Abel L. Toriola, Bridget M. Dolan, Chris Evans and Olawafunhe Adetimole
This study investigates intergenerational and cross-cultural aspects of body weight
satisfaction in 103 female Yoruba students, 48 adult women in Western Nigeria and
68 Nigerian women living in Britain. The groups were compared on anthropometric
parameters of height, body weight and desired weight. It was found, in line with
predictions, that the students living in Nigeria had a lower current and desired body
mass index (BMI), than the adult Nigerian group. The students also showed a
discrepancy between their current and desired BMI which was greater than the adult
sample although, in conmast to Western populations, neither group showed a marked
overall dissatisfaction with their cuwent BMI. The 68 Nigerian wa n living in
Britain were matched pairwise for age, marital status and parity with 68 women from
the Nigerian sample. Despite a similar current BMI the sample living in Britain
showed significantly lower desired weight than their matched counterparts in Nigeria
and significantly greater discrepancy between their current and desired BMI. These
findings suggest that younger women in Nigeria may be mowing towards a Western
body image dissatisfaction which is already evident in their peers in Britain. The study
also supports the contention that culture exposure may cause immigrants fiom cultures
where thinness is not highly valued to adopt Western positiwe valuations of thinness.
Eating disorders are a paradigm area for the study of socio-cultural influences on
psychological problems. One major feature which has been attributed to
sociocultural factors is dissatisfaction with body size, specifically the drive to be
thinner (Gordon, 1990; Dolan and Gitzinger, 1994). In Westernized and
industrialized countries thinness has come to be associated with competency,
self-control, success, and beauty. However, anthropological studies have
indicated that this is not universal. Until recently in some non-Western
cultures, plumpness has been considered attractive and associated with
sexuality, fertility and caring (Hamadi, 1960; El Sarrag, 1968; Rudolfsky,
1972). This may reflect the positive correlation between body weight and
socio-economic class or wealth in many non-Western countries, the inverse
correlation to that found in Western societies (Sobal and Stunkard, 1989).
From studies of cross-cultural aspects of eating disorders (for review see
Dolan, 1991; Davis and Yager, 1992; Pate et al., 1992), it is clear that, although
supported by anecdotal reports and qualitative anthropological surveys, cultural
and generational differences in attitudes to eating and weight have not been
widely empirically and statistically tested. One study has investigated changing
socio-cultural pressures and dissatisfaction with body size. Furnham and Alibhai
(1983) compared how Kenyan Asian, British, and immigrant Kenyan British
women evaluated female body shapes. The Kenyan Asian group have a higher
rating to the larger figures either of the British groups, and the Kenyan British
group was even more extreme than the British group in their positive
evaluation of slimness. Further support of the view that exposure to
Westernized society can lead to adoption of Western attitudes towards eating
and body shape is provided by two samples of North African, Arab women one
attending university in Britain, the other in Cairo (Nasser, 1986). Women who
had lived in Britain for at least 4 years had more abnormal eating attitudes and
higher rates of eating problems than the students living in Cairo. The
prevalence of clinical bulimia nervosa was 12 per cent in the London group,
the highest rate reported for any group of women living in Britain (Meadows et
al. 1986). No clinical cases were found in the Cairo group, however there was a
relatively high percentage of abnormal eating attitudes in the Cairo group,
leading Nasser (1986) to argue that this was a reflection of the rapid socio-
economic changes in Egypt (see also Nasser, 1994a, b, c, for further evidence).
The present study was devised to investigate intergenerational and cross-
cultural aspects of eating attitudes and behaviours in Nigerian women in
Western Nigeria and in Britain. Nigeria was chosen as its indigenous culture has
been influenced by Western colonization and by more recent independence.
Nigeria is accepted to be the most populous country in Africa although the
precise population has been disputed and was put at 88.5 million in its 1991
census which suggested that 35 per cent of the population lived in urban areas.
