Just how much viewpoint diversity do we have in social psychology? In 2011 nobody knew, so I asked 30 of my friends in the field to name a conservative. They came up with several names, but only one suspect admitted, under gentle interrogation, to being right of center.

A few months later I gave a talk at the annual convention of the Society for Personality and Social Psychology in which I pointed out the field’s political imbalance and why this was a threat to the quality of our research. I asked the thousand-or-so people in the audience to declare their politics with a show of hands, and I estimated that roughly 80% self-identified as “liberal or left of center,” 2% (I counted exactly 20 hands) identified as “centrist or moderate,” 1% (12 hands) identified as libertarian, and, rounding to the nearest integer, zero percent (3 hands) identified as “conservative or right of center.” That gives us a left:right ratio of 266 to one. I didn’t think the real ratio was that high; I knew that some conservatives in the audience were probably afraid to raise their hands.

Some of my colleagues questioned the validity of such a simple and public method, but Yoel Inbar and Yoris Lammers conducted a more thorough and anonymous survey of the SPSP email list later that year, and they too found a very lopsided political ratio: 85% of the 291 respondents self-identified as liberal overall, and only 6% identified as conservative.

That gives us our first good estimate of the left-right ratio in social psychology: fourteen to one. It’s a much more valid method than my “show of hands” (which was intended as a rhetorical device, not a real study). But still, we need more data, and we need to try more ways of asking the questions.

A new data set has come in. Bill von Hippel and David Buss surveyed the membership of the Society for Experimental Social Psychology. That’s a professional society composed of the most active researchers in the field who are at least five years post-PhD. It’s very selective – you must be nominated by a current member and approved by a committee before you can join. Von Hippel and Buss sent a web survey to the 900 members of SESP and got a response rate of 37% (335 responses). So this is a good sample of the mid-level and senior people (average age 51) who produce most of the research in social psychology. Von Hippel and Buss were surveying the members’ views about evolution, to try to understand the reasons why many social psychologists distrust or dislike evolutionary psychology. At the end of the survey, they happened to include a very good set of measures of political identity. Not just self-descriptions, but also whom the person voted for in the 2012 US Presidential election. And they asked nine questions about politically valenced policy questions, such as “Do you support gun control?” “Do you support gay marriage?” and “Do you support a woman’s right to get an abortion?”

In a demonstration of the new openness and transparency that is spreading in social psychology, Von Hippel and Buss sent their raw data file and a summary report to all the members of SESP, to thank us for our participation in the survey. They noted that their preliminary analysis showed a massive leftward tilt in the field – only four had voted for Romney. I then emailed them and asked if I could write up further analyses of the political questions and post them at HeterodoxAcademy. They generously said yes, and then went ahead and made all the relevant files available to the world at the Open Science Framework (you can download them all here).

So here are the results, on the political distribution only. (Von Hippel and Buss will publish a very interesting paper on their main findings about evolution and morality in a few months). There are three ways we can graph the data, based on three ways that participants revealed their political orientation.

1) Self descriptions of political identity: 36 to one.

One item asked “Where would you put yourself on a continuum from liberal to conservative?” The 11 scale points were labeled “very liberal” on the left-most point and “very conservative” on the right-most point. If we do a simple frequency plot (a graph of how many people chose each of the 11 possible responses) we get the following:

Figure 1: Self-ratings of politics

The graph shows that 291 of the 326 people who responded to this question picked a left-of-center label (that’s 89.3%), and only 8 people (2.5%) picked a right of center label, giving us a Left to Right ratio of 36 to one. This is much higher than that found by Inbar and Lammers. The main source of political diversity appears to be the 27 people (including me) who self-identified as centrists.

2) Presidential voting: 76 to one.

Another item asked: “Who did you vote for in the last presidential election (if you are not a US citizen, or if you did not vote, who would you have voted for if you had voted)? The options were: “Obama,” “Romney,” or “Other.” If we do a frequency plot of the 3 possible choices we get this:

vonhippel.figure2

Figure 2: Presidential vote in 2012 US Election

The graph shows that 305 of the 322 people (94.7%) who responded to this question voted for Obama, 4 (1.2%) voted for Romney, and 13 (4.0%) said they voted for another candidate. This gives us a Democrat to Republican ratio of 76 to one.

