288 J Epidemiol Community Health 2001;55:288
It seems equally probable that the non- underestimated sex diVerences in reactivity.
manual and manual groups may have experi- She also speculates that the low reactivity in
LETTERS TO enced the stressor in very diVerent ways. the non-manual group might have reﬂected
Thus, for example, the lower reactivity in the disengagement from the task.
THE EDITOR manual group could easily reﬂect lesser We have no information on the subjective
engagement with the task in this group. impact of mental arithmetic in our study, but
Indeed, in the absence of any plausible given the general tendency for men to be less
biological mechanism to suggest that higher declarative than women when it comes to
socioeconomic groups should show more negative aVect,4 one can guess that women
Measuring demographic cardiovascular reactivity to stress, an explana- might respond more negatively. It remains to
be determined, however, what impact such
group diVerences in tion based on diVering subjective reactions to
reporting tendencies have on cardiovascular
the arithmetic task seems far more probable.
cardiovascular reactivity Without data on subjective experience of the reactivity. It would be rash to assume, as Dr
Pollard does, that they have a substantial
to stress using a mental test, I suggest that general conclusions about
impact. I have spent a number of years
demographically deﬁned diVerences in
challenge in the cardiovascular reactivity to stress are diYcult exploring cardiovascular reactions to chal-
lenges such as mental arithmetic and have yet
laboratory: problems of to justify from the data presented; the
to ﬁnd inter-individual correlations between
conclusions should be limited to a descrip-
interpretation tion of responses to a mental arithmetic test. subjective report and magnitude of reaction.
My experience is far from unique.
EDITOR,—In their recent paper Carroll et al1 TESSA M POLLARD We do have data on performance; women
showed that men and women in non-manual Department of Anthropology, University of Durham, did less well in mental arithmetic than men,
occupations exhibited greater diastolic blood 43 Old Elvet, Durham DH1 3HN, UK and manual did less well than non-manual
pressure and heart rate responses to a mental Correspondence to: Dr Pollard groups. If the latter ﬁnding is considered
arithmetic test than did people in manual (email@example.com) consistent with the disengagement hypoth-
occupations, a ﬁnding that they contrasted esis, so must the former. Dr Pollard cannot
with known socioeconomic patterning of have it both ways. Indeed, controlling for per-
cardiovascular disease. They also demon- 1 Carroll D, Harrison LK, Johnston DW, et al. formance variations attenuated the sex diVer-
strated lower systolic blood pressure and Cardiovascular reactions to psychological
stress; the inﬂuence of demographic variables. J ences in reactivity. I am not unsympathetic to
heart rate reactivity to the test in women, the notion that social class diVerences in
Epidemiol Community Health 2000;54:876–7.
which they considered more consistent with 2 Stoney CM, Davis M, Matthews KA. Sex reactivity stem, at least in part, from diVer-
sex diVerences in cardiovascular disease. It diVerences in physiological responses to stress ences in task engagement, even though they
does indeed seem likely that lower catecho- and in coronary heart disease. Psychophysiology
survived adjustment for performance in the
lamine and blood pressure reactivity to stress
3 Girdler SS, Light KC. Hemodynamic stress present analyses. However, at the very least,
in women is one of the explanations for their
lower rates of cardiovascular disease, an eVect responses in men and women examined as a our data suggest that could we equate
function of female menstrual cycle phase. Int J engagement, by using a task that is less likely
that has been ascribed largely to the eVects of Psychophysiol 1994;17:233–48.
oestrogen.2 to privilege higher social class participants,
4 Girdler SS, Jamner LD, Jarvik M, et al. Smoking
However, I suggest that Carroll et al cannot status and nicotine administration diVerentially lower social class participants would still not
claim to have tested cardiovascular reactivity modify hemodynamic stress reactivity in men show higher reactions and, thus, our conclu-
to stress, but only cardiovascular reactivity to and women. Psychosom Med 1997;59:294–306. sion remains appropriate. Finally, it is worth
a mental arithmetic test. Presumably the pointing out that cardiovascular reactions to
authors’ aim was to use a mental arithmetic mental arithmetic are correlated with reac-
task to cause feelings of stress among partici- Author’s reply tions to other commonly used stress chal-
pants, but they provide no information on lenges,5 and that our participants were tested
whether this was indeed the eVect of the task, EDITOR,—In analyses of three large age at home and not “in the laboratory”.
and, most crucially, whether its eVects were cohorts of men and women, we found that DOUGLAS CARROLL
the same in all the demographic groups women showed lower cardiovascular reac- School of Sport and Exercise Sciences, University of
tested. Girdler and Light3 showed that tions to a mental arithmetic challenge than Birmingham, Birmingham B15 2TT, UK
women reported greater feelings of helpless- men and manual lower reactions than non-
ness than men in response to a variety of manual social class groups.1 We concluded
laboratory stressors, but particularly after a that diVerences in cardiovascular disease
mental arithmetic task, as well as reporting 1 Carroll D, Harrison LK, Johnston DW, et al.
among men and women could be mediated Cardiovascular reactions to psychological
being less calm after all tasks, despite being by variations in cardiovascular reactivity, but stress: the inﬂuence of demographic variables. J
more satisﬁed with their performance. In a Epidemiol Community Health 2000;54:876–7.
that this sort of mechanism was unlikely to
diVerent sample, Girdler et al4 showed that 2 Girdler SS, Light KC. Hemodynamic stress
underlie diVerences in the prevalence of responses in men and women examined as a
women reported signiﬁcantly more tension
during a mental arithmetic task than did cardiovascular disease among diVerent social function of female menstrual cycle phase. Int J
class groups. Dr Pollard challenges these Psychophysiol 1994;17:233–48.
men. It is certainly possible that similar 3 Girdler SS, Jamner LD, Jarvik M, et al. Smoking
eVects may have occurred in the experiments conclusions, largely on the basis of our use of status and nicotine administration diVerentially
conducted by Carroll et al, in which case we mental arithmetic as a stress task. She cites modify hemodynamic stress reactivity in men
might be even more impressed by the fact scattered ﬁndings by Girdler and colleagues2 3 and women. Psychosom Med 1997;59:294–306.
that women feel more helpless after but 4 Niven CA, Carroll D, eds. The health psychology
that women showed less cardiovascular re- of women. Reading: Harwood, 1993.
sponse, and more likely to ascribe this to bio- tenser during mental arithmetic, and argues 5 Turner JR. Inter-task consistency: An integra-
logical diVerences between men and women. that our study may have, if anything, tive review. Psychophysiology 1988;25:235–8.