Yours is an excellent example of how correlative observational studies (however formal or informal) are practically useless. Once you have experience interacting with the unhoused population, you will quickly learn that they are largely using things like crack, meth, and opioids (if at all -- many are in fact sober), not psychedelics.

I don't find that surprising, but I'd still be interested to see a survey of what proportion of California's homeless population have used psychedelics extensively _in their past_. Because my experience leads me to believe that someone who does a lot of psychedelics in their younger years is more likely to wind up on a bad track later in life (even if they don't continue with those drugs later on).

Bayesian thinking can pay here, as well as a little self-critique.

Simply put, confirmation bias will convince you that people who do psychedelics end up on a bad track if your sample only includes people who are on bad tracks, rather than a sample composed of those who use psychedelics.

More probably (and this in my experience is definitely true) many more people use psychedelics than you might naively assume, and in fact probably the proportion of people who use psychedelics and end up on good tracks is way higher than that of those who use psychedelics and end up on bad tracks.

On a more meta-level, your searching and inquisition is still based on a more fundamental assumption, that homeless people are homeless for reasons stemming from their own decisions, rather than some confluence of bad events that leads one to ruin or another possible path to homelessness (e.g., they have mental health conditions which prevent integration into "normal" society without proper treatment, or that they simply want to be homeless and nomadic as a lifestyle choice). Empathy pays off here to temper un-credible assumptions and analyses.

> the proportion of people who use psychedelics and end up on good tracks is way higher than that of those who use psychedelics and end up on bad tracks

I'll agree with that, but the real question is whether people who use psychedelics more frequently end up on bad tracks than those who don't, which in my experience is definitely true.

> homeless people are homeless for reasons stemming from their own decisions

No, my bias is that many homeless are homeless because they live in a place that enables and even encourages actions that lead people to become homeless (psychedelics use being one example).

Psychedelic use in SF is not primarily among the homeless population. That population is using very different substances.