di

Raeleen D’Agostino Mautner, Ph.D.

University of New Haven

Una nota ai miei lettori: Il mio campo di specializzazione e’ la ricerca transnazionale (quella che confronta i dati Americani e Italiani), sui tema della psicologia delle scienze educative, la psicologia della pedagogia/apprendimento, la psicologia della salute, behaviorismo, ecc.. mi piacerebbe mettermi in contatto con chi ha interessa simile, per poter realizzare una collaborazione ricercativa.

Potete scrivere a : Mautner@scsud.ctstateu.edu

Body image disturbance (BID) is related to eating disorders and other psychological maladies, and multiple factors appear to be related to it. It is especially prevelent in females, throughout junior high, high school, and college years. Prior to this study, cross-cultural research on BID focused on Western versus non-Western and Caucasian versus non-Caucasian comparisons. This is the first study that compares BID data from the US, Italy, and England. Also, none of the extant cross-cultural studies have used the necessary variety of measures , capable of assessing the multidimensionality of BID (Thompson, 1990). Such measures should include knowledge, affective, and behavioral components of BID.

Very few serious attempts have been made to replicate US findings regarding BID in other Western cultures. Nonetheless, European nations have generally relied on US research findings in most areas of psychology. The current investigation sought to discover if the findings of Stormer and Thompson’s (1996) multidimensional model for BID assessment could be generalized for use either in areas of further transnational investigation, or for enhancement of current prevention and treatment protocols in and out of clinical settings. Stormer and Thompson (1996) simultaneously examined the four strongest correlates of BID: societal values about appearance, physical appearance comparison, teasing history, and pubertal timing. Specifically, this study focused on BID in English and Italian female students, as well as in a in new US student group.

Two research questions were addressed. The first research question asked if there were differences across the three national samples in the relationships between BID and the four correlates mentioned above. The second research question more deeply probed specific relationships between each of the four predictors and FRS, the only BID measure to differentiate among the three groups.

North America and Europe are considered to be culturally similar "Western" nations (Triandis, 1994), and researchers sometimes erroneously assume that U.S. findings are valid for European countries (e.g., Mangweth, Pope Jr., & Hudson, 1995; Matacin, 1994; MuKay & McCloskey, 1996). Thus there is a notable paucity of research that compares the correlates of BID in the U.S. with those in European nations; perhaps an assumption of cultural similarity lessens the perceived value of such comparisons (Berry, Poortinga, Segall, & Dasen, 1995). The goal of cross-cultural psychology is to go beyond ethnocentric perspectives, and look for etics, or universals, in order to arrive at a clearer and more accurate understanding of behavior (Triandis, 1994). Also, comparisons made among Western countries help to clarify what aspects of these cultures promote thin ideals (Altabe, 1996). Without cross-cultural research it is impossible to differentiate among studies that report local, as opposed to more universal phenomena (Triandis, 1994).

Knowledge gained by comparing different cultural groups can lead to a clearer understanding of the etiology and symptoms of body image disturbance and eating disordered behavior (Altabe, 1996), and there is an urgent need for theoretical models that guide cross-culture researchers in this area (Matacin, 1994). Some studies (e.g., Furnham & Baguma, 1994) have compared single components of body image satisfaction between Western and non-Western societies, where standards of ideal beauty differ in an obvious way. But it is also useful to compare BID similarities across cultures, in order to establish differences. Differences embedded into a framework of similarities (such as between two similar cultures) assure that many of the rival hypotheses, such a response sets, are not relevant (Triandis, 1994).

Cross-national studies, in which data from relatively similar cultures are compared, help to further this goal. To date, cross-national comparisons of body image are relatively scarce. Furthermore, none of the extant studies has compared samples from different Western cultures on multiple correlates of body dissatisfaction in a single cross-cultural design. Additionally, the use of several assessment instruments, as was implemented in this study, is also advantageous in cross-cultural comparisons, because it offers evidence about their convergent validity (Triandis, 1994).

Body Image Disturbance, has been a priority for U.S. and European researchers in recent years. A similar pattern of body dissatisfaction and eating disordered behavior is also emerging in non-Western nations, as exposure to Western society’s body shape ideals increases (Apter, Abu Shah, Lancu, Abramovitch, Weizman & Tyano, 1994; Wade, Tiggemann, Heath, Abraham, Treloar & Martin, 1996).

