What's With the Emerging Gender Gap in Social Psychology?

Disparity in leadership positions.

Posted Jul 18, 2017

Do you believe that "gaps" necessarily reflect discrimination? Do you believe that backlash against strong, powerful women is so strong, that women can almost never attain important leadership positions?  If so, then consider this.

The Society for Personality and Social Psychology (SPSP) is one of the two main professional organizations for social psychologists, and it is the only one open to undergraduates and graduate students (there are lots of specialty organizations that many social psychologist belong to that I am not including here).

Here are the last two Presidents and the President Elect (Diane Mackie, Wendy Wood, Lynne Cooper):

SPSP
Source: SPSP

It is perhaps worth noting that the next President after M. Lynne Cooper has also been elected -- and that is Linda Skitka, who appears below.

That is four women in a row.  Many of my colleagues write as if "gap = discrimination against the under-represented group."1  In a perfectly gapless world, which many of my colleagues seem to argue for, the probability of four women in a row being elected President would be about 6% (according to the binomial theorem, which sounds very statistically sophisticated but anyone can compute these odds with an online calculator such as this one; in this case, though, you just need 7th grade math: .54 = .0625).  This is a nearly statistically significant bias in favor of women in the last four elections (traditional cutoff for considering something "statistically significant" is a probability of 5% or less). Of course, if you go back further, there are more men than women, but that is the point of this blog -- this gap is emerging, the point is not that it has been everpresent.

Here are the members of the current executive board, most of whom are elected.

SPSP
Source: SPSP

The last member (Rummel) is not elected, so let's exclude him.  6 of 8 elected members to the leadership positions are women.  The probability of this occurring by chance in a gapless world, is about 1 in 7.  Not a "statistically significantly different than a 50-50" gap, and yet ... 

If we combine the results for the 8 elected leadership positions and the four most recent Presidents, 10 of 12 are women.  The probability of this occurring by chance in a gapless world is about 2 in 100, again using the binomial, which assumes independent probabilities -- an assumption which may not be viable, as I discuss below.  

Here is a screenshot of the membership.  The totals appear on the right in red, but what makes this an "emerging" gap is the steadily increasing size of the gap as members go from older and more senior (on the left) to younger and more junior (on the right).

SPSP
Source: SPSP

(If that appears difficult to see, most browsers will allow you to zoom in; it is clear when larger).  Note that is data from 12/31/16.  SPSP also has data from February, 2015 here. There is a larger proportion of women in every category on 12/31/16 than in February 2015, reflecting an emerging gender gap in SPSP membership. Even the trend among full professors is moving towards such a gap (among full professors, men outnumbered women 42 to 36% in 2/15, numbers which are now nearly dead even; this trend, combined with the much larger differences among people early in their careers renders it likely that the overall gender gap will emerge among full professors some time in the next few years).  

If the gap was reversed, the typical author would, about now, be telling compelling narratives about the power of prejudice (whether intended or not) and discrimination to create such gaps (see my previous post for some linked examples).

But I have no evidence of discrimination.  Well, that is not quite true.  There is quite a lot of evidence of pro-female bias in academia (again, see my previous post). The problem with the pro-female discrimination interpretation here is that there is also plenty of evidence of pro-male bias, and plenty of evidence of no bias.  But if I wanted to tell a "compelling narrative" I would do so by ignoring the latter evidence and only focus on the evidence of pro-female bias.  And I do want to tell a compelling narrative.  I just do not want to tell one that is not actually true.  In fact, I do not even want to tell a compelling narrative based on equivocal evidence, either.

 Gender Gap in Carrying of Bulldogs in Backpacks
Source: Lee Jussim: Gender Gap in Carrying of Bulldogs in Backpacks

Here is just one concrete reason this "gap" may not reflect discrimination. The President of SPSP is often a person with a long history of involvement and service to the organization.  If, for (undetermined reasons, which may but do not necessarily involve discrimination), women in social psychology have been very active in nonpresidential leadership roles (as they are now in the elected leadership), the run of female presidents may simply reflect a larger pool of female social psychologists highly active in the organization.  

The run of female presidents, then, would be meritocratic, not "bias." Indeed, this is why the binomial theorem producing a probability of 2 in 100 may not really be justified either.  Combining the elected presidents and leadership may NOT involve "independent probabilities."  That is, if the presidents tend to come from the lower levels of elected leadership, and those lower levels of leadership are mostly women, then the "gapless world of 50-50 women and men" is not a justified assumption; the probabilities are no longer independent.  (However, the binomial probability is still valid for answering the question, "What is the likelihood of 10 of 12 leadership positions being women, if the odds of electing a woman exactly equaled the odds of electing a man for each position?").

Now, it is possible that bias of some sort leads to the gender gap in the leadership at lower levels.  For example, my previous post links several papers showing ingroup biases on the part of women faculty (i.e., studies showing women favor women, holding quality constant).  But just because some papers have shown such biases, one cannot presume such biases characterize the women of SPSP, because none was a recent study of the leadership gender preferences of women of SPSP.

