Gender Differences in Food Choice: The Contribution of Health Beliefs and Dieting

Jane Wardle, Ph.D., Anne M. Haase, Ph.D., and Andrew Steptoe, D.Phil.
University College London

Maream Nillapun, Ph.D. and Kiriboon Jonwutiwes, Ph.D.

Silpakorn University

France Bellisle, Ph.D.

INSERM U341
Paris, France

ABSTRACT other countries show similar results (4,5). Curiously, behavioral

Background: Gender differences in health behaviors have
been reported in many studies but causal mechanisms have
been neglected. Purpose and Methods: This study examines 4
food choice behaviors in a large sample of young adults from
23 countries and tests 2 possible explanatory mechanisms for
the gender differences—women’s greater likelihood of dieting
and women’s greater beliefs in the importance of healthy diets.
Results: Women were more likely than men to report avoiding
high-fat foods, eating fruit and fiber, and limiting salt (to a
lesser extent) in almost all of the 23 countries. They were also
more likely to be dieting and attached greater importance to
healthy eating. Dieting status explained around 22% of the
gender difference in fat choices, 23% of fiber choices, and 7%
of fruit, but none of the gender difference in salt. Health be-
liefs explained around 40% of the differences in each of the di-
etary behaviors and together they explained almost 50%. Gen-
der differences in food choices therefore appear to be partly
attributable to women’s greater weight control involvement
and partly to their stronger beliefs in healthy eating. Conclu-
sions: Further research is needed to understand the additional
factors that could promote men’s participation in simple
healthy eating practices.
(Ann Behav Med 2004, 27(2):107–116)
INTRODUCTION
Men live an average of 7 years fewer than women (1).
This results from a combination of higher male mortality in in-
fancy, more accidental deaths in young adult men, and a
higher risk of the major causes of death in middle-aged men
than middle-aged women. In the United States, for example,
the age-adjusted death rate for heart disease is twice as high in
men as women (2), and men’s cancer death rate is 1.5 times
higher (3). Comparisons between male and female deaths in
This research was supported by Cancer Research UK.
Reprint Address: J. Wardle, Ph.D., Cancer Research UK, Health Be-
haviour Unit, Dept. of Epidemiology & Public Health, University Col-
lege London, Gower Street, London WC1E 6BT, United Kingdom.
E-mail:j.wardle@ucl.ac.uk
© 2004 by The Society of Behavioral Medicine.
medicine investigators, along with many clinical researchers,
have paid little attention to this striking phenomenon. A recent
report by the Institute of Medicine (6) noted that “being male
or female is an important fundamental variable that should be
considered when analyzing basic and clinical research” (p. 7).
The report also stated that “the understanding of sex differ-
ences in health and illness merits serious scientific enquiry in
all aspects of biomedical and health-related research” (p. 4).
This “call to arms” was focused particularly on biological dif-
ferences, but sex differences in behavior and cognition were
also noted to play a part in differences in health. Observations
that the female survival advantage is a relatively modern phe-
nomenon, that it has varied considerably over time, and that it
differs between countries add to the case for the importance of
environmental and behavioral factors in sex differences in
health (7–10).
The fact that the major diseases of adult life have a strong
behavioral element (11) means that studies of sex differences in
health behaviors could play an important role in understanding
sex difference in health and illness. A study of cardiovascular
disease mortality in 24 countries found that gender differences
in five risk factors explained over 40% of the gender difference
in mortality (12). Many studies of health behaviors have noted
that men have higher rates of risky behaviors and lower rates of a
range of healthy and hygienic practices than women (13–17). A
recent review confirmed that across more than 30 health behav-
iors, ranging from wearing seatbelts and smoking to physician
visits and nutrition, men demonstrated fewer healthy choices
than women (18). Such extensive behavioral differences could
make a significant contribution to gender differences in morbid-
ity and mortality. The ubiquity of observations of gender differ-
ences in health behaviors might be expected to have generated
an extensive literature on the causal mechanisms, but as ob-
served in a recent review, gender issues, along with other demo-
graphic factors, have been relatively neglected in behavioral
medicine research (19).
Courtenay (20) argued that socially constructed gender
roles, which identify health risk behaviors with maleness and
health protective behavior with femaleness, are part of the
mechanism for gender differences in health behaviors. He
pointed out that men on television smoke more, drink more, and
are more overweight, yet they are much less likely to be diag-

