Friday, 27 April 2012

The Numbers on the Game

Perhaps one of the biggest "sex myths" making the rounds these days is to what extent experiences like mine are (or are not, as it is claimed) representative of sex workers in general.

(Of course, I personally make no claim to be an "average" sex worker and certainly not to speak for them as a whole. If anything, I would love to see more and more voices claim the title of sex worker publically so we can demonstrate once and for all what a diverse group it is.)

But when it comes to studies and statistics, this is an issue that comes up a lot. Addressing the question of how representative a sample of people is of a population in general is one of the cornerstones of good study design. It's one of those things that, if wrong at the start of a research project, is a devil to try to correct in retrospect. In many cases it's impossible.

The Sex Myth discusses at length how this affects virtually all studies relating to prostitution. A large number of researchers assume street-based sex workers to be the majority of sex workers, which has the tendency to skew (and sometimes fully invalidate) their results. Because they often obtain access to research subjects who come to them via outreach or addiction programs, their sample is necessarily biased towards sex workers whose lives are chaotic.

Streetwalkers may often be the most visible face of sex work but it's far from the whole story.

So how representative is the cliche of the drug-addicted streetwalker leaning into a punter's car anyway? This is in some ways difficult to answer. But I highly suggest if you are interested in the topic to check out Maggie McNeill's excellent summary of what we do know. She makes the case far more succinctly that I ever could here, here and here. Go on and read those then come back when you're done.

In general, the data seem to agree that in most Western countries the percentage of sex workers who are streetwalkers is about 15%.

So whichever way you look at it, the kinds of studies recruit their subjects almost exclusively among street-based sex workers, and do so through crisis centre referrals, are not going to be representative of sex workers at large. They can at best be said to be a study of those people at that time which makes the results non-generalisable.

Promising studies do exist which try to address the problem of imbalance of numbers in counting sex workers. Again, while it's hard to describe 'sex work' as an experience from what is necessarily a very diverse group, I found this study from Suzanne Jenkins at Keele [pdf] to be a good example of how to build a better sex worker survey.

Now on the the problem of what you do with your subjects.

A particular bugbear of mine is the tendency to present scare stats about sex workers in isolation and not to involve a control group. So say, for instance, you heard a statistic that claimed the majority of strippers experienced unwanted attention from clients. Presented without context it may sound impressive, but is meaningless.

What would be a control group here? People with similar working hours in licensed establishments might be one - barmaids at non-strip clubs for example. Or people of a similar age in service industry professions involving tipping - like waitresses. Have a think: do you know any barmaids or waitresses who haven't had difficult and at times physically aggressive customers? I don't. An stud that doesn't even address the possibility that their results come from the service industry and alcohol rather than sex work per se has not fully examined the evidence in a way that should be taken seriously.

Control groups, when it comes to population statistics like these, are hard. I get it. But that is, as we say where I come from, hard cheese. So there's no perfect control like a group of homeless, drug-addicted nuns somewhere we can use to see whether it's the sex that drives people to despair or not. But you still have to make an effort. And you recruit and match your controls up front, not after the fact.

Finally there is the matter of where data originates. As a scientist I know that it is damned difficult, if not impossible, to do work that is totally free of any external conflict of interest or internal hope for a particular outcome. But there are ways we can help sieve the believable from the unbelievable: if a study comes from a source with a strong ideology and a financial interest in promoting this stance it is right to question whether that affects both study design and interpretation of results. These generally fall into the category Laura Agustin has dubbed "the Rescue Industry".

This is not to say that academic publishing is always right and self-publishing or internal reports always wrong. But there is a significant grain of salt we should take when the people who present themselves as experts on the topic of sex workers are from the same stable of folks better at generating press coverage than at reporting their mistakes.

Do I expect laying things out like this will please everyone? No, not at all. There are a lot of people with a big investment in keeping the myths about what sex work is supposedly "really" like alive. As well, there are people whose opposition to sex work isn't affected by the many well-adjusted people who do it anyway. It's fair to say my particular bias is to prefer the quantitative over the qualitative: for as "Uncle Joe" Stalin so elegantly put it, quantity has a quality all its own.

But if you are the sort of person to whom the evidence is more important than the anecdote - and if you're a reader of this blog, I assume you are - then take the numbers seriously. The next time someone tries to sell you the poor-addicted-hooker myth, call it for the nonsense it so clearly is.