The present state has arisen by combination of a number of distinct kingdoms
and states by British colonial rule in the late 19th century culminating in the
present boundaries by 1906. That colonial rule followed Portuguese, British and
other countries’ slave trading from the Niger basin since the 17th century. The
country became independent on 1 October, 1960 since when it has been
through phases of intense civil war and military rule. About 45 per cent of the
population are Muslim and 40 per cent Christian and there are about 250
distinct ethnic groups falling into a number of major groupings. Before the 1960s
the principal wealth of the country was agricultural although there is an
historical tradition of sophisticated use of mineral wealth in the Benin culture.
Between 1965 and 1980 it is estimated that the Nigerian Gross Domestic
Product rose by an average of 6.9 per cent per year but, following the collapse of
the oil boom, the GDP fell by an average of 1.1 per cent per year between 1980
and 1988. The results of all these factors create a remarkably interesting country
profoundly aware of its transcultural issues.
Attitudes towards body shape and weight have varied across the major
Nigerian ethnic groups, largely determined by people’s cultural, religious or
socio-economic backgrounds. However in traditional Yoruba society an
endomorphic build is positively regarded as a mark of affluence, healthful
living and motherhood. A Yoruba woman is regarded as attractive if she is
plump (although certainly not obese). However, recent research in Nigeria
suggests that appearance and weight control are becoming a major concern for
many Nigerian adolescents and that educated adolescent and adult Yoruba
women are coming to regard slimness positively. Salokun and Toriola (1985)
measured the association of perceived somatotype and behavioural stereotypes
among 160 Nigerian schoolboys and 140 schoolgirls. They found that both
groups preferred a linear physique and associated traits of positive social
character, athletic ability and leadership capability with ectomorphic build.
The Nigerian adolescents associated undesirable physical and psychosocial
characteristics with endomorphic build. The findings further supported a widely
noted tendency for adolescents to associate positive traits with the body shape
they most desire. These stereotypes suggest that body weight control could
affect social and psychological adjustment of adolescents. In another study of
Nigerian children Toriola and Igbokwe (1985) found that self-perceived
physique did not reliably correlate with anthropometric somatotype. Generally,
adolescent girls held more erroneous beliefs about their weight than boys. This
finding was further supported in a more recent survey in which as many as 43
per cent of 16-year-old Yoruba girls judged themselves as heavy although only
13 per cent were actually overweight (Toriola, 1989).
These Nigerian findings are concordant with the findings of Furnham and
Alibhai (1983) in Kenyans and of Nasser (1986) in Arab women and suggest
there have been changes from the earlier anecdotal and qualitative
anthropological descriptions of non-Western samples. The results suggest
that, over the last few decades, there may have been historical and generational
changes in the connotations of body image in non-Western countries exposed
to Western influence. We aimed to test this empirically in the Nigerian sample
by testing the general proposition that adult Nigerian women accepted a larger
body image than Nigerian students. We specifically tested the substantive
predictions that compared to adult Nigerian women Nigerian students will
have: a lower current Body Mass Index (BMI=weight/heightz); a lower
maximum and minimum adult BMI; a lower desired adult BMI; and a greater
mean preference for lower weight than their current weight (body weight
discrepancy: current BMI - desired BMI).
The work of Furnham and Alibhai (1983) and of Nasser (1986) also suggests
that non-Western women living in Western settings show more ‘Western’
positive evaluation of low body weight than do non-Western women in non-
Western settings. We aimed to test the prediction that, using a matched
sample, compared to Nigerian women living in Nigeria, Nigerian women in
Britain will have: a lower current BMI; a lower maximum and minimum adult
BMI; a lower desired adult BMI; and a greater body weight discrepancy.
The Nigerian student sample consisted of 103 female students attending
Obafemi Awolowo University in Ile-Ife, Western Nigeria. Eighty-three women
(81 per cent) were from Western Nigeria whilst 20 (19 per cent) were from
Eastern Nigeria. The majority (95 per cent) were born in Nigeria, 1 per cent
were born in another African country and 4 per cent were born in Europe. The
first language of 60 (58 per cent) of the students was Yoruba whilst 38 (37 per
cent) gave English as their first language, the remaining five subjects speaking
Hausa, Igbo or French at home. All subjects were fluent in English, which is the
teaching medium at the University. Ninety-two (89 per cent) of the women
were Christian, eight were Muslim and three gave no religion. The mean age of
the students was 23.1 years (SD=3.3; range 16-34). The majority of the women
(95 per cent) had never been married and only five women (5 per cent) had
The Nigerian adult sample consisted of 48 Yoruba women aged 21 to 50
years (mean 35.0, SD=6.5) all of these women were married. The majority of
the sample (98 per cent) had children. Only one woman had no children and
two women (4 per cent) had eight children. The mean number of children was
3.6 (SD=1.8). All women were employed as teachers, their average age on
leaving full-time education was 19.3 (SD=3.3; range 12-28 years). All of the
women considered their first language to be Yoruba, and normally spoke in
Yoruba in their home. All declared themselves fluent in English. One woman
also spoke Hausa. Forty-three (89.6 per cent) of the women were Christian, five
(10.4 per cent) were Muslim.