107
108 Wardle et al.
nosed with illnesses (21–24)—thus appearing to be citizens who
can take health risks without paying health costs (20,25). Men
who subscribe to the socially constructed gender roles would
therefore be particularly at risk of premature death from dis-
eases where health behaviors play an important role.
Food choices are an area in which consistent behavioral dif-
ferences have been observed. Numerous studies have reported
that boys and men eat fewer fruits and vegetables, choose fewer
high-fiber foods, eat fewer low-fat foods, and consume more
soft drinks than do women (26–32). Differences in reported
food choices have not always been reflected in differences in the
proportion of energy consumed as fats, or fiber intake (33–37),
but this is partly because of gender differences in alcohol con-
sumption, which may add a substantial sum of energy as “drink
calories.” Adjusted for energy intake, women’s diets do tend to
be higher in micronutrients than men’s diets (35,38).
Men rate many health behaviors, including food choice be-
haviors, as less important than women rate them (18,39,40).
Courtenay’s (20) analysis made special mention of nutrition:
“[The socially prescribed male] would not be interested in learn-
ing about health and nutrition, or cooking, and he would be un-
concerned about his weight, diet or hygiene” (p. 10). It has also
been shown that men give lower priority to health compared
with other considerations such as taste or convenience in mak-
ing food choices (4,41–45). Differences in beliefs about the im-
portance of eating in accordance with health recommendations
could therefore contribute to explaining the gender differences
in food choice.
Another factor that might contribute to gender differences
in diet is women’s greater concern about weight control and
their higher frequency of dieting. In almost every study of atti-
tudes to body weight or dieting, women report more worries
about weight and make more attempts to control weight than do
men (41,46–48). This applies equally in pre-adolescence and in
old age (49,50), to the poor as well as the rich (51), and in devel-
oping as well as developed countries (52–54). Selecting lower
fat and lower sugar foods, as well as increasing intake of fruits
and vegetables, are important aspects of many weight-control-
ling diets (31,55,56), and therefore differences in dieting
might partly explain gender differences in food choices, particu-
larly food choices such as fat restriction, which are linked with
dieting.
Closely related to the dieting issue is the sociological analy-
sis of women’s eating practices, which suggests that femininity
is associated with “lightness” and delicacy in appetite (57,58).
According to this view, the volume of foods eaten should appear
to be less and the type of foods selected should appear to be less
substantial and more “healthy” for women than men (59–62).
Gender differences in food choice would therefore be most ap-
parent where lightness or “low energy” were important differen-
tiating factors, even where dieting, per se, is not an issue.
This study examines gender differences in four food choice
behaviors that were measured in the International Health Behav-
iour Survey (IHBS): avoiding fat, eating fiber, eating fruit, and
limiting salt (63). The same food choice behaviors had also been
assessed in the predecessor study (the European Health Behav-
Annals of Behavioral Medicine
iour Survey [EHBS]; 17). They had been chosen to represent as-
pects of food choices on which there was consensus about the
health benefits, which could be simply phrased, and would be
meaningful across the young adult sample drawn from many
different countries. On the basis of existing work, we expected
to see gender differences in food choices in most if not all of the
participating countries. The study also tested the hypotheses that
gender differences would be most marked in food choices rele-
vant to dieting or “light” food consumption (fat and fruit), that
differences in intake would be partly mediated by dieting status,
that beliefs about the importance of healthy food choices would
be lower in men, and that differences in food choices would be
partly mediated by gender differences in beliefs. Finally, we ex-
amined how much of the gender differences in food choices
were left unexplained after taking account of dieting status and
health beliefs.
METHOD
Study Sample and Design
The IHBS is a self-report, questionnaire-based study of a
wide range of health behaviors and attitudes carried out in uni-
versity students from 23 different countries between 1999 and
2001. The international perspective permits both cross-cultural
comparisons and examination of the predictors of health behav-
iors in diverse cultural contexts. Data collection was carried out
at one university in each of 19 countries and at two universities
in the other 4 countries. Questionnaire items were based on the
forerunner of the IHBS, the EHBS; a very similar survey was
carried out in Europe between 1989 and 1991 (17,63). Data col-
lection was carried out by a network of collaborators from one
or two universities in each country, who distributed the IHBS
questionnaires to students on non-health-related courses, usu-
ally during lectures. Although questionnaire completion was
voluntary, participation exceeded 90% in most countries. A total
of 19,298 university students (men, n = 8,482; women, n =
10,816), aged 17 to 30, completed the questionnaire. The ques-
tionnaire items for the IHBS were deliberately selected to be
simple and were translated and back-translated into 17 lan-
guages (Bulgarian, Czech, Dutch, Flemish, French, German,
Greek, Hungarian, Icelandic, Italian, Japanese, Korean, Polish,
Portuguese, Romanian, Spanish, and Thai). This study is con-
cerned specifically with aspects of food choice, and therefore
only items related to food choice behaviors, related beliefs, diet-
ing status, gender, age, and smoking are reported.
Measures
Four food choice behaviors were assessed using single
items: “Do you make a conscious effort to avoid foods that con-
tain fat and cholesterol?” (yes, no), “Do you make a conscious
effort to eat foods that are high in fiber?” (yes, no), “How often
do you eat fruit?” (daily, 2–3 times a week, once a week, less
than once a week, never), and “Do you add salt to your meals?”
(usually, sometimes, very occasionally, never). Reliability of
these items was reported earlier (64). The fruit and salt items are
very similar to fruit and salt items in other food frequency ques-
tionnaires (65,66), but the other two items represent subjective
Volume 27, Number 2, 2004 Gender and Healthy Food Choices 109

perspectives on food choices, similar to those included in food
choice measures (67,68) rather than nutritional evaluations.
Comparisons were conducted between respondents who re-
ported avoiding and not avoiding high-fat foods, eating or not
eating high-fiber foods, eating fruit daily or less than daily, and
usually or less often adding salt.
Health beliefs that corresponded approximately to the four
food choices were rated on a scale from 1 (of very low impor-
tance) to 10 (of very great importance), each with the stem
“How important is it …?” The selected items to use were not to
eat too much animal fat, to eat enough fiber, to eat enough fruit.
and not to add too much salt. For some analyses, a median split
of the belief scores was used to characterize high and low be-
liefs. The belief items were not an exact match to the behaviors
because it was deemed important to keep all items the same in
the IHBS as they had been in the EHBS.
Dieting status was assessed by asking participants to re-
spond to the single question “Are you dieting to lose weight?”
with a yes–no response format.
Smoking was included as a control variable in the analyses
in the light of evidence that smokers tend to eat less healthy diets
than nonsmokers and that the men have higher levels of smoking
than women in the IHBS sample (44). Smoking was measured
by asking participants to select one of eight statements: “I have
never smoked a cigarette, not even a puff,” “I have only ever
tried one or two cigarettes”; “I used to smoke sometimes, but I
don’t now”; “I don’t smoke cigarettes, but smoke a pipe or ci-
gars”; “I smoke cigarettes, but not as many as one per day”; “I
usually smoke between 1 and 10 cigarettes per day”; “I usually
smoke between 10 and 20 cigarettes per day”; and “I usually
smoke more than 20 cigarettes per day.” For these analyses, two
categories were created, with the first three response items des-
ignated nonsmokers and the remaining five responses desig-
nated smokers.
Demographic data including gender and age were collected
from all participants.
Statistical Analyses
Age-adjusted prevalence of food choice behaviors, mean
health belief scores, and smoking and dieting frequencies, for
both men and women, were computed for each country sepa-
rately and also for all 23 countries combined, controlling for any
clustering effect by country. Results are presented as percent-
ages or means with 95% confidence intervals. Gender differ-
ences in the dependent variables (reported food choices) and the
potential mediating factors (health beliefs and dieting status) as
well as control variables (smoking status) were analyzed using
chi-square tests and analysis of variance with confidence inter-
vals adjusted for age and clustering by country. Associations be-
tween food choice behaviors and the potential mediating vari-
ables were analyzed with univariate logistic regression. Where
the conditions for mediation were met (i.e., the potential mediat-
ing variables were related both to the independent and the de-
pendent variables) (69), successive logistic regression models
were used to assess the extent to which adding the potential me-
diators (dieting status and health beliefs) to the basic model
modified the odds ratio associated with gender. All statistical
analyses were done with STATA 6 (STATA Corporation, Col-
lege Station, TX).
RESULTS
The frequencies of each food choice behavior in men and
women, adjusted for age and within-country clustering effects,
are presented in Table 1. Across the whole sample, women were
50% more likely than men to report avoiding high-fat foods and
eating high-fiber foods and 25% more likely to eat fruit at least
daily. They were, on average, 6% less likely to add salt, which

