.
• The probability, computed assuming H0 is true, that the
statistic would take a value as extreme as or more
extreme than the one actually observed is called the P-
value of the test.
• The smaller the P-value, the stronger the evidence against
H0 provided by the data.
Example: Studying Job Satisfaction

H0: µ = 0 English Math Does the job satisfaction of assembly-line workers differ when
their work is machine-paced rather than self-paced?
Small P- Evidence Unlikely to We reject H0 One study chose 18 subjects at random from a company with
value against occur if H0 is over 200 workers who assembled electronic devices. Half of
the workers were assigned at random to each of two groups.
true Both groups did similar assembly work, but one group was
allowed to pace themselves while the other group used an
assembly line that moved at a fixed pace.
Large P-value Not Could occur if We fail to
After two weeks, all the workers took a test of job satisfaction.
convincing H0 is true reject H0 Then they switched work setups and took the test again after
evidence two more weeks. The response variable is the difference in
satisfaction scores, self-paced minus machine-paced.
a) Describe the parameter of interest in this setting.
b) State appropriate hypotheses for performing a
significance test.
4
2017. 12. 10.
Example: Studying Job Satisfaction Example: Studying Job Satisfaction

a) Describe the parameter of interest in this setting.
The parameter of interest is the mean µ of the differences For the job satisfaction study, the hypotheses are
(self-paced minus machine-paced) in job satisfaction scores H0: µ = 0
in the population of all assembly-line workers at this Ha: µ ≠ 0
company. (FYI- Matched pairs!!!) Data from the 18 workers gave x = 17 and sx = 60. That is, these workers rated the
self - paced environment, on average, 17 points higher. Researchers performed a
b) State appropriate hypotheses for performing a significance test using the sample data and found a P - value of 0.2302.
significance test.
Because the initial question asked whether job satisfaction c) Explain what it means for the null hypothesis to be true
differs, the alternative hypothesis is two-sided; that is, in this setting.
either µ < 0 or µ > 0. For simplicity, we write this as µ ≠ 0.
That is, d) Interpret the given P-value in context.
H0: µ = 0
Ha: µ ≠ 0

d) Interpret the P-value in context.
For the job satisfaction study, the hypotheses are
The P-value is the probability of observing a sample result as
H0: µ = 0
extreme as (or more extreme) by pure chance given that the null
Ha: µ ≠ 0
hypothesis is actually true.
c) Explain what it means for the null hypothesis to be true
in this setting. Since the test is two-sided, we have a 23% chance of observing a
In this setting, the null hypothesis is that there is no mean value that is 17 points or more from the mean, in either
difference in employee satisfaction scores (self-paced - direction. An outcome that would occur so often just by chance
machine-paced) for the entire population of assembly- (almost 1 in every 4 random samples of 18 workers) when the
line workers at the company. If null hypothesis is true, null is true is not convincing evidence against null.
then the workers don’t favor one work environment over
the other, on average. We fail to reject H0: µ = 0.
5
2017. 12. 10.
Conclusion: Statistical Significance Statistically Significant

We will make one of two decisions based on the strength of
the evidence against the null hypothesis (and in favor of the • There is no perfect rule for how small a P-value we should
alternative hypothesis) -- reject H0 or fail to reject H0. require in order to reject H0 — it’s a matter of judgment and
depends on the specific circumstances.
ü If our sample result is too unlikely to have happened by • We can compare the P-value with a fixed value (typically α =
chance assuming H0 is true, then we’ll reject H0. 0.05), called the significance level (alpha α).
ü Otherwise, we will fail to reject H0. • When our p-value is greater than the chosen α, there is no
statistically significance. We “fail to reject” the null.
A fail-to-reject H0 decision in a significance test doesn’t • When our P-value is less than the chosen α, we say that the
mean that H0 is true. For that reason, you should never result is statistically significant.
“accept H0” or use language implying that you believe H0 – In that case, we reject the null hypothesis H0 and conclude
is true. that there is convincing evidence in favor of the
alternative hypothesis Ha.
General conclusion in a significance test : Example: Better Batteries

A company has developed a new deluxe AAA battery that is supposed
P-value small → reject H0 → conclude Ha to last longer than its regular AAA battery. However, these new
batteries are more expensive to produce, so the company would like to
be convinced that they really do last longer. Based on years of
P-value large → fail to reject H0 → cannot conclude Ha experience, the company knows that its regular AAA batteries last for
30 hours of continuous use, on average. The company selects an SRS of
15 new batteries and uses them continuously until they are completely
drained. A significance test is performed using the hypotheses
Conclusion with fixed level of significance : H0 : µ = 30 hours
Ha : µ > 30 hours
P-value < α → reject H0 → conclude Ha where µ is the true mean lifetime of the new deluxe AAA batteries. The
resulting P-value is 0.0276.
P-value ≥ α → fail to reject H0 → cannot conclude Ha a) What conclusion can you make for the significance level α = 0.05?
b) What conclusion can you make for the significance level α = 0.01?
6
2017. 12. 10.
Example: Better Batteries

a) What conclusion can you make for the significance level α = 0.05?
Type I and Type II Errors
Since the P-value, 0.0276, is less than α = 0.05, the sample Type I error
result is statistically significant at the 5% level. We have
sufficient evidence to reject the null hypothesis and have
Reject H0 when H0 is true
sufficient evidence to conclude that the company’s deluxe
AAA batteries last longer than 30 hours, on average. Type II error
b) What conclusion can you make at significance level α = 0.01?
Fail to reject H0 when H0 is false
Since the P-value, 0.0276, is greater than α = 0.01, the sample
result is not statistically significant at the 1% level. We fail to
reject the null hypothesis; therefore, we cannot conclude that Double F = Type II
the deluxe AAA batteries last longer than 30 hours, on
average.
Type I & II Errors Type I & II Errors

American Justice System Example:
• Ho: innocent
• Ha: guilty
• Type I error: punish an innocent person

• Type II error: let a not innocent (guilty)
person go free
More: http://www.intuitor.com/statistics/T1T2Errors.html
7
2017. 12. 10.
Type I & II Errors Type I and II Errors

Quality Control Example:
• Ho: the product is acceptable to the

customer
• Ha: the product is unacceptable to the
customer
• type I error: reject acceptable product and
don't ship it.
• type 2 error: ship unacceptable product to
the customer
Example: Perfect Potatoes Example: Perfect Potatoes

A potato chip producer and its main supplier agree that each Describe a Type I and a Type II error in this setting, and explain
shipment of potatoes must meet certain quality standards. If the the consequences of each.
producer determines that more than 8% of the potatoes in the
shipment have “blemishes,” the truck will be sent away to get
• A Type I error would occur if the producer concludes that the
another load of potatoes from the supplier. Otherwise, the entire
truckload will be used to make potato chips. To make the decision, a
proportion of potatoes with blemishes is greater than 0.08
supervisor will inspect a random sample of potatoes from the when the actual proportion is 0.08 (or less). Consequence: The
shipment. The producer will then perform a significance test using potato-chip producer sends the truckload of acceptable
the hypotheses potatoes away, which may result in lost revenue for the supplier.
H0 : p = 0.08
Ha : p > 0.08 • A Type II error would occur if the producer does not send the
where p is the actual proportion of potatoes with blemishes in a truck away when more than 8% of the potatoes in the shipment
given truckload.
have blemishes. Consequence: The producer uses the truckload
Describe a Type I and a Type II error in this setting, and explain the of potatoes to make potato chips. More chips will be made with
consequences of each. blemished potatoes, which may upset consumers.
8
2017. 12. 10.
Power
The probability of
making a Type 1 • The probability of NOT making a Type II error.
error is ALWSYS
equal to the • The higher the power, the less likely the
significance level
(𝛼).
mistake is.
The probability of making a Type

1 error is 1 - 𝛽. Beta is the
“power” of the test.
Factors that Increase Power Increase Power, Decrease Type II

• Sample Size
– The larger the sample size, the higher the power.
• Alpha Significance Level
– Increasing alpha (from 0.01 to 0.05) increases the power,
because a less conservative alpha increases the chance
of (correctly) rejecting the null.
• Value of the Alternative Parameter
– The greater the difference between the Hypothesized
and True Mean the more obvious the result and
therefore the greater the power.
More:
https://onlinecourses.science.psu.edu/stat414/book/export/html/245
9
2017. 12. 10.
Increase Power, Decrease Type II What’s Worse? Type I or II

Blue = Power • It depends.
Red = Type II error
• It’s impossible to minimize both error types
completely.
Error Probabilities
We can assess the performance of a significance test by looking at the probabilities of
the two types of error. That’s because statistical inference is based on asking,
“What would happen if I did this many times?”
For the truckload of potatoes in the previous example, we were testing

H0 : p = 0.08
Ha : p > 0.08
where p is the actual proportion of potatoes with blemishes. Suppose
that the potato-chip producer decides to carry out this test based on a
random sample of 500 potatoes using a 5% significance level (α = 0.05).
More on Type 1 and 2 Errors
10
2017. 12. 10.
Error Probabilities
Assuming H 0 : p = 0.08 is true, the sampling distribution of pÙ will have : Type 2 Errors Investigation WS
Shape : Approximately Normal because 500(0.08) = 40 and www.rossmanchance.com/applets
500(0.92) = 460 are both at least 10.
Center : µ pÙ = p = 0.08
Spread : s pÙ =
p(1- p)
=
0.08(0.92)
= 0.0121 The shaded area in the right
Select: Improved Batting Averages (Power)
n 500 tail is 5%. Sample
proportion values to the
right of the green line at Or direct link:
0.0999 will cause us to
reject H0 even though H0 is http://statweb.calpoly.edu/chance/applets/p
true. This will happen in 5% ower/power.html
of all possible samples. That
is, P(making a Type I error)
= 0.05.
Error Probabilities • Error Probabilities

The potato-chip producer wonders whether the significance test of H0 : p = 0.08 The potato-chip producer wonders whether the significance test of H0 : p = 0.08 versus Ha : p
versus Ha : p > 0.08 based on a random sample of 500 potatoes has enough > 0.08 based on a random sample of 500 potatoes has enough power to detect a
power to detect a shipment with, say, 11% blemished potatoes. In this case, a shipment with, say, 11% blemished potatoes. In this case, a particular Type II error is to
particular Type II error is to fail to reject H0 : p = 0.08 when p = 0.11. fail to reject H0 : p = 0.08 when p = 0.11.
Earlier, we decided to reject H0
at α = 0.05 if our sample
What if p = 0.11? yielded a sample proportion to
the right of the green line.
( pÙ= 0.0999)
Since we reject H0 at α= 0.05 if

Power and Type II Error our sample yields a proportion
The power of a test against any alternative is > 0.0999, we’d correctly reject
1 minus the probability of a Type II error the shipment about 75% of the
for that alternative; that is, power = 1 - β. time.
11
2017. 12. 10.
Section 9.2
Section 9.2 Tests About a Population Proportion
Tests About a
ü CHECK conditions for carrying out a test about a

population proportion.
Population ü CONDUCT a significance test about a population

proportion.
Proportion ü CONSTRUCT a confidence interval to draw a conclusion

about for a two-sided test about a population proportion.
Carrying Out a Significance Test Theory: One-Sample z Test for a Proportion

Recall our basketball player who 현재 이미지를 표시할 수 없습니다.
The z statistic has approximately the standard Normal
claimed to be an 80% free-throw distribution when H0 is true.
shooter. In an SRS of 50 free-
throws, he made 32. His sample pÙ- p
z=
proportion of made shots, 32/50 p0 (1 - p0 )
= 0.64, is much lower than what n
he claimed.
Does it provide convincing
evidence against his claim? We
use z-scores and p-values to
evaluate.
12
2017. 12. 10.
Theory: Test Statistic and P-value Carrying Out a Significance Test

P Parameters
A significance test uses sample data to measure the strength
of evidence against H0.
H Hypothesis
• The test compares a statistic calculated from sample data
A Assess Conditions
with the value of the parameter stated by the null
hypothesis.
N Name the Test
• Values of the statistic far from the null parameter value in T Test Statistic (Calculate)
the direction specified by the alternative hypothesis give
evidence against H0. Obtain P-value
O
• A test statistic measures how far a sample statistic
M Make a decision
diverges from what we would expect if the null hypothesis
H0 were true.
S State conclusion
Carrying Out a Significance Test Carrying Out a Significance Test

Parameters & Hypothesis Assess Conditions: Random, Normal & Independent.
Parameter: p = the actual proportion of free throws the ü Random

shooter makes in the long run. We can view this set of 50 shots as a simple random sample
from the population of all possible shots that the player
takes.
Hypothesis: ü Normal
H0: p = 0.80 Assuming H0 is true, p = 0.80. then np = (50)(0.80) = 40 and
Ha: p < 0.80 n (1 - p) = (50)(0.20) = 10 are both at least 10, so the
normal condition is met.
ü Independent
In our simulation, the outcome of each shot does is
determined by a random number generator, so individual
observations are independent.
13
2017. 12. 10.

Name the Test, Test Statistic (Calculate) & Obtain P-value
Name the Test
Name Test: One-proportion z-test TINspire: Menu, 6: Statistics, 7: Stats Tests, 5: 1-Prop z Test
p(1 - p) (0.8)(0.2)
µ pÙ = p = 0.80 and standard deviation s pÙ = = = 0.0566
n 50
statistic - parameter
test statistic =
standard deviation of statistic
0.64 - 0.80
z= = -2.83
0.0566
Z- score: -2.83
P- value: 0.0023

Make a Decision & State the Conclusion
Test Statistic (Calculate) & Obtain P-value
Make a Decision: P-value is 0.0023. Since the p-value is so
small we reject the null hypothesis.
State the Conclusion: We have convincing evidence that the
basketball player does not make 80% of his free throws.
14
2017. 12. 10.
Example: One Potato, Two Potato Example: One Potato, Two Potato
A potato-chip producer has just received a truckload of State Parameter & State Hypothesis
potatoes from its main supplier. If the producer determines
that more than 8% of the potatoes in the shipment have α = 0.10 significance level
blemishes, the truck will be sent away to get another load
from the supplier. A supervisor selects a random sample of Parameter: p = actual proportion of potatoes in this shipment
500 potatoes from the truck. An inspection reveals that 47 of with blemishes.
the potatoes have blemishes. Carry out a significance test at
the α= 0.10 significance level. What should the producer Hypothesis:
conclude? H0: p = 0.08
Ha: p > 0.08
Example: One Potato, Two Potato

Example: One Potato, Two Potato
Assess Check Conditions
üRandom Name the Test
Random sample of 500 potatoes
üNormal
Assuming H0: p = 0.08 is true, the expected numbers of
blemished and unblemished potatoes are np0 = 500(0.08)
= 40 and n(1 - p0) = 500(0.92) = 460, respectively. Because
both of these values are at least 10, we should be safe
doing Normal calculations.
üIndependent
Because we are sampling without replacement, we need
to check the 10% condition. It seems reasonable to
assume that there are at least 10(500) = 5000 potatoes in
the shipment.
15
2017. 12. 10.
Example: One Potato, Two Potato Example: One Potato, Two Potato
Test Statistic (Calculate) and Obtain P-value Make a Decision and State Conclusion
Make a Decision: Since our P-value, 0.1251, is greater than

the chosen significance level of α = 0.10, so we fail to reject
the null hypothesis.
State Conclusion: There is not sufficient evidence to

conclude that the shipment contains more than 8%
blemished potatoes. The producer will use this truckload of
potatoes to make potato chips.
Smoking in High School State Parameter & State Hypothesis

According to the Centers for Disease Control and Perform at test at the α = 0.05 significance level
Prevention (CDC) Web site, 50% of high school students
have never smoked a cigarette. Taeyeon wonders whether Parameter: p = actual proportion of students in Taeyeon’s
this national result holds true in his large, urban high school who would say they have never smoked cigarettes
school. For his AP Statistics class project, Taeyeon surveys
an SRS of 150 students from his school. He gets responses Hypothesis:
from all 150 students, and 90 say that they have never H0: p = 0.50
smoked a cigarette. What should Taeyeon conclude? Give Ha: p ≠ 0.50
appropriate evidence to support your answer.
16
2017. 12. 10.
Assess Conditions Carrying Out a Significance Test

üRandom
Name the Test, Test Statistic (Calculate) & Obtain P-value
Taeyeon surveyed an SRS of 150 students from his
school. Name: One Proportion Z- Test
üNormal
P0: 0.50
Assuming H0: p = 0.50 is true, the expected numbers of
smokers and nonsmokers in the sample are np0 = x: 90
150(0.50) = 75 and n(1 - p0) = 150(0.50) = 75. Because n: 150
both of these values are at least 10, we should be safe
Test Statistic: z = 2.449
doing Normal calculations.
üIndependent Obtain p-value: p = 0.0143
We are sampling without replacement, we need to
check the 10% condition. It seems reasonable to assume
that there are at least 10(150) = 1500 students a large
high school.
Make a Decision & State Conclusion Confidence Intervals Give More

