I have been skeptical of assertions that Jerome Kagan’s longitudinal study, begun in 1989, demonstrates that high-reactive infants turn into introverts, and low-reactive infants into extraverts. I purchased Jerome Kagan and Nancy Snidman’s book (The Long Shadow of Temperament) to find out. It turns out that it isn’t about introverts and extraverts, nor does it make any claims about them!
The study examines the relation between levels of reactivity in infancy (that is, reactivity to unfamiliar visual, auditory, and olfactory stimuli) and subsequent levels of inhibition. Inhibition and introversion are not the same. There is some overlap between them, but one cannot draw conclusions about introverts from a study of inhibition.
Five hundred four-month-old infants were tested for their reactions to stimuli. Of the 237 children who returned for a follow-up study at age 11, only 33 percent of the former high- and low-reactives showed behavior consistent with their infant temperament (Kagan and Snidman, p. 19). About a fourth showed both behavior and biology consistent with their infant profile. (But only a small percentage showed the opposite of their infant profile.) For the purposes of the study, these results are interesting; they do suggest a relation between infant reactivity and later temperament. Still, three points stand out: (a) first, while the study considers all levels of reactivity, it focuses on the high- and low-reactive infants; (b) most of the high- and low-reactive infants under study did not retain the expected behavioral profile at age 11 (though few moved to the opposite profile); and (c) the profile of inhibition does not match, point by point, with profiles of introversion and extraversion. Thus any conclusions about introverts and extraverts are incorrect and unwarranted.
I imagine Kagan and Snidman would agree. They take pains to dispel any simplistic conclusions about the predictability of adolescent and adult temperament; in addition, they distinguish between inhibition and introversion. They note on p. 218 that “Carl Jung’s descriptions of the introvert and extrovert, written over 75 years ago, apply with uncanny accuracy to a proportion of our high- and low-reactive adolescents.” They do not specify the proportion, but the very statement suggests a distinction between high-reactivity and introversion.
Nonetheless, people continually cite the study as evidence that high-reactive babies turn into introverts.
Susan Cain states in Quiet: The Power of Introverts in a World that Can’t Stop Talking, p. 99, “For one of those studies, launched in 1989 and still ongoing, Professor Kagan and his team gathered five hundred four-month-old infants in his Laboratory for Child Development at Harvard, predicting they’d be able to tell, on the strength of a forty-five-minute evaluation, which babies were more likely to turn into introverts or extroverts.”
No, that was not the goal of the study. But that did seem to be her takeaway; in an interview with NPR, she stated that introverts and extraverts have “literally, different nervous systems.” Whether she was referring to Kagan’s study or something else, the statement needs clarification.
Others have seized on the takeaway and taken it even farther. In an opinion piece on PsychCentral, Neil Thompson claims that “Kagan found that those who reacted strongly to the stimuli were introverts, exhibiting serious and careful personalities at each age. The children with minimal reaction to the stimuli were confident and relaxed; they were extroverts (Kagan and Snidman, 2004).” It doesn’t seem that Thompson looked at the book. Moreover, he is equating introversion with inhibition.
When discussing scientific findings on introversion and extraversion, it is essential to define terms clearly, interpret the studies accurately, and apply them carefully to the topic of discussion. (I don’t mean one should be “inhibited” in this regard; one probably needs a mix of intellectual caution and boldness.)
Kagan’s study says nothing about whether infants’ reaction to stimuli predicts their later introversion or extraversion.
Note: I made a few additions and edits to this piece after posting it. In particular, I changed “showed temperaments consistent with their infant profile” to “showed behavior consistent with their infant temperament,” since the latter wording reflects the authors’ findings more accurately.
Update: See my review of Kagan and Snidman’s book.