Replication of research can be valuable in its own right. Borg and Gall (1989) have advocated replication studies in order to check the findings of a major or milestone study, or to check the validity of research findings across different populations. The present study addressed both of these issues. Sometimes U.S. studies fail to replicate in cross-cultural settings, not so much because of cultural differences, but because results may have differed (e.g., due to sampling error) if those same studies were repeated in the U.S. (Triandis, 1994). The current study used a new U.S. sample to estimate if the relationships of maturational status, negative verbal commentary, social comparison, and sociocultural factors with body image dissatisfaction, were the same as in Stormer and Thompson’s (1996) study. Results were then compared across two European university samples. This study also extends Matacin’s (1994) cross-cultural study, which did not examine the current variables for similarities and differences in body satisfaction between participants in the United States and Italy.

Methodology:

BMDP 1R was the software used to test the equality of regression equations (using standard regression analyses) across groups. Along with the standard regression results, this sofware also calculates an ANOVA for the grouping variable. Then, upon finding a difference for one of the criterion measures, a series of hierarchical regression analyses was performed to more fully examine the contributions of the individual predictors. Levels of obesity and self esteem were used as covariates, in order to concentrate on the 4 target variables..

College aged female volunteers from London, Padova, and Connecticut, completed a packet of surveys; one for each of the four predictors, and one for each of the 6 measures for the criterion variable, BID.

The solution to Question 1 could also be interpreted as a conditional analysis to Question 2. To address this question the BMDP program’s 1R procedure was performed. This program tests equality of regression lines across different groups; in this study, across the three national samples According to Tabachnick and Fidell (1996), the method of choice is standard regression when the researcher simply wants to assess the relationships among variables. The present study sought to determine if the relationship between BID variables would be different across three different national groups. BMDP 1R, which performs both a standard regression, as well as a test of the equality of regression equations across groups, was the most appropriate method for this design. The suggested ratio of 20 subjects per independent variable was met, with 297 subjects (with complete data) in the total sample.

In a standard regression model, all predictor variables (IVs) enter the analysis simultaneously, and each regression weight is calculated according to its unique contribution to the variance of the criterion variable (DV), as if it had entered the regression after all others had been entered (Tabachnick & Fidell, 1996). Consequently, an IV may appear to contribute very little to the equation when some of its shared variance with the DV is parsed out by the other IVs. Thus it is important to consider both the full correlation as well as each IV’s unique contribution when interpreting the data. A review of the correlation matrices revealed little multicollinearity among IVs in the variables to be analyzed; nevertheless, squared semipartial correlations were calculated as another way to assess each IV’s importance. The squared semipartial correlation (sr²) for an IV represents the amount that R² would be reduced if that IV is deleted from the regression equation.

In addition to statistical significance of the regression analyses, the practical significance or effect sizes were also estimated (Cohen, 1988). In a regression equation, R² represents the effect size. According to Cohen (1988), an R² of .02 is considered a small effect, .13 a medium effect, and .26 a large effect.

Covariates. Thompson and Heinberg (1993) noted two factors that historically have explained a large amount of variance associated with BID: overall self-esteem and obesity. As in the study that this investigation sought to replicate, self-esteem and obesity were treated as covariates in order to determine the unique contributions of the four primary explanations of body image disturbance.

Self-Esteem. The Rosenberg Self-Esteem Inventory (Rosenberg, 1965), has a test-retest reliability of .85 and is often used as a covariate in body image studies (Thompson & Heinberg, 1993). This is a 10-item index to which respondents indicate their level of agreement on a four-point Likert-type scale, from strongly agree to strongly disagree. Sample items include "I feel that I have a number of good qualities," and "I am able to do things as well as most other people."

Obesity. Quetlet’s Index is a widely used measure of body mass index (BMI), which correlates highly with other measures of fatness (Garrow & Webster, 1985). One calculates BMI by weight in kilograms by height in meters squared. A BMI of 30 or greater is categorized as obese. In the US, where the metric system is not used, students will be asked to state their height in feet in inches and weight in pounds, along with other biographic information on the first page of the test packet. Computer software was used to perform the conversions. Information on participants’ height and weight was requested on the initial instruction/biographics page of the survey packet (Appendix A).

Social Comparison. Physical Appearance Comparison Scale (PACS) is a brief, five-item index designed to assess the tendency to compare oneself to others. Data collected from a sample resulted in a coefficient alpha of .78, and a test-retest coefficient of .72 (Thompson, Heinberg, & Tantleff, 1991). The PACS also correlates significantly with other measures, such as the EDI-DT; EDI-B; and EDI-BD. A slightly different version of the PACS was used in the 1996 Stormer and Thompson study, however this version used in the present study was selected based on Thompson’s suggestion (personal communication, September 3, 1996). A typical example of item content is: "At parties or other social events, I compare my physical appearance to the physical appearance of others".

