NunoSempere comments on 2021 EA Mental Health Survey Results

[NunoSempere]

So as you discuss, this survey suffers from selection bias. At the time, I suggested^[1] looking at SlateStarCodex results instead, filtering by self-reported EA affiliation . I can’t find the results for 2021 or 2022, but using results from 2020:

## Helpers
formatAsPercent <- function (float){
  return(sprintf("%0.1f%%", float * 100))
}

## Body
data <- read.csv("2020ssc_public.csv", header=TRUE, stringsAsFactors = FALSE)
data_EAs <- data[data["EAID"] == "Yes",]
n=dim(data_EAs)[1]
n ## 993 EAs answered the survey.

mental_illnesses <- colnames(data)[56:68]
cat("Illness, rate diagnosed, rate suspected")
for(x in mental_illnesses){
  rate_diagnosed=sum(data_EAs[x] == "I have a formal diagnosis of this condition")/n ## 18%
  rate_suspected=sum(data_EAs[x] == "I think I might have this condition, although I have never been formally diagnosed")/n ## 18%
  cat(paste(x, formatAsPercent(rate_diagnosed), formatAsPercent(rate_suspected), sep=", "), "\n")
}

returns

   count           type  frequency
1    179     Depression 18.0261833
2    128        Anxiety 12.8902316
3     20            OCD  2.0140987
4     18 Eatingdisorder  1.8126888
5     20           PTSD  2.0140987
6      6     Alcoholism  0.6042296
7      3  Drugaddiction  0.3021148
8      3     Borderline  0.3021148
9     17        Bipolar  1.7119839
10    57         Autism  5.7401813
11   133           ADHD 13.3937563
12     5  Schizophrenia  0.5035247

Plotting this:

Code to replicate this here, source for the data here^[2]. There are other things one could look at, like whether self-assessed EAs are more likely to be mentally ill, or whether this is mediated by income, student status, etc.

1. ^

   This point was originally suggested by David Moss in the first iteration of the survey; I just remembered.

2. ^

   I downloaded the .xlsx data and converted it to a csv, because the csv data doesn’t seem to be available.