Make Decision: Since our P-value, 0.0143, is less than the Information
chosen significance level of α = 0.05. we have sufficient The result of a significance test is basically a decision to reject
evidence to reject H0. H0 or fail to reject H0. When we reject H0, we’re left
wondering what the actual proportion p might be. A
State Conclusion: We have convincing evidence to conclude confidence interval might shed some light on this issue.
that the proportion of students at Taeyeon’s school who say
they have never smoked differs from the national result of
0.50. High School Smoking; our 95% confidence interval is:
pÙ(1 - pÙ) 0.60(0.40)
pÙ± z * = 0.60 ± 1.96 = 0.60 ± 0.078 = (0.522,0.678)
n 150
We are 95% confident that the interval from 0.522 to 0.678

captures the true proportion of students at Taeyeon’s high school
who would say that they have never smoked a cigarette.
17
2017. 12. 10.
Confidence Intervals and Two- Confidence Intervals and Two-

Sided Tests Sided Tests
Basketball- Carrying Out a Significance Test

Step 3a: Calculate Mean & Standard Deviation
If the null hypothesis H0 : p = 0.80 is true, then the player’s
sample proportion of made free throws in an SRS of 50 shots
would vary according to an approximately Normal sampling
distribution with mean
p(1 - p) (0.8)(0.2)
One Proportion Z-Test by Hand
n 50
18
2017. 12. 10.
Basketball- Carrying Out a Significance Test Example: One Potato, Two Potato
Step 3b: Calculate Test Statistic Step 3: Calculations
p(1 - p) (0.8)(0.2)
The sample proportion of blemished potatoes is
n 50
pˆ = 47 / 500 = 0.094.
test statistic = pÙ- p0 0.094 - 0.08
Test statistic z = = = 1.15
z=
0.64 - 0.80
= -2.83
p0 (1- p0 ) 0.08(0.92)
0.0566
n 500
Then, Using Table A, we find that the P-value is P(z ≤ – 2.83) =
0.0023.
P-value Using Table A the
desired P-value is
P(z ≥ 1.15) = 1 – 0.8749 = 0.1251
High School Smoking

The sample proportion is pÙ= 60 /150 = 0.60.
Test statistic z =
pÙ- p0
=
0.60 - 0.50
= 2.45
Section 9.3
Tests About a
p0 (1- p0 ) 0.50(0.50)
n 150
Population Mean
P-value To compute this P-
value, we find the area in one
tail and double it. Using Table A
or normalcdf(2.45, 100) yields
P(z ≥ 2.45) = 0.0071 (the right-
tail area).
So the desired P-value
is2(0.0071) = 0.0142.
19
2017. 12. 10.
Section 9.3 Introduction

Tests About a Population Mean
Confidence intervals and significance tests for a population
After this section, you should be able to… proportion p are based on z-values from the standard
Normal distribution.
ü CHECK conditions for carrying out a test about a Reminder: Inference about a population mean µ uses a t
population mean. distribution with n - 1 degrees of freedom, except in the
rare case when the population standard deviation σ is
ü CONDUCT a one-sample t test about a population mean. known.
ü CONSTRUCT a confidence interval to draw a conclusion
for a two-sided test about a population mean.
ü PERFORM significance tests for paired data.
The One-Sample t Test Carrying Out a Significance Test

Choose an SRS of size n from a large population that contains an for µ
unknown mean µ. To test the hypothesis H0 : µ = µ0, compute
ABC company claimed to have developed a new AAA
the one-sample t statistic x - µ0
t= battery that lasts longer than its regular AAA batteries.
sx
n Based on years of experience, the company knows that its
Find the P-value by calculating the probability of getting a t regular AAA batteries last for 30 hours of continuous use,
statistic this large or larger in the direction specified by the on average. An SRS of 15 new batteries lasted an average of
alternative hypothesis Ha in a t-distribution with df = n - 1 33.9 hours with a standard deviation of 9.8 hours. Do these
data give convincing evidence that the new batteries last
longer on average?
20
2017. 12. 10.
Assess Conditions
State Parameter & State Hypothesis Random, Normal, and Independent.
ü Random: The company tests an SRS of 15 new AAA batteries.
Parameter: µ = the true mean lifetime of the new deluxe ü Normal: With such a small sample size (n = 15), we need to
AAA batteries. inspect the data for any departures from Normality.
The boxplot show slight right-

Hypothesis: skewness but no outliers. We should
H0: µ = 30 hours be safe performing a t-test about
the population mean lifetime µ.
Ha: µ > 30 hours
ü Independent Since the batteries are being sampled without
replacement, we need to check the 10% condition
ü 10% Condition: There must be at least 10(15) = 150 new
AAA batteries. This seems reasonable to believe.
Name Test, Test statistic (Calculation) and Obtain P- Make a Decision & State Conclusion
value Make a Decision: Since the p-value of 0.072 exceeds our α =
0.05 significance level, we fail to reject the null hypothesis
One sample t- test and
t: 1.5413 df = 14 Make a Conclusion: we can’t conclude that the company’s
p-value: 0.072771 new AAA batteries last longer than 30 hours, on average.
21
2017. 12. 10.
Details: Normal Condition Details: T- score table

• T-score table gives a range of possible P-values for a
• The Normal condition for means is either population significance. We can still draw a conclusion by comparing the
distribution is Normal or sample size is large (n ≥ 30). range of possible P-values to our desired significance level.
• If the sample size is large (n ≥ 30), we can safely carry out • T-score table only includes probabilities only for t distributions
a significance test (due to the central limit theorem). with degrees of freedom from 1 to 30 and then skips to df = 40,
50, 60, 80, 100, and 1000. (The bottom row gives probabilities for
• If the sample size is small, we should examine (create df = ∞, which corresponds to the standard Normal curve.)
graph on calculator and then DRAW on paper) the
sample data for any obvious departures from Normality, • If the df you need isn’t provided in Table B, use the next
such as skewness and outliers. lower df that is available.
• T-score table shows probabilities only for positive values of t. To

find a P-value for a negative value of t, we use the symmetry of
the t distributions.
Example: Healthy Streams Example: Healthy Streams

The level of dissolved oxygen (DO) in a stream or river is an State Parameters & State Hypothesis
important indicator of the water’s ability to support aquatic
life. A researcher measures the DO level at 15 randomly α = 0.05
chosen locations along a stream. Here are the results in
milligrams per liter: Parameters: µ = actual mean dissolved oxygen level in this
stream.
4.53 5.04 3.29 5.23 4.13
Hypothesis:
5.50 4.83 4.40 5.55 5.73 H0: µ = 5
Ha: µ < 5
5.42 6.38 4.01 4.66 2.87
A dissolved oxygen level below 5 mg/l puts aquatic life at risk.
22
2017. 12. 10.

Assess Conditions
Name Test
üRandom The researcher measured the DO level at 15
randomly chosen locations. Name Test: One Sample T-Test
üNormal With such a small sample size (n = 15), we need to
look at (and DRAW) the data.
The boxplot shows no outliers; with no

outliers or strong skewness, therefore we
can use t procedures.
üIndependent There is an infinite number of possible locations

along the stream, so it isn’t necessary to check the 10%
***Enter data into list first and name “stream”
condition. We do need to assume that individual measurements
are independent.

Test Statistic (Calculate) and Obtain P-value Make a Decision & State Conclusion
Make Decision: Since the P-value is 0.1806 and this is

greater than our α = 0.05 significance level, we fail to
reject H0.
State Conclusion: We don’t have enough evidence to
conclude that the mean DO level in the stream is less
than 5 mg/l.
Test Statistic: t= - .9426 with df= 14

P-value: 0.1809
23
2017. 12. 10.
Pineapples: Two-Sided Tests State Parameters & State Hypothesis:

At the Hawaii Pineapple Company, managers are interested in Parameters: µ = the mean weight (in ounces) of all
the sizes of the pineapples grown in the company’s fields. Last pineapples grown in the field this year
year, the mean weight of the pineapples harvested from one
large field was 31 ounces. A new irrigation system was installed Hypothesis:
in this field after the growing season. Managers wonder H0: µ = 31
whether this change will affect the mean weight of future Ha: µ ≠ 31
pineapples grown in the field. To find out, they select and
weigh a random sample of 50 pineapples from this year’s crop.
Since no significance level is given, we’ll use α = 0.05.
The Minitab output below summarizes the data.
Assess Conditions Name Test, Test Statistic (Calculate) and

üRandom The data came from a random sample of 50 Obtain P-value
pineapples from this year’s crop.
üNormal We don’t know whether the population
distribution of pineapple weights this year is Normally
distributed. But n = 50 ≥ 30, so the large sample size (and
the fact that there are no outliers) makes it OK to use t
procedures.
üIndependent There need to be at least 10(50) = 500
pineapples in the field because managers are sampling
without replacement (10% condition). We would expect
many more than 500 pineapples in a “large field.”
24
2017. 12. 10.
Make Decision & State Conclusion Confidence Intervals Give More

Make Decision: Since the P-value is 0.0081 it is less than our Information
α = 0.05 significance level, so we have enough evidence to Minitab output for a significance test and confidence
reject the null hypothesis. interval based on the pineapple data is shown below.
The test statistic and P-value match what we got earlier
State Conclusion: We have convincing evidence that the (up to rounding).
mean weight of the pineapples in this year’s crop is not 31
ounces; meaning the irrigation system has an effect.*
* We only KNOW that there is an effect. We did not test

The 95% confidence interval for the mean weight of all
whether the effect was positive (bigger pineapples) or
the pineapples grown in the field this year is 31.255 to
negative.
32.616 ounces. We are 95% confident that this interval
captures the true mean weight µ of this year’s pineapple
crop.
Confidence Intervals and Two-Sided Tests

Inference for Means: Paired Data
The connection between two-sided tests and confidence intervals is
even stronger for means than it was for proportions. That’s because
both inference methods for means use the standard error of the • Study designs that involve making two observations on
sample mean in the calculations. the same individual, or one observation on each of two
similar individuals, result in paired data.
x - µ0 sx
Test statistic : t = Confidence interval: x ± t *
sx n • When paired data result from measuring the same
n quantitative variable twice, we can make comparisons by
ü A two-sided test at significance level α (say, α = 0.05) and a 100(1 – α)% analyzing the differences in each pair.
confidence interval (a 95% confidence interval if α = 0.05) give similar
information about the population parameter. • If the conditions for inference are met, we can use one-
ü When the two-sided significance test at level α rejects H0: µ = µ0, the
sample t procedures to perform inference about the
100(1 – α)% confidence interval for µ will not contain the hypothesized value mean difference µd.
µ0 .
• These methods are sometimes called paired t
ü When the two-sided significance test at level α fails to reject the null procedures.
hypothesis, the confidence interval for µ will contain µ0 .
25
2017. 12. 10.
Results of a caffeine deprivation study

Caffeine: Paired Data Subject Depression Depression Difference
Researchers designed an experiment to study the effects of caffeine (caffeine) (placebo) (placebo – caffeine)
withdrawal. They recruited 11 volunteers who were diagnosed as 1 5 16 11
being caffeine dependent to serve as subjects.
2 5 23 18
Each subject was barred from coffee, colas, and other substances
with caffeine for the duration of the experiment. During one two- 3 4 5 1
day period, subjects took capsules containing their normal caffeine 4 3 7 4
intake. During another two-day period, they took placebo capsules. 5 8 14 6
The order in which subjects took caffeine and the placebo was 6 5 24 19
randomized. At the end of each two-day period, a test for 7 0 6 6
depression was given to all 11 subjects.
8 0 3 3
Researchers wanted to know whether being deprived of caffeine
9 2 15 13
would lead to an increase in depression. A higher value equals
higher levels of depression. 10 11 12 1
Data on next slide. 11 1 0 - 1
State Parameters & State Hypothesis: Assess Conditions:

üRandom researchers randomly assigned the treatment order—
If caffeine deprivation has no effect on depression, then we placebo then caffeine, caffeine then placebo—to the subjects.
would expect the actual mean difference in depression üNormal We don’t know whether the actual distribution of difference
scores to be 0. in depression scores (placebo - caffeine) is Normal. So, with such a
small sample size (n = 11), we need to examine the data
Parameter: µd = the true mean difference (placebo –
caffeine) in depression score.
The boxplot shows some right-
skewness but no outliers; with no
Hypotheses: outliers or strong skewness, the t
H0: µd = 0 procedures are reasonable to use.
Ha: µd > 0
Since no significance level is given, we’ll use α = 0.05. üIndependent We aren’t sampling, so it isn’t necessary to check the
10% condition. We will assume that the changes in depression scores
for individual subjects are independent. This is reasonable if the
experiment is conducted properly.
26
2017. 12. 10.
Name Test, Test Statistic (Calculate) and Make Decision & State Conclusion:
Obtain P-value Make Decision: Since the P-value of 0.0027 is much less
than our chosen α = 0.05, we have convincing evidence to
reject H0: µd = 0.
State Conclusion: We can therefore conclude that

depriving these caffeine-dependent subjects of caffeine
caused* an average increase in depression scores.
*Since the data came from a well-designed experiment

we can use the word “caused”
Using Tests Wisely: Statistical Using Tests Wisely: Don’t Ignore Lack of
Significance and Practical Importance Significance
Statistical significance is valued because it points to an effect There is a tendency to infer that there is no difference
that is unlikely to occur simply by chance. whenever a P-value fails to attain the usual 5% standard. In
some areas of research, small differences that are detectable
When a null hypothesis (“no effect” or “no difference”) can be only with large sample sizes can be of great practical
rejected at the usual levels (α = 0.05 or α = 0.01), there is good significance. When planning a study, verify that the test you
evidence of a difference. But that difference may be very plan to use has a high probability (power) of detecting a
small. When large samples are available, even tiny deviations difference of the size you hope to find.
from the null hypothesis will be significant.
27
2017. 12. 10.
Using Tests Wisely: Statistical Inference

Is Not Valid for All Sets of Data
Badly designed surveys or experiments often produce invalid

results. Formal statistical inference cannot correct basic flaws
in the design. Each test is valid only in certain circumstances,
with properly produced data being particularly important.
T-Tests by Hand
Crap in = crap out
Example: Healthy Streams Pineapples

The sample mean and standard deviation are x = 31.935 and sx = 2.394
Step 3: Calculations
The sample mean and standard deviation are
x = 4.771 and sx = 0.9396
x - µ0 x - µ 0 31.935 - 31
Test statistic t = =
4.771- 5
= -0.94 Test statistic t = = = 2.762
sx 0.9396 sx 2.394
n 15 n 50
Upper-tail probability p P-value The P-value is the area to the left of Upper-tail probability p P-value The P-value for this two-sided test is
df .25 .20 .15 t = -0.94 under the t distribution curve with df .005 .0025 .001 the area under the t distribution curve with 50
13 .694 .870 1.079
df = 15 – 1 = 14. 30 2.750 3.030 3.385
- 1 = 49 degrees of freedom. Since Table B
14 .692 .868 1.076 40 2.704 2.971 3.307
does not have an entry for df = 49, we use the
15 .691 .866 1.074 50 2.678 2.937 3.261
50% 60% 70% 99% 99.5% 99.8%
more conservative df = 40. The upper tail
Confidence level C Confidence level C
probability is between 0.005 and 0.0025 so the
desired P-value is between 0.01 and 0.005.
28
2017. 12. 10.
Caffeine Calculations Carrying Out a Significance Test

The sample mean and standard deviation are Name Test, Test statistic (Calculation) and Obtain
x d = 7.364 and sd = 6.918 P-value
test statistic =
x d - µ 0 7.364 - 0 For a test of H0: µ = µ0, our statistic is the sample mean. Its
Test statistic t = = = 3.53 standard deviation is s
sd 6.918 sx =
n 11 n
P-value According to technology, the area to the right of t Because the population standard deviation σ is usually unknown, we
use the sample standard deviation sx in its place. The resulting test
= 3.53 on the t distribution curve with df = 11 – 1 = 10 is
statistic has the standard error of the sample mean in the
0.0027. denominator x - µ0
t=
sx
n
When the Normal condition is met, this statistic has a t distribution
with n - 1 degrees of freedom.
Carrying Out a Hypothesis Test

