IQ and Politics
Michelle Chan, Derek Chen, Ashley Hwang, Benjamin Sadun
San Marino High School, AP Statistics, Period 2
Advanced Statistics Project III:
IQ and Politics
“The great thing about democracy is that it gives every voter a chance
to do something stupid – Art Spander.” When the Constitution was drafted,
the first Republicans like Alexander Hamilton worried about mob rule, the
idea that the common people would be unable to make intelligent votes and
consequently vote for the wrong man. After the recent controversial
election of November 2004, web sites popped up all over the web denouncing
the supposed stupidity of Republicans in voting for Bush. Was this another
angry, sour grapes attempt to belittle the winners? Or is there truth in
After the 2004 presidential election, studies were conducted to find
relations between the general characteristics of the voters and which
political party they represented. Eager to discover whether these theories
are indeed valid, AP Statistics Group 8 decided to investigate the possible
correlation between the average IQ of each state and which party they voted
for in the last presidential election.
Although the choice of using the
2004 election does not model a simple random sample, the latest IQ of each
state would be most accurate to this year’s election, and we will also be
using the results of the past ten elections as the population to compare
with the data from the sample. While Internet rumors claim that the votes
for the presidential candidates were biased and skewed towards one party
because the Democrats are predominantly intelligent and the Republicans
predominantly of lower intelligence, Group 8 hypothesizes that these
reports are incorrect and that there is no correlation between the IQ of
the voters from a state and the party for which they voted because we
believe that IQ does not determine one’s intellect and that intelligence
does not reflect one’s choice of who to vote for. To verify the falsehood
of these Internet studies, the group needed to gather a set of accurate
data without bias.
An important factor in gathering accurate data was to collect different
samples of data. This bivariate data compared the average state IQ’s with
their dominant voting preferences. This data was numerical, discrete, since
it compared the scores of IQ to the percents of people who voted Democrat.
The measures of central tendency which include the mean, median, or mode of
the IQ’s, reveal the prevailing tendencies of that election, our sample.
Similarly, the measures of dispersion, range, and interquartile range, have
a real significance here as well since it would make sense to find out how
divisive the United States was when its citizens voted for Bush or Kerry.
This study was empirical in nature as it relied on observation to make
assumptions. In this observational study, we took existing data and drew
forth relationships between the two variables of interest.
Each website had a different method of measuring the IQ, and some sites
were more reliable than others; one site used the mean SAT and ACT scores
to convert them to an IQ scale, another site used high school degrees and
college degrees. We didn’t want to place blind faith in one site, so we
averaged the IQ’s given to get a more comprehensive data.
We decided to use
the mean since the IQ’s were close to each other, and there appeared to be
no outliers. The median is generally used when outliers persist in the
data. We didn’t stratify our data into the top 10 IQ’s and the last 10
IQ’s, as we felt it would not give the best overview of the issue. However,
we did choose to look explicitly at the percentage of voters who voted
Democrat, to narrow and simplify our data. This technique was known as
blocking, since it blocked out the extraneous factor of people voting for
other parties. This basically led to dichotomous data, since people would
either be voting Democrat, or not be voting Democrat.
generally allows us to calculate probability of an outcome using binomial
distribution. However, in this statistics project, the dichotomous data
created a sample space of Democrat or non-Democrats. We also decided to get
rid of the confounding variable of popular votes and electoral votes,
because the IQ is directly related the popular votes. Electoral votes only
represent a minority, and a highly educated minority at that, of the whole
In order to ensure the absence, or as little of as .