Hi, I'm obviously a new poster on this forum. I've been involved in sex worker demographic research both informally and formally. I thought now is a good time to talk about my experience.
I first saw the CL ers section in 2007. I quickly realized that some people using that forum literally advertised every hour ... So, like many people at the time I started downloading ads - but with a twist - I wanted to be able to tell if an advertiser was someone I had never seen before. This was not easy but over time I managed to come up with a way to do it.
At around the same time I thought this might be a useful tool for sociologists and criminologists who wanted to get more comprehensive data on how common sex work was in Canada. I knew a well known researcher and told him about what I was doing. I can say that he was thrilled - prior to that, this type of research usually involved something similar to going to the stroll at 2 in the morning and trying to count who was standing around on street corners generally for weeks at a time. This was a brutal time to be a criminology grad student I think.
He wasn't really able to help but recommended someone else. We got together and discussed what I was doing. It was pretty much the same, they thought this was a great idea and could be a major methodological improvement. I thought, naively, great, maybe this will eventually get published ... Boy was I wrong.
2008 rolled around and I hadn't really heard anything. I decided to keep collecting data for another year. That transpired, still nothing. I contacted the original researcher and provided him with a copy of the data. I thought maybe the whole thing was dead in the water.
Then in 2014 I got an email from the second researcher. They wanted more data collection and, even better, had compiled a list of the main websites people were using at the time to advertise. So I adapted my work from 2007 to work with these sites (they are all different) and started downloading ads. After about a year I had over 1.5 million ads - yeah, even for a small country this is an involved task. I analyzed them similarly to how I had done with CL before and provided the researcher with the data. Once again, I waited, ... and waited.
Part of the reason for the wait was that someone was hand reviewing every ad. Yes, that really happened. It took a year and a half to review a year's worth of ads using excel. If this sounds insane to you, I would agree but they were adamant that this was the only way to do this. I decided, while this was going on, that I should grab another year's worth of data. This turned out to be fortuitous as they were having difficulty getting published and the reviewers were asking for more data. This was around 2018 (I can't remember exactly). I kept in touch periodically but nothing was published. In fact, I'm pretty sure nothing has been published even informally.
Anyway, 2021 rolled around and like many people I had lots of time on my hands. I thought maybe now was a good time to poke the hibernating researchers and see whether we couldn't try and get anything published. Once again this seemed to go well. We were going to "bang out" a few papers ... still nothing. However, I got a great reading list from this researcher. This was very helpful because some of these papers are difficult to access ($40-$60 to download just one paper). I later got access to a university library - which is essential as there's a large body of research on sex work now both academic and sponsored by NGOs. After a while I started to have misgivings about not only about the lack of progress but also how labor intensive analyzing the data was turning out to be...
So I decided to rethink what we were doing and see whether I could make some headway. I took the opposite approach, instead of trying to look at every ad (now over 3 million) I started to look at how I was identifying advertisers. I weeded out many invalid contacts (we were using contacts to group posts) and also looked at how often people changed contacts. No method in a dataset this messy is ever going to be perfect.
I was somewhat surprised by the results. It looks like, at the time, over the whole two year period about 2% of Canadian 20-49 year old women had some involvement in the industry but in any week only about 0.2% were actually active. Keep in mind this is after controlling for over-counting advertisers who change phone numbers etc. Another thing that some political pundits in the field may find surprising is that the industry is dominated by white cis-women. Other gender identities represent about 10% - there were even female to male trans workers at that time (I could identify around 35). White identified people represent at least 62% of advertisers (probably more if most of the "mixed" advertisers are partly Caucasian). Think about this, most Canadians know a friend of a friend who is a sex worker and probably have no idea because they only participated in the industry for one to three months.
PM me if you are interested in looking at the work and commenting. In particular, please contact me if you are a worker or a statistician (even better if you are both). I am currently editing the paper for a journal submission and would like more feedback. No guarantees it will be published. A pre-print is available that can be viewed on SocArXiv (look for "The silent majority: the typical Canadian sex worker may not be who we think").
Note that all data is secured and is anonymized - including the source websites most of which are currently defunct. Also note that I am not using my real name on the paper.