Name Test, Test statistic (Calculation) and Obtain P-value
x = 33.9 hours and sx = 9.8 hours.
test statistic =
x - µ 0 33.9 - 30
t= = = 1.54
sx 9.8
n 15
The P-value is the probability of getting a

result this large or larger in the direction
indicated by Ha, that is, P(t ≥ 1.54).
Upper-tail probability p
df .10 .05 .025 ü Go to the df = 14 row.
13 1.350 1.771 2.160
ü Since the t statistic falls between the values 1.345
14 1.345 1.761 2.145
15 1.341 1.753 3.131
and 1.761, the “Upper-tail probability p” is between
80% 90% 95%
0.10 and 0.05.
Confidence level C ü The P-value for this test is between 0.05 and 0.10.
29
2017. 12. 10.
Section 10.1
10.1 Comparing Two Proportions
Comparing Two ü DETERMINE whether the conditions for performing

inference are met.
Proportions ü CONSTRUCT and INTERPRET a confidence interval to

compare two proportions.
ü PERFORM a significance test to compare two proportions.
ü INTERPRET the results of inference procedures in a
randomized experiment.
Introduction
• Helps us compare the proportions of individuals
with a certain characteristic in two different
populations.
• We can compare proportions in both surveys and
experiments.
• Sample sizes nor population sizes need to be the
same.
1
2017. 12. 10.
Theory: The Sampling Distribution Theory: The Sampling Distribution of a

of a Difference Between Two Difference Between Two Proportions
Proportions To explore the sampling distribution of the difference between
two proportions, let’s start with two populations having a known
The sampling distribution of a sample proportion has the proportion of successes.
following properties: ü At School 1, 70% of students did their homework last night
ü At School 2, 50% of students did their homework last
Shape Approximately Normal if np ≥ 10 and n(1 - p) ≥ 10
night.
Suppose the counselor at School 1 takes an SRS of 100 students
Center µ pÙ = p
and records the sample proportion that did their homework.
School 2’s counselor takes an SRS of 200 students and records
p(1 - p) the sample proportion that did their homework.
Spread s pÙ = if the sample is no more than 10% of the population
n
Theory: The Sampling Distribution of a

Difference Between Two Proportions
Using Fathom software, we generated an SRS of 100 students from School 1 and a
separate SRS of 200 students from School 2. The difference in sample proportions was
then calculated and plotted. We repeated this process 1000 times. The results are below:
What do you notice about the shape, center, and spread

of the sampling distribution of pÙ1 - pÙ2 ?
2
2017. 12. 10.
Formula: Confidence Intervals for p1 – p2 Conditions: Two Proportion Z-

statistic ± (critical value) × (standard deviation of statistic)
Interval CI
pÙ(1- pÙ1 ) pÙ2 (1- pÙ2 ) 1) Random: Both sets of data should come from a well-designed
( pÙ1 - pÙ2 ) ± z * 1 +
n1 n2 random samples or randomized experiments.
3) Independent: Both sets of data must be independent. When

sampling without replacement, the sample size n should be no
more than 10% of the population size N (the 10% condition).
Must check the condition for each separate sample.
Teens and Adults on Social Networks

Let’s Practice: Teens and Adults on Social
Networks Parameters: p1 = proportion of teens using social media
p2 = proportion of adults using social media
As part of the Pew Internet and American Life Project, Assess Conditions:
researchers conducted two surveys in late 2009. The first ü Random The data come from two separate random samples
survey asked a random sample of 800 U.S. teens about ü Normal: the Normal condition is met since all counts are all at
their use of social media and the Internet. A second survey least 10:
posed similar questions to a random sample of 2253 U.S. n1 pÙ1 = 800(0.73) = 584 n1(1- pÙ1) = 800(1- 0.73) = 216
adults. In these two studies, 73% of teens and 47% of n2 pÙ2 = 2253(0.47) =1058.91 Þ1059 n2 (1- pÙ2 ) = 2253(1- 0.47) =1194.09 Þ1194
adults said that they use social-networking sites. Use these ü Independent We clearly have two independent samples—one
results to construct and interpret a 95% confidence of teens and one of adults. Individual responses in the two
interval for the difference between the proportion of all samples also have to be independent.
U.S. teens and adults who use social-networking sites. ü10% condition: We can reasonably assume that there are at
least 10(800) = 8000 U.S. teens and at least 10(2253) = 22,530
U.S. adults.
3
2017. 12. 10.
Teens and Adults on Social Networks n1p1 = 800*.73

Name the Interval: 2-proprtion z-Confidence Interval
Interval:
pÙ1 (1- pÙ1) pÙ2 (1- pÙ2 ) 0.73(0.27) 0.47(0.53)
( pÙ1 - pÙ2 ) ± z * + = (0.73 - 0.47) ± 1.96 + n2p2 = 2253*.47
n1 n2 800 2253
= 0.26 ± 0.037 = 1058.91
= (0.223, 0.297)
Round to nearest
whole number.
We are 95% confident that the interval from 0.223 to 0.297
captures the true difference in the proportion of all U.S. teens
and adults who use social-networking sites.
Conclude: This interval suggests that more teens than adults

in the United States engage in social networking by between
22.3 and 29.7 percentage points.
Significance Tests for p1 – p2

• An observed difference between two sample proportions can
reflect an actual difference in the parameters, or it may just be
due to chance variation in random sampling or random
assignment.
• Significance tests help us decide which explanation makes

more sense.
• The null hypothesis has the general form: H0: p1 = p2
• The alternative hypothesis says what kind of difference we

expect.
Ha: p1 > p2 or Ha: p1 < p2 or Ha: p1 ≠ p2
4
2017. 12. 10.
Significance Tests for p1 – p2 Significance Tests for p1 – p2

count of successes in both samples combined X1 + X 2
pÙC = =
count of individuals in both samples combined n1 + n 2
Use pÙC in place of both p1 and p2 in the expression for the denominator of the test
statistic :
( pÙ1 - pÙ2 ) - 0
z=
pÙC (1 - pÙC ) pÙC (1 - pÙC )
+
n1 n2
Why Do We Pool? School Breakfast

“We have merged the data from the two samples to obtain
what is called the "pooled" estimate of the standard Researchers designed a survey to compare the proportions of
deviation. We have done this not because it is more children who come to school without eating breakfast in two
convenient (it isn't -- there's more calculation involved) nor low-income elementary schools. An SRS of 80 students from
because it reduces the measurement of variability (it doesn't School 1 found that 19 had not eaten breakfast. At School 2,
always -- often the pooled estimate is larger*) but because it an SRS of 150 students included 26 who had not had
gives us the best estimate of the variability of the difference breakfast. More than 1500 students attend each school. Do
under our null hypothesis that the two sample proportions these data give convincing evidence of a difference in the
came from populations with the same proportion.” population proportions? Carry out a significance test at the α
= 0.05 level to support your answer.
For more details see:
http://apcentral.collegeboard.com/apc/members/courses/te
achers_corner/49013.html
5
2017. 12. 10.
Hungry Children
Parameters & Hypotheses:
H0: p1 = p2
Ha: p1 ≠ p2
p1 = the true proportion of students at School 1 who did

not eat breakfast
p2 = the true proportion of students at School 2 who did
not eat breakfast.
Name Test: Two-proportion z test
Test Statistic: z = 1.17
Obtain p-value: p = 0.2426
Colon Cancer: FRQ 2015 # 4

Make a Decision: Since our P-value, 0.2420, is greater than
the chosen significance level of α = 0.05,we fail to reject H0.
Conclude: There is not sufficient evidence to conclude that

the proportions of students at the two schools who didn’t eat
breakfast are different.
6
2017. 12. 10.
Name Test: Two-proportion z test Scoring:

Test Statistic: z = -1.75 1. Parameters & Hypothesis
Obtain p-value: p = 0.0396
– Direction must be correct
– Must be in present tense (NO: developed cancer,
took aspirin, had cancer, etc.)
2. Name of Test, Conditions (random & normal)
Make a Decision: Since our P-value, 0.0396, is less than the
chosen significance level of α = 0.05,we reject the null. – Random Assignment (NO: stated or SRS)
– Normal must have formula, values and sentence
Conclude: There is sufficient evidence to conclude that taking
low does aspirin each day reduces the chance of developing
3. Test Statistic & P-value
colon cancer among all people similar to volunteers. 4. Decision, p-value linked to alpha & Context
– NO proves, proven or for ALL people, etc.
7
2017. 12. 10.
Confidence Intervals for p1 – p2

When data come from two random samples or two groups in a randomized
experiment, the statistic pÙ1 - pÙ2 is our best guess for the value of p1 - p2 . We
can use our familiar formula to calculate a confidence interval for p1 - p2 :
statistic ± (critical value)× (standard deviation of statistic)
When the Independent condition is met, the standard deviation of the statistic
pÙ1 - pÙ2 is :
p1(1 - p1) p2 (1 - p2 )
s pÙ - pÙ = +
1 2
n1 n2
Standard Error Because we don't know the values of the parameters p1 and p2, we replace them
in the standard deviation formula with the sample proportions. The result is the
pÙ1(1 - pÙ1) pÙ2 (1 - pÙ2 )
standard error of the statistic pÙ1 - pÙ2 : +
n1 n2
If the Normal condition is met, we find the critical value z* for the given confidence
level from the standard Normal curve. Our confidence interval for p1 – p2 is:
statistic ± (critical value) × (standard deviation of statistic)
pÙ(1- pÙ1 ) pÙ2 (1- pÙ2 )
( pÙ1 - pÙ2 ) ± z * 1 +
n1 n2
Section 10.2
Section 10.2
Comparing Two Means
Comparing Two
ü DESCRIBE the characteristics of the sampling distribution of the
difference between two sample means
ü CALCULATE probabilities using the sampling distribution of the
difference between two sample means
Means
ü DETERMINE whether the conditions for performing inference are
met
ü USE two-sample t procedures to compare two means based on
summary statistics or raw data
ü INTERPRET computer output for two-sample t procedures
ü PERFORM a significance test to compare two means
ü INTERPRET the results of inference procedures
8
2017. 12. 10.
Theory: The Sampling Distribution of a Theory: The Sampling Distribution of a

Difference Between Two Means Difference Between Two Means
The sampling distribution of a sample mean has the following properties: Using software, we generated an SRS of 12 girls and a separate SRS of 8 boys and
calculated the sample mean heights. The difference in sample means was then
calculated and plotted. We repeated this process 1000 times. The results are below:
Shape Approximately Normal if the population distribution is Normal or
n ≥ 30 (by the central limit theorem).
Center µx = µ
s
Spread sx = if the sample is no more than 10% of the population
n
What do you notice about the shape, center, and spread

of the sampling distribution of x f - x m ?
Formula: Two Mean T-Interval
2 2
s1 s2
( x1 - x2 ) ± t * +
n1 n2
The degrees of freedom is determined by smaller of n1 - 1

and n2 – 1.
9
2017. 12. 10.
Conditions: Two Mean T- Interval Big Trees, Small Trees, Short Trees, Tall Trees
The Wade Tract Preserve in Georgia is an old-growth forest of longleaf
pines that has survived in a relatively undisturbed state for hundreds of
years. One question of interest to foresters who study the area is “How
1) Random: Both sets of data should come from a well-designed do the sizes of longleaf pine trees in the northern and southern halves
random samples or randomized experiments. of the forest compare?” To find out, researchers took random samples
of 30 trees from each half and measured the diameter at breast height
(DBH) in centimeters. Comparative boxplots of the data and summary
2) Normal: Both sets of data must meet the Central Limit statistics from Minitab are shown below. Construct and interpret a 90%
confidence interval for the difference in the mean DBH for longleaf
Theorem (CLT) with sample sizes greater than 30 or graph values pines in the northern and southern halves of the Wade Tract Preserve.
that are less than 30 to check normality.
3) Independent: Both sets of data must be independent. When

sampling without replacement, the sample size n should be no
Parameters:
µ1 = the true mean DBH of all trees in the northern half of the Name Test: Two-sample t interval for the difference µ1 – µ2
forest df = 30-1 = 29 OR df = 55.72 (calculator)
µ2 = the true mean DBH of all trees in the southern half of the
forest. Interval:
(-17.7238 to -3.93617)
Assess Conditions:
ü Random: Random samples of 30 trees each from the We are 90% confident that the
northern and southern halves of the forest. interval from -17.7238 to -3.93617
ü Normal: Reasonable to assume normality, since the samples centimeters captures the difference
sizes are each 30. in the actual mean DBH of the
southern trees and the actual mean
üIndependent Researchers took independent samples from the
DBH of the northern trees.
northern and southern halves of the forest.
ü10 % Condition: Since we are sampling without Conclude: This interval suggests that the mean diameter of the
replacement, there have to be at least 10(30) = 300 trees in southern trees is between 3.83 and 17.83 cm larger than the
each half of the forest. This is pretty safe to assume. mean diameter of the northern trees.
10
2017. 12. 10.
Significance Tests for µ1 – µ2 Formula: Significance Tests for

• An observed difference between two sample means can reflect µ1 – µ2
an actual difference in the parameters, or it may just be due to
chance variation in random sampling or random assignment. ( x1 - x2 ) - ( µ1 - µ 2 )
t=
2 2
• Significance tests help us decide which explanation makes s1 s2
+
more sense. n1 n2
• The null hypothesis has the general form: H0: µ1 = µ2

The degrees of freedom is determined by smaller of n1 - 1
• The alternative hypothesis says what kind of difference we and n2 – 1.
expect:
Ha: µ1 > µ2 OR Ha: µ1 < µ2 OR Ha: µ1 ≠ µ2
Conditions: Two Mean T- Using the Two-Sample t Procedures:

Significance Test The Normal Condition
1) Random: Both sets of data should come from a well-designed • Sample size less than 15: Use two-sample t procedures if the
random samples or randomized experiments. data in both samples/groups appear close to Normal (roughly
symmetric, single peak, no outliers). If the data are clearly
skewed or if outliers are present, do not use t.
2) Normal: Both sets of data must meet the Central Limit
Theorem* with sample sizes greater than 30 or graph values • Sample size at least 15: Two-sample t procedures can be
that are less than 30 to check normality. No crazy outliers! used except in the presence of outliers or STRONG skewness.
• Large samples: The two-sample t procedures can be used
3) Independent: Both sets of data must be independent. When even for clearly skewed distributions when both
sampling without replacement, the sample size n should be no samples/groups are large, roughly n ≥ 30.
11
2017. 12. 10.
Calcium and Blood Pressure Calcium and Blood Pressure

Does increasing the amount of calcium in our diet reduce blood An increase (which is bad for blood pressure) appears as a negative
response.
pressure? Examination of a large sample of people revealed a
relationship between calcium intake and blood pressure. The
relationship was strongest for black men. Such observational
studies do not establish causation. Researchers therefore
designed a randomized comparative experiment. The subjects
were 21 healthy black men who volunteered to take part in the
experiment. They were randomly assigned to two groups: 10 of
the men received a calcium supplement for 12 weeks, while the
control group of 11 men received a placebo pill that looked
identical. The experiment was double-blind. The response
variable is the decrease in systolic (top number) blood pressure
for a subject after 12 weeks, in millimeters of mercury. Data on
next slide.
Parameters & Hypotheses: Assess Conditions:

• Random The 21 subjects were randomly assigned to the two
treatments.
H0: µ1 = µ2
Ha: µ1 > µ2 • Normal: Since the sample sizes are less than 15, we must
check and draw graphs.
µ1 = the true mean decrease in systolic blood pressure for
healthy black men who take a calcium supplement The boxplots show no
clear evidence of
µ2 = the true mean decrease in systolic blood pressure for skewness and no
healthy black men who take a placebo. outliers, therefore we
can use t procedures.
We will use α = 0.05.
• Independent: Due to the random assignment, these two

groups of men can be viewed as independent.
12
2017. 12. 10.
Name the Test: Two-sample t test for the difference µ1 – µ2.

