Friday, 27 March 2015

An Outbreak Of Zombie Prostitutes

Good news everyone! The press have found themselves a study to report. Data journalism! It’s topical as well, a great big bomb into the middle of the SWERF wars. For the uninitiated, SWERF stands for Sex Worker Exclusionary Radical Feminist, and the SWERF wars are closely related to the TERF wars for reasons beyond my ken. 

The study is looking into the numbers of students who are involved in the sex industry. While the survey’s definition of sex work is relatively broad, encompassing everything from prostitution to glamour modelling and online cam shows, it is inevitably illustrated with pictures of women’s fish-netted legs on street corners, and will find itself slap bang in the middle of that debate. 

So every site in the land is running with the figure of 5% of students being involved in the sex industry. This is patently nonsense. Anyone who has ever been to university knows it is nonsense. It in no way passes the sniff test. Things get worse when we dig into things a little further. 5% of men are involved, while only 3.5% of women. Quite how this sums to 5% in total I’m not sure, but hey. It also contradicts almost everything we know about sex work. It’s a tricky area to get figures for, but there’s pretty much nobody claiming anything in the ballpark of 50% more men. This study looking into online profiles on a sex work website gives us a figure of about 42% of sex workers being male. I'm a little sceptical it is that high. The male users seem considerably less active. Could their profiles be a joke, or a bit of a fishing expedition, while they never actually go any further? It also seems to suggest an improbable number of men signed up as straight and hoping to attract female clients. This contradicts other reports on the subject, suggesting that most male sex workers sell services to other men.

In addition, there is the claim that 1/5 students have 'considered' going in to sex work. I'm not sure how much 'considering' you need to do to count in the figures here. If someone asks me if I've considered becoming a sex worker, I immediately consider it. I've considered lot of things in my life, jumping off a cliff, becoming an astronaut, running away to join the peshmerga. That's what people do. They don't suddenly become sex workers because of it, so I'm not too sure we can read much into this.

The most likely explanation of the survey findings is simple self-selection bias. It’s a pretty decent sample size, with over six thousand respondents, but was conducted through a simple mass e-mail. Anyone want to hazard a guess as to whether students working in the sex industry are more or less likely to respond to a survey about students working in the sex industry?

“Hey guys, I’d really appreciate it if you could complete my study looking into World of Warcraft players among the student population. Here’s a link.”

“New survey shows over 90% of students play World of Warcraft.”

We’re still left with a bit of a conundrum though. While self-selection bias is probably enough to account for the whopping overall figures, does it explain the crazy discrepancy between men and women? It could do. Two thirds of the people who responded to the survey were female. This could indicate that in general men were less interested in the topic, which may lead to a stronger self-selection bias.

That may not be the whole story though. The final puzzle piece is something explained with far greater style than ever I could by Scott Alexander, the idea of a lizardman constant. Essentially this is the noise in any poll, especially notable in small numbers. It’s made up of a combination of people who misunderstood the question, those who agree with anything a researcher says, and people who are deliberately sabotaging the poll results. Alexander speculates it adds up to about 4%, about the number of US residents who will agree, when prompted, that lizardmen are running the world. This isn’t to say we should discount any polling figure that is so low, 3% of the UK population said they were black at the last census and that’s probably correct, but on slightly less official online surveys it could well be an issue.

Could it be a big part of the results here? It seems useful, especially in explaining the discrepancy between male and females. Are university aged men more likely than university aged women to say they’re a prostitute for a joke? Would your typical uniLAD think it is totally top bantz? Seems pretty plausible to me, certainly more plausible than there being more male student sex workers than female. It doesn’t have to explain the whole figure, undoubtedly there are some male student sex workers, but it could easily account for a few percentage points, and at what is such a low figure to start with it could even make up most of it. Lots of people really are total dicks.


Andrew Sabisky adds an interesting side point point here about the lizardman constant versus social desirability bias, and how we shouldn't trust what people say on surveys. 0.5% of mothers will claim their pregnancy was in spite of their virginity.

I'm not aware of any research into gender differences here, but it may be something to consider in the future. I doubt self-selecting female sex workers answering a survey about sex work will be influenced much by social desirability to deny it though.


My final major point of contention is that the researchers performing the study should have realised this. In addition to their quantitative survey, they added in some qualitative interviews with participants.

“Semi-structured face-to-face interviews were performed with student sex workers. Eight students (one male and seven female) participated in the interviews.”

In the light of their study findings, this seems strange. They are claiming that there are 50% more male student sex workers in the UK than there are female. If so, why have they only managed to find one to interview, compared to seven women?

I support legalised sex work, but this work seems to be more designed to push a narrative than to find reliable numbers. It seems pretty plausible that if people believe there are more student sex workers, they will be more supportive of the case for their acceptance. Indeed, the survey makes a big point of this. Before we get on to any of the actual findings of the survey, the authors present us with data telling us that these findings make people more accepting of sex work. Research with such a blatant ideological aim is unlikely to have much relation to the truth. Their incentives are all to inflate the numbers, and the structure of their study means it is almost guaranteed to.

The most interesting question here is whether anyone reporting on or tweeting about the study actually believes it. I'm not convinced anyone really believes 5% of guys on campus are sex workers, it's obviously implausible. Instead it seems more likely that by sharing it people are saying 'yay, sex workers!' and showing their loyalty to the cause. 

The one in twenty figure is a great statistic, surely destined to be quoted in a thousand student union debates. It may join the ranks of the great zombie stats, those that simply refuse to die in spite of repeated refutation, such as one in five women are raped at university, or several about women’s wealth. In terms of what we really can take from the study, I think it's safe to say that there are some students out there working in the sex industry. Past that I'm not confident of anything, and it would take a far better and more difficult to perform study than this to say anything more.

No comments:

Post a Comment