Schilder (1958) defined body image as "the three-dimensional mental representation that each person has of oneself." This image is continually constructed and determines a physical, psychological and social gestalt. Slade (1988) defines body image as "the picture we have in our mind of the size, outline and shape of our body and the feelings we have about these characteristics and parts that make them up." Body shape includes the following components: cognitive (thoughts and beliefs about the body), perceptual (tactile, kinaesthetic, proprioceptive, visual, olfactory, auditory), affective (feelings about one’s own body), behavioural and social. The development of body image is influenced by events affecting the body (accidents, illnesses, sexual abuse), relationships with others, self-esteem and socialization.

According to Gardner (1996), body image includes two components: one perceptual, which refers to estimation of size and appearance, and another attitudinal, which relates to feelings and attitudes towards one’s own body. Cash (1994) distinguishes three dimensions in attitudes towards the body: a) evaluation (e.g., body satisfaction, cognitive appreciation and evaluative thoughts that derive from discrepancies with the ideal self); b) cognitive-behavioural factors (e.g., schemata about one’s appearance, strength of internalized ideas about physical appearance), c) affect (e.g., situational dysphoria about body image).

Body distortion refers to the discrepancy between perception of and beliefs about one’s figure and the observable reality. Body dissatisfaction refers to discomfort with one’s own body. Garner and Garfinkel(1981) point out that alterations in body image involve two aspects that are related but different: perceptual distortion of the figure that leads to overestimation of the parts of the body, and the cognitive-affective alteration associated with dissatisfaction and worry with regard to one’s figure.

Body image disorder includes perceptual, affective and cognitive aspects, and is defined as an exaggerated preoccupation with some imaginary or overestimated defect in physical appearance. This leads to having a low opinion of one’s own appearance, worrying excessively about the opinion of others and believing that one is inadequate, and that one cannot be loved due to one’s physical appearance. Thus, people with body image disorders will tend to "hide" their bodies, submit them to dieting and exercise and avoid social relationships. As Hutchinson (1982) points out, body image refers to the description of the body where body, mind and culture meet. It is the space that brings together thoughts, feelings, perceptions, attitudes, values and judgements about our body.

Preoccupation with weight in the female population should not be considered a body image disorder, since western culture encourages women to be uncomfortable with their figure, resulting in the development of a discredited body image in a very high percentage of women, some of whom will have a body image disorder. Cash and Henry (1995) found that 48% of adult women evaluated their appearance negatively, 63% were dissatisfied with their weight and 495 were worried about being overweight. Distortion of body figure is only a symptom of body image, which is a multidimensional problem involving perceptual, attitudinal and behavioural alterations (Garfinkel and Garner, 1982; Cash and Brown, 1987).

The magnitude of the discrepancy between self-perception of one’s body and internalized ideas about it are associated with body dissatisfaction and eating disorders (ED) (Jacobi and Cash, 1994; Strauman and Glenberg, 1994; Cash and Szymanski, 1995). Bruch (1962) stressed that in patients with eating disorders, their perception of their own body does not coincide with reality. Cash and Brown (1987) and Cooper and Taylor (1988) found overestimation of figure size in patients with ED, especially those with bulimia nervosa, though such overestimation was also found, to a lesser extent, in the control group, which is attributed by Rodin, Silberstein and Striegel-Moore (1985) to the influence of culture. Patients with ED present greater discrepancies with the ideal self than control patients, and place excessive emphasis on weight and appearance in self-judgements (Cash and Deagle, 1997). In Spain, several studies show the incidence of eating disorders and body dissatisfaction in young people (Toro, Castro, García, Pérez and Cuesta, 1989; Raich, Deus, Muñoz, Pérez and Requena, 1991aa, 1991b; Raich et al., 1992; Martínez, Toro, Salamero and Blecua, 1993; Rodríguez, 1997; De Gracia, Marcó, Fernández and Juan, 1999; Toledo et al., 1999)

The objective of this study is to detect degree of body dissatisfaction in relation to body image and ED symptoms in a non-representative sample of students in the Basque Country. The present work is an exploratory epidemiological study, with statistical significance, that forms part of a more extensive study currently in progress.