Make Decision: Because the P-value, 0.06442, is greater than
Test Statistic: t = 1.60372 α = 0.05, we fail to reject the null hypothesis.
Obtain p-value: p-value: 0.06442
` df= 15.5905 (calculator) or 9 (by hand) State Conclusion: The experiment provides some evidence
that calcium reduces blood pressure, but the evidence is not
convincing enough to conclude that calcium reduces blood
Byhand : pressure more than a placebo.
( x1 - x2 ) - ( µ1 - µ 2 ) [5.000 - (-0.273)] - 0
t= = = 1.604
2 2
s1 s2 8.7432 5.9012
+ +
n1 n2 10 11
Confidence Interval vs. Significance Test Using Two-Sample t Procedures Wisely

To get results that are consistent with the one-tailed test at α ü In planning a two-sample study, choose equal sample sizes
= 0.05 from the example, we’ll use a 90% confidence level. if you can.
We are 90% confident that the ü Do not use “pooled” two-sample t procedures!
interval from -0.4766 to 11.022
captures the difference in true
ü We are safe using two-sample t procedures for comparing
mean blood pressure reduction
two means in a randomized experiment.
on calcium over a placebo.
Because the 90% confidence ü Do not use two-sample t procedures on paired data!
interval includes 0 as a
plausible value for the ü Beware of making inferences in the absence of
difference, we fail to reject the randomization. The results may not be generalized to the
null hypothesis. larger population of interest.
13
2017. 12. 10.
The Sampling Distribution of a Difference

Between Two Means
Both x1 and x 2 are random variables. The statistic x1 - x 2 is the difference
of these two random variables. In Chapter 6, we learned that for any two
independent random variables X and Y,
µX -Y = µX - µY and s X2 -Y = s X2 + s Y2
Therefore,
µx1 -x2 = µx1 - µx2 = µ1 - µ2 s x2 -x = s x2 + s x2
Formula Derivations
1 2 1 2
æs ö æs ö
2 2
= çç 1 ÷÷ + çç 2 ÷÷
è n1 ø è n 2 ø
s 12 s 12
= +
n1 n2
s 12 s 12
s x -x = +
1 2
n1 n2
The Two-Sample t Statistic

When data come from two random samples or two groups in a randomized
experiment, the statistic x1 - x 2 is our best guess for the value of µ1 -µ2 .
When the Independent condition is met, the standard deviation of the statistic
x1 - x 2 is :
s12 s2 2
sx -x 2 = +
1
n1 n2
Since we don't know the values of the parameters s1 and s2, we replace them
in the standard deviation formula with the sample standard deviations. The result
s12 s2 2
is the standard error of the statistic x1 - x 2 : +
n1 n 2
If the Normal condition is met, we standardize the observed difference to obtain

a t statistic that tells us how far the observed difference is from its mean in standard
deviation units:
(x1 - x 2 ) - ( µ1 - µ2 ) The two-sample t statistic has approximately a t distribution.
t= 2 2 We can use technology to determine degrees of freedom
s1 s2 OR we can use a conservative approach, using the smaller
+
n1 n 2 of n1 – 1 and n2 – 1 for the degrees of freedom.
14
11.1 Chi-Square Goodness-of-Fit Tests
Chapter 11:
Inference for
Distributions of
Categorical Data

Chi-Square Goodness-of-Fit Tests
Input/Desired Result Test Required
ü COMPUTE expected counts, conditional distributions, and
contributions to the chi-square statistic One Categorical Variable Chi-Square Goodness of Fit
(GOF)
ü CHECK the Random, Large sample size, and Independent
conditions before performing a chi-square test Two Categorical Variables Chi-Square Test for
ü PERFORM a chi-square goodness-of-fit test to determine Homogeneity
whether sample data are consistent with a specified
distribution of a categorical variable Study Relationship between Chi-Square Test for
ü EXAMINE individual components of the chi-square Two Categorical Variables Association/Independence
statistic as part of a follow-up analysis
1
Categorical Data WHY? Chi-Square Goodness-of-Fit Tests

Mars, Incorporated makes milk chocolate candies. Here’s what The one-way table below summarizes the data from a sample bag of
the company’s Consumer Affairs Department says about the M&M’S Milk Chocolate Candies.
color distribution of its M&M’S Milk Chocolate Candies: On
average, the new mix of colors of M&M’S Milk Chocolate Color Blue Orange Green Yellow Red Brown Total
Candies will contain 13 percent of each of browns and reds, 14 Count 9 8 12 15 10 6 60
percent yellows, 16 percent greens, 20 percent oranges and 24
9
percent blues. The sample proportion of blue M& M's is pÙ= = 0.15.
60
Since the company claims that 24% of all M&M’S Milk Chocolate Candies
are blue, we might believe that something fishy is going on. We could use
the one-sample z test for a proportion from Chapter 9 to test the
hypotheses
H0: p = 0.24
Ha: p ≠ 0.24
However, performing a one-sample z test for each proportion

would be pretty inefficient and would lead to the problem of
multiple comparisons.
WHY? Chi-Square Goodness-of-Fit Tests Idea of the Chi-Square Goodness-of-Fit Test

More important, performing one-sample z tests for each color We compare the observed counts from our sample with the
wouldn’t tell us how likely it is to get a random sample of 60 counts that would be expected if H0 is true. The more the
candies with a color distribution that differs as much from the observed counts differ from the expected counts, the more
one claimed by the company as this bag does (taking all the evidence we have against the null hypothesis.
colors into consideration at one time).
In general, the expected counts can be obtained by
For that, we need a new kind of significance test, called a
multiplying the proportion of the population distribution in
chi-square goodness-of-fit test.
each category by the sample size.
Hypothesis:
H0: The company’s stated color distribution for
M&M’S Milk Chocolate Candies is correct. The chi-square statistic is a measure of
how far the observed counts are from
Ha: The company’s stated color distribution for
M&M’S Milk Chocolate Candies is not correct. the expected counts.
2
The Chi-Square Statistic Calculating Expected Values

To see if the data give convincing evidence against the null For random samples of 60 candies, the average number of
hypothesis, we compare the observed counts from our blue M&M’s should be (0.24)(60) = 14.40. This is our
sample with the expected counts assuming H0 is true. If expected count of blue M&M’s.
the observed counts are far from the expected counts, Using this same method, we can find the expected counts
that’s the evidence we were seeking. for the other color categories:
Big Chi-Square values means the data is far from what we
are expecting, giving us evidence against the null (reject). Proportions According to Mars
Orange: (0.20)(60) = 12.00
The chi-square statistic is a measure of how far the observed
counts are from the expected counts. The formula for the Green: (0.16)(60) = 9.60
statistic is Yellow: (0.14)(60) = 8.40
(Observed - Expected) 2
c2 = å Red: (0.13)(60) = 7.80
Expected
Brown: (0.13)(60) = 7.80
We computed the chi - square statistic for our sample of 60 M & M’ s to be
Calculating Expected Values c 2 = 10.180. Because all of the expected counts are at least 5, the c 2
statistic will follow a chi - square distribution with df = 6 (number of categories) - 1 = 5 reasonably
well when H 0 is true.
(Observed - Expected) 2 To find the P - value, use Table C
c =å
2
and look in the df = 5 row.
Expected
P
df .15 .10 .05
(9 -14.40) 2 (8 -12.00) 2 (12 - 9.60) 2 4 6.74 7.78 9.49
c =
2
+ +
14.40 12.00 9.60 5 8.12 9.24 11.07
(15 - 8.40) 2 (10 - 7.80) 2 (6 - 7.80) 2 6 9.45 10.64 12.59

+ + +
8.40 7.80 7.80
The value c 2 = 10.180 falls between the critical values 9.24 and 11.07. The
corresponding areas in the right tail of the chi - square distributi on with df = 5
are 0.10 and 0.05.
c 2 = 2.025 + 1.333 + 0.600 + 5.186 + 0.621 + 0.415 Since our P-value is between 0.05 and 0.10, it is greater than α = 0.05.
Therefore, we fail to reject H0. We don’t have sufficient evidence to
= 10.180 conclude that the company’s claimed color distribution is incorrect.
3
Theory: The Chi-Square Distributions and

P-Values
Conditions for Chi-Square
• The sampling distribution of
Chi-squared statistic is not • Random: Samples must be randomly taken.
Normal.
• Large Sample Size: All expected counts must
• It is a right-skewed
distribution that allows only be at least 5.
positive values because x2 can • Independent: All observations are
never be negative.
• When the expected counts independent. When sampling without
are all at least 5, the sampling replacement, the sample size must be less
distribution of x2 statistic is than 10% of the population size.
close to a chi-square
distribution with degrees of
freedom equal to the number
of categories minus 1.
Using your TI-Nspire: 3. Menu, Statistics, Stat Tests, 7: X2 GOF

Enter info:
1. Enter the observed and expected data in the
spreadsheet.
2. Label the 1st column: mmobs and 2nd column:
mmexp.
4
4. Results When Were You Born?

Are births evenly distributed across the days of the week?
The one-way table below shows the distribution of
births across the days of the week in a random sample
of 140 births from local records in a large city. Do these
data give significant evidence that local births are not
equally likely on all days of the week?
Day Sun Mon Tue Wed Thu Fri Sat

Births 13 23 24 20 27 18 15
Parameters & Hypothesis: Assess Conditions:

We want to perform a test of • Random The data came from a random sample of local
H0: Birth days in this local area are evenly distributed across births.
the days of the week. • Large Sample Size Assuming H0 is true, we would expect
Ha: Birth days in this local area are not evenly distributed one-seventh of the births to occur on each day of the week.
across the days of the week. For the sample of 140 births, the expected count for all 7
The null hypothesis says that the proportions of births are days would be 1/7(140) = 20 births. Since 20 ≥ 5, this
the same on all days. In that case, all 7 proportions must be condition is met.
1/7. So we could also write the hypotheses as • Independent Individual births in the random sample
H0: pSun = pMon = pTues = . . . = pSat = 1/7. should occur independently (assuming no twins). Because
Ha: At least one of the proportions is not 1/7. we are sampling without replacement, there need to
be at least 10(140) = 1400 births in the local area. This
We will use α = 0.05. should be the case in a large city.
5
Name Test & (Calculate) Test Statistic Obtain P Value

Name Test: Chi-Square goodness-of-fit test. P-value = 0.269.
Test statistic :
c2 = å
Expected
Chi - Square = 7.60
Make a Decision & State Conclusion Inherited Traits

Biologists wish to cross pairs of tobacco plants having genetic
Because the P-value, 0.269, is greater than α = 0.05, we makeup Gg, indicating that each plant has one dominant gene
fail to reject H0. These 140 births don’t provide enough (G) and one recessive gene (g) for color. In other words, the
evidence to say that all local births in this area are not biologists predict that 25% of the offspring will be green, 50%
evenly distributed across the days of the week. will be yellow-green and 25% will be albino.
Do these data differ significantly from what the biologists have

predicted? Carry out an appropriate test at the α = 0.05 level to
help answer this question.
6
Parameters & Hypothesis: Assess Conditions

H0: The biologists’ predicted color distribution for tobacco • Random The data came from a random sample of local
plant offspring is correct. births.
That is, pgreen = 0.25, pyellow-green = 0.5, palbino = 0.25 • Large Sample Size We check that all expected counts
Ha: The biologists’ predicted color distribution isn’t are at least 5. Assuming H0 is true, the expected counts
correct. That is, at least one of the stated proportions is for the different colors of offspring are green: (0.25)(84)
incorrect. = 21; yellow-green: (0.50)(84) = 42; albino: (0.25)(84) =
We will use α = 0.05. 21
• Independent Individual offspring inherit their traits
independently from one another. Since we are sampling
without replacement, there would need to be at least
10(84) = 840 tobacco plants in the population. This
seems reasonable to believe.
Name Test, (Calculate) Test Statistic & Make a Decision & State Conclusion
Obtain P-value Because the P-value, 0.0392, is less than α = 0.05, we will
reject H0. We have convincing evidence that the biologists’
Name: Chi-square goodness-of-fit test. hypothesized distribution for the color of tobacco plant
offspring is incorrect.
c2 = å
Expected
(23 - 21) 2 (50 - 42) 2 (11 - 21) 2

= + +
21 50 21
= 6.476
7
WARNING!!!! 11.2 Inference for

In order to perform a Chi- Relationships
Square Test, the data must
be counts!!!!

Inference for Relationships
ü COMPUTE expected counts, conditional distributions, and Input/Desired Result Test Required
contributions to the chi-square statistic
ü CHECK the Random, Large sample size, and Independent conditions One Categorical Variable Chi-Square Goodness of Fit
before performing a chi-square test (GOF)
ü PERFORM a chi-square test for homogeneity to determine whether
the distribution of a categorical variable differs for several populations Two Categorical Variables Chi-Square Test for
or treatments
Homogeneity
ü PERFORM a chi-square test for association/independence to
determine whether there is convincing evidence of an association
between two categorical variables Study Relationship between Chi-Square Test for
Two Categorical Variables Association/Independence
ü EXAMINE individual components of the chi-square statistic as part of a
follow-up analysis
ü INTERPRET computer output for a chi-square test based on a two-way
table
8
Introduction Expected Counts and the Chi-Square

The two-sample z procedures of Chapter 10 allow us to Statistic
compare the proportions of successes in two populations The problem of how to do many comparisons at once with
or for two treatments. an overall measure of confidence in all our conclusions is
What if we want to compare more than two samples or common in statistics. This is the problem of multiple
groups? comparisons. Statistical methods for dealing with multiple
What if we want to compare the distributions of a single comparisons usually have two parts:
categorical variable across several populations or
– 1. An overall test to see if there is good evidence of any
treatments?
differences among the parameters that we want to
compare.
– 2. A detailed follow-up analysis to decide which of the
parameters differ and to estimate how large the
differences are.
Example: Comparing Conditional Distributions

Market researchers suspect that background music may affect the mood
and buying behavior of customers. One study in a supermarket compared
three randomly assigned treatments: no music, French accordion music,
and Italian string music. Under each condition, the researchers recorded
the numbers of bottles of French, Italian, and other wine purchased.
Here is a table that summarizes the data:
Are the distributions of wine purchases under the three music

treatments similar or different?
9
Hypothesis: Expected Counts and the Chi-

H0: There is no difference in the distribution of a categorical Square Statistic
variable for several populations or treatments. • To find the expected counts, we start by assuming that
H0 is true. We can see from the two-way table that 99 of
Ha: There is a difference in the distribution of a categorical the 243 bottles of wine bought during the study were
variable for several populations or treatments. French wines.
• If the specific type of music that’s playing has no effect on
wine purchases, the proportion of French wine sold under
each music condition should be 99/243 = 0.407.
Expected Counts and the Chi-Square Statistic

The expected count in any cell of a two-way table when H0 is Finding the expected counts is not that difficult, as the following example
true is The overall proportion of French wine bought during the study was 99/243 =
illustrates.
row total × column total 0.407. So the expected counts of French wine bought under each treatment
expected count = are: null hypothesis in the wine and music experiment is that there’s no
table total The
difference
No music : in the
99
× 84 =distribution
34.22 Frenchof music
wine: purchases
99 in the
× 75 = 30.56 store
Italian when
music : no× 84
99 music,
= 34.22
French accordion
243 music, or Italian string
243 music is played. 243
To find the expected counts, we start by assuming that H0 is true. We can see
from the two-way table that 99 of the 243 bottles of wine bought during the
The overall proportion of Italian wine bought during the study was 31/243 =
study were French wines.
0.128. So the expected counts of Italian wine bought under each treatment
Ifare:
the specific type of music that’s playing has no effect on wine purchases,
the proportion31 of French wine sold under 31 each music condition should 31 be
No music : × 84 = 10.72 French music : × 75 = 9.57 Italian music : × 84 = 10.72
99/243 = 0.407.
243 243 243
The overall proportion of Other wine bought during the study was 113/243 =
0.465. So the expected counts of Other wine bought under each treatment
are:
113 113 113
No music : × 84 = 39.06 French music : × 75 = 34.88 Italian music : × 84 = 39.06
243 243 243
10
Calculating The Chi-Square Statistic Calculating The Chi-Square Statistic

The tables below show the observed and expected counts for the For the French wine with no music, the observed count is 30 bottles and the
wine and music experiment. expected count is 34.22. The contribution to the c2 statistic for this cell is
(Observed - Expected)2 (30 - 34.22) 2

= = 0.52
Expected 34.22
The c 2 statistic is the sum of nine such terms:
(Observed - Expected)2 (30 - 34.22) 2 (39 - 30.56) 2 (35 - 39.06) 2

c2 =å = + + ... +
Expected 34.22 30.56 39.06
= 0.52 + 2.33 + ... + 0.42 = 18.28
Does Music Influence Purchases? Follow-up Analysis

We use df = 4 because: P
(wine -1)(music -1) = df .0025 .001
(3-1)(3-1) = 4 4 16.42 18.47
The small P-value gives us convincing evidence to reject H0 and

conclude that there is a difference in the distributions of wine
purchases at this store when no music, French accordion music,
or Italian string music is played.
Furthermore, the random assignment allows us to say that the

difference is caused by the music that’s played. Looking at the output, we see that just two of the nine components that make up
the chi-square statistic contribute about 14 (almost 77%) of the total χ2 = 18.28.
We are led to a specific conclusion: sales of Italian wine are strongly affected by
Italian and French music.
11
Hypothesis: Chi-Square Test for Conditions: Chi-Square Statistic

Homogeneity
üRandom: Random assignment
H0: There is no difference in the distribution of a categorical üLarge Sample Size: All the expected counts must be at
variable for several populations or treatments. least 5. (Use expected matrix on calculator and copy)
üIndependent:
Ha: There is a difference in the distribution of a categorical
variable for several populations or treatments. üIndividual observations (studied groups are not
connected)
üLess than 10% of population
Using Your TI- Nspire: Using your TI-Nspire:

