Case Study One: Dumb College Students

Early in The Bell Curve, Herrnstein and Murray report that they have come across "two little-known sources of test data" that enable them to report that the entering classes of the Ivy League and Seven Sisters colleges in 1926 "had a mean IQ of about 117," which is "barely above the 115 level that has often been considered the basic demarcation point for prime college material." This is important evidence for their theory that inherent brainpower is more concentrated in elite institutions today than it was in the past.

Let's look at this number a little more closely. Herrnstein and Murray say they came up with it by using a table they found "for converting the SAT of that era to IQ scores." When you go to the source, though, you find that they do not actually have the mean SAT scores for Ivy League students. What they have are the SAT scores of applicants to these colleges--and not all applicants, just those who happened to take the SAT in its first year of administration, when none of the colleges required it. Also, the IQ formula that Herrnstein and Murray used is an extremely primitive measure, obtained by adding the number of questions that the test-taker got right to a number for the student's age. Ivy Leaguers may well have been relatively dumb in 1926, but for Herrnstein and Murray to say that they have accurately calculated these students' IQ back then is a real stretch.

(Robert Hauser of the University of Wisconsin reran Herrnstein and Murray's calculation using actual students at four top women's colleges, rather than applicants, and got a considerably higher average IQ of 121.)

One more example: Herrnstein and Murray have a chart showing the likelihood that the top quarter of high-school graduates, in terms of IQ, will go to college; the likelihood zooms upward after 1950. Go back to the source, and you find that the chances of going to college rose steeply after 1950 at all IQ levels. Why? Because higher education expanded rapidly after 1950. In short, the authors simply leave out basic information that would sever any connection between the evidence they offer and what it is intended to prove.

{{{{Case Study Two#50883}}#2:CaseStudyTwo.asp}}: Smart Rich People