TABLE 1
Prevalence of Healthy Food Choices, Mean Ratings of Associated Beliefs, and Dieting and Smoking in Men and Women (95% CI)

Variable Womena Menb Tests of Significance

Age (years) 20.63 (20.53–20.73) 20.93 (20.82–21.04)

Dietary behavior (% of sample)
Avoiding fat 46% (42–58) 26% (23–29) χ2 = 825, p < .001
Eating high-fiber foods 40% (33–47) 27% (22–32) χ2 = 349, p < .001
Eating fruit at least daily 47% (41–54) 37% (31–43) χ2 = 225, p < .001
Never adding salt to meals 71% (64–77) 68% (61–74) χ2 = 20.8, p < .001
Health belief (rating = 1–10)
(Mean belief rating and 95% CI)
Avoiding high-fat foods 6.87 (6.83–6.92) 5.73 (5.67–5.78) F = 1025, p < .001
Eating enough fiber 7.06 (7.02–7.10) 6.18 (6.13–6.23) F = 718, p < .001
Eating enough fruit 8.13 (8.10–8.17) 7.37 (7.32–7.42) F = 700, p < .001
Avoiding adding salt 6.61 (6.56–6.66) 5.79 (5.74–5.85) F = 482, p < .001
Dieting (% dieting) 19% (15–23) 5% (4–7) χ2 = 803, p < .001
Smoking (% current smokers) 27% (25–35) 34% (30–38) χ2 = 120, p < .001

Note. Prevalence levels and mean ratings are age-adjusted for the total population, as well as adjusting the CIs for any clustering effect of data by country.
For the health beliefs, scores range from 1 (very low importance) to 10 (very high importance). CI = confidence interval.
an = 10,816. bn = 8,482.
110 Wardle et al. Annals of Behavioral Medicine

though small was still a significant effect (see Table 1). Effect
sizes (binomial effect size display) for the gender difference of
the four behaviors ranged from .21 for fat (conventionally a
small effect) through .13 for fiber, .11 for fruit, and .03 for salt.
Table 2 shows the gender differences in food choice behav-
iors in each country separately. Gender differences were ob-
served for all 23 countries for avoiding fat and eating fruit daily
and in 21 out of 23 countries for eating fiber. This consistency
was not maintained for limiting salt, where women reported lim-
iting salt more than men in 13 countries, and men more than
women in 9 countries. However, the gender differences in salt
avoidance were significant only in 6 countries, and all of these
showed a female advantage.
Mean ratings of the related dietary beliefs over all countries
combined are shown in Table 1, and individual country scores
are shown in Table 3. Food choices were accorded moderately
high levels of importance for health; that is, all belief ratings av-
eraged over the midpoint of 5 and most were over 6. Both men
and women rated eating fruit as the most important, followed by
eating fiber. Fat avoidance and limiting salt were ordered differ-
ently for men and women. Strikingly, women reported signifi-
cantly stronger beliefs in the importance of each of the aspects
of diet than men in almost every country in the sample.
Percentages and confidence intervals for dieting and smok-
ing are also shown in Table 1. As expected, almost four times as
many women as men reported dieting to lose weight. Smoking
was more prevalent in men (34%) than women (27%).
Having shown significant gender differences in each of
food choice behaviors and in the potential mediators, the next
set of analyses used logistic regression to quantify the gender
differences for each food choice behavior and to see if the two
potential mediators (dieting and beliefs) were related to the food
choice behaviors. The first column of Table 4 shows the odds ra-
tios for each food choice behavior, in relation to gender, dieting
status, and beliefs in the importance of each behavior (high vs.
low beliefs), controlling for age and within-country clustering.
The odds of healthier behavior in women compared with men
ranged from 2.44 for avoiding fat, 1.80 for fiber, and 1.56 for
fruit to 1.15 for salt, and in all cases the confidence intervals
were higher than 1. The effects of dieting status were significant
for fat (odds ratio [OR] = 3.31), fiber (2.29), and fruit (1.31) but
were not significant for salt. OR for being in the highest versus
the lowest half of the population for the related belief showed
strong and significant effects for all four behaviors, ranging
from 2.27 for salt up to 6.08 for fiber.
Given that beliefs were associated with both gender and
each of the four food choices, they met Baron and Kenny’s (69)
conditions for being potential mediators of the gender/food
choice association. Dieting met criteria for mediation for fat,
fruit, and fiber. Subsequent analyses therefore examined the
gender effect after controlling for dieting (Model 2), beliefs
(Model 3), and both dieting and beliefs (Model 4). The same
analyses are illustrated for all four food choices to maintain con-
sistency of reporting, even though the absence of an association

TABLE 2
Prevalence of Food Choice Behaviors (% Avoiding Fat, Eating Fiber, Eating Fruit Daily, and Limiting Salt Intake) in
Men and Women Across 23 Different Countries

Country Avoiding Fat Men/Women Eating Fiber Men/Women Eating Fruit Daily Men/Women Limiting Salt Men/Women