Maturational Status. Age of Menarche (AM) attempts to pinpoint the subject’s exact age of menarche through several detailed questions and corresponding ratings of certainty of response (Stormer & Thompson, 1996). In the Stormer and Thompson (1996) study, data were not used if participants did not score 4 or 5 on certainty ratings (1=not certain at all; 5=absolutely certain). In the present study a subscale score was determined by calculating the average of two five-point Likert-type questions, pertaining to one’s pubertal timing.

Negative Verbal Commentary. Perception of Teasing Scale (POTS) weight subscale is a six-item questionnaire. Thompson et al. (1995) reported an alpha estimate of .83 with 130 college females. Along a three-point scale (1=never, 2=sometimes, 3=often), the questionnaire assesses the frequency of teasing or negative commentary by others regarding one’s weight that occurred during adolescence. It also inquires how upset the individual was at the time teasing occurred (1=not upset, 2=upset, 3=very upset). Though the measure has two subscales: Weight-Related teasing and Competency teasing. Sample items include "People made fun of you because you were heavy," and "People made jokes about you being too heavy."

Sociocultural Influence. Sociocultural Attitudes Toward Appearance Questionnaire (SATAQ) is a 14-item self-report measure that reflects awareness and internalization of societal attitudes about thinness and attractiveness (Heinberg et al., 1995). Alphas of .71 for the six-item Awareness scale, and .88 for the eight-item Internalization Scale, were reported. Sample items include the following: (Awareness) "It’s important for people to work hard on their figures/physiques if they want to succeed in today’s culture"; (Internalization)—"Music videos that show thin women make me wish I were thin." Although the internalization subscale has a stronger relationship to body image disturbance that does the awareness scale, both subscales were explored in the present analysis, as had been in the Stormer and Thompson (1996) study.

Body Image Disturbance. The Figure Rating Scale (FRS) is widely-used measure designed to assess satisfaction with one’s own size by calculating discrepancy scores for actual and ideal body shapes (Stunkard et al., 1983). A series of nine female schematic body shapes, which graduate in weight, are depicted. Test-retest reliability coefficients for the ratings of self-feel (i.e., how you look versus how you feel you look) and self-ideal, were .83 and .71 respectively, in a sample of 204 females (Thompson & Altabe, 1991). Participants are asked to choose a shape that represents their "ideal" figure; how they "feel" they appear; the figure that represents "society’s ideal" female figure. A discrepancy score is calculated by subtracting the ideal rating from the feel rating. In the present study the subscale score of FRS was calculated as the mean differences between feel-personal ideal, and feel-society’s ideal.

Physical Appearance State and Trait Anxiety Scale (PASTAS) is a 16-item inventory that measures anxiety regarding weight and non-weight-related aspects of appearance. Internal consistencies based on sample data were .82 for non-weight and .88 for weight (Reed, Thompson, Brannick & Sacco, 1991). Participants rate how anxious, tense, or nervous, they feel about weight and non-weight-related body parts, such as thighs, buttocks, waist, ears, lips, and feet. Ratings are made on a five-point response scale; 0= never, to 4 = almost always.

Multidimensional Body-Self Relations Questionnaire-Physical Appearance Subscale (MBSRQ-PAE) is a widely used seven-item Likert-type instrument that measures satisfaction with overall physical appearance. The Cronbach’s alpha coefficient was .88 for a female sample, aged 17-88 (Brown, Cash, & Mikulka, 1990). Participants rate their level of agreement, from 1 (definitely disagree) to 5 (definitely agree) with such items as the following: "Most people would consider me good-looking"; and "I like my looks just the way they are."

Body Image Avoidance Questionnaire (BIAQ) consists of 19 items that deal with the avoidance of situations (i.e., the behavioral component of BID) that are likely to provoke anxiety about physical appearance (e.g., weighing oneself, physical intimacy, etc.). After a "dry run" was conducted to check item comprehension and timing of test completion, I removed the two extreme ends of the scale ("never" and "always"), which made several items difficult to understand. Rosen, Srebnik, Saltzberg, and Wendt (1991) reported a Cronbach’s alpha of .89 and a test-retest reliability coefficient of .87 on a sample of 145 female undergraduates. This test purports to measure behaviors that reflect body image disturbance (Rosen et al.,1991). On a four-point scale (1=Rarely; 4=Usually), participants are asked to rate how items such as the following, applies to them: "I wear a special set of ‘fat clothes’"; "I do not go out socially if it involves eating."