METHOD

The sample was of an intentional non-probabilistic type, and was taken from the populations of four municipalities in Guipúzcoa, a province of the Basque Country (San Sebastián, Hernani, Lasarte and Zarauz). The choice of municipalities was incidental, and the sample was obtained by the cluster sampling method. The number of schools from which the sample was taken was 5, of which 4 were public and 1 private (that of San Sebastián). In two of the schools (San Sebastián and Lasarte) the questionnaires were filled out in Spanish, and in the other three (the two in Hernani and the one in Zarauz) in Euskera (the Basque language). The total sample was made up of 1023 subjects and the definitive one, after elimination of those subjects that did not answer all the items, comprised 969 subjects: 508 girls and 461 boys. The sample that filled out the BDS (EDI-2) questionnaire and was evaluated reliably in the Body Mass Index (BMI) was made up of 330 girls and 336 boys.

The first step was to make contact with the schools and inform of them of the study’s objectives. After obtaining authorization from the school, the parents and the pupils, the questionnaire application sessions were arranged. The questionnaires were distributed individually in the classroom by members of the research team, after they had given an explanation of their content and cleared up any doubts. The researchers remained in the classroom, dealt with any problems arising and checked that the questionnaires had been filled out correctly as they were handed in. The order of administration of the questionnaires was as follows: sociodemographic variables, EAT, BSQ, and BDS of the EDI-2.

We first carried out descriptive analyses of some sociodemographic variables and scores in the tests used. We studied the internal consistency of each test and the correlations between them. Subsequently, we carried out variance and covariance analyses, chi-squared, correlations and t-tests for comparing different groupings of the sample.

The internal consistency index of the BSQ, obtained by means of Cronbach’s alpha coefficient, was α =.97 for girls, and α =.96 for boys, the same as that found by Mora and Raich (1993) with students, and similar (.93) to that of the instrument’s authors (Cooper et al., 1987). As far as the EAT is concerned, Cronbach’s alpha for the girls was α =.87, and for the boys, α =.79. The alpha coefficient obtained with the BDS of the EDI-2 was, for girls, α =.89 , and for boys, α =.87. Convergent validity was measured by correlating total scores in the BDS with those of the EAT. For girls, a correlation of r = .67 (p = .000) was obtained, and for boys, r =.58 (p= .000) (the authors of the questionnaire obtained .61 with students and .35 with bulimia patients, while Mora and Reich obtained .71). We also correlated the BSQ with the BDS subscale of the EDI-2, obtaining a correlation for girls of r = .64 (p = .000) and for boys, of r = .44 (p = .000). Cooper et al. (1987) obtained .66 with bulimia patients, and Mora and Raich (1993) obtained .78. Finally, on correlating the EAT with the BDS subscale of the EDI-2, we found r = .35 (p = .000) in girls and r = .44 (p = .000) in boys. Gross, Rosen, Leitenberg and Willmuth (1986) found .50.

The distribution of the sample by schools and ages can be seen in Tables 1 and 2.

With regard to the percentages of pathological cases in the EAT and the different categories of the BSQ, we can see in Table 3 that 7.15 of girls and 2.4% of boys present ED symptoms, but if we lower the cut-off point to 27 in the boys, due to the elimination of item 23, we find 20 (4.35%) cases with ED symptoms. Table 4 shows that 32.5% of the girls and 8.9% of the boys are worried about body image (BSQ>80), with moderate or extreme levels of concern in 16.8% of the girls and 4.4% of the boys.

The differences in the BSQ and EAT categories between boys and girls are statistically significant: EAT, with cut-off point at 30, chi-squared = 11.570, df 1, p = .001; with the cut-off point at 29 the results did not differ. BSQ, chi-squared = 81.708, df 3, p = .000. On comparing the sexes in two categories of the BSQ (<111/>110), no concern/mild concern against moderate/extreme concern, we found a significant difference between the sexes, chi-squared = 38.29, df 1, p = .000. We observed the same on comparing the two groups in the means (eliminating item 23), as Table 5 shows. The EAT with the 40 items shows in girls a mean of 13.39 (s.d. = 12.13).

On comparing by sexes the possible "criterion items" of anorexia and bulimia nervosa, we found significant differences, with higher scores in girls, for the items: 4 ("I am terrified about being overweight"), 15 ("I’m preoccupied with a desire to be slimmer"), 25 ("I’m preoccupied with the thought of having fat on my body"), 28 ("I take laxatives"), 37 ("I engage in dieting behaviour"); there were no significant differences in the items: 7 ("I have gone on eating binges"), 13 ("I vomit after I have eaten"), 30 ("I eat diet foods") and 40 ("I have the impulse to vomit after meals"). There were no differences in the items referring to binge eating and vomiting, while those relating to worry over one’s weight and the use of laxatives differed between the sexes, with girls scoring higher.

