Biases; Surveys

I've been reading papers on prostitution again, both papers whose conclusions support criminalization and papers whose conclusions support decriminalization of prostitution. What I've gotten from that is not better arguments for the position I support (decriminalization), but a reminder of some of the ways in which researchers, and the rest of us, bias our data.

Surveys

Surveys are a good way to get biased data. Every one who conducts surveys knows that the answers you get depend on how you word the questions. One way of biasing a survey is to ask questions about an ideal world, and use the answers to support concrete action. For example, many prostitutes don't consider their job ideal. They would prefer to be doing something that paid more, or had more prestige, or was less dangerous. Ask them if they think prostitution should be eliminated, without giving the question any context, and they're very likely to put the question in the context of their own aspirations and say yes. In an ideal world there they would not have to practice prostitution.

However, ask them if they think they should be forced to give up prostitution and take up one of the currently available alternatives, and they would probably say no. In general, they are prostitutes because it is more attractive than any currently available alternative. If they've been forced into prostitution by the patriarchy, it's still better than any of the alternatives the patriarchy is offering. And if they're informed agents exercising free will, as liberal theory claims, then they've chosen prostitution because it's the best choice available.

Ideal world questions are fine if you want to create alternatives for prostitutes that are superior to prostitution. If prostitutes are generally satisfied with their jobs, creating alternatives isn't going to cause a lot of them to leave prostitution. Or if you create the wrong alternatives, prostitutes won't find them attractive enough to make a switch. So if you are creating alternatives, you want to know if prostitutes want alternatives, and what sort of alternatives they want.

However, I don't know of a single debate that hinges on whether to provide prostitutes with alternatives. In every case that I know of, the debate is whether to make prostitutes current jobs easier (decriminalization), or to try to drive prostitutes out of their current jobs without offering an alternative (increase penalties for prostitutes and/or clients). In that situation, ideal world questions produce misleading results. Prostitutes generally don't want to be driven from their current jobs when the job market isn't offering good alternatives. Even if they hate prostitution, they hate the best alternative more.