Twitter roundup

Comments

[Rob]

Brendan points us to Vavreck and Enos's essay on YouGov in which they address Gingrich's statement about Obama being "the greatest food stamp president in history." The authors tell us that Gingrich "is trying to link Obama to policies that help Black Americans" and that Gingrich's "strategy" is "to link Obama to food stamps in order to remind them that Obama favors policies that help Blacks --- and, oh yeah, that Obama is Black. Racial prejudice will do the rest."

Unless Gingrich has described these things as his intention and his strategy, Vavreck and Enos are engaging in a little mind-reading. It's bad enough when journalists and pundits engage in mind-reading, for which Brendan routinely breaks out his swami graphic. But at least pundits and columnists might be said to be paid to speculate about such things. When political scientists like Vavreck and Enos make unequivocal assertions about somebody's state of mind, they not only engage in speculation, they cheapen the currency of political science. They've earned not only swami hats but dunce caps as well.

The authors debase science in another way as well. They tell us, "Marty Gilens has shown definitively that food stamps are associated with African Americans in popular imagination," linking to the Amazon page for his book Why Americans Hate Welfare. Focus please on that word "definitively." Gilens book was published in 2000, based, it appears, on research conducted years or even decades before then, and certainly subject to all the vagaries inherent in seeking by questionnaires or interviews of limited samples to establish the attitudes of the general population. How could such survey results establish "definitively" what is in the "popular imagination" in 2012?

Last month sixteen scientists signed an opinion piece in the Wall Street Journal that included this statement:

In September, Nobel Prize-winning physicist Ivar Giaever, a supporter of President Obama in the last election, publicly resigned from the American Physical Society (APS) with a letter that begins: "I did not renew [my membership] because I cannot live with the [APS policy] statement: 'The evidence is incontrovertible: Global warming is occurring. If no mitigating actions are taken, significant disruptions in the Earth's physical and ecological systems, social systems, security and human health are likely to occur. We must reduce emissions of greenhouse gases beginning now.' In the APS it is OK to discuss whether the mass of the proton changes over time and how a multi-universe behaves, but the evidence of global warming is incontrovertible?"

Incontrovertible. Definitive. These are not the words of science, they're the words of polemics, calculated to foreclose inquiry rather than foster it. Professors Vavreck and Enos should know better.

[David in Cal]

Re: Can Mitt Romney seal the deal with the Christian right?

This article shows a regresion line based on 5 state primaries. It shows that Romney's percentage of votes received correlates negatively with the state's percentage of evangelical Christian voters. It then deduces that Romney is less popular with evangelical Christians than with other Republicans.

From a statistical POV this analysis is almost worthless. There are so many other differences among these 5 primary elections, and the number of points used is so small, that it's invalid to deduce that percentage of evangelicals is the dominant cause of the different results.

What makes this analysis even dumber is that exit polls are available that specifically measure Romney's popularity among various subgroups of voters. E.g., there's no need to use a dubious regression analysis to guess whether Romney's poor showing in South Carolina was due to evangelicals voting against him or to some other cause. Exit polls directly show that evangelical Christians accounted for nearly two-thirds of the electorate and Gingrich easily beat Romney in that group: 44 percent to 21 percent.

[David in Cal]

I agree with Brendan and with Timothy Ogden that academic incentives are less than ideal to get policy & practice right. Ogden pointed out several reasons why the focus on publishable research limits the knowledge gained. Here are three others:

1. Research in politically incorrect areas will have difficultly getting published. E.g., an academic wouldn't get far with research (hypothetically) showing that blacks or hispanics are genetically inferior.

2. Funding drives research. E.g., in the field of climate change, funding for studies that will support a warmist POV dwarfs funding for skeptical studies. It's therefore no surprise that most studies support the warmist POV.

Another example: During the Nixon Administration, there was a "War on Cancer." In practice this meant greater funding for cancer-related studies, some of little value. So, cancer studies increased, while potentially more valuable basic science studies decreased.

3. A paper that uses sophisticated statistical methods has a greater chance of being published. So, academic researchers look for areas where they can conveniently obtain enough data to do a nice analysis. Meanwhile, more important questions for which convenient data doesn't exist may be ignored.

Take election fraud, for example. It's a potentially significant problem, but it's hard to study. Most instances of voting fraud are probably never made public. The ones that do get investigated and prosecuted have dissimilar characteristics. IMHO a good study of voting fraud would have great value to society and to our lawmakers, but the difficulties involved discourage academics from doing such studies.

[Rob]

I too was happy to see Timothy Ogden's essay and posted this comment about it at The Monkey Cage:

Timothy Ogden makes excellent points, though his suggestions for what should be done seem to assume that academic institutions are incapable of reforming themselves. The biases he describes on the part of academic journals and the disincentives for junior faculty to engage in different kinds of research can be changed if tenured faculty will step up to the plate and address these problems. Is that really too much to expect?

So far none of the tenured faculty members who host The Monkey Cage have responded.