Its been a long 15 years but maybe something will come of this yet ... ?
I first saw the CL ers section in 2007. I quickly realized that some people using that forum literally advertised every hour ... So, like many people at the time I started downloading ads - but with a twist - I wanted to be able to tell if an advertiser was someone I had never seen before. This was not easy but over time I managed to come up with a way to do it.
At around the same time I thought this might be a useful tool for sociologists and criminologists who wanted to get more comprehensive data on how common sex work was in Canada. I knew a well known researcher and told him about what I was doing. I can say that he was thrilled - prior to that, this type of research usually involved something similar to going to the stroll at 2 in the morning and trying to count who was standing around on street corners generally for weeks at a time. This was a brutal time to be a criminology grad student I think.
He wasn't really able to help but recommended someone else. We got together and discussed what I was doing. It was pretty much the same, they thought this was a great idea and could be a major methodological improvement. I thought, naively, great, maybe this will eventually get published ... Boy was I wrong.
2008 rolled around and I hadn't really heard anything. I decided to keep collecting data for another year. That transpired, still nothing. I contacted the original researcher and provided him with a copy of the data. I thought maybe the whole thing was dead in the water.
Then in 2014 I got an email from the second researcher. They wanted more data collection and, even better, had compiled a list of the main websites people were using at the time to advertise. So I adapted my work from 2007 to work with these sites (they are all different) and started downloading ads. After about a year I had over 1.5 million ads - yeah, even for a small country this is an involved task. I analyzed them similarly to how I had done with CL before and provided the researcher with the data. Once again, I waited, ... and waited.
Part of the reason for the wait was that someone was hand reviewing every ad. Yes, that really happened. It took a year and a half to review a year's worth of ads using excel. If this sounds insane to you, I would agree but they were adamant that this was the only way to do this. I decided, while this was going on, that I should grab another year's worth of data. This turned out to be fortuitous as they were having difficulty getting published and the reviewers were asking for more data. This was around 2018 (I can't remember exactly). I kept in touch periodically but nothing was published. In fact, I'm pretty sure nothing has been published even informally.
Anyway, 2021 rolled around and like many people I had lots of time on my hands. I thought maybe now was a good time to poke the hibernating researchers and see whether we couldn't try and get anything published. Once again this seemed to go well. We were going to "bang out" a few papers ... still nothing. However, I got a great reading list from this researcher. This was very helpful because some of these papers are difficult to access ($40-$60 to download just one paper). I later got access to a university library - which is essential as there's a large body of research on sex work now both academic and sponsored by NGOs. After a while I started to have misgivings about not only about the lack of progress but also how labor intensive analyzing the data was turning out to be...
So I decided to rethink what we were doing and see whether I could make some headway. I took the opposite approach, instead of trying to look at every ad (now over 3 million) I started to look at how I was identifying advertisers. I weeded out many invalid contacts (we were using contacts to group posts) and also looked at how often people changed contacts. No method in a dataset this messy is ever going to be perfect.
I was somewhat surprised by the results. It looks like, at the time, over the whole two year period about 2% of Canadian 20-49 year old women had some involvement in the industry but in any week only about 0.2% were actually active. Keep in mind this is after controlling for over-counting advertisers who change phone numbers etc. Another thing that some political pundits in the field may find surprising is that the industry is dominated by white cis-women. Other gender identities represent about 10% - there were even female to male trans workers at that time (I could identify around 35). White identified people represent at least 62% of advertisers (probably more if most of the "mixed" advertisers are partly Caucasian). Think about this, most Canadians know a friend of a friend who is a sex worker and probably have no idea because they only participated in the industry for one to three months.
PM me if you are interested in looking at the work and commenting. In particular, please contact me if you are a worker or a statistician (even better if you are both). I am currently editing the paper for a journal submission and would like more feedback. No guarantees it will be published. A pre-print is available that can be viewed on SocArXiv (look for "The silent majority: the typical Canadian sex worker may not be who we think").
Note that all data is secured and is anonymized - including the source websites most of which are currently defunct. Also note that I am not using my real name on the paper.
Its been a long 15 years but maybe something will come of this yet ... ?
Last edited:
