Academics tend to be liberal, this we all know. But it is interesting to contemplate the finer details of this association. When we do so, we discover some interesting differences among different kinds of academics, and one group in particular — “Health Scientists” — can be singled out as bucking the trend. Indeed, if Health Scientists were not allowed to vote in the last election, but all else was held constant, it is likely that John Kerry, not George Bush, would be president todayIn truth, I have no idea if this is true. I’m just remembering that the election was very close, so this is a reasonable guess..

A recently released Working Paper by Neil Gross and Salon SimmonsGross, Neil and Salon Simmons. 2007. The Social and Political Views of American Professors. Working Paper. Available here. provides what the Chronicle of Higher Education calls “arguably the best-designed survey of American faculty beliefs since the early 1970s.” This paper has been discussed at various points in the blogosphere.“The Liberal (and Moderating) Professoriate,” on Inside Higher Ed and “Do health sciences professors really prefer George Bush?” Marginal Revolution.

This study differs from earlier work in that the methodology is improved in ways that the study’s authors argue (convincingly) to be important (see pages 21-25 of their report). Here, I reproduce a selection of tables from their study. Most of these results are not at all surprising, but are nonetheless meaningful. Other results are understandable yet perhaps somewhat disturbing.

A description of the study sample is beyond the scope of this humble blog post, but suffice it to say that this represents college teachers working in programs that offer undergraduate degrees. An overview of the sample indicates a liberal bias:

To simplify: Although about one in five college profs are “middle of the road,” 62% are liberal and 20% are conservative. To put this in some context, a recent nationwide Washington Post-ABC News Poll “This Washington Post-ABC News poll was conducted by telephone September 4-7, 2007, among a random national sample of 1,002 adults. The results from the full survey have margin of sampling error of plus or minus three percentage points. Sampling, data collection and tabulation by TNS of Horsham, Pa.” this poll asks the question: “Would you say your views on most political matters are liberal, moderate, or conservative?” measures the Liberal - Moderate - Conservative spread as 20 - 39 - 39.

The following table gives the L - M - C spread for different academic areas. Not surprisingly, the most liberal are the social sciences and humanities (and Other, of course). The Physical and Biological sciences are most representative of the sample overall. It is also not surprising that Business shows up as most conservative, but with a very high proportion of “Moderate” respondents. I’m a little surprised that the “Conservative” category for business is not larger.

The next most conservative group is “Health Sciences.” This sample does not include professors of medicine … they are not teaching college students, but rather, post-graduates and MD’s. According to the report, Health Sciences “in our sample means mostly professors of nursing.”

The conservative nature of nursing professors and business professors may be explained by, or at least linked to, this next tabulation:

There is a general trend of conservative to liberal as one goes from community college through more elaborate “higher-level” institutions on to the PhD University type programs. This may reflect an overall trend we see in other contexts of education level and liberalism, which could be interpreted as a link between overall intelligence and liberalism (liberals being smarter, on average). However, we should not overlook the linkage between in-group and out-group effects and the culture of liberalism. Hiring and tenure within any given institutional framework is linked as much to cultural factors as it is to ability and performance. I once actually witnessed a tenured professor deciding to support a particular candidate based on his answer to this series of questions:

“When you have a drink, what do you drink” (Answer: Martini, of course.)
“Dry?” (Answer: Yes, of course.)
“Gin or Vodka?” (Answer: Vodka, of course.)]
“Very good. But, what brand of Vodka?” (Answer: I do not recall the answer. By this time I was busy gagging.)I am not making this up. I can produce witnesses.

One might wonder about the apparent reversal in the above mentioned trend when shifting from “non-Elite” to “Elite” programs. This is in fact difficult to account for, and the study’s authors attribute it to a trend, with the presumption that the “Elites” lead (or at least, that is how I read the report … it is a little vague in this area). The authors also site the role of liberal arts colleges on the liberal landscape as important.

Consider reformulating the data as follows: From Community College to Elite we find Liberal and Moderate combined to follow the sequence 81,87,96,96,90, with the most “non-conservative” being the Liberal Arts and Non-Elite institutions. (Or, more exactly, professors with said affiliation.) This is in accord with the idea that hiring practices involving consideration of political orientation rather than, say, teaching ability or pure research ability help shape the political character of these institutions. Yet another consideration is that the “Elite” institutions probably include a higher percentage of relatively wealthy individuals. This is not because you get paid more to work at the “Elite” institutions. This is often, in fact, not the case. However, there may be a culture of hiring and tenure in the Elite institutions that incorporates in-group vs. out-group effects that are not based on politics as much as they are on social factors that correlate with wealth. Wealth, in turn, correlates even among the academic elite with conservatism (maybe).

Now, consider party affiliation:

Remember that the Liberal - Moderate - Conservative spectrum (above) was 62 - 18 - 20. Assume that the corresponding Party Affiliation Spectrum would be Democratic - Independent - Republican, where Indy-Dems and Indy-Rep’s will be placed with the nominal party. This produces the following statistic: 71 - 8 - 21. By adding “Independent Republican” into “Independent” we get, essentially, a replica of the Liberal - Moderate - Conservative spectrum. This confirms what has always been suspected: Moderate independents are conservative.

The following table is the reported voting pattern for this group in the last presidential election:

This roughly mirrors the statistics for political affiliation. Here we see that the Health Sciences (Nursing) did marginally prefer Bush over Kerry. Does this imply that Nursing professionals or instructors have a view of the health care system and potential reform that runs counter to the idea of Universal Health Care? Or is this because most Nursing instructors are in the less-educated group (Masters Degrees)? One thing I suspect is this: Socially, culturally, there is a “metagroup” of nursing professionals and business professionals that internally divides itself into those two professions. I predict that if you look at the demographics of these two groups there will be considerable overlap, reflecting the idea that they are really one and the same cultural/social class, by and large. Why do I say that? This is nothing other than raw ethnographic observation, so don’t take it too seriously without further confirmation.

I find it fascinating that a relatively large percentage of the Business group voted for Nader, with the second largest group being the Social Sciences. I’m not surprised that some of the social science people chose to throw away their vote even if the consequences would be (and were) dire. Note that this group also has a relatively high percentage of “Other” candidates. But the business group going for Nader seems to me enigmatic. The authors of the study are silent on this issue. Perhaps they are dumbstruck, which is a reasonable response.

On a local note, we also see that Computer Science and Engineering voted in relatively high proportion for “Other.” My guess is that this subset … subculture … with relatively high geekiosity measures overlaps considerably with the large non-voting population that occassinally comes out of the woodwork and actually does vote, as happened here in Minnesota when Jesse Ventura was elected governor. He was voted into office mostly by male non-voters ranging in age from 20-40 or so … people who had never voted before (and likely, since). Interestingly, 96 percent of the respondents in this survey reported having voted in the last election. My prediction is that much of that 4% is among the Engineers.

The present discussion is a very cursory overview, just looking at the first-order analysis provided by these tables of percentages. The original study, of course, carries out more detailed analysis including regression models of various factors, and has much more to say. I recommend reading it. The conclusion of Gross and Simmons study is simple: The nature and character of political attitudes and actions among academics is more complex than previous studies have indicated.