Sampling error. Psychosis is not an ongoing thing, yielding many fewer chances to observe it than months or years long depression or anxiety. Psychosis often manifests when people are already isolated due to worsening mental health, whereas depression and anxiety can be exactly exacerbated by the situations in which you would observe it i.e. socializing. Nor would people volunteer their experience due to much greater stigma.
I am not comparing “number of psychosis among my friends” to “number of depression episodes among my friends”. I am comparing “number of psychosis among my friends” to “number of psychosis among rationalists”. Any sampling errors should apply equally to the rationalists (or if not, that demands an explanation).
The observation is that there’s a lot more reported psychosis among rationalists than reported psychosis among (say) CS grad students. I don’t have an explanation (and maybe there’s an innocuous one), but I don’t think people should be denying this fact.
A hypothesis is that rationalists are a larger gossip community, so that e.g. you might hear about psychosis from 4 years ago in people you’re nth-degree socially connected with, where maybe most other communities aren’t like that?
Certainly possible! I mentioned this hypothesis upthread.
I wonder if there are ways to test it. For instance, do non-Bay-Arean rationalists also have a high rate of reported psychosis? I think not (not sure though), though perhaps most of the gossip centers on the Bay Area.
Are Bay Area rationalists also high in reported levels of other gossip-mediated things? I’m trying to name some, but most sexual ones are bad examples because of the polyamory confounder. How about: are Bay rationalists high in reported rates of plastic surgery? How about abortion? These seem like somewhat embarrassing things that you’d normally not find out about, but that people like to gossip about.
Or maybe people don’t care to gossip about these things on the internet, because they are less interesting the psychosis.
I’m someone with a family history of psychosis and I spend quite a lot of time researching it—treatments, crisis response, cultural responses to it. There are roughly the same number of incidences of psychosis in my immediate to extended family than are described in this post in the extended rationalist community. Major predictive factors include stress, family history and use of marijuana (and, to a lesser extent, other psychedelics). I don’t have studies to back this up but I have an instinct based on my own experience that openness-to-experience and risk-of-psychosis are correlated in family risk factors. So given the drugs, stress and genetic openness, I’d expect generic Bay Area smart people to have a fairly high risk of psychosis compared to, say, people in more conservative areas already.
(Sort of; you did say “more gossipy → more widely known”, but I wanted to specifically add the word “larger”, the point being that a small + extra gossipy community would have a higher that usual report rate, and so would a large + extra gossipy (+ memory-ful) community; but the larger one would have more raw numbers, so you’d get a wrong estimate of the proportional rate if you estimated the size of the relevant reference class using intuitions based on small gossip communities. And maybe even a less gossipy but larger network would still have this effect; like, I *never* hear gossip about people in communities I’m not a part of, even if I talk to some people from those communities, so there’s more structure than just the rate of gossip. It’s more a question of how large is the “gossip-percolation connected component”.)
See PhoenixFriend’s comment, there were multiple cases I didn’t know about, so a lot of people’s thoughts about this post are recapitulating sampling bias from my own knowledge (which is from my own social network, e.g. oversampling trans people and people talking with Michael). This confirms that people are avoiding volunteering the information that they had a psychotic break.
PhoenixFriend alleges multiple cases you didn’t know about, but so far no one else has affirmed that those cases existed or were closely connected with CFAR/MIRI.
I think it’s entirely possible that those cases did exist and will be affirmed, but at the moment my state is “betting on skeptical.”
Sampling error. Psychosis is not an ongoing thing, yielding many fewer chances to observe it than months or years long depression or anxiety. Psychosis often manifests when people are already isolated due to worsening mental health, whereas depression and anxiety can be exactly exacerbated by the situations in which you would observe it i.e. socializing. Nor would people volunteer their experience due to much greater stigma.
I am not comparing “number of psychosis among my friends” to “number of depression episodes among my friends”. I am comparing “number of psychosis among my friends” to “number of psychosis among rationalists”. Any sampling errors should apply equally to the rationalists (or if not, that demands an explanation).
The observation is that there’s a lot more reported psychosis among rationalists than reported psychosis among (say) CS grad students. I don’t have an explanation (and maybe there’s an innocuous one), but I don’t think people should be denying this fact.
A hypothesis is that rationalists are a larger gossip community, so that e.g. you might hear about psychosis from 4 years ago in people you’re nth-degree socially connected with, where maybe most other communities aren’t like that?
Certainly possible! I mentioned this hypothesis upthread.
I wonder if there are ways to test it. For instance, do non-Bay-Arean rationalists also have a high rate of reported psychosis? I think not (not sure though), though perhaps most of the gossip centers on the Bay Area.
Are Bay Area rationalists also high in reported levels of other gossip-mediated things? I’m trying to name some, but most sexual ones are bad examples because of the polyamory confounder. How about: are Bay rationalists high in reported rates of plastic surgery? How about abortion? These seem like somewhat embarrassing things that you’d normally not find out about, but that people like to gossip about.
Or maybe people don’t care to gossip about these things on the internet, because they are less interesting the psychosis.
I’m someone with a family history of psychosis and I spend quite a lot of time researching it—treatments, crisis response, cultural responses to it. There are roughly the same number of incidences of psychosis in my immediate to extended family than are described in this post in the extended rationalist community. Major predictive factors include stress, family history and use of marijuana (and, to a lesser extent, other psychedelics). I don’t have studies to back this up but I have an instinct based on my own experience that openness-to-experience and risk-of-psychosis are correlated in family risk factors. So given the drugs, stress and genetic openness, I’d expect generic Bay Area smart people to have a fairly high risk of psychosis compared to, say, people in more conservative areas already.
(Sort of; you did say “more gossipy → more widely known”, but I wanted to specifically add the word “larger”, the point being that a small + extra gossipy community would have a higher that usual report rate, and so would a large + extra gossipy (+ memory-ful) community; but the larger one would have more raw numbers, so you’d get a wrong estimate of the proportional rate if you estimated the size of the relevant reference class using intuitions based on small gossip communities. And maybe even a less gossipy but larger network would still have this effect; like, I *never* hear gossip about people in communities I’m not a part of, even if I talk to some people from those communities, so there’s more structure than just the rate of gossip. It’s more a question of how large is the “gossip-percolation connected component”.)
See PhoenixFriend’s comment, there were multiple cases I didn’t know about, so a lot of people’s thoughts about this post are recapitulating sampling bias from my own knowledge (which is from my own social network, e.g. oversampling trans people and people talking with Michael). This confirms that people are avoiding volunteering the information that they had a psychotic break.
PhoenixFriend alleges multiple cases you didn’t know about, but so far no one else has affirmed that those cases existed or were closely connected with CFAR/MIRI.
I think it’s entirely possible that those cases did exist and will be affirmed, but at the moment my state is “betting on skeptical.”