Education Level & Political Party…A Conspiracy That Reaches to the Highest Levels of Our Government?
A Statistical Investigation By Matt, Katelen, Anna, and Drew
A Description
• We were interested in political preference and education level. • We also collected extra data on age level
Experimental Design
• Randomly selected a number 1-350
– Split page into approx. 21 groups of 20 numbers – Selected another random number 1-21 to get sample.
Survey
• Script for Survey: ―Hi, I am from Ferris High School and am conducting a survey for my statistics class. Would you like to participate?‖
Survey (Cont.)
1. Of some high school, high school graduate, some college, 2 yr. Degree, 4. Yr. Degree, Masters degree or a Doctorate Degree, what is the highest level of education you have completed?
2. With which political party—democratic, republican, or other—do you consider yourself associated? 3. Which age group are you in?
~ 18-30 ~ 31-43 ~ 44-56 ~ 57+
Data Analysis
•In order to analyze data, we had to further stratify some of the categories. •Education level was split into low ( groups 1, 2, 3) and high ( groups 4, 5, 6, 7).
•Age group was split into young (18-43) and old (44+).
Data
Dem. Low Ed. 5 Rep. 10 Total 15
High Ed. 14
Total 19 Dem.
17
27 Rep.
31
46 Total
Young Old Total
7 12 19
12 15 27
19 27 46
Interesting Charts
Education Level/ Political Party
18 16 14 12 10 8 6 4 2 0 Dem Rep
Low Ed. High Ed.
More Interesting Charts
Age/ Political Party
16 14 12 10 8 6 4 2 0 Dem Rep Young Old
Confidence Intervals
1 prop. Z-int: Democrats w/ high education c = .99 47.6% - 99.7
1 prop. Z-int: Republicans w/ High education
c = .99 36.9% - 84.4%
1 prop. Z-int: Old Democrats c = .99 34.6% - 91.6%
1 prop. Z-int: Old Republicans c = .99 30.9% - 80.1%
More Data Analysis
Because all counts are greater than 1 and less than 20% of counts are less than 5, Chi-squared procedures are reasonable to test education level against political party preference. Is there an association between political party and education level? Chi-sq. Test of Independence
X² = .5833
-Fail to reject the null.
p = .4450
Data Analysis…Again
Because all counts are greater than 1 and less than 20% of counts are less than 5, X² procedures are reasonable to test political party preference against age. Does an association exist between age level and political party preference?
Chi-sq. Test of Independence
X² = .2658 Fail to reject the null. p= .6061
Conclusions
•Both p-values were well above the significance level of .05.
•There is not enough evidence to suggest an association between education level and political party preference.
•There is not enough evidence to suggest an association between age and political preference. •Response bias is always an issue, but because we properly incorporated randomization into our study, that bias should have been accounted for. •Originally expected a strong association between in both instances.