The sample of Nigerian women living in Britain consisted of 68 Yoruba
women aged 17 to 33 years (mean 24.3, SD=3.6), eight (11.7 per cent) of
whom were married. The women had come to Britain at an average age of 8.7
years old (range 3-28 years). Sixteen women (23.5 per cent) considered their
first language to be Yoruba, and their second language to be English. The
remaining 49 (72 per cent) said their first language was English. Three subjects
also spoke Egbo, one subject also spoke Hausa, and one spoke French. Fifty-one
(75 per cent) of the women said that English was the language usually spoken
in the home they currently lived in, the remainder usually spoke Yoruba at
home. The majority of the sample (80.1 per cent) did not have any children.
Two woman had three children; five had two children and six women had one
child. Forty-three (63.2 per cent) of the women were currently full-time
students, a further 22 (32 per cent) were in employment and three were
An age cut-off point was used to prune the adult and student samples from
Nigeria so as to produce non-overlapping samples suitable for an
intergenerational comparison. The cut-off of 25 years, which minimized data
loss and seemed meaningful in terms of our hypotheses about changing
attitudes, results in 80 students and 43 adult women for the comparison. For
the cross-cultural comparison the 68 Nigerian Yoruba women living in Britain
(Nigerian British group) were case matched with Nigerian women from the
total Nigerian adult and student sample of 151 women. Individuals were
matched for occupation (i.e. a worker or a student), age and parity. An exact
match on all three variables was obtained for 49 (72 per cent) women. It was
necessary to match the remaining British Nigerian women for age and parity,
but not occupation, with Nigerian students.
Questionnaires were given to 122 female Yoruba students who were
approached in the communal halls of residence and therefore attended a
wide variety of university courses. Of the 122 students, 19 (15.6 per cent)
either refused to participate or missed the majority of the questionnaire items,
leaving usable data from 103 women, a response rate of 84.4 per cent. The
researcher (AT) measured each woman’s current height and weight (using
domestic weighing scales).
Questionnaires were given to 75 adult Yoruba women who were primary
school teachers at three schools in Ile-Ife. All women had obtained a National
Certificate in Education which is the basic Nigerian teaching qualification.
Women were approached at the school following contact through a senior co.
ordinator of teachers. Forty-eight women (64 per cent) returned complete data
on the questionnaires. Current height and weight was measured as with the
Nigerian women in Britain were contacted by snowball sampling starting
from female undergraduates and postgraduate students registered at University
of London and Middlesex University who were contacted through the Students
Union and at the School of Oriental and African Studies. Women were
contacted personally by a fellow Nigerian student and asked to complete the
questionnaires. In all 75 women were asked to participate in the study of whom
68 (90.7 per cent) responded. All heights and weights were self-reported.
The questionnaire elicited information on demographic and anthropometric
parameters of body shape and cultural background. Respondents were asked
their weight, height, desired weight, maximum and minimum weights, age,
marital status, number of children, main language, religion, birthplace and
socio-economic indicators. These questions were developed from earlier British
studies (Dolan et al. 1990) and were intended to provide categorical predictor
variables for further analysesl. All subjects were also administered an English
language version of the 26-item version of the Eating Attitudes Test (EAT-26;
Gamer et al. 1982) and the Hospital Anxiety and Depression Scale (HAD;
Zigmond and Snaith, 1983). However, psychometric investigation of
questionnaire responses showed that the data would be unreliable for
generational or cross-cultural comparisons and these data are not reported
here (details are available from the authors on request).