3) Views on political issues: 314 to one.

A third way of graphing the viewpoint diversity of these senior social psychologists is by computing an average score across all 9 of the politically valenced policy items. For each one, the 11 point response scale was labeled “strongly oppose” on the left-most point and “strongly support” on the right-most point. I converted all responses to the same 11 point scale used in figure 1 so that “strongly supporting” the progressive position (e.g., pro-choice) was scored as -5 and “strongly supporting” the conservative position (e.g., prayer in school) was scored as +5. That puts the leftists on the left and the rightists on the right of the graph. Here’s the graph:

vonhippel.figure3

Figure 3: Views on 9 political issues

I counted anyone whose average score fell between -1.0 and +1.0 (inclusive) as a centrist. The graph shows that 314 of the 327 participants (96.0%) had an average score below -1.0 (i.e., left of center), one had an average score above +1.0 (i.e., right of center), and 12 were centrists. That gives us a Left to Right ratio of 314 to one.

What does this mean?

However you measure it, and for all samples measured so far, social psychology leans heavily to the left and has very few people right of center. Von Hippel and Buss’s new data confirms the story that a few of us told in a recent paper (Duarte, Crawford, Stern, Haidt, Jussim & Tetlock, 2015) in which we created the graph below, which shows just how fast psychology has been moving to the left since the 1990s. The ratio of Democrats to Republicans (diamonds) and liberals to conservatives (circles) was roughly 3 to 1 for most of the 20th century. But it skyrockets beginning in the 1990s as the Greatest Generation retires and the Baby Boomers take over.

diversity graph

Figure 4: The rising ratio of Left to Right among academic psychologists since the 1920s. (Details in Duarte et al., 2015)

Why does this matter?

Most people know that professors in America, and in most countries, generally vote for left-leaning parties and policies. But few people realize just how fast things have changed since the 1990s. An academic field that leans left (or right) can still function, as long as ideological claims or politically motivated research is sure to be challenged. But when a field goes from leaning left to being entirely on the left, the normal safeguards of peer review and institutionalized disconfirmation break down. Research on politically controversial topics becomes unreliable because politically favored conclusions receive less-than-normal scrutiny while politically incorrect findings must scale mountains of motivated and hostile reasoning from reviewers and editors.

I consider the rapid loss of political diversity, over the last 20 years, to be the second-greatest existential threat to the field of social psychology, after the “replication crisis.” The field is responding constructively to the replication crisis. Will it also attend to its political diversity crisis? Or will it continue to think of diversity only in terms of the demographic categories that most matter to people on the left: race, gender and sexual orientation?

I don’t mean to single out social psychology. It is the field that I know best, but what we have learned at Heterodox Academy is that this problem, this rapid shift to political purity, has happened to most fields in the humanities and social sciences in just the last 2 decades.

An optimistic ending

I would like to end by thanking my colleagues. I have been raising a fuss about these issues since 2011. In that time I also moved from the left to the center, politically. I am no longer a progressive. So you might expect that I’ve been ostracized, but I have not. Nothing bad has happened to me. Some of my colleagues believe that the political imbalance is not a problem. But the majority response has been, roughly: “This is really interesting. We really truly value diversity, and we agree with you and your co-authors that diversity of viewpoints is the kind that confers the most benefits on groups. But gosh, how are we going to get more?”

That’s our mission at Heterodox Academy – to figure out how to get more. It will be hard, but it can and must be done. Please see our “solutions page.”

—————————

Post script: Paul Krugman recently referred to us at HeterodoxAcademy as “outraged conservatives,” and he said that the leftward shift in the academy was really just the rightward shift of the Republican Party since the 1990s. He suggests that professors didn’t change their views on policy, they just stopped identifying as Republicans as the party went off the deep end. There is surely some truth to Krugman’s argument, but that doesn’t negate our claim that the makeup of the professoriate really did change after the Greatest Generation retired. Krugman’s argument could not explain graph #3, for example, which shows just a single person with views on social issues that are right of center. Also, I should point out that most of us at Heterodox Academy are not conservatives, and if you read everything on our site, it will be hard to find evidence of “outrage.”