Eating Disorder Inventory (EDI) is a widely used and psychometrically sound measure for the assessment of characteristics of eating disorders (Garner, 1991). Three scales from this instrument were used in this study, as they were in the Stormer and Thompson (1996) study:

1. The Body Dissatisfaction scale (EDI-BD) contains nine items pertaining to body site relevant to weight issues. Garner (1991) reported an alpha of above .80 for eating-disordered and normal control groups of various ages.
2. The Drive for Thinness scale (EDI-DT). Garner (1991) reported an internal consistency of .83 for eating disordered samples and .81 and above for several non-patient samples, for the DT scale. It measures fear of weight gain, restrictive tendencies and desire to lose weight.

The Bulimia scale (EDI-B) measures tendencies to binge and purge. Garner (1991) reported an internal consistency coefficient of .86, for a combined sample of eating disordered individuals; and .69-.82 for several samples of non-patient female controls.

Of the six separate measures of the dependent variable, BID, in this analyses,only one measure, the Figure Rating Scale (FRS) showed a significant difference among the regression equations for the three national samples. A"supervariable" was created as a way to summarize overall findings, as had been done in the original study by Stormer and Thompson (1996). Because of a high degree of overlap among DV measures, they calculated a composite score by transforming scale scores to z scores and then summing. The four IV measures of interest were entered into the analyses, along with Self-esteem and BMI, which, because of their established relationship to BID were considered as covariates. The two subscales of the SATAQ (SATAQ-A and SATAQ-I), an instrument used to determine the influences of social attitudes toward appearance, were examined separately, in order to examine the influence of one’s awareness versus one’s internalization of societal perspectives.

In addition to individual regression analyses for each subsample, the BMDP 1R procedure offers an analysis of regression planes across groups. The reduction of residuals due to grouping serves as a test of the equality of regression equations (Tabachnick & Fidell, 1996). The Figure Rating Scale (FRS) was the only measure to report a small but significant difference in the regression equations across the grouping variable country F (16, 273) =1.9, p<.02. According to Tabachnick and Fidell (1996), the eta squared formula (ì ²) is often used as an estimate of the effect size or strength of association between the variables. The eta² statistic for the ANOVA using group x FRS was .11, which can be interpreted as a small to medium effect. The results of this ANOVA are illustrated in Table 1.

_________________________________________

Source df Mean sq F

Reg over groups 16 1.79 1.9 *

Res within groups 273 0.95

__________________________________________

* p <.02

In the first of these analyses, the relationship between FRS, and all of the IVs were examined for the US sample. Results indicated a statistically significant model F (7, 117) =17.34, p <.001, explaining approximately 50 % of the variance in the Figure Rating Scale scores. According To Cohen (1988), this indicates a very large practical effect. The relationship between FRS and internalized social attitudes (SATAQ-I) was the only statistically significant relationship,other than that of body mass index (BMI). The squared semipartial correlation for SATAQ-I was .03. These results are summarized in Table 2.

Table 3 reports the results of the standard regression results for FRS regressed on all of the IVs for the English sample. A statistically significant F (7, 92) =7.65, p<.001, indicated that 33 % of the variance in the Figure Rating scale scores was explained by the overall model; a large effect according to Cohen’s (1988) guidelines. The only statistically significant relationship, beyond BMI and SE, was between FRS and SATAQ-I. The unique importance of this variable was estimated by calculating its squared semipartial correlation, .02.

The next simultaneous regression analysis shows the results of FRS regressed on the independent variables for the Italian sample. The overall regression equation explained approximately 25 percent of the variance in FRS, F (7, 64 = 4.11, p<.001, which, according to Cohen’s (1988) guidelines indicates a medium to large effect size. The only statistically significant relationship was between FRS and BMI. The corresponding squared semipartial correlation was .03. The other independent variables were not individually important in explaining the variance in the BID. Results are summarized in Table 4.

Rý adj=.25, F 7,64=4.11,p<.01

The following six analyses include the five remaining BID measures, and the created supervariable as the dependent variables, and regressed them on the four independent variables and two covariates. Because none of the analyses reported significance in equations across groups, results will only be reported for the entire group, with the three samples combined.