On comparing means in the BSQ between the groups constituted according to BMI category (normal, low weight, severe underweight and overweight), we found significant differences between the groups in girls (see Table 7). On carrying out multiple comparisons with the Scheffé test we found that in the BSQ there were significant differences for the normal weight group compared to those with low weight (p = .000), severe underweight (p = .000) and overweight (p= .000). The highest score was in overweight, followed by normal weight. Significant differences also appeared between low weight and overweight (p = .033), and between overweight and severe underweight (p = .002). In the boys we also found significant differences between groups (see Table 7): between normal BMI and low weight (p = .021), between low weight and overweight (p = .024) and between overweight and severe underweight (p = .043). In both boys and girls, the higher the BMI, the greater the body dissatisfaction.

On analyzing the relationship between BMI and body image dissatisfaction evaluated with the BDS (EDI-2), we found results similar to those found with the BSQ: in girls, F(3, 329) = 11.79, p = .000, and in boys, F(3, 332) = 6.935, p = .000. The differences are between the following groups: in girls, between normal weight and low weight (p = .002), normal weight and severe underweight (p = .000), and between overweight and severe underweight (p = .003); in boys, between normal weight and low weight (p = .002), and between low weight and overweight (p = .019). In both sexes we found that the higher the BMI, the greater the body dissatisfaction. On carrying out the same analyses with the EAT we found that there was no relationship between BMI and EAT score.

In order to determine whether BMI influenced EAT, BSQ and BDS (EDI-2), we carried out covariance analyses separately for each scale (EAT, BSQ and BDS of the EDI-2), with sex as independent variable and BMI as covariable. After confirming that the interaction effect "sex X covariable" was non-significant and eliminating the interaction term, the BMI was found to be significantly related to the BSQ and the BDS (EDI-2). After adjusting for the covariable BMI, we found that there were no significant differences in the EAT between sexes. In the BSQ and BDS (EDI-2), after controlling the BMI, we continued to find significant differences between the sexes: BSQ, F(1, 662) = 158.84, p = .000 and an effect size of .194; BDS (EDI-2), F (1, 662) = 96.09, p = .000 and an effect size of .127. We also checked correlations between BSQ, EAT and BDS (EDI-2) with BMI, finding a positive correlation between BMI and BSQ (r = .21, p = .000) and between BMI and the BDS of EDI-2 (r = .21, p = .000).

On comparing the EAT, BSQ and BDS (EDI-2) scores between the different schools, we found that, in boys, there were significant differences between schools, F(3, 332) = 11.181, p = .000; body dissatisfaction was significantly greater in the Hernani schools compared to the San Sebastián school (p = .000 and p = .003). In the Lasarte school this questionnaire was not administered. In girls, we also found significant differences for the BSQ, F(4, 503) = 3.256, p = .012, but on applying the post hoc Scheffé test we found only a tendency to significance (p = .063) in the difference between the San Sebastián school (with lower score in the BSQ) and one of the Hernani schools.

On comparing the means for EAT, BSQ and BDS (EDI-2) between ages, separating the sexes, we found that, in girls, there are statistically significant differences in the BSQ, F(7, 500) = 7.313, p = .000. On carrying out the Scheffé post hoc test we found that these differences are between subjects aged 11 and those aged 15 (p = .03), 16 (p = .001), 17 (p = .001) and 18 (p = .02), and between those aged 12 and those aged 16 (p = .01) and 17 (p = .01). As age increases, so does concern over appearance. In the BDS (EDI-2) we also found statistically significant differences, F(7, 322) = 3.599, p = .001; however, on carrying out the post hoc Scheffé test we found only a tendency to significance between those aged 11 and 18 (p = .055), with greater worry at age 18. In boys we found statistically significant differences in the BDS (EDI-2), F(7, 328) = 2.530, p = .01, but on carrying out the post hoc Scheffé test we did not find significant differences.

