Tired old ideologues flogging a still dead horse

From Twitter's Gender Imbalance by Colleen Flaherty. Reporting on the original study Gender Differences in Twitter Use and Influence Among Health Policy and Health Services Researchers by Jane M. Zhu, Arthur P. Pelullo, and Sayed Hassan.

Kind of astonishingly shallow and obviously ideologically driven and interpreted. From the Abstract:

Ample research has documented the lower visibility and success of women compared with men in academic medicine. Against this setting, social media platforms such as Twitter offer academics opportunities to promote their research, network professionally, gain visibility, and, in turn, foster opportunities for career advancement. These opportunities are particularly critical in health policy and health services research, in which dissemination of policy-relevant research and engagement with health care decision-makers impacts academic influence, recognition, and promotion. Herein, we describe gender differences in Twitter use and influence among health services researchers.

That very first sentence is the tell. There are all sorts of disparate impacts in every field based on almost any variable you choose - gender, religion, height, education attainment, handedness, income, etc. What tends not to vary so much are results when you control for duration in the field, hours invested, etc. People who work the hardest, the longest and most consistently tend to rank the highest in that field. Not always, but on average.

Not only that, but the rewards in almost all fields also follow a log or Pareto Distribution. The top 20% of the performers obtain 80% of the rewards. And often with minuscule differentials in absolute performance. A runner with a 400 meter speed half a second faster time than their nearest competitor will win most the rewards.

Performance and Pareto Distribution seem to be the drivers of the results reported here. And setting aside the discussion of whether twitter performance is a suitable measure of academic performance.

We have to go to the article for greater detail as the research is gated.

The article also starts with a clanger.

Women on social media face disproportionate levels of harassment compared to men.

It is an article of faith among the true believers of postmodernist beliefs but most the replicated studies and meta-analyses I have seen indicate that men receive slightly more harassment but that the nature of the harassment tends to be different (women suffering more sex-related taunts, men suffering more raw ad hominem attacks). But let's set aside this tell as well.

Flaherty goes pretty much from sketchy summary of findings to full blown gender theory interpretation in no time flat.

A new study says that female academics also have disproportionately fewer Twitter followers, likes and retweets than their male counterparts on the platform, regardless of their Twitter activity levels or professional rank.

Women were also more likely than men to reciprocate relationships with followers and follow back, and to follow other women.

Observers of gender dynamics, casual or expert, probably won’t be surprised by the findings. But they do have implications for scientific impact and careers and the general sharing of information. And because this study, in particular, involved health policy and health services researchers, there are implications for public health. So it’s important to understand what’s happening, and why -- as best we can, since the study was more quantitative analysis than a deep dive into the psychology of Twitter.

The short answer, said lead author Jane M. Zhu, assistant professor of medicine at Oregon Health and Sciences University and a senior adjunct fellow in health economics at the University of Pennsylvania, is that the “same power dynamics that exist in the real-world office settings seem to exist online.”

The slightly longer answer, she said, is that while women may be supporting other women and “amplifying each other’s professional voices online, men may still be considered the more authoritative voices, even within the same academic rank.”

Additionally, Zhu said, “Men may not find the content of women’s tweets as compelling as other men’s, or they may not feel obligated to reciprocate relationships and follow women who follow them.”

There is a lot of speculation and ideological projection in a piece that is supposed to be science-based. You have to get deep into the article and the social justice muck before a clearer picture begins to emerge.

For their study, the researchers identified the names and institutional affiliations of authors and speakers at the 2018 AcademyHealth annual research meeting. They excluded trainees and those without medical degrees or doctorates and then pulled Twitter data -- including the most recent 3,200 tweets -- for the remaining sample. About one-third of that group had a Twitter account, some 492 women and 427 men.

The men followed about 375 people, on average, while women followed 332. In addition to following fewer people, women had been on Twitter for less time, on average -- about 4.5 years, versus 5.1 years for men. They also had fewer original tweets, at about 71 per year, compared to 98 for men. But the study asserts that that doesn’t account for the dramatic difference in followers: 567 for women, on average, versus 1,162 for men.

