Sunteți pe pagina 1din 4

EXPERIENCING AUSTINS POPULATION

CHANGE THROUGH CRIME


WRITTEN BY CAROLINA A. DUN CAN
SUBMITTED: APRIL 27, 2015

Introduction
The capital of Texas, Austin, is known for being
the live music capital of the world. Recently, Austin has
been experiencing growth that has been affecting the
city. According to CNNMoney.com, Austin is the thirdfastest-growing large city in the nation from 2000 to
2006 (Christie 2007). For all residents in Austin, crime
is an important factor for knowing where to move or
where is the safest place to be. Specifically, Part I
Crimes are analyzed mainly for residents of an area.
Part I Crimes are defined as the first part of what is
called Index Crimes, which are crimes for which reports
of offenses committed are collected (TxDPS 2005). The
Part I Crimes consist of eight types: murder, rape,
robbery, burglary, theft, automobile theft, aggravated
assault, and arson. With Austin divided by zip codes, the
population change and the alteration in crime rates
within those zip codes will be analyzed to deduce if
there is a difference between the means of both
variables.
Hypothesis: The averages in change of both variables
are not equal.

Null Hypothesis:
There is no difference between the mean of change in
population, from 2000 2010, and the mean of change in Part I
crime, from 2003-2013, in the zip codes of Austin.
Ho: 1 - 2 = 0

Alternative Hypothesis:
There is a difference between the mean of change in
population, from 2000 2010, and the mean of change in Part I
crime, from 2003 2013, in the zip codes of Austin.
HA: 1 - 2 0

Photo of Study Area By:


http://www.city-data.com/zipmaps/AustinTexas.html

Photo Credit: James.


Photography-onthe.net

Data
Secondary data was gathered and input into
Excel by hand to discover more about these two
variables. A sample size of fifty zip codes was obtained
of central Austin and its surrounding area zip codes.
The attainment of the Part I Crime data was discovered
through austintexas.gov. The data was separated by
each of the eight Part I Crimes and by fifty zip codes in
Austin. The austintexas.gov website only provided Part I
Crimes from 2003 to 2013. The experimented value will
be the difference of Part I Crime, in the fifty zip codes,
from 2003 to 2013. The population data was primarily
found at city-data.com, and, for a few missing values,
clrsearch.com also was used. The population values for
the years 2000 to 2010, segregated by fifty zip codes in
Austin, will be tested since that is when the Census
Bureau conducts their count.
These are the differences of Part I crime and
population change, over the years, from the data
obtained:
Difference In Part I Crime

2013
2003

Difference In
Part I Crime

40,000 42,000 44,000 46,000


Graph and map created by: Carolina A. Duncan

Difference in Population

2010

Difference in
Population

2000
0

500000 1000000 1500000


Graph and map created by: Carolina A. Duncan

rejected. If the values were distributed with a common


rank, by percentage for example, a histogram could be
more easily created. The kurtosis of the histogram
would be leptokurtic, due to the very high and very low
ranges both variables have. The skewness would be
difficult to show due to the varying zip code values for
the change in crime and the change in population; both
have extreme loss (negative) and gain (positive) values.
However, after running a correlation test on Excel using
the CORREL function, the answer displayed was 0.25, a
weak positive relationship. Therefore, as one value
increases so does the other and so the skewness would
be the left.

Photo Citation: Jones, James

After inputting all the values into Excel, the


results for the descriptive statistics were:

Mean
Median
Mode
Minimum
Maximum
Standard Dev.

Population
4988
3035.5
No Mode
-4477
34361
7111.38

Crime
55
13.5
0
-730
1177
282.39

50,571,774.29

79,744.24

( / s)

Sample Variance
(s2)

A standard error is a measure of uncertainty


and the confidence interval is an interval of values
within which the true mean of the population is
believed to be located (Townsend 2015). Both are
important equations to indicate precision. Before the
confidence interval is figured out, the standard error of
the mean must be done first. The steps are as follows:
Standard Error of the Mean:
( ) = /
Pop = 7111.38
Crime=282.39
n = 50
( ) Pop = 7111.38/50 = 1,005.70

Methods Used
The level of significance, also known as the
alpha ( ) level or the probability for a sampling error,
which will be employed, is 0.05. To define the regions of
rejection and non-rejection of the null hypothesis, the
application of the decision rule is utilized. As a result of
our data and the level of significance, the test will be a
non-directional test, with /2 = 0.025 or 2.5%. If -1.96
z 1.96, then the null hypothesis would fail to be

( ) Crime= 282.39/50 = 39.94


Confidence Interval:
CI =

z ( )

Due to the level of significance being 0.05, the z-value associated with a 95%
confidence interval is 1.96.