Belgium 20.6 / 44.8* 17.5 / 40.4* 32.8 / 44.0* 81.4 / 76.9

Bulgaria 22.9 / 53.7* 12.5 / 25.4* 33.8 / 34.7 70.5 / 70.5
Columbia 36.3 / 55.7* 40.5 / 50.3* 38.1 / 51.7* 49.7 / 52.2
England 29.8 / 49.5* 26.0 / 38.7* 35.8 / 44.1* 70.3 / 73.0
France 16.6 / 38.2* 10.9 / 28.7* 36.4 / 53.5* 60.8 / 67.7*
Germany 21.6 / 54.2* 25.6 / 43.1* 31.4 / 56.1* 90.1 / 89.1
Greece 27.0 / 46.5* 34.8 / 49.0* 39.2 / 47.3* 50.3 / 60.9*
Hungary 21.4 / 36.5* 12.7 / 31.3* 29.5 / 39.8* 74.4 / 71.2
Iceland 25.2 / 50.1* 25.9 / 39.7* 22.7 / 30.1* 89.1 / 93.6*
Ireland 25.7 / 47.9* 26.2 / 41.9* 43.5 / 46.8 65.1 / 74.7*
Italy 23.2 / 46.6* 18.3 / 34.4* 57.9 / 63.8* 60.2 / 69.0*
Japan 21.6 / 35.7* 27.3 / 38.0* 11.9 / 12.1 97.8 / 98.5
Korea 11.3 / 13.2 38.7 / 58.8* 16.0 / 33.3* 91.7 / 90.8
Netherlands 23.2 / 41.0* 32.0 / 46.6* 37.3 / 53.9* 53.8 / 55.9
Poland 23.1 / 46.1* 24.0 / 37.8* 42.4 / 59.6* 51.8 / 46.9
Portugal 21.7 / 42.4* 21.2 / 39.6* 54.7 / 66.4* 89.6 / 95.3*
Romania 36.5 / 47.9* 25.5 / 30.8 24.3 / 32.3* 52.8 / 49.9
Slovak Republic 19.1 / 44.0* 24.9 / 41.6* 44.5 / 63.7* 68.2 / 72.7
South Africa 37.3 / 50.6* 38.1 / 35.3 23.1 / 34.2* 46.3 / 46.5
Spain 31.5 / 52.3* 31.1 / 49.6* 57.1 / 64.0 52.5 / 54.9
Thailand 34.5 / 59.2* 64.6 / 78.6* 28.2 / 49.5* 58.9 / 58.4
USA total 33.1 / 50.4* 21.2 / 20.4 31.7 / 35.5 84.5 / 83.0
Venezuela 30.9 / 51.2* 31.8 / 46.3* 33.3 / 38.2 63.0 / 58.6

*p < .01, indicating that the prevalence rates are significantly different between men and women.
Volume 27, Number 2, 2004 Gender and Healthy Food Choices 111

TABLE 3
Age-Adjusted Mean Ratings of Food Choice Beliefs (Avoiding Fat, Eating Fiber, Eating Fruit Daily, and Limiting Salt Intake) in
Men and Women Across 23 Different Countries

Country Avoiding Fat Men/Women Eating Fiber Men/Women Eating Fruit Daily Men/Women Limiting Salt Men/Women

Belgium 4.29 / 6.18 5.14 / 6.52 6.35 / 7.60 5.35 / 6.26

Bulgaria 5.60 / 7.06 5.49 / 6.58 7.95 / 8.77 5.94 / 6.47
Columbia 5.83 / 7.21 5.86 / 6.54 7.08 / 8.09 5.90 / 6.82
England 5.72 / 6.73 6.30 / 7.13 7.40 / 8.23 5.87 / 6.61
France 5.07 / 6.28 4.79 / 5.79 6.27 / 7.39 4.88 / 5.30
Germany 5.06 / 6.72 5.79 / 6.97 7.27 / 8.40 5.23 / 5.96
Greece 6.18 / 7.26 6.84 / 7.87 7.94 / 8.59 5.00 / 5.77
Hungary 5.20 / 6.80 5.13 / 6.81 7.58 / 8.90 5.61 / 6.81
Iceland 5.46 / 7.07 6.57 / 7.91 6.72 / 8.41 5.72 / 7.19
Ireland 5.96 / 7.05 6.36 / 7.23 7.46 / 8.06 5.95 / 7.20
Italy 5.72 / 6.92 5.93 / 6.96 7.21 / 7.90 5.64 / 6.63
Japan 8.00 / 8.52 8.29 / 8.62 7.34 / 7.46 8.19 / 8.59
Korea 6.10 / 6.98 7.14 / 7.99 7.75 / 8.23 6.34 / 7.44
Netherlands 4.47 / 5.50 5.94 / 6.60 6.97 / 7.89 4.70 / 5.40
Poland 4.88 / 6.50 5.07 / 6.26 7.63 / 8.37 4.58 / 4.48
Portugal 6.11 / 7.33 6.35 / 7.67 7.45 / 8.40 6.32 / 7.58
Romania 6.53 / 7.87 7.48 / 8.63 8.73 / 9.34 6.66 / 7.31
Slovak Republic 5.98 / 7.48 6.68 / 7.77 8.17 / 8.86 5.34 / 6.18
South Africa 6.10 / 6.51 6.98 / 7.28 8.10 / 8.50 6.57 / 7.38
Spain 6.24 / 7.24 6.39 / 7.21 7.68 / 8.02 5.66 / 6.59
Thailand 5.59 / 6.73 6.62 / 7.49 6.64 / 7.58 6.70 / 7.39
USA total 5.45 / 6.28 5.46 / 5.90 6.73 / 7.43 5.66 / 6.21
Venezuela 6.12 / 7.00 6.23 / 6.66 7.08 / 7.72 6.20 / 6.86

Note. The mean ratings were significantly different between men and women in most cases.

