Estimating the IQ of Jordan Peterson

According to pumpkin person’s blog, JP scored above 150 on an IQ test, then scored in the 99th percentile on the verbal and in the 70th percentile on the quantitative portion.

According to simulations, within people with an bachelors or advanced degree, the 99th percentile IQ is roughly 2.6, and the 70th percentile IQ is roughly 0.95. I will use these as rough estimates for how good those GRE scores are.

Realistically he probably took the test between the ages of 6 and 25 - averaging out the potential stability coefficients from ages 6-15 the table here yields a regression of 0.7 - after that I will assume the coefficent is 0.9. This indicates that I should regress him by about 0.8.

Using a simulation generates an estimate of 138 with a standard error of 7.

Code:

latented <- rnorm(1000000)
g <- rnorm(1000000) + latented*.6
g = g/sd(g)

daf <- data.frame(latented, g)

daf$ed = NA
daf$ed[daf$latented > qnorm(.91)] <- 4
daf$ed[daf$latented < qnorm(.91)] <- 3
daf$ed[daf$latented < qnorm(.58)] <- 2
daf$ed[daf$latented < qnorm(.47)] <- 1
daf$ed[daf$latented < qnorm(.06)] <- 0
GREtakers <- subset(daf, daf$ed>2)
mean(GREtakers$g)
sd(GREtakers$g)
quantile(GREtakers$g, probs = seq(.01, .99, by = .01))

g <- rnorm(60000000, mean=0)
iq <- 0.8*g + rnorm(60000000)*sqrt(1-0.8^2)
cs <- 0.78*g + rnorm(60000000)*sqrt(1-0.78^2)
gr <- 0.78*g + rnorm(60000000)*sqrt(1-0.78^2)

subby1 <- data.frame(iq, cs)
subby1$g = g
subby1$gr = gr
subby2 <- subset(subby1, (subby1$iq > 3.2) & (subby1$cs > 2.4 & subby1$cs < 2.8) & (subby1$gr > 0.85 & subby1$gr < 1.05))
mean(subby2$g)*15
sd(subby2$g)*15