Similarly, women’s tweets generated fewer average likes than men, at 316 per year versus 578, respectively. Same for retweets, at 207 per year on average for women and 400 for men. Per tweet, not just per year, the same was true. Most gender differences were between full professors.

The study, published in JAMA Internal Medicine, notes that men and women are using Twitter at equal rates, and that social media offers women “opportunities for engagement, perhaps with fewer barriers than may be present in day-to-day academic interactions.” The gender disparity was also less pronounced at lower ranks, suggesting things may be changing.

Yet while some have hoped that social media would “help level the playing field in academic medicine by giving women an accessible and equitable platform on which to present themselves,” it says, the danger is that these forums “may do little to improve gender parity and may instead reinforce disparities.”

So their goal becomes clearer. They expect to see men and women deriving the same level of exposure benefit from Twitter even though there are disparate usage patterns.

Men have been using Twitter 13% longer than women.

Men generate 38% more original content than women.

Men follow 16% more Twitter accounts than women.

So the male users are more experienced using Twitter, generate significantly more original content, and engage with others more.

In every other field this greater engagement on a sustained basis is what generates differential outcomes. It appears that Twitter is the same. The more engaged, high productivity, more experienced males experience higher results, consistent with the well established Pareto Distribution found in virtually all human endeavors.

Men have 105% more followers.

Men get 83% more likes.

Men get 93% more retweets.

So it appears that Twitter, like all other fields, rewards engaged, high productivity, and experienced participants disproportionately more than those who are less engaged, lower productivity, and less experienced.

The researchers profess to be shocked by this outcome and, ignoring Occam's Razor, leap to more imaginative explanations. Explanations which are consistent with feminist and social justice theory.

The final tell in this "research" is the dog that's not barking. There is an easy test for gender bias here and that is to compare apples to apples. Compare males who perform at the lower levels (4.5 years instead of 5.1, following 332 instead of 375, 71 original tweets instead of 98) and see if they have more followers, more likes, and more retweets because they are male than women with that profile, or whether, as should be expected, they have the same lower outcomes as those women.

The fact that this comparison is not done would seem to indicate that either the researchers are poor at designing their research or that they did such a comparison and it showed that lower productive twitters users have the same results whether they are male or female.

This is well traveled ground. Forty years ago, it was excusable to do crude comparisons of men and women's incomes and conclude that possibly women were being discriminated against. For thirty years now we have known that when you compare like-with-like (profession, education attainment, years in service, hours per week, etc.) men and women earn the same income.

Which isn't especially surprising since that is what cultural values would predict, economics would predict, and the law requires.

There is one further piece of evidence that this is a productivity/experience/engagement effect, not a bias effect.

In the article they include a single graph.

Click to enlarge.

This seems to indicate that the correlation between original content and number of followers is pretty weak. In aggregate, the overwhelming majority of the 919 participants have virtually no relationship between how much they are tweeting and how many followers they have. Indeed, the most prolific original content tweeter of all, with some 3,100 tweets, is a male with perhaps at most a few hundred followers.

This weak correlation between original content volume and number of followers suggests that perhaps number of years and level of engagement might actually be the stronger predictors of number of followers.

The other thing that leaps out about the graph is that women seem to be heavily concentrated in the low volume tweeting quadrant. They don't tweet much and they have few followers. Just as is the case for low tweeting men.

If we accept that there is some base level of presence and activity, say at least 1,000 tweets, the correlation between volume of tweets and number of followers strengthens substantially but is still overall pretty weak.

Finally, look at the twelve accounts which have more than 10,000 followers. 35% of them are women. As I have documented elsewhere on this blog, in virtually any competitive field in the US (authors, artists, CEOs, surgeons, partners in law firms, etc.), virtually all of them have between 15-30% of their top performers who are women. And this level of representation is higher in the US than every other OECD country. Which suggests that the issue is not bias but choices.