1. Create matrix. Menu, 7: Matrix, 1: Create 4. Menu, 6: Statistics, 7: Stat Test, 8: X2 2-
way. Then, type variable name.
2. Fill in data.
3. Store matrix. Blue “ctrl” button, “var” and
then type name of the table.
12
Using Your TI-Nspire: Cell-Only Telephone Users

Random digit dialing telephone surveys used to exclude cell
To see the expected counts: phone numbers. If the opinions of people who have only cell
Press Vars. phones differ from those of people who have landline service, the
poll results may not represent the entire adult population. The
Select: stat. expmatrix Pew Research Center interviewed separate random samples of
cell-only and landline telephone users who were less than 30
years old. Here’s what the Pew survey found about how these
people describe their political party affiliation.
Parameters & Hypothesis Assess Conditions

H0: There is no difference in the distribution of party • Random The data came from separate random samples of 96
affiliation in the cell-only and landline populations. cell-only and 104 landline users.
• Large Sample Size: Since all expected counts are greater than
Ha: There is a difference in the distribution of party 5, we met the condition. (Use your calculator!! Then list counts
affiliation in the cell-only and landline populations. in a table.)
• Independent Researchers took independent samples of cell-

only and landline phone users. Sampling without replacement
was used, so there need to be at least 10(96) = 960 cell-only
users under age 30 and at least 10(104) = 1040 landline users
under age 30. This is safe to assume.
13
Name Test & (Calculate) Test Statistic Obtain p-value, Make a Decision & State
Since the conditions are satisfied, we can a perform chi-
Conclusion
test for homogeneity.
Test statistic :
c2 = å
Expected
USE
CALCULATOR!!! Because the P-value, 0.20, is greater than α = 0.05, we fail to
reject H0. There is not enough evidence to conclude that the
= 3.22
distribution of party affiliation differs in the cell-only and
landline user populations.
Cocaine Addiction is Hard to Break Parameters & Hypothesis

Cocaine addicts need cocaine to feel any pleasure, so
perhaps giving them an antidepressant drug will help. A H0: there is no difference in the relapse rate for the three
three-year study with 72 chronic cocaine users treatments.
compared an antidepressant drug called desipramine
with lithium (a standard drug to treat cocaine addiction) Ha: there is a difference in the relapse rate for are different
and a placebo. One-third of the subjects were randomly the three treatments.
assigned to receive each treatment. Here are the results:
14
Name Test, (Calculate) Test Statistic &

Assess Conditions Obtain P-value
• Random: The subjects were randomly assigned to the
treatment groups. Name: Chi-test for homogeneity
• Large Sample Size: All the expected counts are ≥ 5 so the
condition is met. (Use expected matrix; must copy)
• Independent : The random assignment helps create three

independent groups. If the experiment is conducted properly,
then knowing one subject’s relapse status should give us no
information about another subject’s outcome. So individual
observations are independent.
Make a Decision and State Conclusion The Chi-Square Test for

Association/Independence
Because the P-value, 0.0052, is less than α = 0.01, we reject
H0. We have sufficient evidence to conclude that the true Another common situation that leads to a two-way table
relapse rates for the three treatments are not all the same. is when a single random sample of individuals is chosen
from a single population and then classified according to
two categorical variables. In that case, our goal is to
analyze the relationship between the variables.
15
More About the Chi-Square Test for The Chi-Square Test for
Association/Independence Association/Independence
Hypothesis:
We often gather data from a random sample and arrange
H0: There is no association between two categorical
them in a two-way table to see if two categorical variables
are associated. The sample data are easy to investigate: variables in the population of interest.
turn them into percents and look for a relationship Ha: There is an association between two categorical
between the variables. variables in the population of interest.
Our null hypothesis is that there is no association between Or, alternatively

the two categorical variables. The alternative hypothesis is H0: Two categorical variables are independent in the
that there is an association between the variables. For the population of interest.
observational study of anger level and coronary heart Ha: Two categorical variables are not independent in
disease, we want to test the hypotheses
the population of interest.
The Chi-Square Test for Association/Independence Background: Angry People and Heart
A study followed a random sample of 8474 people with normal Disease
blood pressure for about four years. All the individuals were
free of heart disease at the beginning of the study. Each person
took the Spielberger Trait Anger Scale test, which measures
how prone a person is to sudden anger. Researchers also
recorded whether each individual developed coronary heart
disease (CHD). This includes people who had heart attacks and
those who needed medical treatment for heart disease. Here is
a two-way table that summarizes the data:
There is a clear trend: as the anger score

increases, so does the percent who suffer heart
disease. A much higher percent of people in the
high anger category developed CHD (4.27%)
than in the moderate (2.33%) and low (1.70%)
anger categories.
16

• Random The data came from a random sample of 8474
H0: There is no association between anger level and people with normal blood pressure.
heart disease in the population of people with normal • Large Sample Size All the expected counts are at least 5, so
blood pressure. this condition is met.
• Independent Knowing the values of both variables for one
Ha: There is an association between anger level and person in the study gives us no meaningful information about
heart disease in the population of people with normal the values of the variables for another person. So individual
blood pressure. observations are independent. Because we are sampling
OR without replacement, we need to check that the total number
of people in the population with normal blood pressure is at
H0: Anger and heart disease are independent in the least 10(8474) = 84,740. This seems reasonable to assume.
population of people with normal blood pressure.
Ha: Anger and heart disease are not independent in the

population of people with normal blood pressure.
Name Test & (Calculate) Test Statistic Obtain P-value, Make a Decision and
Name: Chi-test for association/independence. State Conclusion
P-Value:
0.00032.
Because the P-value is clearly less than α = 0.01, we reject

H0 and conclude that anger level and heart disease are
associated in the population of people with normal blood
pressure.
17
Which Chi-Square Test Do I Use?!?! Which Chi-Square Test Do I Use?!?!

Both the chi-square test for homogeneity and the chi-
Instead of focusing on the question asked, it’s much
square test for association/independence start with a
two-way table of observed counts. They even calculate easier to look at how the data were produced.
the test statistic, degrees of freedom, and P-value in the üIf the data come from two or more independent
same way. The questions that these two tests answer random samples or treatment groups in a randomized
are different, however. experiment, then do a chi-square test for
§A chi-square test for homogeneity tests whether the homogeneity.
distribution of a categorical variable is the same for each üIf the data come from a single random sample, with
of several populations or treatments. the individuals classified according to two categorical
variables, use a chi-square test for
§The chi-square test for association/independence
association/independence.
tests whether two categorical variables are associated
in some population of interest.
Chi- Square Test Key Elements
Goodness of Fit One variable, one population
One variable, two or more

Homogeneity populations/groups
Association/Independence Two variables, one population
18
Chapter 12
Section 12.1
12.1: Inference for Linear Inference for Linear Regression
Regression ü CHECK conditions for performing inference about the
slope β of the population regression line
ü CONSTRUCT and INTERPRET a confidence interval for the
slope β of the population regression line
ü PERFORM a significance test about the slope β of a
population regression line
ü INTERPRET computer output from a least-squares
regression analysis
Introduction Review: General Form of Linear

When a scatterplot shows a linear relationship
between a quantitative explanatory variable x and a Regression Equation
quantitative response variable y, we can use the least-
squares line fitted to the data to predict y for a given
value of x. If the data are a random sample from a
larger population, we need statistical inference to
y = a + bx
answer questions like these:
• Is there really a linear relationship between x and y in y = dependent variable
the population, or could the pattern we see in the
scatterplot plausibly happen just by chance?
a = intercept
b = slope
• In the population, how much will the predicted value of x = independent variable
y change for each increase of 1 unit in x? What’s the
margin of error for this estimate?
1
Chapter 12
Review: Interpreting Computer Review: Interpret s, r2 and standard error

Regression Output of slope
“Standard Error”
s : The typical error when using the regression line Theory: Inference for Linear Regression
to predict calorie consumption is about 23.4
calories. The least-squares regression line for this population of data has
been added to the graph. It has slope 10.36 and y-intercept
33.97. We call this the population regression line (or true
r2: Approximately 42.1% of the variation in calorie regression line) because it uses all the observations that month.
consumption can be explained by the linear
Suppose we take an SRS of 20
relationship with the time spent at the table. eruptions from the population and
calculate the least - squares
regression line yˆ = a + bx for the
Standard Error of Slope: If samples like this were sample data. How does the slope
observed many times, the typical distance that of the sample regression line
(also called the estimated
the estimated slope would differ from the regression line) relate to the slope
population slope by an average of 0.8498. of the population regression line?
2
Chapter 12
Sampling Distribution of b Sampling Distribution of b

The figures below show the results of taking three different SRSs of 20 Old Confidence intervals and significance tests about the slope
Faithful eruptions in this month. Each graph displays the selected points and of the population regression line are based on the sampling
the LSRL for that sample.
distribution of b, the slope of the sample regression line.
Fathom software was used to simulate choosing 1000 SRSs of
n = 20 from the Old Faithful data, each time calculating the
equation of the LSRL for the sample. The values of the slope b
for the 1000 sample regression lines are plotted. Describe this
approximate sampling distribution of b.
Shape: We can see that the distribution of b-

values is roughly symmetric and unimodal. A
Normal probability plot of these sample
Notice that the slopes of the sample regression regression line slopes suggests that the
approximate sampling distribution of b is close
lines – 10.2, 7.7, and 9.5 – vary quite a bit from
the slope of the population regression line, to Normal.
10.36.
Center: The mean of the 1000 b- Spread: The standard deviation of the 1000 b-
The pattern of variation in the slope b is values is 10.32. This value is quite values is 1.31. Later, we will see that the
described by its sampling distribution. close to the slope of the population standard deviation of the sampling distribution
(true) regression line, 10.36. of b is actually 1.30.
Conditions for Regression Inference How to Check the Conditions for

L I N E R Inference
• Linear The (true) relationship between x and y is linear. For • Linear Examine the scatterplot to check that the overall
any fixed value of x, the mean response µy falls on the L pattern is roughly linear. Look for curved patterns in the
population (true) regression line µy= α + βx. The slope b and residual plot. Check to see that the residuals center on the
intercept a are usually unknown parameters. “residual = 0” line at each x-value in the residual plot.
• Independent Individual observations are independent of each
I • Independent Look at how the data were produced. Random
other.
sampling and random assignment help ensure the
• Normal For any fixed value of x, the response y varies
independence of individual observations. If sampling is done
according to a Normal distribution.
without replacement, remember to check that the
• Equal variance The standard deviation of y (call it σ) is the
population is at least 10 times as large as the sample (10%
same for all values of x. The common standard deviation σ is
condition).
usually an unknown parameter.
• Random The data come from a well-designed random sample
or randomized experiment.
3
Chapter 12
How to Check the Conditions for TI-Nspire: Residuals Plot

Inference, cont.
1. Enter data in list.
• Normal: Make a stemplot, histogram, or Normal
N
probability plot of the residuals and check for clear
2. Create scatterplot.
skewness or other major departures from Normality. 3. Run Regression. Menu, 4: Analyze, 6:
• Equal variance: Look at the scatter of the residuals above Regression.
E
and below the “residual = 0” line in the residual plot. The 4. Menu, 4: Analyze, 7: Residuals, 2: Show
amount of scatter should be roughly the same from the
smallest to the largest x-value.
Residual Plot.
R • Random: See if the data were produced by random
sampling or a randomized experiment.
TI-Nspire: Histogram of Residuals

1. Enter data in list. Then, create scatterplot.
2. Menu, 4: Analyze, 7: Residuals, 2: Show
Residual Plot.
4
Chapter 12
TI-Nspire: Histogram of Residuals TI-Nspire: Normal Probability Plot

1. Enter data in list. 1. Enter data in list.
2. Create scatterplot. 2. Create scatterplot.
3. Run Regression. Menu, 4: Analyze, 6: 3. Run Regression. Menu, 4: Analyze, 6:
Regression. Regression.
4. Click purple graph. 4. Click purple graph.
5. Select “stat.resid” as x-axis. 5. Select “stat.resid” for x-axis.
6. Menu, 1: Plot Type, 4: Normal Probability
Plot
Practice Checking Conditions: The Helicopter Experiment The Helicopter Experiment

Students randomly assigned 14 helicopters to each of five drop heights: ü Linear The scatterplot shows a clear linear form. For each drop height
152 centimeters (cm), 203 cm, 254 cm, 307 cm, and 442 cm. Teams of used in the experiment, the residuals are centered on the horizontal line
students released the 70 helicopters in a predetermined random order at 0. The residual plot shows a random scatter about the horizontal line.
and measured the flight times in seconds. The class used Minitab to ü Independent Because the helicopters were released in a random
carry out a least-squares regression analysis for these data. A order and no helicopter was used twice, knowing the result of one
scatterplot, residual plot and histogram. observation should give no additional information about another
observation.
ü Normal The histogram of the residuals is single-peaked, unimodal,
and somewhat bell-shaped.
ü Equal variance The residual plot shows a similar amount of scatter
about the residual = 0 line for the 152, 203, 254, and 442 cm drop
heights. Flight times (and the corresponding residuals) seem to vary
more for the helicopters that were dropped from a height of 307 cm.
ü Random The helicopters were randomly assigned to the five
possible drop heights.
Except for a slight concern about the equal-variance condition, we should be safe
performing inference about the regression model in this setting.
5
Chapter 12
Estimating the Parameters The Helicopter Experiment

ˆ
We use y = a + bx to estimate µ y = a + bx
• The slope b of the LSRL is an unbiased estimator of

the true slope b.
• The intercept a of the LSRL is an unbiased estimator The least - squares regression line for these data is
of the true intercept a.

flight time = -0.03761 + 0.0057244(drop height)
• The standard error s is an unbiased estimator of the
true standard deviation of y (sy).
å (y - yˆ)
2
Remember “S” is
standard error and s = = å residuals 2
formula is : (n - 2) n -2
The Helicopter Experiment: The Helicopter Experiment:

Interpreting Values Interpreting Values
• Because b = 0.0057244 estimates the unknown β, we
estimate that, on average, flight time increases by about • We need the intercept a = -0.03761 to draw the line and
0.0057244 seconds for each additional centimeter of drop make predictions, but it has no statistical meaning in this
height. example. No helicopter was dropped from less than 150 cm,
so we have no data near x = 0.
• Our estimate for the standard deviation σ of flight times
• We might expect the actual y-intercept α of the true
about the true regression line at each x-value is s = 0.168
regression line to be 0 because it should take no time for a
seconds.
helicopter to fall no distance.
• This is also the size of a typical prediction error if we use the
• The y-intercept of the sample regression line is -0.03761,
least-squares regression line to predict the flight time of a
which is pretty close to 0.
helicopter from its drop height.
6
Chapter 12
Helicopter Experiment
Constructing a Confidence Interval for the
Construct and interpret a 95% confidence interval for the
Slope slope of the population regression line.
Because we use the statistic b as our estimate, the
confidence interval is
b ± t* SEb
We call this a t interval for the slope.
In this formula, the standard error of the slope is

s
SE b =
sx n -1
and t* is the critical value for the t distribution with df = n - 2
having area C between -t* and t*.
Helicopter Experiment Does Fidgeting Keep you Slim?