A supervariable, which combined the z-scores of the dependent variable measures, was created as a way to confirm what previous analyses in this study had indicated. Stormer and Thompson (1996) had originally created this variable for their investigation after performing a factor analysis of the BID scales, which showed in a high degree of overlap among the measures. In the present study an internal consistency score was calculated by finding the average interitem correlation (.61), then estimating the reliability of the 6-item z-score scale by using the general form of the Spearman-Brown Prophecy Formula (Gable & Wolf, 1993). The Supervariable data showed an internal consistency score of .90.

The analysis of variance of regression coefficients reported no significant differences in statistical equations across the three national samples, as had been expected on the basis of the other analyses, F (16, 273) = 1.51, p > .05. Overall, the standard regression model explained 53 % of the variance in the supervariable, F (7, 289) = 48, p< .01. In conformance with the other analyses, the two covariates shared a strong relationship with the dependent measure, followed by SATAQ-I (t = 5.02, p< .01). An equally strong relationship was found for the POTS (t = 2.59, p< .01) and SATAQ-A (t = 2.59, p< .01). Finally, a statistically significant relationship between PACS and the supervariable was uncovered (t = 2.18, p< .03). The squared semipartial correlations were as follows:.04 for SATAQ-I, and .01 for POTS, SATAQ-A, and PACS. As was found in the original study by Stormer and Thompson (1996), pubertal timing appeared to have no association with body image disturbance. These results are summarized in Table 5.

Results of the regression analyses used to evaluate the first research question uncovered two surprising findings. First, of the six criterion measures, only the Figure Rating Scale (Stunkard et al., 1983) differentiated among US, English, and Italian samples. Second, for each of the analyses, pubertal timing was the only predictor to apparently have no unique association to BID. All models using combined samples showed strong relationships among the predictor and criterion variables, as indicated by the total R² and the large effect sizes for each of the standard regression analyses

The second research question was addressed using BMDP 2R for a hierarchical regression analysis using the Figure Rating Scale (FRS) as the most appropriate criterion measure. The FRS was the chosen as the BID measure for this analysis because of its ability to detect group differences with respect to the predictor variables. Also, researchers have consistently found the FRS to be highly correlated to the other BID measures that were used in this study (Stormer & Thompson, 1996; personal communication with Joel Thompson, March, 1998). Four analyses were performed which regressed FRS on PACS, POTS, SATAQ (I and A), and AM; one for the combined sample, and one for each of the three national samples. The covariates, BMI and SE were stepped in first in all instances. The decision for the subsequent order of variable entry was based on prior literature findings.

Predictor variables for the following hierarchical analyses were entered in the order indicated in the table titles. Table 6 contains the results for the regression analysis using data from the three data samples combined. For the FRS, an overall adjusted R² of .35 was obtained, with internalization (SATAQ-I) and awareness of societal values (SATAQ-A) contributing unique variance (.08, .01 respectively).

R² step 1 = .18; @ R² step 2 = .09; @ R² step3 = .08; @ R² step4 = .01

Table 7 contains the results for the regression analysis using data from the US sample. For the FRS, an overall adjusted R² of .48 was obtained, with internalization of societal values (SATAQ-I) contributing unique variance (.17).

Table 7.

R² step1 = .24; @ R² step2 = .08; @ R² step3 = .17

Table 8 contains the results for the regression analysis using data from the English sample. For the FRS, an overall adjusted R² of .32 was obtained, with awareness of (SATAQ-A) and internalization of societal values (SATAQ-I) contributing unique variance (.07, .03 respectively).

R² step1 = .10; @ R² step2 = .14; @ R² step3 = .07; @ R² step4 = .03

Table 9 contains the results for the regression analysis using data from the Italian sample. For the FRS, an overall adjusted R² of .25 was obtained, with frequency of comparison (PACS) contributing unique variance (.05).

R² step1 = .22; @ R² step2 = .05; @ R² step3 = .05

There appears to be little difference in the relationships among variables in the original as compared to the present study. In general, the internalization of societal appearance ideals was the most important contributor to Body Image Disturbance, as was found by Stormer and Thompson (1996). Pubertal timing, as in the Stormer and Thompson investigation, did not play an important role in any of the analyses. Stormer and Thompson (1996) reasoned that developmental theories of BID are hard to evaluate in a cross-sectional research design. Although the timing of puberty showed some variation of importance across groups, it was not statistically different. Several of the participants for this study indicated they could not remember just when puberty occurred, or the events surrounding it. Perhaps as time goes on, the detrimental effects of a girl’s early physical development become confounded with other influential variables, such as society’s influence of beauty ideals, as seen in print and televised media. A suggestion that might enhance estimation of the strength of the effects of pubertal timing on the development of BID, is to examine the variable of pubertal timing alone or with other developmental factors such as teasing, during the time of life in which they occur.

