Red Voter, Blue Voter
The American National Election Studies is a collaboration between the U of Michigan and Stanford. They do polling of individuals during election years, asking them various questions about their political views, views of candidates, and demographics data. They make the full dataset available for download, so one can do data analysis.
N=939 (everyone who gave valid answers to the pertinent questions)
Independant variables:
B1: gave answer #1 to the bible origin question (roughly, the Bible is the word of God) [0-1]
B3: gave answer #3 to the bible origin question (roughly, the Bible was just written by men) [0-1]
M: answered "yes" to married or attached [0-1]
P: has a post-grad degree [0-1]
A: Self-identified as african-american [0-1]
H: Self-identified as hispanic [0-1]
I: Answer to household income level, on a 23-point scale [1-23]
Dependant variables:
DT: Bush temp + Rep party temp - Kerry temp - Dem party temp (temp is a "temperature" gauge from 0-100 on how much the respondent likes the entity)
Adjusted R-squared = .17
Intercept = -26.9
B1 coefficient = 21.7
B3 coefficient = -51.6
M coefficient = 19.6
P coefficient = -25.5
A coefficient = -76.1
H coefficient = -26.0
I coefficient = 1.92
Notes:
All coefficients were significant to 99%
No other way of chopping up education produced 95% significant results
No other ethnicity responses were 95% significant
No other way of chopping up income produced better results than simple linear
The questions asking about military service were not 95% significant
I failed to include any variables for children in the household
There's lots more that can be done with this dataset. This was just the result of me goofing around with the data for a couple of hours on Saturday night.
N=939 (everyone who gave valid answers to the pertinent questions)
Independant variables:
B1: gave answer #1 to the bible origin question (roughly, the Bible is the word of God) [0-1]
B3: gave answer #3 to the bible origin question (roughly, the Bible was just written by men) [0-1]
M: answered "yes" to married or attached [0-1]
P: has a post-grad degree [0-1]
A: Self-identified as african-american [0-1]
H: Self-identified as hispanic [0-1]
I: Answer to household income level, on a 23-point scale [1-23]
Dependant variables:
DT: Bush temp + Rep party temp - Kerry temp - Dem party temp (temp is a "temperature" gauge from 0-100 on how much the respondent likes the entity)
Adjusted R-squared = .17
Intercept = -26.9
B1 coefficient = 21.7
B3 coefficient = -51.6
M coefficient = 19.6
P coefficient = -25.5
A coefficient = -76.1
H coefficient = -26.0
I coefficient = 1.92
Notes:
All coefficients were significant to 99%
No other way of chopping up education produced 95% significant results
No other ethnicity responses were 95% significant
No other way of chopping up income produced better results than simple linear
The questions asking about military service were not 95% significant
I failed to include any variables for children in the household
There's lots more that can be done with this dataset. This was just the result of me goofing around with the data for a couple of hours on Saturday night.