Recently a study investigated why some people don’t gain
weight even when they overeat. Perhaps fidgeting and other
“nonexercise activity” (NEA) explains why. Researchers
deliberately overfed a random sample of 16 healthy young
SEb = 0.0002018, from the “SE Coef ” column in the computer output. adults for 8 weeks. They measured fat gain (in kilograms) and
df = n - 2 = 70 - 2 = 68. Using the more conservative df = 60 from Table B gives t* = change in energy use (in calories) from activity other than
2.000. deliberate exercise:
The 95% confidence interval is
b ± t* SEb = 0.0057244 ± 2.000(0.0002018)
= 0.0057244 ± 0.0004036
= (0.0053208, 0.0061280)
We are 95% confident that the interval from 0.0053208 to

0.0061280 seconds per cm captures the slope of the true
regression line relating the flight time y and drop height x of
Construct and interpret a 90% confidence interval for
paper helicopters.
the slope of the population regression line.
7
Chapter 12
Does Fidgeting Keep you Slim? Does Fidgeting Keep you Slim?
Assess Conditions:
• Linear The scatterplot shows a clear linear pattern. Also, the residual plot
shows a random scatter of points about the “residual = 0” line.
Parameter:
β = slope of true regression
line relating fat gain and NEA.
• Independent Individual observations of fat gain should be independent if

the study is carried out properly. Because researchers sampled without
replacement, there have to be at least 10(16) = 160 healthy young adults in
the population of interest.
• Normal The histogram of the residuals is roughly symmetric and single-
peaked, so there are no obvious departures from normality.
• Equal variance It is hard to tell from so few points whether the scatter of
points around the residual = 0 line is about the same at all x-values.
• Random The subjects in this study were randomly selected to participate.
Does Fidgeting Keep you Slim? Performing a Significance Test for the Slope
Name the Interval: T-Interval for β We can also perform a significance test to determine whether
a specified value of β is plausible. The null hypothesis has the
We use the t distribution with 16 - 2 = 14 degrees of freedom to find general form H0: β = hypothesized value. To do a test,
the critical value. For a 90% confidence level, the critical value is t* = standardize b to get the test statistic:
1.761. So the 90% confidence interval for β is
b ± t* SEb = −0.0034415 ± 1.761(0.0007414) test statistic =
= −0.0034415 ± 0.0013056
= (−0.004747,−0.002136)
b - b0
t=
SEb
Interval: We are 90% confident that the interval from -0.004747 to To find the P-value, use a t distribution with n - 2 degrees of
-0.002136 kg captures the actual slope of the population regression freedom. Here are the details for the t test for the slope.
line relating NEA change to fat gain for healthy young adults.
8
Chapter 12
Crying and IQ
Infants who cry easily may be more easily stimulated than others. This may be
a sign of higher IQ. Child development researchers explored the relationship
between the crying of infants 4 to 10 days old and their later IQ test scores. A
snap of a rubber band on the sole of the foot caused the infants to cry. The
researchers recorded the crying and measured its intensity by the number of
peaks in the most active 20 seconds. They later measured the children’s IQ at
age three years using the Stanford-Binet IQ test. A scatterplot and Minitab
output for the data from a random sample of 38 infants is below.
Do these data provide convincing evidence that there is a positive linear

relationship between crying counts and IQ in the population of infants?

• Linear The scatterplot suggests a moderately weak positive linear relationship
H0 : β = 0 between crying peaks and IQ. The residual plot shows a random scatter of points
about the residual = 0 line.
Ha : β > 0
β = true slope of the population regression line relating
crying count to IQ score.
No significance level was given, so we’ll use α = 0.05.

• Independent Later IQ scores of individual infants should be independent. Due
to sampling without replacement, there have to be at least 10(38) = 380 infants in
the population from which these children were selected.
• Normal The Normal probability plot of the residuals shows a slight curvature,
which suggests that the responses may not be Normally distributed about the
line at each x-value. With such a large sample size (n = 38), however, the t
procedures are robust against departures from Normality.
• Equal variance The residual plot shows a fairly equal amount of scatter around
the horizontal line at 0 for all x-values.
• Random We are told that these 38 infants were randomly selected.
9
Chapter 12
Name Test, (Calculate) Test Statistic & Obtain Make Decision & State Conclusion
P-value
T-Test for the slope of B The P-value, 0.002, is less than our α = 0.05 significance level, so
b - b0 1.4929 - 0
we have enough evidence to reject H0 and conclude that there is
t= = = 3.07 a positive linear relationship between intensity of crying and IQ
SE b 0.4870
score in the population of infants.
The Minitab output gives P =

0.004 as the P-value for a two-
sided test. The P-value for the
one-sided test is half of this,
P = 0.002.
10
Chapter 12
Conditions for Regression Inference The Sampling Distribution of b

Let’s return to our earlier exploration of Old Faithful eruptions. For all 222 eruptions in a
The figure below shows the regression model when the single month, the population regression line for predicting the interval of time until the
next eruption y from the duration of the previous eruption x is µy = 33.97 + 10.36x. The
conditions are met. The line in the figure is the standard deviation of responses about this line is given by σ = 6.159.
population regression line µy= α + βx. The Normal curves show
If we take all possible SRSs of 20 eruptions from
how y will vary when x is
For each possible value held fixed at different values.
the population, we get the actual sampling
of the explanatory All the curves have the same distribution of b.
variable x, the mean of standard deviation σ, so the
the responses µ(y | x) variability of y is the same for Shape: Normal
moves along this line. all values of x.
Center : µb = β = 10.36 (b is an unbiased
estimator of β)
s 6.159
Spread : s b = = = 1.30
sx n -1 1.083 20 -1
The value of σ determines In practice, we don’t know σ for the population regression line. So we
whether the points fall close
to the population regression
estimate it with the standard deviation of the residuals, s. Then we estimate
line (small σ) or are widely the spread of the sampling distribution of b with the standard error of the
scattered (large σ). slope: s
SE b =
sx n -1
The Sampling Distribution of b

What happens if we transform the values of b by standardizing? Since the
sampling distribution of b is Normal, the statistic
b-b
z=
sb
has the standard Normal distribution.
Replacing the standard deviation σb of the sampling distribution with its
standard error gives the statistic b-b
t=
SE b
which has a t distribution with n - 2 degrees of freedom.
The figure shows the result of

standardizing the values in the sampling
distribution of b from the Old Faithful
example. Recall, n = 20 for this example.
The superimposed curve is a t

distribution with df = 20 – 2 = 18.
11
Mrs. Daniel- AP Stats
1.1 WS
1. What variables are measured? Identify each as categorical or quantitative. In what units were the quantitative
variables measured?
Marital
State Number of Family Members Age Gender Total Income Travel time to work
Status
Kentucky 2 61 Female Married 21000 20
Florida 6 27 Female Married 21300 20
Wisconsin 2 27 Male Married 30000 5
California 4 33 Female Married 26000 10
Michigan 3 49 Female Married 15100 25
Virginia 3 26 Female Married 25000 15
Pennsylvania 4 44 Male Married 43000 10
Virginia 4 22 Male Never married/ single 3000 0
California 1 30 Male Never married/ single 40000 15
New York 4 34 Female Separated 30000 40
2. A sample of 200 children from the United Kingdom ages 9-17 was selected from the CensusAtSchool website
(www.censusatschool.com). The gender of each student was recorded along with which super power they would most
like to have: invisibility, super strength, telepathy (ability to read minds), ability to fly, or ability to freeze time. Here are
the results:
Female Male Total

Invisibility 17 13 30
Super Strength 3 17 20
Telepathy 39 5 44
Fly 36 18 54
Freeze Time 20 32 52
Total 115 85 200
a. What proportion of males want the power of invisibility?
b. What proportion of females want the power of freeze time?
c. What proportion of children that want the power of telepathy are male?
d. What proportion of children that want the power of fly are female?
Female Male Total
Invisibility 17 13 30
Super Strength 3 17 20
Telepathy 39 5 44
Fly 36 18 54
Freeze Time 20 32 52
Total 115 85 200
3. Create a well labeled segmented bar graph of the marginal distributions of power preference and gender.
Be sure to include a key.
Female:
Males:
Key:
4. Based on the graphs above, can we conclude that boys and girls differ in their preference of superpower? Give
appropriate evidence to support your answer.
1.2 WS
Smart Phone Battery Life

Smart Phone Battery Life (minutes)
Apple iPhone 300
Motorola Droid 385
Palm Pre 300 Collection 1 of the data:
Here is a dotplot Dot Plot
Blackberry Bold 360

Blackberry Storm 330
Motorola Cliq 360
Samsung Moment 330
300 340 380 420 460
Blackberry Tour 300
HTC Droid 460 BatteryLife (minutes)
1. Describe the shape, center, and spread of the distribution. Are there any (potential) outliers?
Dotplot of EnergyCost vs Type
Top vs. Bottom Freezers

How do the annual energy costs (in dollars) compare for refrigerators with top freezers and refrigerators with
bottom freezers? The data below is from the May 2015 issue of Consumer Reports.
Type
bottom
top
56 70 84 98 112 126 140
EnergyCost
2. Compare the distributions of annual energy costs for these two types of refrigerators.
Who’s Taller?
Which gender is taller? A sample of 14-year-olds from the United Kingdom was randomly selected using the
CensusAtSchool website.
Here are the heights of the students (in cm):
Male: 154, 157, 187, 163, 167, 159, 169, 162, 176, 177, 151, 175, 174, 165, 165, 183, 180
Female: 160, 169, 152, 167, 164, 163, 160, 163, 169, 157, 158, 153, 161, 165, 165, 159, 168, 153, 166, 158, 158,
166
Here is a back-to-back stemplot comparing male and female heights:
Female Male
332 15 14
98887 15 79 Key: 15|1 represents a
433100 16 23 student who is 151 cm tall.
99876655 16 5579
17 4
17 567
18 03
18 7
3. Compare the distributions of height for females and males.

1.3 WS

McDonald’s Beef Sandwiches
Here are data for the amount of fat (in grams) for McDonald’s beef sandwiches:

Sandwich Fat (g)
Hamburger 9 g
Cheeseburger 12 g
Double Cheeseburger 23 g
McDouble 19 g
Quarter Pounder® 19 g
Quarter Pounder® with Cheese 26 g
Double Quarter Pounder® with Cheese 42 g
Big Mac® 29 g
Big N' Tasty® 24 g
Big N' Tasty® with Cheese 28 g
Angus Bacon & Cheese 39 g
Angus Deluxe 39 g
Angus Mushroom & Swiss 40 g
McRib ® 26 g
Mac Snack Wrap 19 g

Use your graphing calculator to find the following:

Mean

Median

5 Number Summary

IQR

Are there any outlier/s
using the IQR*1.5 Rule?

The Previous Home Run King

Using your graphing calculator, create a box plot using the data below. Be sure to identify each number in a five number
summary and any outliers using the IQR*1.5 Rule.

Number of home runs that Hank Aaron hit in each of his 23 seasons:
13 27 26 44 30 39 40 34 45 44 24 32 44 39 29 44 38 47 34 40 20 12 10

Who Has More Contacts—Males or Females?
The following data show the number of contacts that a sample of high school students had in their cell phones. Do the
data give convincing statistical evidence that one gender has more contacts than the other? You need both graphical
and numerical evidence.

Male: 124 41 29 27 44 87 85 260 290 31 168 169 167 214 135 114 105 103 96 144
Female: 30 83 116 22 173 155 134 180 124 33 213 218 183 110

2.1 WS
1. The stemplot below shows the number of wins for each of the 30 Major League Baseball teams in 2009.
5 9
6 2455 Key: 5|9 represents a
7 00455589 team with 59 wins.
8 0345667778
9 123557
10 3
Find the percentiles for the following teams:
(a) The Colorado Rockies, who won 92 games. (b) The New York Yankees, who won 103 games.
(c) The Kansas City Royals and Cleveland Indians, who both won 65 games.
2. Here is a table showing the distribution of median household incomes for the 50 states and the District of Columbia.
Calculate the relative frequency and cumulative relative frequency.
Median Cumulative
Relative Cumulative
Income Frequency Relative
Frequency Frequency
($1000s) Frequency
35 to < 40 1 1
40 to < 45 10 11
45 to < 50 14 25
50 to < 55 12 37
55 to < 60 5 42
60 to < 65 6 48
65 to < 70 3 51
3. Use the cumulative relative frequency graph for the state income data to answer each question.
(a) At what percentile is California, with a median income of $57,445?
(b) Estimate and interpret the first quartile of this solution.
4. Miami-Dade County Public Schools employs teachers at salaries between $40,000 and $71,000. The teachers’ union
and the school board are negotiating the form of next year’s increase in the salary schedule.
(a) If every teacher is given a flat $1000 raise, what will this do to the mean salary? To the median salary? Explain your
answers.
(b) What would a flat $1000 raise do to the extremes and quartiles of the salary distribution? To the standard deviation
of teachers’ salaries? Explain your answers.
2.2 WS
Wins in Major League Baseball
1. In 2009, the mean number of wins was 81 with a standard deviation of 11.4 wins.
Find and interpret the z-scores for the following teams.
(a) The New York Yankees, with 103 wins.
(b) The New York Mets, with 70 wins.
Batting Averages
2. In the previous alternate example about batting averages for Major League Baseball players in 2009, the mean of the
432 batting averages was 0.261 with a standard deviation of 0.034. Suppose that the distribution is exactly Normal with
 = 0.261 and  = 0.034.
(a) Sketch a Normal density curve for this distribution of batting averages. Label the points that are 1, 2, and 3 standard
deviations from the mean.
(b) What percent of the batting averages are above 0.329? Show your work.
(c) What percent of the batting averages are between 0.227 and .295? Show your work.
3. According to the CDC, the heights of 3 year old females are approximately Normally distributed with a mean of 94.5
cm and a standard deviation of 4 cm. Be sure to draw curves for each calculation!
(a) What is the third quartile of this distribution?
(b) What is the median of this distribution?
(c) If a 3 year old female was 91.7 cm tall, what percentile would she be in?
(d) If a mother knew her daughter was at the 91st percentile in height, how tall is her daughter?
(e) If a 3 year old female was 96.4 cm tall, what percentile would she be in?
4. Scores on the Wechsler Adult Intelligence Scale (a standard IQ test) for the 20 to 34 age group are approximately
Normally distributed with μ = 110 and σ = 25. For each part, follow the four-step process.
(a) At what percentile is an IQ score of 150?
(b) What percent of people aged 20 to 34 have IQs between 125 and 150?
(c) MENSA is an elite organization that admits as members people who score in the top 2% on IQ tests. What score on
the Wechsler Adult Intelligence Scale would an individual have to earn to qualify for MENSA membership?
3.1 WS
1. The table below shows data for 13 students in a statistics class. Each member of the class ran a 40-yard sprint and
then did a long jump (with a running start).
Sprint Time (s) 5.41 5.05 9.49 8.09 7.01 7.17 6.83 6.73 8.01 5.68 5.78 6.31 6.04
Long Jump Distance (in) 171 184 48 151 90 65 94 78 71 130 173 143 141
A. Create and label a scatterplot of the data:
B. Describe and interpret the scatterplot above.
C. What is the correlation coefficient? What does it mean?

2. A student wonders if tall women tend to date taller men than do short women. She measures herself, her dormitory
roommate, and the women in the adjoining rooms. Then she measures the next man each woman dates. Here are the
data (heights in inches):
A. How would r change if all the men were 6 inches shorter than the heights given in the table? Does the correlation tell
us if women tend to date men taller than themselves?
B. If heights were measured in centimeters rather than inches, how would the correlation change? (There are 2.54
centimeters in an inch.)
3. Consider each of the following relationships:
A. the heights of fathers and the heights of their adult sons

B. the heights of husbands and the heights of their wives
C. the heights of women at age 4 and their heights at age 18.
Rank the correlations between these pairs of variables from highest to lowest. Explain your reasoning.
3.2 WS
1. The following data shows the number of miles driven and advertised price for 11 used Honda CR-Vs from the 2005-
2009 model years (prices found at www.carmax.com). The scatterplot below shows a strong, negative linear association
between number of miles and advertised cost. The line on the plot is the regression line for predicting advertised price
based on number of miles.
Thousand Cost
Miles Driven (dollars)
22 17998
29 16450
35 14998
39 13998
45 14599
49 14988
55 13599
56 14599
69 11998
70 14450
86 10998
A. Calculate the correlation. What does this value mean in plain English? What is the relative strength of the association?
B. What is the least squares regression equation for this association? Define any variables used.
C. Determine the y-intercept of the regression equation and interpret the value in context. Does the value have any real-
world implications?
D. Determine the slope of the regression equation and interpret the value in context. Does the value have any real-
world implications?
2. For a project, two AP Statistics students decided to investigate the effect of sugar on the life of cut flowers. They
went to the local grocery store and randomly selected 12 carnations. All the carnations seemed equally healthy when
they were selected. When they got home, the students prepared 12 identical vases with exactly the same amount of
water in each vase. They put 1 tablespoon of sugar in three vases, 2 tablespoons of sugar in three vases, and 3
tablespoons of sugar in three vases. In the remaining 3 vases, they put no sugar. After the vases were prepared and
placed in the same location, the students randomly assigned one flower to each vase and observed how many hours
each flower continued to look fresh. A scatterplot, residual plot, and computer output from the regression are shown.
Only 10 points appear on the scatterplot and residual plot since there were two observations at (1, 204) and two
observations at (2, 210).
Predictor Coef SE Coef T P

Constant 181.200 3.635 49.84 0.000
Sugar 15.200 1.943 7.82 0.000
S = 7.52596 R-Sq = 86.0% R-Sq(adj) = 84.5%
A. What is the equation of the least-squares line? Be sure to define any variables you use.
B. Is a line an appropriate model for these data? Justify your answer. (You need at least two sentences.)
C. Interpret the value of “s” in the context of this problem.
D. Interpret the value of r2 in the context of this problem.

