Is Social Science Politically Biased?
Political bias troubles the academy
By Michael Shermer on March 1, 2016
In the past couple of years imbroglios erupted on college campuses across the U.S. over trigger warnings (for example, alerting students to scenes of abuse and violence in The Great Gatsby before assigning it), microaggressions (saying I believe the most qualified person should get the job), cultural appropriation (a white woman wearing her hair in cornrows), speaker disinvitations (Brandeis University canceling plans to award Ayaan Hirsi Ali an honorary degree because of her criticism of Islam's treatment of women), safe spaces (such as rooms where students can go after a talk that has upset them), and social justice advocates competing to signal their moral outrage over such issues as Halloween costumes (last year at Yale University). Why such unrest in the most liberal institutions in the country?
Although there are many proximate causes, there is but one ultimate causelack of political diversity to provide checks on protests going too far. A 2014 study conducted by the University of California, Los Angeles, Higher Education Research Institute found that 59.8 percent of all undergraduate faculty nationwide identify as far left or liberal, compared with only 12.8 percent as far right or conservative. The asymmetry is much worse in the social sciences. A 2015 study by psychologist José Duarte, then at Arizona State University, and his colleagues in Behavioral and Brain Sciences, entitled Political Diversity Will Improve Social Psychological Science, found that 58 to 66 percent of social scientists are liberal and only 5 to 8 percent conservative and that there are eight Democrats for every Republican. The problem is most relevant to the study of areas related to the political concerns of the Leftareas such as race, gender, stereotyping, environmentalism, power, and inequality. The very things these students are protesting.
How does this political asymmetry corrupt social science? It begins with what subjects are studied and the descriptive language employed. Consider a 2003 paper by social psychologist John Jost, now at New York University, and his colleagues, entitled Political Conservatism as Motivated Social Cognition. Conservatives are described as having uncertainty avoidance, needs for order, structure, and closure, as well as dogmatism and intolerance of ambiguity, as if these constitute a mental disease that leads to resistance to change and endorsement of inequality. Yet one could just as easily characterize liberals as suffering from a host of equally malfunctioning cognitive states: a lack of moral compass that leads to an inability to make clear ethical choices, a pathological fear of clarity that leads to indecisiveness, a naive belief that all people are equally talented, and a blind adherence in the teeth of contradictory evidence from behavior genetics that culture and environment exclusively determine one's lot in life.
Duarte et al. find similar distortive language across the social sciences, where, for instance, certain words are used to suggest pernicious motives when confronting contradictory-- evidence deny, legitimize, rationalize, justify, defend, trivializewith conservatives as examples, as if liberals are always objective and rational. In one test item, for example, the endorsement of the efficacy of hard work was interpreted as an example of rationalization of inequality. Imagine a study in which conservative values were assumed to be scientific facts and disagreement with them was treated as irrational, the authors conjecture counterfactually. In this field, scholars might regularly publish studies on ... the denial of the benefits of a strong military or the denial of the benefits of church attendance. The authors present evidence that embedding any type of ideological values into measures is dangerous to science and is much more likely to happenand to go unchallenged by dissentersin a politically homogeneous field.
Political bias also twists how data are interpreted. For instance, Duarte's study discusses a paper in which subjects scoring high in right-wing authoritarianism were found to be more likely to go along with the unethical decisions of leaders. Example: not formally taking a female colleague's side in her sexual harassment complaint against her subordinate (given little information about the case). Maybe what this finding really means is that conservatives believe in examining evidence first, instead of prejudging by gender. Call it left-wing authoritarianism.
The authors' solution to the political bias problem is right out of the liberal playbook: diversity. Not just ethnic, race and gender but viewpoint diversity. All of us are biased, and few of us can see it in ourselves, so we depend on others to challenge us. As John Stuart Mill noted in that greatest defense of free speech, On Liberty, He who knows only his own side of the case, knows little of that.