between dieting and salt intake meant that dieting was not a po-
tential mediator of this effect. Controlling for dieting reduced
the odds associated with gender for avoiding fat (22% reduction
in OR), eating fiber (23% reduction), and eating fruit (7% reduc-
tion). As expected, there was no reduction in the OR for salt in-
take. Controlling for health beliefs (Model 3) had stronger ef-
fects, eroding the odds favoring women’s healthier choices for
fat (31% reduction), fiber (38%), fruit (39%), and salt (87%). In
the final model, both dieting and health beliefs were included.
This further reduced the female advantage in healthy eating over
either one alone for fat (47% reduction in OR), fiber (53%), and
fruit (43%), suggesting that dieting and health beliefs exert
largely independent effects.
The results shown in Model 4 show that there are still sig-
nificant gender differences in three of the four food choices,
even after health beliefs and dieting status have been taken into
account. This indicates that there must be other factors, not mea-
sured in this study, which make men and women behave differ-
ently in relation to food choices.
DISCUSSION
The results of this study demonstrate the predicted gender
differences in four food choice behaviors, confirming that in this
well-educated, young, healthy, and reasonably affluent group,
men are less likely than women to report that they try to follow
important healthy eating recommendations. Although the ef-
fects are small, if they translate into nutritional differences, they
could have an important influence on health over the longer
term. Recent research has indicated that adoption and adherence
to basic healthy eating guidelines is associated with signifi-
cantly better health outcomes (70–74). Given that we also know
that dietary preferences tend to track over the lifespan (75), our
results suggest that young men and women are on different di-
etary trajectories, which may ultimately contribute to the 7-year
differences in life expectancy.
As predicted, the effects were strikingly consistent interna-
tionally, with women being more likely than men to avoid fat
and eat fiber in all 23 countries and to eat more fruit in 21 out of
23 countries. Differences in salt intake were less consistent but
nevertheless showed a significant female advantage in six coun-
tries. The fact that similar patterns are seen across such disparate
cultures with varied cuisines indicates that gender differences in
food choice behaviors cannot be attributed to any particular id-
iosyncrasy of a Euro-American research participant bias. This
study was not concerned with cross-cultural differences per se
but rather with whether gender differences were observed across
these different settings. However, it is worth noting that there
were differences in both the average level of each reported food
choice and in the extent of the gender differences across coun-
tries, which might be an interesting focus for future research.
The main objective of these analyses was to consider
whether beliefs about the importance of each of the food choices
or dieting status (a prominent aspect of gender differences in
eating behavior) “explained” the gender differences in food
 TABLE 4
Age-Adjusted Odds (With 95% CI) of Reporting Healthy Food Choices in Relation to Gender, Dieting, and Health Beliefs
(Controlling for Age, Smoking Status, and Within-Country Clustering)

Multivariate Odds Ratios Multivariate Odds Ratios Multivariate Odds Ratios Including
Univariate Odds Ratios Including Gender and Dieting Including Gender and Beliefs Gender and Dieting and Beliefs

Avoiding high-fat foods

Gender (women vs. men) 2.44 (2.14–2.79) 2.12 (1.85–2.43) 2.00 (1.75–2.29) 1.77 (1.54–2.02)
Dieting (dieting vs. not dieting) 3.31 (2.43–4.51) 2.71 (1.99–3.69) 2.46 (1.81–3.33)
Beliefs (high vs. low beliefs) 4.52 (3.73–5.46) 4.10 (3.42–4.91) 3.99 (3.33–4.77)
Eating high-fiber foods
Gender (women vs. men) 1.80 (1.50–2.16) 1.62 (1.37–1.93) 1.50 (1.29–1.74) 1.38 (1.19–1.60)

112
Dieting (dieting vs. not dieting) 2.29 (1.82–2.88) 2.00 (1.63–2.45) 1.87 (1.50–2.34)
Beliefs (high vs. low beliefs) 6.08 (5.14–7.20) 5.81 (4.91–6.87) 5.73 (4.84–6.78)
Eating fruit at least daily
Gender (women vs. men) 1.56 (1.38–1.75) 1.52 (1.35–1.72) 1.34 (1.19–1.51) 1.32 (1.18–1.49)
Dieting (dieting vs. not dieting) 1.31 (1.12–1.53) 1.16 (0.99–1.34) 1.12 (0.96–1.30)
Beliefs (high vs. low beliefs) 3.11 (2.57–3.75) 2.98 (2.46–3.60) 2.96 (2.45–3.59)
Never adding salt to foods
Gender (women vs. men) 1.15 (1.04–1.28) 1.13 (1.02–1.26) 1.02 (0.91–1.14) 1.03 (0.92–1.15)
Dieting (dieting vs. not dieting) 1.06 (0.83–1.35) 1.02 (0.80–1.30) 0.96 (0.75–1.23)
Beliefs (high vs. low beliefs) 2.27 (1.82–2.82) 2.26 (1.80–2.84) 2.27 (1.81–2.85)

Note. CI = confidence interval.