CI Pop = 4988 1.96 (1,005.70) =


3016.83 6959.17
CI Crime = 55 1.96 (39.94) =
-23.28 133.28

The selected test for the data, the change in


population and the change in Part I crimes, will be the
two sample t-test, or also recognized as the difference
of means test. It was thought to be the best choice for a
test to examine these two independent samples that
are being tested only once. The selection of these cases
for one sample does not affect the selection of cases for
the other sample. The Z-test was not chosen since there
were less than one hundred samples in total;
comprising of the fifty samples, or zip codes, from
Austins population change and the fifty samples, or zip
codes, from Austins Part I crime change. To get the
result of the t-test, it must begin with the standard
error of the difference of means, which originates from
the pooled variance estimate. The process goes as
such:
Pooled Variance Estimate:

Numerator of Equation:
= 50,571,774.29 (50-1) = 2,478,016,940

s22(n2-1)

= 79,744.24 (50-1) = 3,907,467.76


2,478,016,940 + 3,907,467.76 = 2,481,924,407.76
Denominator of Equation:

n1 + n2 2 = 50 + 50 -2 = 98
All together,
2,481,924,407.76
98

= 25,325,759.26 = 5032.47 = PVE

Standard Error Difference of the Means:


1

1- 2

= PVE (1 + 2)

1
1
+ 2
1

= 50 + 50 = 5

5032.47 (5) = 1006.49 =

1- 2

T-Test:
t=

1- 2

(n 1) + (n 1) = (50 1) + (50 1) = 98
Results
Obtainment of the critical value was derived
from MedCalc.org. Using the rounded d number of 100
in the table, the critical value found with a 0.05 level of
significance is 1.984. Subsequently, 4.90 > 1.98, which
results in a rejection of the null hypothesis, so there is a
difference between the mean of Part I crime change
and the mean of population change, in Austins zip
codes. The significance of different means of the two
independent samples is that they are not equal and
there is a difference between the amounts of change of
population and the change in Part I crime. The results
may lack practical significance, because a different
mean for the change of population will not describe any
different mean for the change in crime.
Conclusion

PVE =

2
S1 (n1-1)

degree of freedom (d) will be:

t=

4988 55
1006.49

= 4.90

To compare the test value to the critical value, the

The null hypothesis was rejected and the


alternative hypothesis proved to be true. This signifies
that there is a difference between the population mean
for change in Part I crime from 2003 2013, and the
population mean for change in population from 2000 2010. After further research, a possibly superior test for
this specific type of data would be the Welchs t-test.
The reason why this test might be more appropriate is
because the Welchs t-test does not require the
assumption of equal variance between populations.
However, there were extreme values for both the
change in crime and the change in population, which
would affect the assumption of normality. This data
follows more of an exponential distribution, due to the
values defining a growth or loss.
In closing, the next map displays the
combination of the change in Austin population and the
change in Part I crime. The zip codes to avoid due to
high Part I crime and the biggest increases in population
are in dark red: 78610 (Buda) and 78613 (Cedar Park).
The zip codes that have the safest Part I crime rate and
the largest loss of population change are in dark green:
78728 and 78722. While, the t-test proves there is a
difference of means between the two variables, this
map affirms that the more population increase an

Austin zip code undergoes, the more of Part I crime the


zip code suffers. Residents that have lived in Austin for a
length of time, or people that have just moved in, have
or will experience less or more crime based on a loss or
gain of population in their zip code.

information-listed-zip-code#overlay-context=user
(last accessed 25 April 2015).
Buthmann, Arne. Dealing with Non-normal Data: Strategies
and Tools. iSixSigma.
http://www.isixsigma.com/toolstemplates/normality/dealing-non-normal-datastrategies-and-tools/ (last accessed 26 April 2015)

Christie, Les. CNN Money. 2007. The fastest growing U.S.


cities: From the Empire State to the Lone Star State
the cities that are growing the most.
http://money.cnn.com/2007/06/27/real_estate/fast
est_growing_cities/ (last accessed 25 April 2015).
City-Data.com. 2015. Austin, Texas (TX) Zip Code Map
Locations, Demographics. http://www.citydata.com/zipmaps/Austin-Texas.html (last accessed
25 April 2015).
CLRsearch.com. 2012. Austin, TX Population Growth and
Statistics. http://www.clrsearch.com/78712Demographics/Population-Growth-and-PopulationStatistics?compare=78717 (last accessed 25 April
2015).

Biography of Author
My name is Carolina, but friends
call me Cat. I was born and raised in Austin,
Texas, so the study area for this article is
one close to my heart. I am majoring in
Environmental and Resource Studies, with
a minor in Nature and Heritage Tourism, and also a
certification in Environmental Interpretation. I hope to add on
to my accomplishments, as well as travel to more amazing
places.

MedCalc. 2015. Values of the t-distribution(two-tailed).


Belgium. https://www.medcalc.org/manual/tdistribution.php (last accessed 25 April 2015).
Texas Department of Public Safety. 2005. 2005 Crime in
Texas; Appendices, and Glossary of Terms.
https://www.dps.texas.gov/crimereports/05/appen
dices.pdf (last accessed 25 April 2015).
Townsend, Christi. 2015. Confidence Intervals, Hypothesis
Testing with Two Samples. San Marcos, Texas: Texas
State University.

References

Photo Citations

Albright, Elizabeth. Two Independent Sample (unequal


variance) Comparison of Means Test [Welchs t-test].
Durham, NC: Duke University, Nicholas School of the
Environment.
http://sites.nicholas.duke.edu/statsreview/means/w
elch/ (last accessed 26 April 2015).

James. 2011. Austin Skyline at Night. Photography-onthe.net: Digital Photography Forums.


http://photography-onthe.net/forum/showthread.php?t=1082185 (last
accessed 26 April 2015)

Austintexas.gov, the official website of the City of Austin.


2013. Austin Police Department, Crime Stats listed by
zip code, 2003 2013.
http://www.austintexas.gov/page/crime-

Jones, James. 2004. Technology Exercise 3: Inferential


Statistics. Richland Community College.
https://people.richland.edu/james/fall04/m113/tec
h/tech3-mtb.html (last accessed 26 April 2015).

S-ar putea să vă placă și