Appendix B: Politics and Pollution: National Discoveries from ToxMap

I had imagined, before I came, that the more polluted the place in which people live, the more alarmed they would be by that pollution and the more in favor of cleaning it up. Instead I found Louisiana to be highly polluted, and the people I talked with to be generally opposed to any more environmental regulation and, indeed, regulations in general. So was Louisiana an oddball state in this regard, or not?

According to previous research, the more polluted a state is, the more likely it is to vote red (chapter 5). So far from being an oddball state, what was true for Louisiana was true nationwide. But what went on within each red state? Was it, as journalist Alec MacGillis claimed in the New York Times, that within red states the people facing poverty, poor schools, and broken families didn’t show up in political polls because they didn’t vote at all and meanwhile, others living in the same state, two class levels higher, did vote as Republican? If we apply MacGillis’s logic, we might expect that the people who live near polluting industry wish the polluters were regulated but don’t bother to vote while Republicans, who are richer and live in cleaner places, don’t think there’s a problem and so reject the idea of regulating polluting industry. Maybe.

But a second possibility is more puzzling: did the same person both face pollution and vote against regulating polluters? Rebecca Elliott and I set about finding out by combining two key sets of data. One was from the General Social Survey (GSS). Run by the National Opinion Research Center at the University of Chicago, the General Social Survey is widely regarded by social scientists as one of the best datasets on social trends in the country. The survey asks people to rate their response, on a scale from “strongly agree” to “strongly disagree,” to such statements as “people worry too much about progress harming the environment,” “Industrial air pollution is dangerous to the environment,” “The U.S. does enough to protect the environment,” and “Some people think that the government in Washington is trying to do too many things that should be left to individuals and private business.” We received permission from the National Opinion Research Center to analyze three thousand anonymous answers to these questions in the survey for 2010.

The second source of information was the Toxics Release Inventory of the Environmental Protection Agency. The Toxics Release Inventory (TRI) calculates various measures of exposure to toxic chemical releases and waste disposal. Based on reports from industrial and federal facilities, the most comprehensive of TRI measures are called the Risk-Screening Environmental Indicators (RSEI). For any particular zip code in the country, it gives us a measure of inhabitants’ exposure based on three things—the volume of chemical releases, the degree of toxicity of those chemicals, and the size of the exposed population. We used risk-screening information for the matching year, 2010.

Then we interrelated information on political choice and attitudes about the environment on one hand with the actual risk of toxic releases in the county in which a person lives, on the other hand. With the help of Dr. Jon Stiles, director of U.C. Berkeley’s Data Lab, and consulting with Professor Mike Hout of the Sociology Department at New York University, we used “bridging software” to do this.

Using regression analysis, we tested whether the riskiness of where one lives (RSEI score) “predicts” one’s answer to various environment-related questions (from the GSS). We also reversed the analytical arrow to see if various sociodemographic and political variables predict the riskiness of the places in which a person lives. We also examined the relationship between riskiness associated with one’s place of residence and one’s political orientations generally.

The most interesting findings are these: as the relative riskiness of the county a person lived in increased, the more likely that person was to agree with the statement “People worry too much about human progress harming the environment.” So the higher the exposure to environmental pollution, the less worried the individual was about it—and the more likely that person was to define him- or herself as a “strong Republican.”

Those who identified themselves as male, high income, conservative, Republican, Christian, and “strongly religious” were also likely to believe that air and water pollution were not a danger. Further, the higher the risk within a person’s county, the more likely a person was to agree with the statement that “the U.S. is doing more than enough to protect the environment.” Again, curiously, the higher a person’s exposure to pollution, given their residence, the more likely the individual was to think the United States is, in general, overreacting to the issue.

This is a paradox, but not one born of ignorance. For the greater the risk of exposure to pollution, the more individuals are likely to answer “agree” to the statement “Industrial air pollution is dangerous to the environment.” The better off and more educated among them also expressed the idea that humankind can improve the environment. They disagreed that “it is too difficult to do anything about the environment.”

In the end, red states are more polluted than blue states. And whether an individual does or doesn’t vote, conservative and Republican individuals tend to brush aside the environment as an issue, and to suffer the consequences by living with higher rates of pollution. The Louisiana story is an extreme example of the politics-and-environment paradox seen across the nation.