Volume 27, Number 2, 2004
choice. The dieting hypothesis predicted that behavioral differ-
ences would be greatest for fat and fruit—dieters often change
their intake of these foods—and that controlling for dieting
would reduce this effect. The first part of this prediction was
partly supported in that the gender difference was greatest for
fat, with fruit and fiber close behind, and salt showing only a
very small gender difference. The second part of the prediction
also got mixed support. We found that controlling for dieting re-
duced the gender differences for fat and fiber intake but not the
gender difference in fruit or salt, indicating that dieting partly
explains why women eat less fat and more fiber but not why they
eat more fruit or less salt.
Beliefs in the importance of each aspect of food choice for
health were comparatively high overall, and significantly higher
in women than men, as reported in other studies (42,76). They
also showed striking international consistency, with significant
gender differences for each belief for almost every country in
the sample. Gender difference in beliefs in the importance of
healthy eating explained a substantial amount of the gender dif-
ferences in food choices, eliminating the (modest) gender differ-
ence in salt and considerably reducing women’s greater concern
over fat and fiber. This effect was independent of the dieting mo-
tivation because the dieting and health belief effects together ex-
plained more of the gender difference in food choices than either
one alone. These results are consistent with the view that part of
the reason that men report making less healthy food choices is
that health is a less important motivation to them in the food
domain.
It has been suggested that men feel more ambivalent about
healthy dietary behaviors (77,78), but if that were so, there
should have been a wider spread of belief scores in men than
women, which was not the case in this sample. Courtenay (20)
also argued that healthy eating might be stereotyped as a “femi-
nine” practice, and therefore men who are concerned with con-
firming their masculinity might not choose healthier foods even
if they knew that they should. Again this implies that not only
should mean belief scores be lower in men, but variance should
be greater. In our results, the confidence intervals (an index of
variability) for health beliefs were a similar width in men and
women, which is not consistent with the argument that there are
different subsets of men, some of whom subscribe to the stereo-
type and some who don’t. Our results suggested that men as a
group were simply less enthusiastic about the benefits of healthy
eating. It is also worth noting that although men’s ratings of the
importance to health of dietary behaviors were lower than
women’s, they were not low in absolute terms. Contrary to the
notion that men place little or no importance on healthy eating,
our findings suggest that men believe eating healthily is impor-
tant to some degree, just not as strongly as women. To that ex-
tent, men in this sample did not appear to be signing up to
Courtney’s (20) “socially prescribed male role,” in which men
are supposed to have no interest at all in diet and nutrition. How-
ever, it is worth noting that this sample is composed mainly of
well-educated and higher socioeconomic status (SES) respon-
dents and it is possible that the gender differences might be more
striking in a sample with a wider range of SES.
Gender and Healthy Food Choices 113
Despite a reduction in the size of the gender difference after
controlling for dieting and health beliefs, women were still more
likely than men to report healthier food choices, which means
that other factors must play a part. One possibility is that women
are not only more concerned about health considerations (as in-
dicated by their higher endorsement of the belief items) but are
also more likely to translate these attitudes to actions, perhaps
being in a more advanced “stage of change.” Women’s rela-
tionships with food is often represented as different from men’s
(42,79). Even at young ages, girls are more likely to be responsi-
ble for family meal preparation or to buy foods than are boys.
Young women may therefore use these skills to feed themselves
more healthily.
Smoking was controlled for in these analyses because it of-
ten differs by gender and it has been related to poorer dietary
habits (80,81). However, even taking smoking into account, the
relationships still held for all the food choice behaviors.
There are limitations to the generalizability of these find-
ings. The study participants were all university students and the
patterning of gender differences may be different in older or less
educated groups. Research on other populations will be needed
to address this point. The data were collected in a self-report sur-
vey, and therefore a social desirability bias may be present, as
well as inflated consistency among factors measured at the same
time point. On the other hand, this should not bias the assess-
ment of gender differences, which is the focus of this study. The
study was cross-sectional and this limits any causal interpreta-
tions about interrelations between beliefs and behaviors. The
range of food choice behaviors was very limited, and we had no
items on important behaviors such as limiting sugar or limiting
snacks. This is a drawback of using data from a wide-ranging
health behavior study, where the number of items relating to any
one behavior often has to be limited. We also lacked any specific
evidence that participants’ reports of food choices were re-
flected in real nutritional consequences. However, the aim of
this study was not to compare the diets of men and women as
such but to examine their food choice behaviors. Trying to avoid
fat or to eat higher fiber foods are the kinds of generalizations of
motives for choosing foods, which are often included in instru-
ments examining eating behaviors (67,68).
Notwithstanding these limitations, the results contribute to,
and take forward, the issue of gender differences in health be-
haviors. Many other studies have noted gender differences in as-
pects of diet, and as expected, our results confirm these findings.
In addition, they add to the basic finding by replicating it in 23
countries, many of which have rarely been featured in reports of
health behaviors in the literature. If women report healthier
practices (or at least attempt to make healthier choices) all over
the world, then this suggests that any explanations for the differ-
ences are more likely to reflect underlying biobehavioral char-
acteristics of men and women than local cultural effects. The
data on beliefs showed striking international consistency, indi-
cating that women across the world are more convinced than
men that dietary choices are important—a phenomenon that it-
self demands further investigation. Are women predisposed to
be more careful about what they eat, are they more risk averse in
114 Wardle et al.
this field, or are they more likely to be the nutritional gatekeep-
ers, taking on this role to an extent even before they take on re-
sponsibility for feeding a family?
Given the important role of health behaviors in many of the