The fact that bias is "possible" does not make it "true."  The fact that bias has been found in research in one context does not make it true in some other context, especially if there is research showing no bias or bias in the reverse direction in some contexts.  So many other things may have led to that gap that I cannot even begin to discuss them here.  In fact, it is even possible that pro-male bias has led to the pro-female gap in the leadership! (bonus point for commenters: Can you come up with a possible pro-male bias explanation that could lead to this bias in the SPSP leadership? I am not asking for evidence, just a speculative possibility.  It is not that hard...).

But if pro-male bias has somehow led to this gap, it means this gap favoring women does not reflect pro-female bias.

And that, gentle reader, is a terrific example of why the "gap=discrimination2 against the underrepresented group" assumption ain't necessarily so.

----------------------------------------------------------------------------

All ideas presented here are mine and mine alone.  However, I did run this essay by Linda Skitka, one of the future Presidents of SPSP.  She confirmed that the data on which it is based are accurate, and expressed agnosticism on how to interpret those data.  I completely concur with and admire that agnosticism, something which she constructively brings to much of her own work on political psychology (in contrast, e.g., to an ax-grinding agenda that characterizes some work in political psychology that is plausibly interpretable as setting out to prove how evil, immoral, and incompetent conservatives are).  I would add this, though: More of that same agnosticism by other researchers whose work reveals traditional gender gaps would be invaluable across the board.

1 That is, the "gap=discrimination" assumption is pervasive, unless the underrepresented group is a group they tend to be ideologically opposed to, such as conservatives, nonliberals of any stripe, and people who are religious. Then the alternative explanations for "gap=discrimination" start coming out of the woodwork. For scholarly sources demonstrating the extent to which nondiscrimination hypotheses are deemed not merely plausible but presumptively true, see the many commentaries on my article arguing that there is widespread bias against conservatives in social psychology.  For info on the massive underrepresentation of nonliberals in social psychology, see any of my blog entries here on liberal bias in social psychology, go to the publications page of my Rutgers website, or to the list of member publications and other resources at Heterodox Academy.

Examples of "gap=discrimination against the underrepresented group" assumptions, interpretations, or claims without evidence of actual discrimination occurring in the context in which the gap is presumed to reflect discrimination (i.e., evidence of bona fide discrimination somewhere may be cited, just not where it is being claimed):

Lee Jussim, a nesting snapping turtle.
Source: Lee Jussim, a nesting snapping turtle.

Ledgerwood, Haines, & Ratliff (2015). Not Nutting Up or Shutting Up.  

Leslie et al (2015).  Expectations of brilliance underlie gender distributions across the disciplines.  Science, 347, 262-265.

Pinholster (May 27, 2016).  Journals and funders confront implicit bias in peer review.  American Association for the Advancement of Science News.  (Note: This is the report I exposed  as a drumbeat of presumption over evidence here and here).

2 Discrimination is behavior.  Failure to demonstrate discrimination includes failure to demonstrate intended discrimination, unintended discrimination, half-intended discrimination, conformity based discrimination, and discrimination of any type, no matter its source or motivation.

-------------------------------------------------------------------------------------------

I sent this essay to Alison Ledgerwood, one of the authors cited above.  Here is her reply, posted with her permission, and, as per a personal policy of not seeking to get into a blog-fight, without further commentary from me:

I guess it makes sense as an oversimplified example of what you want to illustrate, completely divorced from any kind of context whatsoever (e.g., the actual gender distribution of the population you're talking about, historical context, etc.). It shows that if all you have is the proportion of one group to another (an outcome), you can't draw strong conclusions about bias (the mechanism). I would add: That's why, when we see gaps in representation, it's important to draw on (a) what we know about likely mechanisms from the well-established psychological literature on intergroup biases and (b) what we know about likely mechanisms in light of historical context. For instance, if I notice that there seems to be an imbalance in the number of white vs. black presidents of our country, I might draw on both psychological research on racism and the deeply racist history of our country to propose likely explanations. If I notice an imbalance in the number of times someone assumes a male vs. female professor must be a secretary rather than a professor during committee meetings at my institution, I might draw on psychological research on gender and the history of our country as well as the history of academic institutions to understand likely explanations. Same thing if I notice that there are more female than male nurses, more male than female surgeons, etc. 

In my view, that's what we did in the blog post you cite. For example, we write: "Here’s a much more compelling explanation for the demographic disconnect, and one that social psychological research can actually tell us a lot about: We suspect that a number of forces have combined to create unintentional barriers that limit the extent to which some scholars are involved and visible in the best practices conversation." And then we take readers through a tour of a sampling of the relevant literature.

Notice that we say we SUSPECT, not that it's a fact, and we look to the relevant literature to inform our discussion of what the most likely mechanisms might be. That's a far cry, in my view, from making a "gap=discrimination against the underrepresented group" assumption. Instead, it's looking at an outcome (the gap) and then examining existing research for information about the most likely mechanisms, without claiming that those mechanisms MUST explain the outcome. That's a pretty common scientific practice, and one that I think is a far cry from a simplistic, context-free, literature-ignoring heuristic like "if [outcome] then it must be [mechanism]."

-------------------------------------------------------------------------------------------

As usual, please read my guidelines for commenting before doing so.  In short, no snark, sarcasm, or insults, and please stay on topic. 

For updates on new blog entries here, and pieces by others (including their blogs, editorials, long form essays, and scientific articles) on this and related topics, follow us on Twitter.