Intergenerational results are shown in Table 1. The Nigerian students showed a
greater preference for lower weight than their current weight than adult
Nigerian women, although the 95 per cent confidence interval for the
difference between the two groups included zero.
A multivariate analysis of variance confirmed a strongly significant difference
between groups. All four BMI measures were strongly positively correlated in
both subsamples and the entire sample. Analyses of covariance confirmed that
controlling for any one variable removed the significant group differences on
any of the other variables owing to this strong intercorrelation. Although the
hypothesis that the groups differed on body weight dissatisfaction, was narrowly
supported, the within-subject discrepancies between current and desired BMI
were not statistically significantly different from zero in the entire sample (95
per cent confidence interval -0.22 to +0.76) nor in either group separately
(confidence intervals students, -0.65 to +0.58; adults, -0.02 to +1.63). This
indicates that neither group showed marked overall body weight dissatisfaction.
Results of the cross-cultural comparison are shown in Table 2. No difference
was found between the groups in their current, maximum or minimum BMI.
However, the desired BMI of the Nigerian women in Britain was significantly
lower than that of the Nigerian women in Nigeria and the discrepancy between
their current and desired BMI was significantly greater.
There are limitations to the study which afflict much cross-cultural survey
work. The samples are samples of convenience and results should not be
generalized beyond the sample frames used. Thus it would be unwise to expect
the results of the samples from Nigeria to apply to ethnically different Nigerian
populations such as Hausa, nor should the British data be expected to predict
likely parameters of older populations who came to Britain other than for
postgraduate university study. The data were incomplete and there were not
untypical levels of refusal to participate so, again, results should not be
extrapolated to entire populations including likely study non-participants. The
samples are also not particularly large though the confidence intervals on the
parameters shown in the tables indicate that moderate precision of estimation
was gained. It is, of course, possible that larger samples would reveal much more
precise estimates and allow the detection of small differences at conventional
statistical significance. Finally, the studies were cross-sectional, not longitudinal
and so cannot tell us anything about the changes that occur in individuals
during ageing, the increasing Westernization (or reactions to this) in their
population nor during migration whether temporary or permanent. However,
our samples are comparable in these respects to the majority of existing cross-
cultural research on eating disorders and body image. One other possible
criticism of the data is that the historical weight limits in the Nigerian samples
and all weights for the British sample were self-reported. We recognize that
verification of these weights would have been valuable, mainly we believe,
because they would indicate the accuracy of the body image perceptions of the
respondents. However, this study is of body image, not actual bodily
measurements and we believe that the quick assessment of body image
through self-reported weights and heights is of value even if more sophisticated
methods such as moving lights, distorting mirrors or camera etc, can provide
other, and often conflicting, perspectives on body image. We will report
elsewhere data from body silhouette repertory grids for these samples, which
throw further light on self-perceptions and stereotyping perceptions of body
images in general.
In the light of these limitations, what more do we know from this study? We
tested two sets of hypotheses, the first concerned intergenerational differences
between samples of Nigerian women; the second concerned differences
between samples of Nigerian women living in Nigeria and Britain after pairing
on certain social parameters.
Comparison of students and adult women in Nigeriu
All five measures of body image supported a difference between the Nigerian
students and the Nigerian adults in the directions predicted. However, the
measure of body image dissatisfaction, the discrepancy between current and
desired BMI showed a difference which was only narrowly significant with a
parametric confidence interval that actually embraced zero. Furthermore, the
dissatisfaction data were interesting in that neither the adult nor the student
groups showed a marked overall group mean dissatisfaction with their current
BMI. This is quite different from that data reported from a number of studies of
white, Western populations (e.g. Dolan et al. 1987).
Comparison of wornen in Nigeria and Britain
It was possible to match on age, marital status, and parity all the 68
Nigerian women living in Britain against women from the Nigerian sample.
