GS 104: Public Science

(on analyzing Internet charts on Democrat/ Republican voting and IQ)

see Gallup Poll data below on college grad voting patterns


I) Apparent Hypothesis/Assertion/Idea: Since this class is about science and the processes of science and physics is just the vehicle for discussion I feel that there is merit in this examination of a current topic. I was also recently (this weekend) gratuitously insulted based on the same basic idea presented here. I would further suggest that some of the same data analysis issues play a role in any implementation of the energy policies you proposed as the Energy Czar or Czarina of the World. So here is what rolled across my desk -- -- take a look at this and think about what your first response is to the data being presented and the questions that come to mind for you and your group. Then answer the following questions.

1) What is the idea or thesis which is being put forward in this data table?

2) What is the source of the data for this particular chart?

3) Assuming for the moment that you are interested in figuring out this potential relationship, list at least 4 different methods you might use to generate the underlying data. I want you to think about how you might generate data to address this question that would seem well thought out and defensible.

II) A better version??: So, following along in the same vein someone else tried to address some of the issues you undoubtedly noticed by taking a different measure of state by state IQ's -- . Take a look at this data and think about your personal and rational responses to the data. Answer the following questions.

1) What is the difference between a correlation in the data and a causal relationship? How does one tell the difference between the two?

2) In what way(s) is this data set "better" in a scientific sense (clear meaning, reproducible, documented sources) than the original data?

3) If you were to accept this data as (more) reliable is the correlation suggested by the data absent, present, stronger, or weaker than that of the original chart? Explain your reasoning.

III) Another measure: So here -- -- is a final version of this ongoing discussion. In this case the hot button language of IQ's has been dropped in favor of a more moderate and quantifiable variable. Again, as you look at this data consider how you respond to the correlation being suggested and whether it is the same as the previous data.

Sloan's note: See the Gallup poll data below on the educational level of Bush/ Kerry voters

1) In what ways is this data seeking to address similar correlations as the previous data and it what ways does it seem to be suggesting different correlations?

2) In what way(s) is this data set better or worse in a scientific sense (clear meaning, reproducible, documented sources) than the previous data?

3) What other possible correlations does this suggest would be of interest to explore? What other features of citizens might be connected with their educational background that might show an equivalent correlation with voting behavior?

IV) Graphing the Data: Graphs are great at helping us more effectively correlate data. Because of the visual processing skills of our brains we can extract a great deal more information from graphical plots than from tables. This being the case, the following are some questions you would need to answer in order to make such a plot.

1) If we are seeking to determine if there is a correlation between educational background what would be reasonable choices for the axes of your plot? What units would you use for each axis?

2) How would the data you require for such a plot be different from, or the same as, the data used to generate the tables we have been looking at in this lab? What would you expect the data to look like if there is a correlation between these factors? What would the data look like if there is not a correlation between the factors?


Gallup's analysis of their final pre-election poll contains some interesting data on the educational level of voters (From Original, restricted data source:

................................................% Bush 2000...Bush, 2004..Change
Postgraduate education.....................43...............47............+4
College graduate (no postgrad).........55...............58............+3
Some college....................................53...............56............+3
High school or less education............45................46...........+1

compare to CNN 2004 election data


Postgrad Stdy (16%) 44% Bush 55% Kerry

College Grad (26%) 52% Bush 46% Kerry

Some College (32%) 54% Bush 46% Kerry

H.S. Graduate (22%) 52% Bush 47% Kerry

No High School (4%) 49% Bush 50% Kerry

All voters (100%)

Go to: Calculating state IQ's from SAT and ACT scores

Go to: What school testing shows about state IQ's

Go to: Basic IQ information

Go to: Home Page

Comments to:

Web (this website, 170+ pages on IQ and Success skills)

The ads below are placed by - they are not necessarily endorsed by this site