busybee

A question keeps on coming up on Mastodon (and the Fediverse in general), namely trying to figure out what the overall age distribution of the userbase is.

However, every single poll that comes about ends up becoming quite contentious in ways that lead to a lot of conflict and drama, primarily due to the nature of how Mastodon polls work. The most common problem is that due to the limited number of options in polls, it is very difficult to make a poll that gives both range and precision. Commonly, the person attempting to run the poll will end up being attacked by people in an age bracket that feel disenfranchised by the specific poll splits, and there’s this… tunnel-vision hyperfocus that a lot of people get into where if they see an “other” or “X or [older|younger]” category, even if there’s a whole thread of polls that dig into splitting up those larger buckets, you get deluged with responses from those who are very angry about the lack of representation.

Attempting to run the poll on a more suitable polling platform (Surveymonkey, Google Forms, Straw Poll, etc.) always ends up with similar concerns, such as worries about privacy or an unwillingness to use the survey platform of choice for whatever technological hill people want to die on. And, again, users who feel put off by this make it their mission to let the perceived offender (and everyone around them) know.

I started out riffing on this with a couple of threads that were not meant to be taken seriously, and which I thought were obvious jokes, but, you know how the Internet can be. Anyway, my third poll (doing it as a median-seeking binary search) was also meant as a joke (which I intended to just be a one-off), but people actually responded to it fairly well. When I posted a followup in which I explained that it was a joke but asked if people wanted me to follow through with it, the overwhelming response was a resounding “yes.”

Clearly there is enough interest in trying to figure out some sort of statistical distribution for the age of Mastodon users (otherwise this wouldn’t keep on happening!), so, #MastoAge was born, for better or worse.

Rebecca Watson made an excellent video about how the recently-released COVID-19 risk map is not particularly helpful, due to the lack of context provided and the public’s general lack of understanding about how statistics work. It’s well worth watching (and also talks about a few other things, like issues with services like 23andme’s genetic risk factor screening kits). People are apparently using this risk map as a means of justifying going to Thanksgiving gatherings based on “only a 5% infection chance” or the like, which is incredibly short-sighted.

Let’s say that 5% of the population is carrying the disease, and assume that this statistical model is completely accurate. (It almost certainly isn’t, but that’s beside the point.) This means that any time you encounter someone there’s a 5% chance that they’re infected. That seems pretty low, right? For a single encounter, sure. But that doesn’t tell the whole story.

A frequent thing that people want to do in making games or interactive applications is to shuffle a list. One common and intuitive approach that people take is to simply sort the list, but use a random number generator as the comparison operation. (For example, this is what’s recommended in Fuzzball’s MPI documentation, and it is a common answer that comes up on programming forums as well.)

This way is very, very wrong.