4.1 WS
1. In May 2015, the Los Angeles City Council voted to ban most travel and contracts with the state of Arizona to protest
Arizona’s new immigration enforcement law. The Los Angeles Times conducted an online poll that asked if the City
Council was right to pass a boycott of Arizona. The results showed that 96% of the 41,068 people in the sample said
“No.” Does this result represent the opinions of all Los Angeles residents? Explain.
2. The manager of a beach-front hotel wants to survey guests in the hotel to estimate overall customer satisfaction. The
hotel has two towers, an older one to the south and a newer one to the north. Each tower has 10 floors of standard
rooms (40 rooms per floor) and 2 floors of suites (20 suites per floor). Half of the rooms in each tower face the beach,
while the other half of the rooms face the street. This means there are (2 towers)(10 floors)(40 rooms) + (2 towers)(2
floors)(20 suites) = 880 total rooms.
A. Explain how to select a simple random sample of 88 rooms.

B. Explain how to select a stratified random sample of 88 rooms.
C. Explain why selecting 2 of the 24 different floors would not be a good way to obtain a cluster sample.
3. Which of the following are sources of sampling error and which are sources of nonsampling error? Explain your
answers.
(a) The subject lies about past drug use.
(b) A typing error is made in recording the data.
(c) Data are gathered by asking people to mail in a coupon printed in a newspaper.
4.2 WS
1. A study published in the New England Journal of Medicine compared two medicines to treat head lice: an oral
medication called ivermectin and a topical lotion containing malathion. Researchers studied 812 people in 376
households in seven areas around the world. Of the 185 randomly assigned to ivermectin, 171 were free from head lice
after two weeks compared to only 151 of the 191 households randomly assigned to malathion.
Identify the experimental units, explanatory and response variables, and the treatments in this experiment.
2. Does adding fertilizer affect the productivity of tomato plants? How about the amount of water given to the plants?
To answer these questions, a gardener plants 24 similar tomato plants in identical pots in his greenhouse. He will add
fertilizer to the soil in half of the pots. Also, he will water 8 of the plants with 0.5 gallons of water per day, 8 of the
plants with 1 gallon of water per day and the remaining 8 plants with 1.5 gallons of water per day. At the end of three
months he will record the total weight of tomatoes produced on each plant.
Identify the explanatory and response variables, experimental units, and list all the treatments.
3. Suppose you have a class of 30 students who volunteer to be subjects in an experiment involving caffeine. Explain
how you would randomly assign 15 students to each of the two treatments.
4. A cell phone company is considering two different keyboard designs (A and B) for its new line of cell phones.
Researchers would like to conduct an experiment using subjects who are frequent texters and subjects who are not
frequent texters. The subjects will be asked to text several different messages in 5 minutes. The response variable will
be the number of correctly typed words.
A. Explain why a randomized block design might be preferable to a completely randomized design for this experiment.
B. Outline a randomized block experiment using 100 frequent texters and 200 novice testers.
5.2 WS
Languages in Canada
Canada has two official languages, English and French. Choose a Canadian at random and ask, “What is your mother
tongue?” Here is the distribution of responses, combining many separate languages from the broad Asia/Pacific region:
(a) What probability should replace “?” in the distribution? Why?
(b) What is the probability that a Canadian’s mother tongue is not English?
(c) What is the probability that a Canadian’s mother tongue is a language other than English or French?
U.S. Senators
The two-way table below describes the members of the U.S Senate in a recent year.
(a) Who are the individuals? What variables are being measured?
(b) If we select a U.S. senator at random, what’s the probability that we choose
i. a Democrat? ii. a female?
iii. a female Democrat? iv. a female or a Democrat?

Phone Usage
According to the National Center for Health Statistics, in December 2008, 78% of US households had a traditional
landline telephone, 80% of households had cell phones, and 60% had both. Suppose we randomly selected a household
in December 2008.
We will define events A: has a landline and B: has a cell phone.

Cell Phone No Cell Phone Total
Landline 0.60 0.18 0.78
No Landline 0.20 0.02 0.22
Total 0.80 0.20 1.00
(a) Construct a Venn diagram to represent the outcomes of this chance process.
(b) Find the probability that the household has at least one of the two types of phones.
(c) Find the probability the household has a cell phone only.
5.3 WS #1
The Probability of a Flush
1. A poker player holds a flush when all 5 cards in the hand belong to the same suit. Remember that a deck contains 52
cards, 4 suits (hearts, diamonds, spades and clubs) with 13 cards of each suit. When the deck is well shuffled, each card
dealt is equally likely to be any of those that remain in the deck.
(a) We will concentrate on spades. What is the probability that the first card dealt is a spade? What is the conditional
probability that the second card is a spade given that the first is a spade?
(b) Continue to count the remaining cards to find the conditional probabilities of a spade on the third, the fourth, and
the fifth card given in each case that all previous cards are spades.
(c) The probability of being dealt 5 spades is the product of the five probabilities you have found. Why? What is this
probability?
(d) The probability of being dealt 5 hearts or 5 diamonds or 5 clubs is the same as the probability of being dealt 5 spades.
What is the probability of being dealt a flush?
2. Internet sites often vanish or move, so that references to them can’t be followed. In fact, 13% of Internet sites
referenced in major scientific journals are lost within two years after publication. If a paper contains seven Internet
references, what is the probability that at least one of them doesn’t work two years later? Show your work!
3. A recent census at Emory University revealed that 40% of its students mainly used Macintosh computers (Macs). The
rest mainly used PCs. At the time of the census, 67% of the school’s students were undergraduates. The rest were
graduate students. In the census, 23% of the respondents were graduate students who said that they used PCs as their
primary computers. Suppose we select a student at random from among those who were part of the census and learn
that the student mainly uses a Mac. Find the probability that this person is a graduate student. Show your work.
4. Many employers require prospective employees to take a drug test. A positive result on this test indicates that the
prospective employee uses illegal drugs. However, not all people who test positive actually use drugs. Suppose that 4%
of prospective employees use drugs, the false positive rate is 5% and the false negative rate is 10%. What percent of
people who test positive actually use illegal drugs?
5.3 WS #2
Picking Two Sneezers

This is a two-way table that classified 40 students according to their gender and whether they had allergies.
Female Male Total

Allergies 10 8 18
No Allergies 13 9 22
Total 23 17 40
Suppose we chose 2 students at random.
a. Draw a tree diagram or chart that shows the sample space for this chance process.
b. Find the probability that both students selected suffer from allergies.
c. Find the probability that neither student selected suffers from allergies.
d. Find the probability that at least one student selected suffers from allergies.
e. Find the probability that only one student selected suffers from allergies.
Media Usage and Good Grades
In January 2010, the Kaiser Family Foundation released a study about the influence of media in the lives of young people
ages 8-18 (http://www.kff.org/entmedia/mh012010pkg.cfm). In the study, 17% of the youth were classified as light
media users, 62% were classified as moderate media users and 21% were classified as heavy media users. Of the light
users who responded, 74% described their grades as good (A’s and B’s), while only 68% of the moderate users and 52%
of the heavy users described their grades as good.
a. According to this study, what percent of young people ages 8-18 described their grades as good? Use a tree
diagram or chart to calculate the probability.
b. According to the tree diagram you constructed above, what percent of students with good grades are heavy
users of media?
c. What percent of students with good grades are moderate users?
d. What percent of light users are students with bad grades?
e. What percent of moderate users are students with bad grades?

6.1 WS
NHL Goals
In 2010, there were 1319 games played in the National Hockey League’s regular season. Imagine selecting one of these
games at random and then randomly selecting one of the two teams that played in the game. Define the random
variable X = number of goals scored by a randomly selected team in a randomly selected game. The table below gives
the probability distribution of X:
Goals: 0 1 2 3 4 5 6 7 8 9
Probability: 0.061 0.154 0.228 0.229 0.173 0.094 0.041 0.015 0.004 0.001
(a) Show that the probability distribution for X is legitimate.
(b) Make a histogram of the probability distribution. Describe what you see.
(c) What is the probability that the number of goals scored by a randomly selected team in a randomly selected game is
at least 6?
NHL Goals, continued.
Previously, we defined the random variable X to be the number of goals scored by a randomly selected team in a
randomly selected game. The table below gives the probability distribution of X:
Goals: 0 1 2 3 4 5 6 7 8 9
Probability: 0.061 0.154 0.228 0.229 0.173 0.094 0.041 0.015 0.004 0.001
(d): Compute the mean of the random variable X and interpret this value in context.
(e): Compute and interpret the standard deviation of the random variable X.
Weights of Three-Year-Old Females
2. The weights of three-year-old females closely follow a Normal distribution with a mean of  = 30.7 pounds and a
standard deviation of 3.6 pounds. Randomly choose one three-year-old female and call her weight X.
A. Find the probability that the randomly selected three-year-old female weighs at least 30 pounds.
B. Should the pediatrician be concerned if a 3 year old girl only weighs 25 pounds?
6.2 WS
Study Habits
The academic motivation and study habits of female students as a group are better than those of males. The Survey of
Study Habits and Attitudes (SSHA) is a psychological test that measures these factors. The distribution of SSHA scores
among the women at a college has mean 120 and standard deviation 28, and the distribution of scores among male
students has mean 105 and standard deviation 35. You select a single male student and a single female student at
random and give them the SSHA test.
(a) Explain why it is reasonable to assume that the scores of the two students are independent.
(b) What are the expected value and standard deviation of the difference (female minus male) between their scores?
(c) From the information given, can you find the probability that the woman chosen scores higher than the man? If so,
find this probability. If not, explain why you cannot.
Swim Team
Alonzo & Tracy Mourning Sr. High School has one of the best women’s swimming team in the region. The 400-meter
freestyle relay team is undefeated this year. In the 400-meter freestyle relay, each swimmer swims 100 meters. The
times, in seconds, for the four swimmers this season are approximately Normally distributed with means and standard
deviations as shown. Assuming that the swimmer’s individual times are independent, find the probability that the total
team time in the 400-meter freestyle relay is less than 220 seconds. Show your work!
6.3 WS #2: Binomial Distributions Practice
For each problem, be sure that the situation fits the criteria for binomial distributions. If so, answer the questions (show
the formula) and then find the mean and standard deviation of the distribution.
1) 80% of the graduates of Northeast High who apply to Penn State University are admitted. Last year, there were 6
graduates from Northeast who applied to Penn State. What is the probability that
a) 4 were admitted b) more than 4 were admitted
Mean of distribution: ____________ Standard deviation of distribution: ____________
2) Tires from the Apex Tire Corp. are traditionally 5% defective. A truck carries 10 tires, 8 in use and 2 spares. If 10 tires
are chosen from Apex, what is the probability that not more than two defective tires are chosen.
3) Studies indicate that in 70% of the families of Blue Bell, both the husband and wife work. If 7 families are randomly
selected from Blue Bell, what is the probability that
a) exactly 4 of them work. b) more than 4 work

4) According to the National Institute of Health, 32% of all women will suffer at least one hip fracture because of
osteoporosis by the age of 90. If 10 women aged 90 are selected at random, find the probability that
a) 2 or more of them suffer/will suffer at least one hip fracture
b) none of them suffer/will suffer a hip fracture
5) According to FBI statistics, only 52% of all rape cases are reported to the police. If 10 rape cases are randomly
selected, what is the probability that at least one is reported to the police?
1
6) In a school, typically only 10 of the student body returns surveys. 20 students are chosen randomly to receive a
survey. What is the probability that
a) they get no surveys back. b) they get more 4 or more surveys back.
7) The probability that a driver making a gas purchase will pay by credit card is 0.60. If 50 cars pull up to the station to
buy gas, what is the probability that at least half of the drivers will pay by credit card?

WS 6.3
1. Patients receiving artificial knees often experience pain after surgery. The pain is measured on a subjective scale with
possible values of 1 to 5. Assume that X is a random variable representing the pain score for a randomly elected patient.
The following table gives part of the probability distribution for X.
(a) Find P(X = 5).
(b) Find the probability that the pain score is less than 3.
(c) Find the probability that the pain score is greater than 3.
(d) Find the mean μ for this distribution.
2. Amarillo Slim, a professional dart player, has an 80% chance of hitting the bull’s-eye on a dartboard with any throw.
Suppose that he throws 10 darts, one at a time, at the dartboard.
(a) Find the probability that Slim hits the bull’s-eye exactly six times.
(b) Find the probability that he hits the bull’s-eye at least four times.
(c) Compute the expected number of bull’s-eyes in 10 throws.

(d) Find the probability that Slim’s first bull’s-eye occurs on the fourth throw.
(e) Find the probability that it takes Amarillo more than 2 throws to hit the bullseye.
3. Harlan comes to class one day, totally unprepared for a pop quiz consisting of ten multiple-choice questions. Each
question has five answer choices, and Harlan answers each question randomly.
(a) Find the probability that Harlan’s gets more than 5 questions right out of 10.
(b) Find the probability that Harlan’s first correct answer occurs after the fourth question.
(c) Find the expected number of questions required for Harlan to get his first correct answer.
(d) Find the probability that Harlan guesses more answers correctly than would be expected by chance.
WS 7.1
Tall Girls
According to the National Center for Health Statistics, the distribution of heights for 16-year-old females is modeled well
by a Normal density curve with mean μ = 64 inches and standard deviation σ = 2.5 inches. To see if this distribution
applies at their high school, an AP Statistics class takes an SRS of 20 of the 300 16-year-old females at the school and
measures their heights. What values of the sample mean X would be consistent with the population distribution being
N(64, 2.5)? To find out, we used Fathom software to simulate choosing 250 SRSs of size n = 20 students from a
population that is N(64, 2.5). The figure below is a dotplot of the sample mean height X of the students in the sample.
(a) Is this the sampling distribution of X? Justify your answer.
(b) Describe the distribution. Are there any obvious outliers?
(c) Suppose that the average height of the 20 girls in the class’s actual sample is X = 64.7. What would you conclude
about the population mean height μ for the 16-year-old females at the school? Explain.
Choosing Cards
We used Fathom (statistical computer software) to simulate choosing 500 SRSs of size 5 from the deck of cards
described in the Alternate Activity on the previous page. The graph below shows the distribution of the sample median
for these 500 samples.
(a) Is this the sampling distribution of the sample median? Justify your answer.
(b) Suppose that another student prepared a different deck of cards and claimed that it was exactly the same as the one
used in the activity. However, when you took an SRS of size 5, the median was 4. Does this provide convincing evidence
that the student’s deck is different?
7.2 WS
The Candy Machine
1. Suppose a VERY large candy machine has 15% orange candies. Imagine taking an SRS of 25 candies from the machine
and observing the sample proportion 𝑝̂ ? of orange candies.
(a) What is the mean of the sampling distribution of 𝑝̂ ? Why?
(b) Check to see if the 10% condition is met.
(c) Find the standard deviation of the sampling distribution of 𝑝̂ .
(d) Is the sampling distribution of 𝑝̂ approximately Normal? Check to see if the Normal condition is met.
(e) If the sample size were 75 rather than 25, how would this change the sampling distribution of 𝑝̂ ? How would this
impact the Normal condition?
Planning for College
2. The superintendent of Miami-Dade County Public Schools wants to know what proportion of middle school students
in his district are planning to attend a four-year college or university. Suppose that 80% of all middle school students in
his district are planning to attend a four-year college or university. What is the probability that a SRS of size 125 will give
a result within 7 percentage points of the true value?
Who owns a Harley?
3. Harley-Davidson motorcycles make up 14% of all the motorcycles registered in the United States. You plan to
interview an SRS of 500 motorcycle owners. How likely is your sample to contain 20% or more who own Harleys?
7.3 WS
Buy Me Some Peanuts and Sample Means
1. At the P. Nutty Peanut Company, dry roasted, shelled peanuts are placed in jars by a machine. The distribution of
weights in the bottles is approximately Normal, with a mean of 16.1 ounces and a standard deviation of 0.15 ounces.
(a) Without doing any calculations, explain which outcome is more likely, randomly selecting a single jar and finding the
contents to weigh less than 16 ounces or randomly selecting 10 jars and finding the average contents to weigh less than
16 ounces.
(b) Find the probability of each event described above. Since the distribution is normal you can use “normalcdf” on your
calculator.
Single jar weighing 16 oz or less:
10 jars weighing 16 oz or less:

