Thank you to True Discipline for gathering the scores.
Ted Kaczynski
He got a score of 167 on an IQ test in 6th grade (cut to 155) (maximum score)
PhD, 1967. Using the Jensen method returns a (maximum) z-score of 2.1.
Assuming longitudinal regression to the mean of 0.8 and a correlation betweeen eduttainment and IQ of 0.55 returns an IQ of 151 with a standard error of 9
########################
set.seed(1)
g <- rnorm(60000000)
c <- 0.55*g + rnorm(60000000)*sqrt(1-0.55^2)
d <- 0.8*g + rnorm(60000000)*sqrt(1-0.8^2)
subby1 <- data.frame(g, d)
subby1$c <- c
subby2 <- subset(subby1, (subby1$d > 3.67) & (subby1$c > 2.1))
mean(subby2$g)
sd(subby2$g)
Alvin Lee King
score of 151.
Adding simple regression to the mean decreases this to 146.
Priscilla Ford
score of 140. Regression to the African mean decreases this to 135.
Charles Whitman
score of 139 at age 6. Adjusting for a longitudinal regression of 0.6 changes this to 123.
Franz Fuchs
score of 139. Regresses to the mean to 135
Edmund Kemper
Scored 136 on one IQ test, 145 on another. Simulation returns an estimate of 140 with a standard error of 4.8.
set.seed(25)
g <- rnorm(60000000)
c <- 0.9*g + rnorm(60000000)*sqrt(1-0.9^2)
d <- 0.9*g + rnorm(60000000)*sqrt(1-0.9^2)
subby1 <- data.frame(g, d)
subby1$c <- c
subby2 <- subset(subby1, (subby1$d > 2.35 & subby1$d < 2.45) & (subby1$c > 2.95 & subby1$c < 3.05))
mean(subby2$g)
sd(subby2$g)
Timothy McVeigh
Score of 126. Regresses to the mean to 123.
David Copeland
Score of 126. Regresses to the mean to 123.
Dylan Roof
Score of 125, but he scored better on g-loaded tests (VCI - 141, but processing speed of 100). For the sake of consistency I’ll just regress him to the mean - (123).
Ted Bundy
IQ of 124. Regresses to the mean to 122.
James Holmes
Graduated in the top 1% of his class. IQ tested twice - 123 and 116. Results in a final estimation of 121 with an SE of 5.
set.seed(25)
g <- rnorm(60000000)
c <- 0.9*g + rnorm(60000000)*sqrt(1-0.9^2)
d <- 0.9*g + rnorm(60000000)*sqrt(1-0.9^2)
e <- 0.6*g + rnorm(60000000)*sqrt(1-0.6^2)
subby1 <- data.frame(g, d)
subby1$c <- c
subby1$e <- e
subby2 <- subset(subby1, (subby1$d > 1.43 & subby1$d < 1.53) & (subby1$c > 1 & subby1$c < 1.12) & (subby1$e > 2.2 & subby1$e < 2.4))
mean(subby2$g)
sd(subby2$g)
George Banks
Score of 121. Biracial, so regresses to a mean of 92.5 → resulting in an estimate of 118.
Nikolas Cruz
one score of 83. Regresses to 85.
Charles Manson
one score of 109, another of 121 page 69 of pdf. Simulations generate average of 115 with a standard error of 5.
Carel J Delport
IQ score of 79, regresses to 81.
Martin Bryant
IQ score of 66, regresses to 69.
I’m a bit of a wordcel, so have some patience— but what is the point of regressing to the mean based off of ethnic or racial group?