On studying the scores in the three tests, separating the sexes and comparing the different ages by pairs, following chronological order (10-11, 11-12, 12-13, etc.), we found that in boys there were significant differences between those aged 13 and those aged 14, with higher scores in the 13-year-olds in the EAT, BSQ and BDS (EDI-2). In the girls, in contrast, we found only one significant difference in the EAT: they scored higher at age 14 (the highest mean of all ages) than at age 13 (the lowest mean of all ages); we also found a tendency to significance in the BSQ (see Table 8).

On comparing the boys aged 14 and 15, significant differences appear in the BDS (EDI-2) (t = 2.600, df 50.03, p = .012); the lowest mean of all ages is for 14-year-olds and the highest for those aged 15. In girls, the difference appears between age 11 and 12, with 12-year-olds scoring higher in the BSQ than 11-year-olds (t = 2.184, df 81, p = .032). In the rest of the age pairs no significant differences were found.

Finally, we made correlations between age and EAT, BSQ and BDS (EDI-2), finding, in girls, positive correlations between age and BSQ (r = 2.94, p = .000) and between age and BDS (EDI-2) (r = 2.63, p = .000), that is, as age increases, so does worry about physical appearance. In boys, on the other hand, we found only a negative correlation between age and EAT (r = -1.47, p = .002), that is, the younger the boys, the more inappropriate the eating behaviour.

The reliability and validity of the scales obtained, similar to that of other studies, shows that they are suitable instruments for carrying out cross-cultural studies, though disparity in the cut-off points makes it difficult to make comparisons between them.

The results of this study show a percentage of cases with anomalous eating behaviours, susceptible to constituting an ED (7.1% of the girls and 2.4% of the boys), similar to that found by Raich et al. (1991b) and lower than those found by Carbajo, Canals, Fernández-Ballart and Doménech (1995) and by Santonastaso et al. (1996) (11%). Toro et al. (1989) found a percentage higher than ours in girls (9.8%) and lower in boys (1.2%). ED symptoms are more frequent in girls, as was found in other studies carried out in Spain (Toro et al., 1989; Raich et al., 1991b; Carbajo et al., 1995; Toledo et al., 1999) and in other countries (Rosen, Silberg and Gross, 1988; Ledoux, Choquet and Flament, 1991; Raich et al., 1992; Wood, Waller, Miller and Slade, 1992; Buddeberg-Fisher, Bernet, Sieber, Schmid and Buddeberg, 1996). Likewise, the mean scores in the EAT in our study are higher in girls than in boys.

The means found in the present work in the EAT (12.41 (s.d. = 11.95) in girls and 11.06 (s.d. = 8.85) in boys) are lower than those of other studies (Garner and Garfinkel, 1979; Leichner, Arnett, Rallo, Srikameswaran and Vulcano, 1986; Raich et al., 1991b; Carbajo et al., 1995; Santonastaso et al., 1996). However, they are somewhat higher than those of the study by Gross et al. (1986), who found a mean of 11.5 (s.d. = 7.5) in girls. In the study by Toro et al. (1989), the mean in girls (16.1) is higher than that of our study, but that of boys (11.4) is similar to ours. Age of subjects in the studies by Toro et al. (1989) and Leichner et al. (1986) is similar to those of our sample, so that the differences are probably due to cultural factors.

Girls scored significantly higher than boys in the EAT items referring to worrying about one’s weight and figure, which influences inappropriate eating behaviours. These results are similar to those of Raich et al. (1991b), though these authors found significant differences between the sexes in all the "criterion items" except item 28 (use of laxatives).

By ages, mean scores in the EAT are highest at age 14 in girls and age 11 in boys. In positive percentages of EAT, for girls this is highest at age 16 (10.8%) and for boys at age 11 (5.3%). These data show that the age at which there is most possibility of presenting an eating disorder is earlier in boys - a fact that should be taken into account in preventive programs.

In the BSQ, our means (73.46; s.d. = 35.63) in girls are lower than those of the studies by Cooper et al. (1987) (mean = 81.5; s.d. = 28.4) and Bunnell, Cooper, Hertz and Shenker (1992) (mean = 91.8; s.d. = 36.2), and quite similar to those of the study by Munford, Whitehouse and Platts (1991) with students from the Far East (mean = 70.5; s.d. = 34.2) and Caucasian students (mean = 70.5; s.d. = 31.6). The comparison with the results of Raich et al. (1991a) is not reliable because they used fewer items, though it is quite probable that, despite this bias, our means are lower. With regard to the BDS (EDI-2), our means are similar to those found by Garner (1998) in young Spaniards.

