Arthur R. Jensen (1999) Evoked Potentials, Testosterone, and g. Psycoloquy: 10(085) Intelligence g Factor (22)

Volume: 10 (next, prev) Issue: 085 (next, prev) Article: 22 (next prev first) Alternate versions: ASCII Summary
PSYCOLOQUY (ISSN 1055-0143) is sponsored by the American Psychological Association (APA).
Psycoloquy 10(085): Evoked Potentials, Testosterone, and g

Reply to Tan on Jensen on Intelligence-g-Factor

Arthur R. Jensen
Educational Psychology
School of Education
University of California
Berkeley, CA 94720-1670


Tan (1999) questions some of the evidence presented in "The g Factor" (Jensen, 1998) showing biological correlates of g, the highest-order common factor in a battery of diverse mental tests. Tan's most interesting comments concern sex differences that affect mental abilities, possibly including the correlations of g with various physiological indices, through hormonal differences between the sexes, particularly testosterone.


behavior genetics, cognitive modelling, evoked potentials, evolutionary psychology, factor analysis, g factor, heritability, individual differences, intelligence, IQ, neurometrics, psychometrics, psychophyiology, skills, Spearman, statistics
1. Tan (1999) summarises some of my book's (Jensen, 1998, 1999) main conclusions about the correlation of brain size with IQ, and agrees that the simple correlation averages about +.40 across a number of studies based on in vivo brain measurements by MRI. Head size, which correlates only about +.50 with brain size, was shown by the method of correlated vectors to be related mostly to the g factor of IQ, but also to a spatial reasoning factor, although to a lesser degree. The issue of whether or not body size should be statistically controlled in the brain-IQ correlations is still regarded as problematic. Because the correlation of brain with body-size is only about +.15 in age-homogeneous groups, partialling body size measures out of the IQ-brain size correlation has little effect. Body size, however, apparently has no intrinsic correlation with IQ. Although in the population the correlation between IQ and stature is about +.20, there is a near-zero correlation between these variables among full siblings, that is, the siblings' differences in height do not in the least predict their differences in IQ (Jensen & Sinha, 1993). The dissenting article about the brain-IQ correlation by Cain & Vanderwolf (1990) cited by Tan is actually a weak and ineffective criticism in light of the total body of evidence for the brain-IQ correlation and the aggregated data on racial differences in brain size, as pointed out in Rushton's (1995, pp.238-241) detailed critique of the Cain & Vanderwolf article.

2. Tan echoes some of Verleger's (1999) comments on evoked potentials; see my reply to Verleger (Jensen, 1999). Tan's additions to Verleger's comments again underline the complexity and virtual chaos in the research on evoked potentials or AEP and IQ. Yet through all the "noise" in the empirical studies of AEP and IQ, the sought correlation often shows up, but also often fails to do so, or is contingent on the subject's sex, or age, or some other variable. One wonders how long it will be before all of the controlling variables in AEP research will be brought sufficiently under experimental control for it to begin yielding consistent findings from one sample to another in the same laboratory and across different labs. No comprehensive review or meta-analysis of the relationship between EEG, ERP, or AEP variables to psychometric variables has yet been able to draw any more satisfying conclusions beyond the fact that there are significant correlations, but these are scarcely understood and don't replicate reliably enough to support the formulation of a coherent explanation or to rule out competing hypotheses. ERP-IQ correlations are highly sensitive to procedural variations and to the specific type of measurements derived from the ERPs. I believe that the ERP studies using the "neural adaptability" and the amplitude habituation measures of E. W. P. Schafer cited in my book are simple and straight-forward enough to provide highly replicable studies, provided Schafer's procedures are followed explicitly. Nothing would please me more than to see Schafer's study of the habituation index repeated in another laboratory, using a subject sample that is fairly homogeneous in age, with a range of IQs ranging from 85 or 90 to 130 or above. Also, a diverse battery of psychometric tests should be used to obtain their g loadings and permit the correlated vectors analysis as well as the simple correlation with the overall total score or Full Scale IQ.

3. Tan's citations of studies showing the effects of testosterone (T) on spatial ability, which may well account for the sex difference on this factor, suggests that we should take a closer look at the effects of T and other hormones on various ability factors. Helmuth Nyborg (1994), one of the leading researchers on T, claims a nonlinear (inverted U) relationship between spatial ability and T level such that men who are very high (or very low) on T have somewhat lower spatial ability than men who are in the middle range of T levels. Therefore it is important to take note of the location of a particular subject sample's T distribution within the full range of T levels. Nyborg & Jensen (2000) have examined the effects of T on g itself in large samples (N=4000) of White and Black U.S. Armed Services veterans and found that g factor scores are low only at the top and bottom extremes of the distribution of T concentrations in the blood, with a somewhat smaller effect in Ws than in Bs, who have a slightly higher mean level of T. For the vast majority falling between the few per cent falling at both the extremes of T level, however, there appears to be no relationship between T and g.

4. One cannot agree with Tan's notion that I am expressing "subjective ideas rather than objective scientific observations and conclusions". I can see no basis for "subjective ideas" in this realm, unless the explicit formulation of hypotheses is considered "subjective". The conclusions I have drawn from the obviously objective evidence presented in my book, though undoubtedly limited by the boundary conditions of the evidence, can hardly be called "subjective" in any sense in which this word is defined in my Webster's dictionary. I think the objective "factual density" of the chapters in my book probably far exceeds the average for psychological books in general.


Cain, D. P., & Vanderwolf, C. H. (1990). A critique of Rushton on race, brain size and intelligence. Personality and Individual Differences, 11, 777-784.

Jensen, A. R. (1998). The g factor: The science of mental ability. Westport, CT: Praeger.

Jensen, A. R. (1999). Precis of: The g Factor: The Science of Mental Ability. PSYCOLOQUY 10(023) psyc.99.10.023.intelligence-g-factor.1.jensen

Jensen, A. R. (1999a). Evoked brain potentials and g. Reply to Verleger on Jensen on Intelligence-g-Factor. PSYCOLOQUY 10(083) psyc.99.10.083.intelligence-g-factor.21.jensen

Jensen, A. R., & Sinha, S. N. (1993). Physical correlates of human intelligence. In P.A. Vernon (Ed.) Biological approaches to the study of human intelligence (pp. 139-242). Norwood, NJ: Ablex.

Nyborg, H. (1994). Hormones, sex, and society: the science of physicology. Westport, CT: Praeger

Nyborg, H., & Jensen, A. R. (2000). Testosterone levels as modifiers of psychometric g. Personality and Individual Differences, 28, 601-607.

Rushton, J. P. (1995). Race, evolution, and behaviour. New Brunswick, NJ: Transaction.

Tan, U. (1999) The biological correlates of the g factor. PSYCOLOQUY 10(049) psyc.99.10.049.intelligence-g-factor.6.tan

Verleger, R. (1999). The g factor and event-related potentials. PSYCOLOQUY 10(039) psyc.99.10.039.intelligence-g-factor.2.verleger

Volume: 10 (next, prev) Issue: 085 (next, prev) Article: 22 (next prev first) Alternate versions: ASCII Summary