Notes on Nittrouer et al. (2017)

This paper reports that “there are gender differences and disparities (relative to available pools) in colloquium speakers” and “that disparity remains even after controlling for obvious alternative explanations (e.g., higher rates of men than women in academia, different levels of interest in and ability to travel for the sake of giving colloquium talks).”

Alas, the authors do not say a word about another very obvious alternative explanation: research productivity. The authors found that women constitute 31% of the speakers, which is supposedly significant underrepresentation. But what proportion of publications are authored by women? Well, surprise surprise, it’s about 31%.

It is not difficult to see why the authors conveniently neglected to mention this well-known fact that is obviously highly relevant. The authors of course want to suggest that “colloquium committees … unwittingly favor men” because of “bias” and “stereotypes.” It would thus be inconvenient to admit that speaking frequency simply mirrors publication productivity, since manuscripts submitted for publication are anonymised. But of course that is only relevant if you are actually trying to investigate facts in an honest way, not if you are trying to concoct evidence for a predetermined conclusion that is ideologically agreeable to you.

The authors’ literature review is of course full of references to many papers I have discussed before. This was a new one for me though:

“Participants who read a lecture, which was posited as having been written and delivered by a male or female professor, rated the lecture by the male (versus the female) professor significantly more positively.”

Sounds like damning evidence! But let’s see what happens if we actually look up the paper cited. Then we find right in the abstract:

“Students … evaluated an identical written lecture by a male and female professor on pay disparities between men and women in the workforce suggesting sex discrimination.”

Are you kidding me?! What the authors deceitfully refer to simply as “a lecture” was in fact a politicised opinion piece that specifically argued that women are discriminated against. Obviously this is an absolutely idiotic way of testing whether students are biased by gender when evaluating “a lecture.” Obviously the students had every reason in the world to be a little more apprehensive when this case came from a female lecturer. It is one of the most elementary principles of critical thinking to be less trusting of information coming from a self-serving source, as everyone knows except gender bias researchers apparently.

Another recent study (featured at Nature’s news blog) makes a similarly dubious case for gender bias. The researchers looked at success rates of grant applications and found that women and men have about equal chances when only the research proposal itself is being evaluated, but that men are favoured when the calibre of the researcher is part of the evaluation criteria.

This of course proves absolutely nothing about gender bias since the higher ratings of male applicants is very plausibly due to better publication records (a possibility the authors themselves admit; pp. 9-10), which is exactly the sort of thing that the evaluation system is designed to take into account, and very reasonably so.

But, as usual, nothing is more biased than gender bias researchers. The title the authors cooked up for their paper is blatantly dishonest and fanatically ideological: “Female grant applicants are equally successful when peer reviewers assess the science, but not when they assess the scientist.”

With this deceitful phrasing the authors manage to insinuate that taking into account people’s scientific track record actually means judging people on factors other than “the science.” It is pathetic that Nature eagerly picks up this propaganda nonsense in their headline. This kind of “research”, and the cheerleading reception it receives, is a disgrace to academia and an insult to critical thought.