Concern about one’s appearance was found in 32.5% of the girls and 8.9% of the boys, with moderate or extreme values in 16.8% of girls and 4.4% of boys. This percentage is lower than that found by Cash and Henry (1995), in whose study 48% of women valued their overall appearance negatively, 46% were dissatisfied with or worried about their weight and 51% had similar negative feelings about their body shape. De Gracia et al. (1999) found that 53.8% of adolescents reported worry over their body image. We found less preoccupation in young people from the Basque Country compared to other samples; however, these results do not permit us to suggest that there are significant differences between cultures, as the age variable is not the same in all the samples, and cut-off points vary from study to study.

Girls are more worried about and dissatisfied with their physical appearance than boys. This fact confirms that social pressure for a slim body model continues to affect girls in a more powerful way. It may also reflect the positive relationship found in Spain between advertising, a feminine aesthetic model focused on slimness and dissatisfaction with body image (Toro et al., 1989; De Gracia et al., 1999).

In both sexes we found that the greater the subject’s weight, the greater the body dissatisfaction. This could be due to the culturally-imposed slimness model, and indicates that overweight is a risk factor for body dissatisfaction, and possibly for the appearance of ED.

The fact that the mean BSQ scores in girls increase with age, reaching their highest levels at age 18, appears to indicate that social pressure with regard to appearance increases with girls’ age. These results for girls coincide with those of Toro et al. (1989) and previous studies (Crisp, 1984; Davies and Furhan, 1986). In boys, on the other hand, we found no relationship between age and concern over appearance, though the age with highest worry levels is 13, after which they decrease, coinciding, perhaps with the physical transformations of puberty. It may be that, in boys, muscular development and the importance given to sport in our social environment make concern about physical appearance and diet less important. These differences show once more that it is necessary to study lower age ranges in boys than in girls, when the aim is to design preventive programs.

An aspect to be taken into account is the age at which there are significant increases in the questionnaire scores. Boys aged 13 have significantly higher scores than those aged 14 in all three questionnaires. Girls score significantly higher in the EAT at age 14 than at age 13, but concern over body image has already made a statistically significant leap at age 12, when worry is considerably greater than at age 11. Some authors have suggested that concern over body image may be a premorbid feature of the appearance of ED, since such concern favours anomalous eating behaviours (Striegel-Moore, Silberstein and Rodin, 1986). On the other hand, Cooper et al. (1987) argue that the BSQ should not be used as a measure of the detection of ED cases, but rather to evaluate an aspect of eating pathology. In our study, on using the BSQ and the EAT, we can observe that, in girls, first concern about appearance increases and subsequently there is an increase in anomalous eating behaviours, probably in an attempt to modify this unsatisfactory appearance. However, we should be cautious about this interpretation, as we found no relationship between age and the EAT. In the boys we found inappropriate eating behaviours to occur at an early age, whilst concern over appearance is not age-related.

With regard to the schools, the fact that dissatisfaction with body image was significantly greater in children from the two public schools in Hernani (Euskera), as against the private school in San Sebastián (Spanish), appears to suggest that social class and other cultural factors may be influencing body image; perhaps after the epidemiological study we can draw clearer conclusions about these differences, if they continue to be found with a larger sample.

Our results reaffirm the importance of the systematic study of eating behaviours and attitudes towards body image in pre-adolescents and adolescents, with a view to the prevention of pathologies with a high prevalence in our society.

References

View Table 1 : [in this window] [in a new window]

Table 1.Distribution of the sample by schools.

View Table 2 : [in this window] [in a new window]

Table 2. Distribution of the sample by age.

View Table 3 : [in this window] [in a new window]

Table 3. Percentage of cases in EAT categories.

View Table 4 : [in this window] [in a new window]

Table 4. Percentage of cases in BSQ categories.

View Table 5 : [in this window] [in a new window]

Table 5. Comparison by sex in EAT, BSQ and BDS (EDI-2).

View Table 6 : [in this window] [in a new window]

Table 6. Comparison by sex of "criterion items" of anorexia and bulimia nervosa of the EAT.

View Table 7 : [in this window] [in a new window]

Table 7. Comparison by sex in the BSQ by Body Mass Index categories.

View Table 8 : [in this window] [in a new window]

Table 8. Comparison by sex in EAT, BSQ and BDS (EDI-2) for ages 13-14.