major causes of morbidity and mortality and the well-estab-
lished gender differences in health, research into the causes of
gender differences in health behaviors seems to have been ne-
glected. In line with the Institute of Medicine’s report (6), we
hope that these results will stimulate other researchers to go be-
yond simply noting gender differences, to seeking to understand
them, and to researching ways of encouraging men to take as
much advantage as women of strategies to promote long-term
health.
REFERENCES
(1) World Health Organization. 1997–1999 World Health Statistics
Annual, Table 3. Retrieved July 16, 2002 from
http://www3.who.int/whosis/whsa.
(2) U.S. Department of Health and Human Services (DHHS):
Deaths: Final data for 1998 (DHHS Publication No. [PHS]
2000–1120). National Vital Statistics Reports. 2000, 48:11.
Hyattsville, MD: National Center for Health Statistics.
(3) American Cancer Society: Cancer Facts and Figures—2001.
Atlanta, GA: Author, 2001.
(4) Lawlor DA, Ebrahim S, Davey-Smith GD: Sex matters: Secular
and geographical trends in sex differences in coronary heart dis-
ease mortality. British Medical Journal. 2001, 323:541–545.
(5) Lowe LP, Greenland PH, Ruth KJ, et al.: Impact of major car-
diovascular disease risk factors, particularly in combination on
22-year mortality in women and men. Archives of Internal Med-
icine. 1998, 158:2007–2014.
(6) Institute of Medicine: Exploring the Biological Contributions
to Human Health: Does Sex Matter? Washington, DC: National
Academy Press, 2001.
(7) Kirkwood T: Time of Our Lives: The Science of Human Ageing.
Oxford: Oxford University Press, 1999.
(8) Perls T: Genetic and environmental influences on exceptional
longevity and the AGE nomogram. Annals of the New York
Academy of Sciences. 2002, 959:1–13.
(9) Waldron I: Contributions of changing gender differences in be-
havior and social roles to changing gender differences in mor-
tality. In Sabo D, Gordon D (eds), Men’s Health and Illness:
Gender, Power and the Body. Thousand Oaks, CA: Sage,
1995:163–179.
(10) Waldron I, Johnston S: Why do women live longer than men?
Part II. Journal of Human Stress. 1976, 3:19–29.
(11) McGinnis JM, Foege WH: Actual causes of death in the United
States. Journal of the American Medical Association. 1993,
270:2207–2212.
(12) Jackson R, Chambless L, Higgins M, et al.: Gender differences
in ischemic heart disease and risk factors in 46 communities:
An ecologic analysis. Cardiovascular Risk Factors. 1998,
7:43–54.
(13) Dean K: Self-care components of lifestyles: The importance of
gender, attitudes and the social situation. Social Science and
Medicine. 1989, 29:137–152.
(14) Kandrack MA, Grant KR, Segall A: Gender differences in
health related behaviour: Some unanswered questions. Social
Science and Medicine. 1991, 32:579–590.
Annals of Behavioral Medicine
(15) Liang W, Shediac-Rizkallah MC, Celentano DD, et al.: A popu-
lation-based study of age and gender differences in patterns of
health-related behaviors. American Journal of Preventive Med-
icine. 1999, 17:8–17.
(16) MacGregor ID, Balding JW, Regis D: Flossing behaviour in
English adolescents. Journal of Clinical Periodontology. 1998,
25:291–296.
(17) Wardle J, Steptoe A: The European health and behaviour sur-
vey: Rationale, methods and results from the United Kingdom.
Social Science and Medicine. 1991, 33:925–936.
(18) Courtenay WH, McCreary DR, Merighi JR: Gender and ethnic
differences in health beliefs and behaviors. Journal of Health
Psychology. 2002, 7:219–231.
(19) Whitfield KE, Weidner G, Clark R, et al.: Sociodemographic
diversity and behavioral medicine. Journal of Consulting and
Clinical Psychology. 2002, 70:463–481.
(20) Courtenay WH: Engendering health: A social constructionist
examination of men’s health beliefs and behaviors. Psychology
of Men and Masculinity. 2000, 1:4–15.
(21) Hazan AR, Glantz SA: Current trends in tobacco use on
prime-time fictional television. American Journal of Public
Health. 1995, 85:116–117.
(22) Hazan AR, Lipton HL, Glanz SA: Popular films do not reflect
current tobacco use. American Journal of Public Health. 1994,
84:998–1000.
(23) Signorielli N: Drinking, sex, and violence on television: The
Cultural Indicators perspective. Journal of Drug Education.
1987, 17:245–260.
(24) Wallack L, Breed W, Cruz J: Alcohol on prime-time television.
Journal of Studies on Alcohol. 1987, 48:33–38.
(25) Signorielli N: Health images on television. In Jackson LD,
Duffy BK (eds), Health Communication Research: A Guide to
Developments and Directions. 1998, 163–179.
(26) Neumark-Sztainer D, Story M, Resnick MD, et al.: Lessons
learned about adolescent nutrition from the Minnesota Adoles-
cent Health Survey. Journal of the American Dietetic Associa-
tion. 1998, 98:1449–1456.
(27) Shimakawa T, Sorlie P, Carpenter MA, et al.: Dietary intake pat-
terns and sociodemographic factors in the atherosclerosis risk and
communities study. Preventive Medicine. 1994, 23:769–780.
(28) Li R, Serdula M, Bland S, et al.: Trends in fruit and vegetable
consumption among adults in 16 US states: Behavioral Risk
Factor Surveillance System, 1990–1996. American Journal of
Public Health. 2000, 90:777–781.
(29) Beer-Borst S, Hercberg S, Morabia A, et al.: Dietary patterns in
six European populations: Results from EURALIM, a collabo-
rative European data harmonization and information campaign.
European Journal of Clinical Nutrition. 2000, 54:253–262.
(30) Patterson RE, Haines PS, Popkin BM: Health lifestyle patterns
of U.S. Adults. Preventive Medicine. 1994, 23:453–460.
(31) Liebman M, Cameron BA, Carson DK, et al.: Dietary fat reduction
behaviours in college students: Relationship to dieting status, gen-
der and key psychosocial variables. Appetite. 2001, 36:51–56.
(32) World Health Organization Health Policy for Children and Ad-
olescents. Health Policy for Children and Adolescents, Series
No. 1. Geneva, Switzerland: World Health Organization, 2000.
(33) Kennedy ET, Bowman SA, Powell R: Dietary-fat intake in the
US population. Journal of American College Nutrition. 1999,
18:207–12.
(34) Ministry of Agriculture, Fisheries and Food (UK): The dietary
and nutritional survey of British adults—Further analysis.
London: HMSO, 1994.
Volume 27, Number 2, 2004
(35) Subar AF, Thompson FE, Kipnis V, et al.: Comparative valida-
tion of the Block, Willett, and National Cancer Institute Food
Frequency Questionnaires. American Journal of Epidemiology.
2001, 154:1089–1099.
(36) Stephen AM, Sieber GM: Trends in individual fat consumption
in the UK 1900–1985. British Journal of Nutrition. 1994,
71:775–788.
(37) Troiano RP, Briefel RR, Carroll MD, et al.: Energy and fat intakes
of children and adolescents in the United States: Data from the
National Health and Nutrition Examination Surveys. American
Journal of Clinical Nutrition. 2000, 72:1343S–1353S.
(38) Zive MM, Nicklas TA, Busch EC, et al.: Marginal vitamin and
mineral intakes of young adults: The Bogalusa Heart Study.
Journal of Adolescent Health. 1996, 19:39–47.
(39) Furnham A, Kirkaldy B: Age and sex differences in health be-
liefs and behaviours. Psychological Reports. 1997, 80:63–66.
(40) Wardle J, Steptoe A, Bellisle F, et al.: Healthy dietary prac-
tices among young Europeans. Health Psychology. 1997,
16:443–450.
