Us Against Them: Part 5, Some Notes on Methodology

In light of some of the comments in this series I thought I'd sketch the methods used in Us Against Them: Ethnocentric Foundations of American Opinion. This post might shed some light on how Kinder and Kam assessed ethnocentrism, where they got their data, and how they drew their conclusions. Obviously, I can only offer a sketch. A fuller examination of their research would involve engaging the book directly.

1. The Data Sets
Kinder and Kam did not collect the data or ask the questions. Rather, they used two publicly available and non-partisan data sets, the General Social Survey (GSS) and the American National Election Studies (NES). Importantly, both the GSS and NES go to great lengths to select a representative and large sample of the American population. All in all, social scientists of every political persuasion--liberal and conservative--mine the GSS and NES as two of the best surveys of American attitudes and behaviors.

In short, we can't accuse Kinder and Kam of using loaded questions or missampling. They simply used the pre-existing and publicly available data of the GSS and NES data sets.

2. Measuring Ethnocentrism
As mentioned in an earlier post, Kinder and Kam assess ethnocentrism though in-group and out-group stereotypes. In both the GSS and NES a section of the interview is devoted to assessing group stereotypes. To give a flavor for this, here is the wording going into the stereotype section of the NES:

Now I have some questions about different groups in our society. I'm going to show you a seven-point scale on which the characteristics of people in a group can be rated. In the first statement a score of 1 means that you think almost all of the people in that group are "hard-working." A score of 7 means that you think almost all of the people in the group are "lazy." A score of 4 means that you think the group is not towards one end or the other, and of course you may choose any number in between that comes closest to where you think people in the group stand.

Where would you rate whites in general on this scale?

The interviewer then asks the same question about blacks, Asian Americans, and Hispanic Americans. After comparing how hardworking the groups are, the interviewer then asks the respondent to compare the groups on intelligence, trustworthiness, patriotism, etc.

Mathematically, Kinder and Kam calculate ethnocentrism as the simple difference between how you see your in-group relative to out-groups. That is, if your overall in-group score is equal to your average out-group score then your ethnocentrism measure is 0.0. If your score is positive then you tend toward ethnocentrism (i.e., you see your in-group as more hard-working, intelligent, trustworthy, etc. relative to out-groups). If your score is negative then you are the opposite of ethnocentric (i.e., you see out-groups in a more positive light compared to your in-group; obviously, this is relatively rare).

In short, the measure for ethnocentrism is simplicity itself: The subtraction of out-group ratings from in-group ratings.

3. The Relationship Between Ethnocentrism and Political Opinions
Once Kinder and Kam calculate ethnocentrism (a simple subtraction) they then correlate the score with other expressed attitudes surveyed that year by the GSS and the NES. For example, one year the NES asked respondents if they "felt sympathy" for the Iraqi civilians affected by American forces, if they "felt disgust" at the killing of Iraqi civilians, and if they "felt it was immoral" to bomb near Iraqi civilians. Using ethnocentrism scores to predict responses to these questions Kinder and Kam found these regression coefficients (taken from page 120):

Sympathy for Iraqi people: -.54
Disgust at Killing: -.19
Immoral to bomb near civilians: -.23

As you see, all the coefficients are negative. That is, as ethnocentrism increased sympathy decreased, disgust decreased, and moral outrage decreased. In short, ethnocentric respondents felt less empathy for Iraqi civilians affected by American forces.

Consider the data regarding the various forms of social welfare. In the NES respondents are asked if the government should "spend more" on various forms of social welfare. Among white respondents the regression coefficients using ethnocentrism to predict attitudes related to Food Stamps, welfare for women with many children, Social Security and Medicare were as follows (pages 186, 187):

Food Stamps: -.48
Welfare for Women with Many Children: -1.19
Social Security: .56
Medicare: .43

As you can see, the coefficients are negative for means-tested welfare (welfare for the poor) but positive for social insurance welfare. In short, as described in my last post, ethnocentrism among whites predicted lower support for means-tested welfare (welfare for them) but higher support for social insurance welfare (welfare for us).

I could go on, this is just a sample of the data I summarized in bullet-point form in the last post.

To summarize:

1. Kinder and Kam did not collect the data so they cannot be accused of using misleading questions or poor sampling to fit a liberal agenda.
2. The GSS and NES data sets are non-partisan and considered two of the best data sets available in tracking American attitudes, beliefs and behaviors.
3. Kinder and Kam's measure of ethnocentrism is simplicity itself: A simple subtraction between in-group and out-group stereotypes (standard sections of the GSS and NES). It is hard to see liberal bias in this method of assessment.
   
4. The bread and butter of Kinder and Kam's analysis is simply taking the ethnocentrism measure (#3) and correlating it with other attitudes/opinions assessed elsewhere in the GSS and NES surveys.
5. Kinder and Kam's findings are replicated in both the GSS and NES data sets, suggesting robust and replicable results.
6. Finally, given that these are publicly available data sets, you can go to the GSS and NES and conduct Kinder and Kam's exact analyses to verify their findings. The results and analyses are transparent and open to evaluation.