Intelligence, education, and mortalityBMJ 2010; 340 doi: https://doi.org/10.1136/bmj.c563 (Published 27 April 2010) Cite this as: BMJ 2010;340:c563
All rapid responses
In a recent editorial (1), Batty and colleagues discussed our work on
intelligence and mortality (2) and came to the conclusion that “efforts to
reduce inequalities should continue to be broadly based, including
educational opportunities and interventions initiated in early life”. We
fully agree with this conclusion. However, we wonder if this conclusion is
consistent with the way they reason, in particular with their defence of
four ideas about intelligence and health that we questioned in our paper,
(A) that intelligence is the fundamental cause of socioeconomic
differences in health,
(B) that the importance of intelligence for mortality is the same for
men and women,
(C) that early intelligence follows on from good health rather than
the other way around, and finally
(D) that intelligence might be a non-malleable trait, as efforts to
improve it “so far [has] yielded disappointing results” (1).This last idea
has often gone hand in hand with the idea that intelligence is, by and
large, an inherited trait or, as it has been summed up, “substantially
heritable” (3, 4).
Do not these propositions suggest that the possibility to reduce
inequalities is fairly small? Luckily, there are empirical findings which
paint a very different overall picture - one that is more promising when
it comes to future public health measures. First, although propositions
(A) and (C) above may have some truth to them, they do not seem to suffice
as explanations. Given the magnitude of the associations between IQ and
mortality (at least among men)(5), this should be enough for the
intelligence-health literature to deserve attention from anyone interested
in promoting health and/or health equity. Secondly, we believe (in
contrast with proposition D) not only that intelligence can be promoted
but that this is already happening, as demonstrated by the secular trend
in intelligence, the so-called Flynn effect (6). The clear effect of
schooling on IQ (7-9) has probably contributed to this trend (3, 10). The
heritability of intelligence is not higher than that of several other
phenomena relevant to public health. Childhood obesity, for example, has a
genetic component that is as high as, or even higher than, the one for
intelligence (11, 12). The prevalence of childhood obesity has tripled or
so during the last 30 years in several countries (13). No one claims that
this is for genetic reasons. Intelligence as well, certainly, is the
result of interaction between genetics and the environment, including the
social environment. Thus, it is only our own ignorance that stops us from
formulating successful policies in this field.
Batty et al also had a number of specific concerns about our paper.
One was that using education as a primary marker of socioeconomic status
is problematic because of collinearity, i.e. that education and
intelligence are so highly correlated that education captures the effect
of intelligence. In fact, the relatively high correlation was one reason
why education was chosen as the marker of socioeconomic status, rather
than occupation and income (see page 2 of our article), since this makes
the testing of the hypotheses as sharp as possible. Given that we found
associations with mortality for one factor net of the other, collinearity
is something that would strengthen the conclusion that (for men) both are
important, in themselves. A second concern was that there could be a lack
of precision for one measurement or the other. Again, we find it difficult
to understand how this could explain the finding that both intelligence
and education are important, net of each other (for men) or the gender
difference in this.
As for the concern that our finding on women is a “spurious chance
finding” based on “subgroup analyses” (re proposition B above), we would
like to point out that the statistical interaction, that is the finding
that the effect of IQ was different for men and women, is comfortably
significant; that this result is based on the whole sample, during the
full follow-up, and that this holds in all our three major models (2).
Also from Figure 1 in the very paper that introduced the field of
cognitive epidemiology, by Whalley and Deary in this journal (14), it
seems to us that gender could modify the effect of IQ on mortality. A
gender difference is suggested since all men (regardless of IQ) seem to be
dying at a similar rate up to 45 years of age, after which age those with
low IQ die at a faster rate – while women with low IQ are dying at a
faster rate until around age 35 (and relatively more so during the war)
while after age 45 they die at a rate that is more similar to women with
The importance of education, and the interaction of early
intelligence and gender suggest to us that intelligence is not the
fundamental cause of health inequalities and that the effect of
intelligence can be modified by social circumstances. Further, other
literature on schooling and intelligence suggests that early intelligence
itself can be promoted in the appropriate social and educational
1. Batty GD, Kivimaki M, Deary IJ. Intelligence, education, and
mortality. BMJ. 2010;340:c563.
2. Lager A, Bremberg S, Vågerö D. The association of early IQ and
education with mortality: 65 year longitudinal study in Malmö, Sweden.
3. Herrnstein RJ, Murray CA. The bell curve: intelligence and class
structure in American life. New York: Free Press; 1994.
4. Deary IJ, Penke L, Johnson W. The neuroscience of human
intelligence differences. Nat Rev Neurosci. 2010 Mar;11(3):201-11.
5. Batty GD, Deary IJ, Gottfredson LS. Premorbid (early life) IQ and
later mortality risk: systematic review. Ann Epidemiol. 2007 Apr;17(4):278
6. Flynn JR. Massive IQ Gains in 14 Nations: What IQ tests Really
Measure. Psychological Bulletin. 1987;101(2):171-91.
7. Husén T. The influence of schooling upon IQ. Theoria. 1951;17(1-
8. Ceci SJ. How much does schooling influence general intelligence
and its cognitive components? A reassessment of the evidence.
Developmental Pscychology. 1991;27(5):703-22.
9. Gustafsson J. Schooling and intelligence: Effects of track of
study on level and profile of cognitive abilities. International Education
10. Ronnlund M, Nilsson LG. The magnitude, generality, and
determinants of Flynn effects on forms of declarative memory and
visuospatial ability: Time-sequential analyses of data from a Swedish
cohort study. Intelligence. 2008 May-Jun;36(3):192-209.
11. Haworth CM, Wright MJ, Luciano M, Martin NG, de Geus EJ, van
Beijsterveldt CE, et al. The heritability of general cognitive ability
increases linearly from childhood to young adulthood. Mol Psychiatry. 2009
12. Silventoinen K, Rokholm B, Kaprio J, Sorensen TI. The genetic and
environmental influences on childhood obesity: a systematic review of twin
and adoption studies. Int J Obes (Lond). 2010 Jan;34(1):29-40.
13. Wang Y, Lobstein T. Worldwide trends in childhood overweight and
obesity. International Journal of Pediatric Obesity. 2006;1(1):11-25.
14. Whalley LJ, Deary IJ. Longitudinal cohort study of childhood IQ
and survival up to age 76. Bmj. 2001 Apr 7;322(7290):819.
Competing interests: No competing interests