The Normal Distribution and Sampling Distributions

PSYC 545

Wednesday, September 11, 13

What is a normal curve?

Wednesday, September 11, 13

What is a normal curve?

The normal curve is a mathematical
distribution with precise properties.

Wednesday, September 11, 13

What is a normal curve?

The normal curve is a mathematical
distribution with precise properties.

Observed data often approximate a

normal distribution, but dont frequently perfectly conform.

Wednesday, September 11, 13

What is a normal curve?

The normal curve is a mathematical
distribution with precise properties.

Observed data often approximate a

normal distribution, but dont frequently perfectly conform. many variables display relatively normal distributions when collected from large samples.

Serves as a foundational model because

Wednesday, September 11, 13

Without further ado...

Wednesday, September 11, 13

Characteristics of the Normal Distribution

Wednesday, September 11, 13

Characteristics of the Normal Distribution

Unimodal

Wednesday, September 11, 13

Characteristics of the Normal Distribution

Unimodal Symmetric

Wednesday, September 11, 13

Characteristics of the Normal Distribution

Unimodal Symmetric Mode, Median, and Mean are Equal

Wednesday, September 11, 13

Characteristics of the Normal Distribution

Unimodal Symmetric Mode, Median, and Mean are Equal Asymptotic

Wednesday, September 11, 13

Characteristics of the Normal Distribution

Unimodal Symmetric Mode, Median, and Mean are Equal Asymptotic Two normal distributions DO NOT
necessarily have same means and variances

Wednesday, September 11, 13

Of probabilities and normality

http://davidmlane.com/hyperstat/ normal_distribution.html

Wednesday, September 11, 13

Standard Scores (z-scores)

Converts raw scores to a common metric

XX z= s

Wednesday, September 11, 13

Characteristics of z-scores

Wednesday, September 11, 13

Characteristics of z-scores

Magnitude & Sign

Wednesday, September 11, 13

Characteristics of z-scores

Magnitude & Sign Mean is always 0, standard deviation

is always 1.

Wednesday, September 11, 13

Standard Normal Distribution

Wednesday, September 11, 13

Standard Normal Distribution

Can use this to answer

Wednesday, September 11, 13

Standard Normal Distribution

Can use this to answer proportion of scores between Mean and a
given score

Wednesday, September 11, 13

Standard Normal Distribution

Can use this to answer proportion of scores between Mean and a
given score raw score

proportion of scores above (below) a given

Wednesday, September 11, 13

Standard Normal Distribution

Can use this to answer proportion of scores between Mean and a
given score raw score scores

proportion of scores above (below) a given proportion of scores between any two raw

Wednesday, September 11, 13

Standard Normal Distribution

Can use this to answer proportion of scores between Mean and a
given score raw score scores

proportion of scores above (below) a given proportion of scores between any two raw what raw score falls above (below) a given
proportion of cases
Wednesday, September 11, 13

Are my data normal?

Visual inspection Q-Q plots Empirical test of skew and kurtosis
Descriptive Statistics N Statistic var2 Valid N (listwise) 21 21 Descriptive Statistics Skewness Std. Error var2 Valid N (listwise) .501 Kurtosis Statistic -.049 Std. Error .972 Minimum Statistic -1.71 Maximum Statistic 1.58 Mean Statistic .0340 Std. Deviation Statistic .81302 Variance Statistic .661 Skewness Statistic .068

Wednesday, September 11, 13

HW
From Howell: 3.1, 3.2, & 3.6

Wednesday, September 11, 13

Using the Normal Distribution

What if we knew the following: The average test score on a given

test for the population of high school seniors is 25.00 and the standard deviation is 5.00. a student who has a score of at least 32?

What is the probability of selecting

Wednesday, September 11, 13

Using the ND
First, stated in terms of probability:
p > 32 = ??

Identify the exact position of X = 32 by

converting to a z-score:

What probability corresponds to z = 1.4?

http://davidmlane.com/hyperstat/ normal_distribution.html
Wednesday, September 11, 13

X X 32 25 7 z= = = = 1.4 5 5

Another Example

Wednesday, September 11, 13

Another Example
In my restaurant, I know that the average
number of diners for my special Sunday brunch is 241 with a standard deviation of 19.

Wednesday, September 11, 13

Another Example
In my restaurant, I know that the average One Sunday, I count 198 diners (maybe it
is the mice in the kitchen??!!!). number of diners for my special Sunday brunch is 241 with a standard deviation of 19.

Wednesday, September 11, 13

Another Example
In my restaurant, I know that the average One Sunday, I count 198 diners (maybe it
is the mice in the kitchen??!!!). number of diners for my special Sunday brunch is 241 with a standard deviation of 19.

If my distribution of Sunday diners is

normal, what is the probability of having only 198 diners (or fewer)?

Wednesday, September 11, 13

Sampling Distribution of Means

We almost never know the true distribution, so
comparing observed data to an ideal normal distribution is often not reasonable.

What we really want to know is, If there is a true population distribution, and From this standpoint, we can determine the
probability of our observed sample mean.
Wednesday, September 11, 13

we draw a given sample from that distribution, how much variability should we expect in sample means drawn from the population?

Sampling Distribution of Means

If we draw samples from a larger population, the

http://onlinestatbook.com/stat_sim/sampling_dist/index.html

means of those samples will form a normal distribution that clusters around the mean of the population.

The variability of the sampling distribution is

less than the population variability.

Sampling distribution variability is a

function of sample size (larger sample = less variability).

Wednesday, September 11, 13

Central Limit Theorem

Wednesday, September 11, 13

Central Limit Theorem

The central limit theorem states that given a
distribution with a particular mean () and variance (), the sampling distribution of the mean approaches a normal distribution with a mean ().

Wednesday, September 11, 13

Central Limit Theorem

The central limit theorem states that given a
distribution with a particular mean () and variance (), the sampling distribution of the mean approaches a normal distribution with a mean (). about the central limit theorem is that no matter what the shape of the original distribution, the sampling distribution of the mean approaches a normal distribution.

The amazing and counter-intuitive thing

Wednesday, September 11, 13

Key Characteristics of Sampling Distribution of Means

Normally distributed Mean = Population Mean Standard Deviation

X =
N

Wednesday, September 11, 13

How Do We Use This Info?

We know that the average score for adults age
25-55 on a national test of Batman trivia is 50 and the standard deviation is 15. individuals

If we were to draw 1000 samples of 25 what would the mean of the 1000 sample be? what would be the standard error?
of samples, but each included 50 individuals?

What would happen if we drew the same number

Wednesday, September 11, 13