Although the group in Nigeria included slightly more students than the
sample in Britain, all of the British sample had very recently been students,
so differences in occupation could also be essentially matched. Interestingly,
the resulting comparison sample did not show a significant difference on
current BMI. However, despite their similar current BMT, the Nigerian
women in Britain expressed a desired weight which was thinner than their
matched counterparts in Nigeria and a greater body image dissatisfaction as
reflected in their current/desired body image discrepancy. They also showed
a narrowly non-significant tendency to lower minimum adult BMI. The
sample of Nigerian women in Britain also showed a group mean body image
dissatisfaction as reflected by a significant and positive difference between
their current weights and the weights they said they would want to be, a
result, if perhaps not a magnitude of discrepancy, in line with the
discrepancy generally found in white, Western samples in other studies.
These findings are also in line with those of Furnham and Alibhai, (1983)
and Nasser (1986) in suggesting that culture exposure may cause immigrants
from cultures where thinness is not particularly highly valued, to adopt
Western positive valuations of thinness and a consequent tension between
the weight that valuation leads them to aspire to, and their perception of
their current weight. Though the findings do support this acculturation
hypothesis, they are less marked than the findings of either of those previous
Since the two groups of women were of similar age, marital status, parity and
current BMI this cross-cultural difference in desired body image is unlikely to
be related to these aspects of social situation and it appears possible that just
living within a British society, in which weight loss and thinness are highly
valued, may cause this intergroup difference. Of course, it is also possible that
the difference reflects other differences than whatever psychological, social and
politico-economic variables lead some Nigerian women to come to Britain and
others to stay in Nigeria. Only a longitudinal study looking at body image
before and after visits to Western settings can explore this.
In summary, our results suggest that exposure to Western cultural norms over
the last 20 years may have led to a detectable difference between generations of
Nigerian women in their body image perceptions and they further suggest that
younger women in Nigeria may be moving toward a Western body image
dissatisfaction and that their peers who have come to this country show that
dissatisfaction already. We believe these results strongly indicate a need for
international agencies and national governments to fund larger studies of these
Information about Nigeria was supplemented by articles from Encarta
(Microsoft Corporation and Funk & Wagnalls Corporation, 1993) and
Bookshelf (Microsoft Corporation and Columbia University Press, 1991/
1994). This work was partially supported by a grant from the Nuffield
Foundation. We are grateful to Ms Isabella Aboderin for her assistance with the
Table 1. Intergenerational comparisons
Students (n=80) Adults (n=43) Difference
Variable Mean 95% C.1. Mean 95% C.I. Mean 95% C.I. P
Maximum BMI 23.5 22.6 to 24.4 27.6 26.5 to 28.8 4.1 -5.6 to -2.6 <0.00005
Minimum BMI 20.8 19.9 to 21.6 24.0 23.2 to 24.9 3.2 -4.6 to -2.0 <0.00005
Current BMI 22.1 21.1 to 23.0 25.7 24.7 to 26.7 3.6 -5.0 to -2.2 <0.00005
Desired BMI 22.1 21.3 to 22.9 24.9 24.2 to 25.6 2.8 -3.9 to -1.7 <0.00005
BMI discrepancy -0.04 -0.65 to +0.58 0.81 -0.02 to 1.63 0.85 -1.86 to +0.17 0.023
BMI, Body Mass Index; C.I., confidence intervals.
Table 2. Cross-cultural comparisons
in Britain (n=68) in Nigeria (n=68) Difference
Variable Mean 95% C.I. Mean 95% C.I. Mean 95% C.I. P
Maximum BMI 24.2 23.1 to 25.3 23.7 22.8 to 24.6 0.5 -0.9 to +1.9 0.5
Minimum BMI 20.1 19.3 to 20.9 21.0 20.3 to 21.7 -0.9 -1.9 to -0.11 0.072
Current BMI 22.2 21.5 to 23.0 22.2 21.4 to 22.9 0.04 -1.05 to +1.13 0.91
Desired BMI 20.9 20.4 to 21.4 22.3 21.6 to 23.0 -1.40 -2.28 to -0.52 0.002
BMI discrepancy 1.9 1.35 to 2.5 - 0.19 -0.75 to +0.37 2.09 +1.31 to +2.9 <0.0005
BMI, Body Mass Index; C.I., confidence intervals.
