Miles Klee highlights a study:
[E]xperimenters set up duplicate “decoy ads” in the Casual Encounters section of Craigslist and the adult/escorts section of Backpage.com for 15 different cities, collecting the phone numbers and texts of everyone who replied. When all was said and done, they had amassed 677 points of contact and 451 numbers. On average, they estimated, under 5 percent of men in a metropolitan city area – supposedly just those “over 18,” as if a curious youngster wouldn’t lie on a terms of use agreement – are soliciting online sex ads. [They found that] some regions are a little hornier than others: “On average, within the 15 markets explored, one out over every 20 males over the age of 18 in a metropolitan city area was soliciting online sex ads. The findings ranged from approximately one out of every 5 males (Houston, 21.4 percent) to less than one of 166 males (San Francisco, .6 percent).
Update from a reader:
Let me assure you, more than 0.6% of men in San Francisco are looking for sex online. But not on Craigslist (what is this, 1998?) We have apps for that now.
Another:
I call bullshit on this study. The results were pretty sensational and the sample size seemed small – 451 numbers in 15 cities. That’s 30 people per city. So I looked up the actual paper to see if it was credible, and was not exactly impressed.
I’ve laid out an example of how they appear to be doing the math below, but short answer is that they appear to both make math errors and also make unrealistic assumptions to make their numbers look bigger. They even appear to have screwed up population data (which seems pretty hard to do).
Their methodology was to post two fake ads, assume the small number of responses was representative and random, do some math and assume every single other ad posted got the same response rate from a totally different set of customers. Their entire methodology is based on using repeat callers to estimate population sizes, but then they assume other ads get completely different “customers” calling in.
It’s ridiculous and they make much too sweeping claims from their research. Now, I was surprised at the levels of activity they cited other (presumably less grandiose) research as showing, but their research doesn’t appear to add much real evidence. I’m not even sure it’s useful as a measure of relative behavior between cities. They may just be capturing relative popularity of the particular sex sites.
I’m going to use NYC as my example because it’s the most obviously ridiculous. Their headline: 3.4% of men over 18 years old solicit sex online in NYC.
Ridiculous underlying data: They received responses from 7, yes SEVEN, people. Three of whom called twice. They extrapolated from this ludicrous sample an entire population estimate. And they made what appear to be two math errors in doing so.
Math error #1: They say there are about 21 thousand sex ad users, which implies NYC adult male population of ~500 thousand. It’s more like 3 million. Maybe they only considered Manhattan under the logic that Johns, like witches, cannot cross moving water.
Math error #2: My best estimate is that their own math would show an estimate of about 4,400 [((8+7)/4-1)*341.5] as the population using the formula from Chapman they provide and maximizing the possible values. Maybe they used one of their other (apparently arbitrary) methodologies, but they don’t actually say.
Combined these errors change their estimate by an order of magnitude so it’s more like 0.25%. Whoops.
Other cities are slightly less absurd, but it pisses me off when sloppy research gets elevated in the media because it tells a good story. I’m all for confirming my bias that Houston sucks, but this one isn’t fair.
The Dish 