Over the years, I’ve become a scholar of Lewis Terman’s Genetic Studies of Genius, which was a study that followed the lives of 1,528 high-IQ children from 1921 to 1999. It’s a landmark study in the history of psychology and was groundbreaking in the fields of methodology, gifted education, intelligence research, and developmental psychology (Warne, 2019).
One criticism of the study that has appeared in recent years is that in Terman’s search for “geniuses” (which he defined as children having an IQ of 140 or higher), he missed two children who would later earn a Nobel Prize in physics: William Shockley and Luis Alvarez. For example, Simonton (2015) believed that this was evidence that intelligence tests could not predict creative achievement. Bond (2014) thought this was a consequence of Terman’s overemphasis on IQ.
While it is certainly true that Alvarez and Shockley were screened for the study (and not accepted), this is not the fatal flaw that some commentators have suggested.
In a newly published article, my coauthors and I ran a simulation study to determine Terman’s probability of selecting one or both future Nobelists for his Genetic Studies of Genius (Warne et al., 2020). Our results show that the odds were stacked against Terman from the beginning.
. . . Terman had a probability of 53.0% to 83.2% of selecting neither Shockley nor Alvarez for the Genetic Studies of Genius. The probability of selecting at least one future Nobelist was 16.2% to 39.7%, and the probability of identifying both individuals was 0.6% to 7.3%.Warne et al., (2020, p. 7, emphasis removed)
Why Terman Missed Two Nobelists
Our simulation showed that the major reason that Terman failed to identify Alvarez and Shockley was because his minimum IQ for inclusion in the Genetic Studies of Genius (140) was too high. IQ scores in the 120s seem to be common among Nobelists (Root-Bernstein, 2015). Lowering the minimum IQ for the study would have increased Terman’s sample size about tenfold and made it too large to study with the resources he had.
We also found that the correlation between IQ and earning a Nobel Prize was too weak for it to be a useful predictor. While we do not know the true correlation between these two variables, the correlation would have to be unrealistically strong (greater than r = .40 and probably greater than r = .80) for IQ to be a good predictor of earning a Nobel Prize.
Why It’s Hard to Predict Future Eminence
If this simulation had only historical implications, then the story would be over. But in the simulation, the Genetic Studies of Genius, and related studies together provide insight into why Terman missed Alvarez and Shockley as geniuses.
First, rare outcomes are–by definition–hard to predict. Statistical procedures work very well with dependent variables that have a lot of variance. But because very few people earn a Nobel Prize, this dependent variable has almost no variance, so almost nothing correlates with it. Plus, eminence takes some luck, which is random.
Second, intelligence tests are the wrong tool to use if one wishes to identify specific outcomes. I agree with Julian Stanley when he stated,
The IQ is a valuable global measure of intellect. It tells us much overall, but not enough specifically.Stanley (1974, p. 7)
IQ is a great predictor of general life outcomes: longevity, socioeconomic status, adult education level, etc. It is not a good predictor of specific life outcomes. And is there any life outcome more specific than earning a Nobel Prize in one particular area? Probably not.
Identifying Future Excellence
In the article, we also discussed how to identify future excellence. Clearly an intelligence test is not sufficient. Instead, we recommend testing children on specific mental abilities that line up with the future goal. For example, if you want to predict which child(ren) will grow up to be a novelist, measure high verbal ability. Are you looking for future mathematicians? Use a math test. This may sound like obvious advice, but too many people ask for general scores (e.g., ACT composite scores, IQ scores, SAT) to do too much. These are global measures that tell us a lot in general terms, “but not enough specifically” (Stanley, 1974, p. 7).
Second, interests matter. Most people select college majors and careers that are in fields where they have some interest. Even if a person has an IQ high enough to accomplish something great in a field, that doesn’t mean they will. They have to be interested in the topic and willing to put in the hundreds (often thousands) of hours necessary to make a great contribution (McCabe et al., 2020).
Third, go where young people with high abilities are. For future Nobelists, this means selecting young people who attend elite universities. For future eminent performers, this means finding young people who are immersing themselves in the performing arts (Warne et al., 2020). Terman’s selection from the general school population was a key decision that made his research more generalizable (which was his goal), but hindered his chances of finding future Nobelists (which was not his goal).
What This Means for the Genetic Studies of Genius
Whether Alvarez and Shockley were “geniuses” probably depends on your definition of a “genius.” If the requirement is a childhood IQ of 140 or higher, then they were not. If “genius” is manifested as high intellectual accomplishment in adulthood, then they were geniuses. I don’t think that the word has any scientific usefulness, so I don’t really care.
What I do think is that Terman’s failure to select Alvarez and Shockley for the Genetic Studies of Genius is not a fatal flaw of the study or of intelligence testing. It’s expected when an outcome is rare and a general ability test is used as the only predictor of a specific outcome. But what is clear in 21st century hindsight was not clear at all in the 1920s. Terman did the best he could to identify “geniuses.” Post hoc armchair criticism greatly underestimates the difficulties in finding future Nobelists as children.
Bond. M. (2014). The secret of success: Blood, or sweat and tears? New Scientist, 221(2959), 30-34. https://doi.org/10.1016/S0262-4079(14)60488-7
McCabe, K. O., Lubinski, D., & Benbow, C. P. (2020). Who shines most among the brightest?: A 25-year longitudinal study of elite STEM graduate students. Journal of Personality and Social Psychology, 119(2), 390-416. https://doi.org/10.1037/pspp0000239
Root-Bernstein, R. (2015). Arts and crafts as adjuncts to STEM education to foster creativity in gifted and talented students. Asia Pacific Education Review, 16(2), 203-212. https://doi.org/10.1007/s12564-015-9362-0
Simonton, D. K. (2015). Collaborating creators still have personal psychologies [Review of The innovators: How a group of hackers, geniuses, and geeks created the digital revolution, by Walter Isaacson]. The American Journal of Psychology, 128(3), 403-407. https://doi.org/10.5406/amerjpsyc.128.3.0403
Stanley, J. C. (1974). Intellectual precocity. In J. C. Stanley, D. P. Keating, & L. H. Fox (Eds.). Mathematical talent: Discovery, description, and development (pp. 1-22). Johns Hopkins University Press.
Warne, R. T. (2019). An evaluation (and vindication?) of Lewis Terman: What the father of gifted education can teach the 21st century. Gifted Child Quarterly, 63(1), 3-21. https://doi.org/10.1177/0016986218799433
Warne, R. T., Larsen, R. A. A., & Clark, J. (2020). Low base rates and a high IQ selection threshold prevented Terman from identifying future Nobelists. Intelligence, 82, Article 101488. https://doi.org/10.1016/j.intell.2020.101488 (preprint version)