StatCrunch logo (home)

Report Properties
Thumbnail:

from Flickr
Created: Mar 23, 2015
Share: yes
Views: 956
Tags:
 
Results in this report
 
Data sets in this report
None
 
Need help?
To copy selected text, right click to Copy or choose the Copy option under your browser's Edit menu. Text copied in this manner can be pasted directly into most documents with formatting maintained.
To copy selected graphs, right click on the graph to Copy. When pasting into a document, make sure to paste the graph content rather than a link to the graph. For example, to paste in MS Word choose Edit > Paste Special, and select the Device Independent Bitmap option.
You can now also Mail results and reports. The email may contain a simple link to the StatCrunch site or the complete output with data and graphics attached. In addition to being a great way to deliver output to someone else, this is also a great way to save your own hard copy. To try it out, simply click on the Mail link.
Social Survey
Mail   Print   Twitter   Facebook

Introduction:

In 2008, a large study was conducted to find out about several factors about the subjects and their demographics. These included education degree, happiness, health, parent’s standard of living, children’s standard of living, and how they compare taxes of the rich. These factors mentioned were all rated on a scale of 1-4, 4 being the highest in most cases. Other characteristics noted were race, political party, political view, and marriage status.

 

Data/Methods:

A total of 4250 subjects were asked each of these questions on the rating scale, and also which categories they fall into. The data was collected to determine if there was any evidence of differences between the different demographics and ratings of the responses. The data collected that was analyzed was at an ordinal level, indicating the use of the Kruskal Wallis test. A significant p-value indicated the need for follow up with the Mann Whitney comparisons to determine where these differences exist.

Analysis:

The Kruskal Wallis test provided a p-value of <0.0001, indicating that the median degree of education differ by political party. In this instance, the medians are all in the same rank category, however, the differences in rank of the total numbers does show us that Republicans are a bit higher than both Democrat and Independent. The significant p-value led us to perform a Mann Whitney test to see where these differences exist. A Bonferroni adjustment was applied to control for type one error; significance level of 0.05/3  means that the possible intervals= 0.0167. 

Result 1: Kruskal-Wallis Table   [Info]
Kruskal-Wallis results:
Responses stored in DEGREE.
Factors stored in PARTY.
Results adjusted for ties
DFChi-SquareP-value
2 67.262716 <0.0001

Summary statistics
PARTYnMedianAve. Rank
Democrat139812138.6395
Independent173311978.1864
Republican111812335.5188

The Mann Whitney results below indicate that there is significant difference in medians between the three types of political parties and level of education, with the p-value less thatn 0.0167 in all 3 analysis. Note the difference estimates is equal to 0 in each pairing, which could potentially mean our sample size is very large, because we know based on the Mann Whitney test that the difference in medians cannot truly be 0.

Result 2: Mann-Whitney DvI   [Info]
Hypothesis test results:
m1 = median of DEGREE where PARTY= Democrat
m2 = median of DEGREE where PARTY= Independent
m1-m2 : m1 - m2
H0 : m1-m2 = 0
HA : m1-m2 ≠ 0
Differencen1n2Diff. Est.Test Stat.P-valueMethod
m1 - m213981733022791510.0001Norm. Approx.

Result 3: Mann-Whitney RvI   [Info]
Hypothesis test results:
m1 = median of DEGREE where PARTY= Republican
m2 = median of DEGREE where PARTY= Independent
m1-m2 : m1 - m2
H0 : m1-m2 = 0
HA : m1-m2 ≠ 0
Differencen1n2Diff. Est.Test Stat.P-valueMethod
m1 - m21118173301758813<0.0001Norm. Approx.

Result 4: Mann-Whitney RvD   [Info]
Hypothesis test results:
m1 = median of DEGREE where PARTY= Republican
m2 = median of DEGREE where PARTY= Democrat
m1-m2 : m1 - m2
H0 : m1-m2 = 0
HA : m1-m2 ≠ 0
Differencen1n2Diff. Est.Test Stat.P-valueMethod
m1 - m21118139801477818<0.0001Norm. Approx.

Conclusion:

 

With the analysis conducted to see the differences in political party with education level, there does appear to be a significant difference in at least two of the medians, but not enough that it could be seen in the median rankings, just on the overall numbers of subjects. This does not take into account that in this study there was many missing responses with subjects which could alter the median outcomes if those subjects were to give their rankings and political party. 

HTML link:
<A href="https://www.statcrunch.com/5.0/viewreport.php?reportid=48987">Social Survey</A>

Comments
Want to comment? Subscribe
Already a member? Sign in.
By taylor.medley
Mar 25, 2015

Stephanie,
This report allowed me to see the difference in analysis of results when the sample size is very large. Since there were significant p-values it allowed us to conclude that there were differences between the political parties in regards to level of degree, however with a sample size this large I agree that it would be possible to detect very small differences. When referencing back to the Kruskal-Wallis test it appears that the Republican party has a slightly higher median education level than both Democrat and Independent. In this case I imagine the large sample size limited our ability to determine if these differences are in fact large enough to be clinically significant. Thanks for sharing!
Taylor Medley
By michele.wilbers99
Mar 24, 2015

Stephanie,

Nice report. Clear, concise and organized. One item that I noted is that the authors failed to take into account is the salary range of the study participants. This may impact the survey results which could present a limitation to the study.

As an election year is closely upon us, researchers could share this data with campaign managers in order to help steer their campaign.

Thanks for sharing,

Michele
By claire.parker123
Mar 24, 2015

Stephanie,

Good report, it was very easy to follow. Your interpretations of all of the parties having differences of means are correct since of the significant p values. I wonder if the other factors such as happiness and standard of living would of yielded different results.
-Claire
By emily.turner20
Mar 23, 2015

Stephanie,
I used the same data set but looked for evidence of difference between degree rating and race. I agree with your analysis that there is evidence of difference in medians in the political party and race. I also agree that the sample size is very large and whether the difference is significant is difficult to determine. Great report, clear results.
Emily

Always Learning