Mean Texts
2. Suppose that the number of texts sent during a typical day by a randomly selected high school student follows a right-
skewed distribution with a mean of 15 and a standard deviation of 35. Assuming that students at your school are typical
texters, how likely is it that a random sample of 50 students will have sent more than a total of 1000 texts in the last 24
hours?
Bad carpet
3. The number of flaws per square yard in a type of carpet material varies with mean 1.6 flaws per square yard and
standard deviation 1.2 flaws per square yard. The population distribution cannot be Normal, because a count takes only
whole-number values. An inspector studies 200 square yards of the material, records the number of flaws found in each
square yard, and calculates X, the mean number of flaws per square yard inspected. Find the probability that the mean
number of flaws exceeds 2 per square yard.
8.1 WS
Losing Weight
A Gallup Poll in November 2014 found that 59% of the people in its sample said “Yes” when asked,
“Would you like to lose weight?” Gallup announced: “For results based on the total sample of national
adults, one can say with 95% confidence that the margin of (sampling) error is ±3 percentage points.”
(a) Explain what the margin of error means in this setting.
(b) State and interpret the 95% confidence interval.
(c) Interpret the confidence level.
The admissions director from Florida International University found that (107.8, 116.2) is a 95%
confidence interval for the mean IQ score of all freshmen. Comment on whether or not each of the
following explanations is correct.
(a) There is a 95% probability that the interval from 107.8 to 116.2 contains μ.
(b) There is a 95% chance that the interval (107.8, 116.2) contains X.
(c) This interval was constructed using a method that produces intervals that capture the true mean in
95% of all possible samples.
(d) 95% of all possible samples will contain the interval (107.8, 116.2).
(e) The probability that the interval (107.8, 116.2) captures μ is either 0 or 1, but we don’t know which.
How do Confidence Intervals Change…
(a) How does the shape of the confidence interval change if the confidence level increases from 90% to
95%?
(b) How would the shape of a confidence change if the sample size was decreased? Assume the new,
smaller sample size still meets all of the normality conditions.
8.2 WS
Kissing the Right Way?
According to an article in the San Gabriel Valley Tribune, “Most people are kissing the ‘right way’.” That
is, according to the study, the majority of couples tilt their heads to the right when kissing. In the study,
a researcher observed a random sample 124 couples kissing in various public places and found that
83/124 (66.9%) of the couples tilted to the right. Construct and interpret a 95% confidence interval for
the proportion of all couples who tilt their heads to the right when kissing.
Tattoos
Suppose that you wanted to estimate the p = the true proportion of students at your school that have a
tattoo with 95% confidence and a margin of error of no more than 0.10. What’s the minimum number
of students you would need to survey?
How much homework?
Ms. Garcia wants to estimate how much time students spend on homework, on average, during a typical
week. She wants to estimate at the 90% confidence level with a margin of error of at most 30 minutes.
A pilot study indicated that the standard deviation of time spent on homework per week is about 154
minutes. What’s the minimum number of students you would need to survey?
8.3 WS
SAT Math Scores
High school students who take the SAT Math exam a second time generally score higher than on their
first try. Past data suggest that the score increase has a standard deviation of about 50 points. How large
a sample of high school students would be needed to estimate the mean change in SAT score to within 2
points with 95% confidence? Show your work.
Travel Time to Work
A study of commuting times reports the travel times to work of a random sample of 20 employed adults
in New York State. The mean (x-bar) is 31.25 minutes, and the standard deviation is 21.88 minutes.
What is the standard error of the mean? Interpret this value in context
Vitamin C Content
Several years ago, the U.S. Agency for International Development provided 238,300 metric tons of corn-
soy blend (CSB) for emergency relief in countries throughout the world. CSB is a highly nutritious, low-
cost fortified food. As part of a study to evaluate appropriate vitamin C levels in this food,
measurements were taken on samples of CSB produced in a factory
The following data are the amounts of vitamin C, measured in milligrams per 100 grams (mg/100 g) of
blend, for a random sample of size 8 from one production run:
Construct and interpret a 95% confidence interval for the mean amount of vitamin C μ in the CSB from
this production run.
9.1 WS
A Better Golf Club?
Mike is an avid golfer who would like to improve his play. A friend suggests getting new clubs and lets Mike try out his 7-
iron. Based on years of experience, Mike has established that the mean distance that balls travel when hit with his old
7-iron is µ = 175 yards with a standard deviation of σ = 15 yards. He is hoping that this new club will make his shots
with a 7-iron more consistent (less variable), so he goes to the driving range and hits 50 shots with the new 7-iron
(a) Describe the parameter of interest in this setting.
(b) State appropriate hypotheses for performing a significance test.
Based on 50 shots with the new 7-iron, the standard deviation was sx = 10.9 yards. A significance test using the sample
data produced a P-value of 0.002.
(c) Interpret the P-value in this context.
(d) Do the data provide convincing evidence against the null hypothesis? Explain.
(e) Interpret the result in part d in context.

State the appropriate null hypothesis H 0 and alternative hypothesis H a in each case. Be sure to define your parameter
each time.
2. A Gallup Poll report on a national survey of 1028 teenagers revealed that 72% of teens said they seldom or never
argue with their friends. Yvonne wonders whether this national result would be true in her large high school. So she
surveys a random sample of 150 students at her school
Hypothesis:
For Yvonne’s survey, 96 students in the sample said they rarely or never argue with friends. A significance test yields a P-
value of 0.0291.
(a) Interpret this result in context.
(b) Do the data provide convincing evidence against the null hypothesis? Explain.
2. Hemoglobin is a protein in red blood cells that carries oxygen from the lungs to body tissues. People with less than 12
grams of hemoglobin per deciliter of blood (g/dl) are anemic. A public health official in Jordan suspects that Jordanian
children are at risk of anemia. He measures a random sample of 50 children.
Hypothesis:
For the study of Jordanian children, the sample mean hemoglobin level was 11.3 g/dl and the sample standard deviation
was 1.6 g/dl. A significance test yields a P-value of 0.0016.
(a) Interpret the P-value in context.
(b) What conclusion would you make if α = 0.05? α = 0.01? Justify your answer
9.1 WS #2
Faster Fast Food?
The manager of a fast-food restaurant want to reduce the proportion of drive-through customers who have to wait
more than 2 minutes to receive their food once their order is placed. Based on store records, the proportion of
customers who had to wait at least 2 minutes was p = 0.63. To reduce this proportion, the manager assigns an
additional employee to assist with drive-through orders. During the next month the manager will collect a random
sample of drive-through times and test the following hypotheses:
H 0 : p = 0.63
H a : p < 0.63
where p = the true proportion of drive-through customers who have to wait more than 2 minutes after their order is
placed to receive their food.
Describe a Type I and a Type II error in this setting and explain the consequences of each.
Suppose that the manager decided to carry out this test using a random sample of 250 orders and a significance level of
 = 0.10. What is the probability of a making a Type I error?
10.1 WS
Presidential approval
Many news organizations conduct polls asking adults in the United States if they approve of the job the president is
doing. How did President Obama’s approval rating change from August 2009 to September 2010? According to a CNN
poll of 1024 randomly selected U.S. adults on September 1-2, 2010, 50% approved of Obama’s job performance. A CNN
poll of 1010 randomly selected U.S. adults on August 28-30, 2009 showed that 53% approved of Obama’s job
performance. Use the results of these polls to construct and interpret a 90% confidence interval for the change in
Obama’s approval rating among all US adults. Based on your interval, is there convincing evidence that Obama’s job
approval rating has changed?
Hearing Loss
In a study of 3000 randomly selected teenagers in 1988-1994, 15% showed some hearing loss. In a similar study of 1800
teenagers in 2012-2013, 19.5% showed some hearing loss. Is there convincing evidence that the proportion of all teens
with hearing loss has increased? Additionally, between the two studies, Apple introduced the iPod/iPhone/Beats
headphones. If the results of the test are statistically significant, can we blame iPods/iPhones/Beats headphones for the
increased hearing loss in teenagers?
10.1 WS # 2
Cash for Quitting
In an effort to reduce health care costs, General Motors sponsored a study to help employees stop smoking. In the
study, half of the subjects were randomly assigned to receive up to $750 for quitting smoking for a year while the other
half were simply encouraged to use traditional methods to stop smoking. None of the 878 volunteers knew that there
was a financial incentive when they signed up. At the end of one year, 15% of those in the financial rewards group had
quit smoking while only 5% in the traditional group had quit smoking. Do the results of this study give convincing
evidence that a financial incentive helps people quit smoking?
10.2 WS Solutions
Plastic Bags
Parameter:
T = the mean capacity of plastic bags from Target (in grams)
µp= the mean capacity of plastic bags from Publix (in grams).
Assess Conditions:
 Random: The students selected a random sample of bags from each store.
 Normal: Since the sample sizes are small, we must graph the data to see if it is reasonable to assume that the
population distributions are approximately Normal.

Publix
Since there is no obvious skewness or outliers, it is safe to use t procedures.

 Independent: The samples were selected independently and it is reasonable to assume that there are more than
10(5) = 50 plastic grocery bags at each store.
Name Interval: two-sample t interval for µt -µp
Interval:
By hand: For these data, xT = 12825.8, sT = 1912.5, xB = 9234, sB = 1474.2. Using the conservative df of 5 – 1 = 4, the
critical value for 99% confidence is t* = 4.604. Thus, the confidence interval is:
1474.22 1912.52
12826  9234   4.604  = 3592  4972 = (–1380, 8564).
5 5
With technology and df =7.5, CI = (–101, 7285).
We are 99% confident that the interval from –101 to 7285 grams captures the true difference in the mean capacity for
plastic grocery bags from Target and from Publix.
Conclude in Context: Since the interval includes 0, it is plausible that there is no difference in the two means. Thus, we
do not have convincing evidence that there is a difference in mean capacity.
The better picker-up?
Solution:
(a) The five-number summary for the Bounty paper towels is (103, 114, 116.5, 124, 128) and the five-number summary
for the generic paper towels is (77, 84, 88, 90, 103). Here are the boxplots:
Both distributions are roughly symmetric, but the generic brand has two high outliers. The center of the Bounty
distribution is much higher than the center of the generic distribution. Although the range of each distribution is
roughly the same, the interquartile range of the Bounty distribution is larger.
Since the centers are so far apart and there is almost no overlap in the two distributions, the Bounty mean is almost
certain to be significantly higher than the generic mean. If the means were really the same, it would be virtually
impossible to get so little overlap.
B.
Parameters:
 B = the mean number of quarters a wet Bounty paper towel can hold
G = the mean number of quarters a wet generic paper towel can hold.
Hypothesis:
H 0 : µb = µ g
H a : µb > µg
Assess Conditions:
 Random: The students used a random sample of paper towels from each brand.
 Normal: Even though there were two outliers in the generic distribution, both distributions were reasonably
symmetric and the sample sizes are both at least 30, so it is safe to use t procedures.
 Independent: The samples were selected independently and it is reasonable to assume there are more than
10(30) = 300 paper towels of each brand.
Name Test: two-sample t test for  B  G .
Test statistic: t 
117.6  88.1  0 = 17.64
6.642 6.302

30 30
Obtain P-value: Using either the conservative df = 30 – 1 = 29 or from technology (df = 57.8), the P-value is
approximately 0.
Make a Decision: Since the P-value is less than 0.05, we reject H 0 .

State a Conclusion: There is very convincing evidence that wet Bounty paper towels can hold more weight, on average,
than wet generic paper towels. Since the P-value is approximately 0, it is almost impossible to get a difference in means
of at least 29.5 quarters by random chance, assuming that the two brands of paper towels can hold the same amount of
weight when wet.
AP Stats- Mrs. Daniel
11.1 WS
Cheap Dice?
Mrs. Daniel purchased a bunch of inexpensive dice from Amazon. She is now wondering if the dice are fair. So, Mrs.
Daniel randomly selects 6 dice to roll 10 times each, for a total of 60 observations to evaluate. Do we have statistically
convincing evidence that the dice are fair?
Outcome Observed
1 13
2 11
3 6
4 12
5 10
6 8
Total 60
Landline surveys
According to the 2010 Census, of all US residents age 20 and older, 19.1% are in their 20’s, 21.5% are in their 30’s, 21.1
% are in their 40’s, 15.5% are in their 50’s, and 22.8% are 60 and older. The table below shows the age distribution for a
sample of US residents age 20 and older. Members of the sample were chosen by randomly dialing landline telephone
numbers.
Category Count
20-29 141
30-39 186
40-49 224
50-59 211
60+ 286
Total 1048
Do these data provide convincing evidence that the age distribution of people who answer landline telephone surveys is
not the same as the age distribution of all US residents?
11.2 WS
Ibuprofen or Acetaminophen?
In a study reported by the Annals of Emergency Medicine, researchers conducted a randomized, double-blind clinical
trial to compare the effects of ibuprofen and acetaminophen plus codeine as a pain reliever for children recovering from
arm fractures. There were many response variables recorded, including the presence of any adverse effect, such as
nausea, dizziness, and drowsiness. Here are the results:
Acetaminophen
Ibuprofen Total
plus Codeine
Adverse effects 36 57 93
No adverse effects 86 55 141
Total 122 112 234
Tide vs New Tide
Before bringing a new product to market, firms carry out extensive studies to learn how consumers react to the product
and how best to advertise its advantages. Here are data from a study of a new laundry detergent. The participants are a
random sample of people who don’t currently use the established brand that the new product will compete with. Give
subjects free samples of both detergents. After they have tried both for a while, ask which they prefer. The answers may
depend on other facts about how people do laundry.
Determine whether or not the sample provides convincing evidence that laundry practices and product preference are
independent in the population of interest.
12.1 WS #2
The Professor Swims
Here are data on the time (in minutes) Professor Moore takes to swim 2000 yards and his pulse rate (beats per minute)
after swimming on a random sample of 23 days:
(a) Calculate and interpret a 95% confidence interval for the slope β of the population regression line. All conditions
have been checked.
(b) Is there statistically significant evidence of a negative linear relationship between Professor Moore’s swim time and
his pulse rate in the population of days on which he swims 2000 yards? Carry out an appropriate significance test at the
α = 0.05 level. All conditions have been checked.

스탯 겨울특강 PDF

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

스탯 겨울특강 PDF

Încărcat de

Drepturi de autor:

Formate disponibile

Mike's AP Statistics

ü DEFINE “Individuals” and “Variables”

What is the Study of Statistics?! Population

Variable - any characteristic of an Distribution

Organizing a Statistical Problem Section 1.1

Distribution & Categorical Variables Displaying Categorical Data

Math 15 Math 30% Oldies

2014 AP Exam Scores

Graphs: Good and Bad Two-Way Tables

Member of No Member of Member of 2 or

Member of No Member of Member of 2 or

Rides the School Bus 55 33 20 108

Does not Ride Bus 16 44 82 142

Comparing Categorical Comparing Categorical

Comparing Categorical Writing to Compare Categorical

Comparing Categorical Comparing Categorical

ü CONSTRUCT and INTERPRET dotplots, stemplots, and

3. Mark a dot above the location on the

How to Describe Quantitative Describing Shape

Symmetric Skewed-left Skewed-right

Other Ways to Describe Shape:

Describe the shape, center, and spread of the

Stemplots (Stem-and-Leaf Plots)

How to Make a Stemplot Stemplots (Stem-and-Leaf Plots)

Two Special Types of Stem Plots

How to Make a Histogram Making a Histogram

Caution: Using Histograms Wisely 1.3: Describing

ü MEASURE spread with standard deviation and sum of observations x1 + x 2 + ...+ x n

ü CALCULATE numerical summaries with technology n

Measuring Center: The Median Comparing the Mean and the

Why is the mean more affected by

Standard Deviation Standard Deviation

Standard Deviation Formula FYI: Why n-1?!

How to Calculate Standard Calculate the Standard Deviation

Calculate the Standard Deviation

x=5 Sum= Sum= Sum=

Calculate the Standard Deviation Two Extreme Examples:

3. With your cursor on the values, press menu

TI-NSpire: Calculate standard deviation and

6. Set screen to:

Interquartile Range (IQR) Interquartile Range (IQR)

Q1 = 15 M = 22.5 Q3= 42.5

1.5 x IQR Rule for Outliers

Any travel time shorter than -26.25 minutes or longer than

TI- Nspire: 5 Number Summary TI- Nspire: 5 Number Summary

3. With your cursor on the values, press menu

Boxplots (Box-and-Whisker Plots)

Construct a Boxplot Construct a Boxplot

Min=5 Q1 = 15 M = 22.5 Q3= 42.5 Max=85

Choosing Best Measures of Center

ü MEASURE position using percentiles

Measuring Position: Percentiles Percentiles

Percentiles Other Percentiles: Deciles and Quartiles

Find the three quartiles. Find the three quartiles.

Other Percentiles: Deciles and Quartiles Cumulative Relative Frequency

40- 2 2/44 = 2 2/44 = 40- 2 2/44 = 2 2/44 =

Cumulative relative frequency (%)

Cumulative relative frequency (%)

Transforming Data Transforming Data: Add/Sub a

Transforming Data: Add/Sub a

Transforming Data Density Curve

Density Curves Continued… Describing Density Curves

ü DESCRIBE and APPLY the 68-95-99.7 Rule