DAVIS, C. and Yager, J. (1992). Transcultural aspects of eating disorders: A critical literature
review. Culture, Medicine and Psychiatry, 16, 377-394.
DOLAN, B. M. (1991). Cross-cultural aspects of anorexia nervosa and bulimia nervosa: A
review. Inteman'onal Journal of Eating Disorders, 10, 67-78.
DOLAN, B. M., Birtchnell, S. A. and Lacey, J. H. (1987). Body image distortion in non-eating
disordered women and men. Journal of Psychosomatic Research, 31, 513-520.
DOLAN, B. M. and Gitzinger, I. (Us) (1994). Why Women? Gender Issues in Eating Discrrders,
2nd edn. London: Athlone.
DOLAN, B., Lacey, J. H. and Evans, C. (1990). Eating behaviour and attitudes to weight and
shape in British women from three ethnic groups. British Joumal of Psychiatry, 157,
EL SARRAG, M. E. (1968). Psychiatry in Northern Sudan: A study in comparative psychiatry.
British Joumal of Psychiatry, 114, 946-948.
FURNHAM, A. and Alibhai, N. (1983). Cross-cultural differences in the perception of female
body shapes. Psychological Medicine, 13, 829-837.
GARNER, D. M., Olmsted, M. P., Bohr, Y. and Garfinkel, P. E. (1982). The Eating Attitudes
Test: Psychometric features and clinical correlates. Psychological Medicine, 12,871-878.
GORDON, R. A. (1990). Anorexia and Bulimia: Anatomy of a Social Epidemic. London: Basil
HAMADI, S. (1960). Temperament and Charactm of the Arabs. Twayne Publications.
MEADOWS, G. N., Palmer, R. L., Newball, E. U. M. and Kenrick, J. M. T. (1986). Eating
attitudes and disorder in young women: A general practice based survey. Psychological
Medicine, 16, 351-357.
MICROSOR CORPORATION and Columbia University Press (1991/1994). Microsoft
Bookshelf incorporating Concise Columbia Encyclopaedia. Redmond, WA: Microsoft
MICROSOR CORPORATION and Funk & Wagnalls Corporation (1993). Microsoft Encarta.
Redmond, WA: Microsoft Corporation.
NASSER, M. (1986). Comparative study of the prevalence of abnormal eating attitudes among
Arab female students of both London and Cairo Universities. P~y~hobgi~~l Medicine
NASSER, M. (1994a). 'Culture and research instruments': The performance of the Eating
Attitudes Test Questionnaire in the Arabic culture. Arab Journal of Psychiatry, 5, 1-1 1.
NASSER, M. (1994b). The psychometric properties of the Eating Attitudes Test in a non-
Western population. Social Psychiatiy and Psychiatric Epidemiology, 29, 8&94.
NASSER, M. (1994~). Screening for abnormal eating attitudes in a population of Egyptian
secondary school girls. Social Psychiatry and Psychiatric Ephiology, 29, 25-30.
PATE, J. E., Pumariega, A. J., Hester, C. and Gamer, D. M. (1992). Cross cultural patterns in
eating disorders: A review. Journal of the American Acudemy of Child and Adokscent
PSYC~Q, 3 1, 802-809.
RUDOLFSKY, B. (1972). The Unfashionable Human Body. New York: Doubleday.
SALOKUM, S. and Toriola, A. L. (1985). Perceived somatotype and stereotypes of physique
among Nigerian school children. loud of Psychology, 119, 587-594.
SOBAL, J. and Stunkard, A. J. (1989). Socio-economic status and obesity: a review of the
literature. Psychological Bulletin, 105, 260-275.
TORIOLA, A. L. (1989). Etiology of Nigerian adolescents’ perception of their weight. Perceptual
and Motor Skills, 69,466.
TORIOLA, A. L. and Igbokwe, N. U. (1985). Relationship between perceived physique and
somatotype characteristics of 10-18-year-old boys and girls. Perceptual and Motor Skills,
ZIGMOND, A. S. and Snaith, R. P. (1983). The Hospital Anxiety and Depression Scale. Acta
Psychtrica Scandinavica, 67, 361-370.
i did my best with the tables, but LJ would have none of it, so they are not really legible. sorry, folks, i tried. :(