Wash. Times news, opinion pieces peddled flawed survey on liberal professorsApril 5, 2005 6:33 PM EDT ››› NICOLE CASTA, SIMON MALOY, & PAUL WALDMAN
Over the past week, The Washington Times ran one news report and columns by Suzanne Fields and Cal Thomas about a recent study indicating that more self-identified liberals than conservatives are serving as professors at U.S. colleges and universities, a conclusion reached by comparing data from faculty surveys taken in 1984 and 1999. Two of the articles repeat the claim that the study demonstrates a profound "ideological shift to the left among college faculty" and a pervasive anti-conservative bias in hiring and tenure decisions. In fact, neither conclusion is warranted based on the study itself.
The study, released March 29, is titled Politics and Professional Advancement Among College Faculty. Its authors are Stanley Rothman, director of the Center for the Study of Social and Political Change and Smith College professor emeritus of government; S. Robert Lichter, George Mason University professor and director of GMU's Center of Media and Public Affairs; and University of Toronto professor Neil Nevitte. The study was sponsored by the Randolph Foundation, a private philanthropy that funds many conservative organizations, such as Americans for Tax Reform, the Independent Women's Forum, and right-wing pundit David Horowitz's Center for the Study of Popular Culture.
The Washington Times reported in a March 30 news article that the study found that "nearly three-quarters" of faculty members describe themselves as liberals, according to 1999 data from the North American Academic Study Survey (NAASS), up from 39 percent in a 1984 survey by the Carnegie Foundation for the Advancement of Teaching. Relying on this comparison, the Times described a "shift to the left among college faculty [that] has become much more pronounced in the past 20 years." In fact, the two surveys examined such dissimilar samples that one cannot draw valid conclusions about a trend.
According to the study, the NAASS "American sample" included 1,643 faculty members from 183 universities and colleges. The responses came from "81 doctoral, 59 comprehensive and 43 liberal arts institutions." The 1984 Carnegie survey, however, contained "data obtained from over 5,000 faculty employed at a variety of institutions from Two-Year Community Colleges to Research Institutions." * Remarking on these two contrasting samples, the weblog Critical Montages observed that "the NAASS's exclusion of two-year colleges and overrepresentation of doctoral institutions is a recipe for accentuating the proportion of liberals":
Research has shown that faculty and students at research institutions are more liberal than those at primarily teaching institutions (see, for instance, Gordon Shepherd and Gary Shepherd, "War and Dissent: The Political Values of the American Professoriate," The Journal of Higher Education 65.5 [September/October 1994], especially p. 586; and Richard F. Hamilton and Lowell L. Hargens, "The Politics of the Professors: Self-Identifications, 1969-1984," Social Forces 71.3 [March 1993], especially pp. 608-609, 613-614, 616), so the NAASS's exclusion of two-year colleges and overrepresentation of doctoral institutions is a recipe for accentuating the proportion of liberals.
On the question of ideological orientation, the study's comparison of the 1984 and 1999 surveys violates a fundamental principle of survey research. As decades of research have shown, altering questions in even subtle ways can produce dramatically different results. Rothman, Lichter, and Nevitte base their conclusion that "a sharp shift to the left has taken place among college faculty in recent years" on questions asked in two entirely different ways in the two studies, one asking respondents to place themselves on a ten-point scale, and one asking them to select from a list of descriptions.
Does this mean that there has been no shift to the left among faculties? Not necessarily -- but with the available data we have no idea whether such a shift has occurred, and neither do Rothman, Lichter, and Nevitte.
The study attempted to depict an epidemic of "liberal bias" on campus by contrasting the alleged "sharp shift to the left" among college faculty to the "relatively stable" ideological makeup of the general public over time. This comparison has little illustrative value, however, since the vast majority of the general public lack the necessary credentials for a professorship at the surveyed schools. Moreover, available data suggest that highly educated Americans may be more left-leaning than the general population. Exit polls from the November 2004 presidential election indicate that 55 percent of voters who have postgraduate study experience voted for Democrat John Kerry, compared to 44 percent for Republican George W. Bush. (Interestingly, when New Yorker staff writer Nicholas Lemann asked Bush adviser Karl Rove how to identify "who's a Democrat" as opposed to a Republican for a 2003 profile, Rove answered: "Somebody with a doctorate.")
Both the news report and an April 4 Times column by Fields quoted Lichter -- whose Center for Media and Public Affairs states on its website that it conducts "scientific studies of the news and entertainment media" but receives funding from numerous conservative organizations -- saying that "this is the first study that statistically proves bias [against conservatives] in the hiring and promotion of faculty members." But Lichter's own study undermines this claim. The study specifically notes: "The results do not definitively prove that ideology accounts for differences in professional standing" [emphasis added]. Rather, the study concluded more modestly that the findings are merely "consistent with the hypothesis" of bias [emphasis added]. According to Lichter's study:
The results do not definitively prove that ideology accounts for differences in professional standing. It is entirely possible that other unmeasured factors may account for those variations. That said, the results are consistent with the hypothesis that political conservatism confers a disadvantage in the competition for professional advancement. ...
Our findings on the more controversial issue of discrimination against conservative faculty should be regarded as more preliminary. [PDF p. 15]
Considering that, according to Lichter's bio on the Center for Media and Public Affairs website, "Dr. Lichter also directs the Statistical Assessment Service (STATS), a nonpartisan organization dedicated to improving the quality of news involving statistical or scientific information," his statement is little short of shocking. Cross-sectional studies like those cited in Lichter's study seldom "prove" anything; at best they can demonstrate associations and relationships.
Furthermore, the study does not even show, much less "prove," that conservatives have been discriminated against in hiring and promotion. Few would doubt that liberals outnumber conservatives among university faculty. But justifying claims about hiring and promotion would require data on the number of conservatives and liberals who applied for various positions or came up for tenure review. Despite Lichter's comments, the study's authors present no data addressing the issue. (Academic promotion is extraordinarily complex; in such a study, researchers would have to determine, for instance, which respondents were denied tenure at a first-tier institution, then received tenure at a second-tier institution, then decide how such a person should be classified.)
The conservative claim of bias (as opposed to mere underrepresentation) rests on the idea that there are significant numbers of conservative Ph.D.s who have been denied faculty positions or tenure because of their political views. Lichter, Rothman, and Nevitte provide no evidence to support this assumption.
Fields similarly claimed that "a left-wing conspiracy -- or something close to it -- flourishes on the campus." But this conclusion assumes, implausibly, that hiring bias is the only conceivable explanation for overrepresentation of liberals on campuses.
Nationally syndicated columnist Thomas also referenced the flawed study in his March 30 column (reprinted in the April 4 Washington Times). He used the study as a springboard to inveigh against "liberalism on college campuses," which he criticized as fostering curricula that is "anti-American, anti-religious, anti-Israel, pro-homosexual rights and pro-abortion, often to the exclusion and ridicule of opposing views."
* Rothman, Lichter, and Nevitte reference the 1984 Carnegie survey in their bibliography as "Carnegie Foundation for the Advancement of Teaching. 1989. The Condition of the Professoriate: Attitudes and Trends. Princeton, NJ: The Carnegie Foundation for the Advancement of Teaching, 1989." Rothman, Lichter, and Nevitte do not provide information on the sample used in the 1984 survey. A Media Matters internet search did not produce the 1984 Carnegie survey, but it did produce references to the same survey (based on the bibliographic information) -- including the one linked above -- that contained information about the 1984 survey sample.