Last night my girlfriend and I participated as volunteers in this:
It was actually quite a lot of fun, although I could have done without the 2am starting time. A whole bunch of volunteers divided up into groups, each of which was assigned a small part of Seattle. Each group thoroughly walked their area and counted homeless people. The intent was to get some good numbers on how many unsheltered homeless there are in Seattle during the winter.
Our group only ended up seeing a single homeless person, but other groups saw quite a few more. Here are the final results:
The gender of most of the homeless people could not be determined (1615 unknown gender). For example, the homeless person we saw was wrapped in many blankets, and given that we were not to disturb them, it was therefore difficult to determine their gender. But of the homeless whose gender could be determined, the large majority were men (897 men vs 205 women). Which shouldn’t be surprising to those familiar with existing homeless statistics.
There could be a variety of reasons for this disparity. But I couldn’t help but notice that the press release phrased this as “2,736 men, women and children had no shelter in King County last night”. There’s nothing particularly wrong with such a phrasing. But I couldn’t help but wonder if the gender ratio were reversed, would they have still used such gender-neutral phrasing? Maybe they would. I don’t know. But it caught my eye regardless.
In any case, I’m very glad I participated, both because it was fun, and because I feel good helping out with data collection. Solid data is important.
 It’s worth noting, of course, that there was likely some amount of mis-gendering that happened for a variety of reasons (e.g. transgender people), but I think it’s unlikely to have biased the gender numbers substantially.
 However, it’s good to be skeptical of numbers where such a large proportion fall under “unknown”. It’s easy for there to be confounding factors–many of which we might not think of–that would make the known ratio not translate to the unknown ratio. So I would call these numbers on the gender of unsheltered homeless “suggestive”, but not definitive.