(41) Wardle J, Griffith, J: Socioeconomic status and weight control
practices in British adults. Journal of Epidemiology of Commu-
nity Health. 2001, 55:185–190.
(42) Fagerli RA, Wandel M: Gender differences in opinions and prac-
tices with regard to a “Healthy Diet.” Appetite. 1999, 32:171–190.
(43) Steptoe A, Pollard T, Wardle J: The development of a measure
of the motives underlying the selection of food: The Food
Choice Questionnaire. Appetite. 1995, 25:267–284.
(44) Steptoe A, Wardle J, Cui W, et al.: Trends in smoking, diet,
physical activity and attitudes to health in young adult Europe-
ans from 13 countries, 1990–2000. Preventive Medicine. 2002,
35:97–104.
(45) Wardle J, Solomons W: Naughty but nice: A laboratory study of
health information and food preferences. Health Psychology.
1994, 13:180–183.
(46) Bellisle F, Monneuse MO, Steptoe A, et al.: Weight concerns
and eating patterns: A survey of university students in Europe.
International Journal of Obesity. 1995, 19:723–730.
(47) French SA, Jeffrey RW: Consequences of dieting to lose
weight: Effects on physical and mental health. Health Psychol-
ogy. 1994, 13:195–212.
(48) Serdula MK, Collins ME, Williamson DF, et al.: Weight control
practices of U.S. adolescents and adults. Annals of Internal
Medicine. 1993, 119:6667–6671.
(49) Gibbons KL, Wertheim EH, Paxton SJ, et al.: Nutrient intake of
adolescents and its relationship to desire for thinness, weight
loss behaviours, and bulimic tendencies. Australian Journal of
Nutrition and Diet. 1995, 52:69–74.
(50) Jeffrey RW, Adlis SA, Forster JL: Prevalence of dieting among
working men and women: The healthy worker project. Health
Psychology. 1991, 10:274–281.
(51) Roos E, Lahelma E, Virtanen M, et al.: Gender, socioeconomic
status and family status as determinants of food behaviour. So-
cial Science and Medicine. 1998, 46:1519–1529.
(52) Afifi-Soweid RA, Kteily MB, Shediac-Rizkallah MC: Preoccu-
pation with weight and disordered eating behaviors of entering
students at a university in Lebanon. International Journal of
Eating Disorders. 2002, 32:52–57.
(53) Ford DA, Dolan BM, Evans C: Cultural factors in the eating
disorders: A study of body shape preferences of Arab students.
Journal of Psychosomatic Research. 1990, 34:501–507.
(54) Lee S, Leung T, Lee AM, et al.: Body dissatisfaction among
Chinese undergraduates and its implications for eating disor-
Gender and Healthy Food Choices 115
ders in Hong Kong. International Journal of Eating Disorders.
1996, 20:77–84.
(55) Kristal AK, Bowen DJ, Curry SJ, et al.: Nutrition knowledge,
attitudes and perceived norms as correlates of selecting low-fat
diets. Health Education Research. 1990, 5:467–477.
(56) Wardle J, Griffith J, Johnson F, et al.: Intentional weight control
and food choice habits in a national representative sample of
adults in the UK. International Journal of Obesity. 2000,
24:534–540.
(57) Bourdieu P: Distinction: A Social Critique of the Judgement of
Taste. London: Routledge & Kegan Paul, 1984.
(58) Caplan P: Why do people eat what they do? Approaches to food
and diet from a social science perspective. Clinical Child Psy-
chology and Psychiatry. 1996, 1:213–227.
(59) Chaiken S, Pliner P: Women, but not men, are what they eat:
The effect of meal size and gender on perceived femininity and
masculinity. Personality and Social Psychology Bulletin. 1987,
13:166–176.
(60) Pliner P, Chaiken S: Eating, social motive, and self-presentation
in women and men. Journal of Experimental Social Psychol-
ogy. 1990, 26:240–254.
(61) Rolls BJ, Fedoroff IC, Guthrie JF: Gender differences in eating
behavior and body weight regulation. Health Psychology. 1991,
10:133–142.
(62) Grogan SC, Bell R, Conner M: Eating sweet snacks: Gender dif-
ferences in attitudes and behaviour. Appetite. 1997, 28:19–31.
(63) Steptoe A, Wardle J: The European Health and Behaviour Sur-
vey: The development of an international study in health psy-
chology. Psychology and Health. 1996, 11:49–73.
(64) Steptoe A, Sanderman R, Wardle J: Stability and changes in
health behaviours in young adults over a one year period. Psy-
chology and Health. 1995, 10:155–69.
(65) Schoenborn, CA: Health promotion and disease prevention:
United States, 1985. Vital Health Statistics, Series 10, 163.
Washington, DC: U.S. Government Printing Office, 1988.
(66) Blaxter M: Health and Lifestyles. London: Tavistock, 1990.
(67) Kristal AR, Shattuck AL, Henry HJ: Patterns of dietary behav-
ior associated with selecting diets low in fat: Reliability and va-
lidity of a behavioral approach to dietary assessment. Journal of
the American Dietetic Association. 1990, 90:214–220.
(68) Johnson F, Wardle J, Griffith J: The Adolescent Food Habits
Checklist: Reliability and validity of a measure of healthy eat-
ing behaviour in adolescents. European Journal of Clinical Nu-
trition. 2002, 56:644–649.
(69) Baron RM, Kenny DA: The moderator–mediator variable dis-
tinction in social psychological research: Conceptual, strategic,
and statistical considerations. Journal of Personality and Social
Psychology. 1986, 51:1173–1182.
(70) Fung TT, Willett WC, Stampfer MJ, et al.: Dietary patterns and
the risk of coronary heart disease in women. Achives of Internal
Medicine. 2001, 161:1857–1862.
(71) Hu FB, Rimm EB, Stampfer MJ, et al.: Prospective study of ma-
jor dietary patterns and risk of coronary heart disease in men.
American Journal of Clinical Nutrition. 2000, 72:912–921.
(72) Liu S, Manson JE, Lee IM, et al.: Fruit and vegetable intake and
risk of cardiovascular disease: The Women’s Health Study.
American Journal of Clinical Nutrition. 2000, 72:922–928.
(73) vanDam RM, Rimm EB, Willett WC, et al.: Dietary patters and
risk for type 2 diabetes mellitus in U.S. men. Annals of Internal
Medicine. 2002, 136:201–209.
(74) Walda IC, Tabak C, Smit HA, et al.: Diet and 20-year chronic
obstructive pulmonary disease mortality in middle-aged men
116 Wardle et al.
from three European countries. European Journal of Clinical
Nutrition. 2002, 56:638–643.
(75) Kelder SH, Perry CL, Klepp KI, et al.: Longitudinal tracking of
adolescent smoking, physical activity, and food choice behav-
iors. American Journal of Public Health. 1994, 84:1121–1126.
(76) Zunft HJ, Friebe D, Seppelt B, et al.: Perceived benefits of
healthy eating among a nationally-representative sample of
adults in the European Union. European Journal of Clinical
Nutrition. 1997, 51:S41–S46.
(77) Povey R, Wellens B, Conner M: Attitudes towards following
meat, vegetarian and vegan diets: an examination of the role of
ambivalence. Appetite. 2001, 37:15–26.
Annals of Behavioral Medicine
(78) Sparks P, James R, Conner M, et al.: Ambivalence about
health-related behaviors: An exploration in the domain of food
choice. British Journal of Health Psychology. 2001, 6:53–68.
(79) Rappoport L, Peters GR, Downey R, et al.: Gender and age dif-
ferences in food cognition. Appetite. 1993, 10:33–52.
(80) Burke V, Milligan RAK, Beilin LJ, et al.: Clustering of
health-related behaviors among 18-year-old Australians. Pre-
ventive Medicine. 1997, 26:724–733.
(81) Ma J, Betts N, Hampl JS: Clustering of lifestyle behaviors: The
relationship between cigarette smoking, alcohol consumption,
and dietary intake. American Journal of Health Promotion.
2000, 15:107–117.