The perception of teasing scale did manage to remain in most of the equations, as it did in the Stormer and Thompson (1996) study, although it is also a developmental measure of BID. Cash (1996) found that the trauma of having been teased as an adolescent often lasts into adulthood. The duration of this negative experience appears to have been reflected in the results of the present analyses.

Whereas cross-cultural researchers look for differences, clinical psychologists look to apply universals. In this study, because most findings were consistent, it makes intuitive sense to assume that a robust model for further transnational investigations—and possibly also for BID treatment and prevention protocols—has been supported. However, these findings do not necessarily generalize beyond the college female populations, and further investigations are needed with other subpopulations. In personal conversations with Italian and English colleagues, it was discovered that participants for this study were all of essentially the same SES. Hence, it is possible that the shared characteristic of being a college student or that of being female, could have overshadowed possible cultural differences that would have been found had the investigation included participants outside of the academic sphere.

Differences in eating disturbances and BID between the US and Europe have been historically documented (Mangweth et al., 1994; Raich, et al, 1992). The US has consistently shown higher rates. These differences were initially attributed to variable diagnostic practices rather than discernable cultural differences between the US and Europe (Nasser, 1997). This study departs from prior research, by finding no differences in BID across the three national groups, in 6 out of the 7 most frequently used BID measures. By ruling out inconsistencies in cross-cultural measurements, and time discrepancies across studies, the current investigation leaves researchers with three viable assumptions. Either the strong influence of transcultural media has succeeded in closing any substantial cultural gaps, or there really is no current difference in BID between the US and industrialized European nations, despite other cultural differences. A third possibility is that although research consistently reports higher incidences of BID in the US, transnational comparisons have never been uniform enough to test for statistical significance as it was in the present study. The present findings are important because they support a robust multidimensional BID assessment model. Prior to this study, no such model. Further, they indicate that extant US prevention and treatment protocols based on this model may possibly be useful when applied to Italian and English females. These protocols should include interventions for counteracting the internalization of multimedia images of thinness. They should also include the implementation of psychoeducational programs during adolescence and pre-adolescence as a preventive measure, to address teasing and negative effects of early pubertal timing.

REFERENCES

Altabe, M.N. (1996). Issues in the assessment and treatment of body image disturbance in culturally diverse populations. In J.K. Thompson (Ed.), Body image, eating disorders, and obesity (pp.129-147). Washington, DC: American Psychological Association.

Apter, A., Abu Shah, M., Iancu, I., Abramovitch, H., Weizman, A., Tyano, S. (1994). Cultural effects on eating attitudes in Israeli subpopulations and hospitalized anorectics. Genetic, Social, and General Psychology Monographs, 120, 83-99.

Berry, J.W., Poortinga, Y.H., Segall, M.H., & Dasen, P.R. (1995). Cross-cultural psychology: Research and applications. New York, NY: Cambridge University Press.

Borg, W.R., & Gall, M.D. (1989). Educational research: An introduction. New York: Longman.

Brown, T.A., Cash, T.F., & Mikulka, P.J. (1990). Attitudinal body-image Assessment: Factor analysis of the Body-Self Relations Questionnaire. Journal of Personality Assessment, 55, 135-144.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2^nd Ed.). Hillsdale, NJ: Lawrence Erlbaum.

Furnham, A., & Baguma, P. (1994). Cross-cultural differences in the evaluation of male and female body shapes. International Journal of Eating Disorders, 15, 81-89.

Gable, R.K., & Wolf, M.B. (1994). Instrument Development in the Affective Domain: Applications to School and Corporate Environments. Boston: Kluwer-Nijhoff Publishing.

Galgan, R., Mable, H., & Balance, W. (1987). The dimensionality of body image disturbance. Psychology—A Quarterly Journal of Human Behavior, 24, 41-43.

Garner, D.M. (1991). Eating Disorders Inventory-2: Professional Manual. Odessa, FL: Psychological Assessment Resources.

Heinberg, L.J. (1996). Theories of body image disturbance: Perceptual, developmental, and sociocultural factors. In J.K. Thompson (Ed.), Body image, eating disorders, and obesity (pp. 27-41). Washington D.C.: American Psychological Association.

Mangweth, B., Pope Jr., H.G., & Hudson, J.I. (1995). Bulimia nervosa in two cultures: A comparison of Austrian and American college students. International Journal of Eating Disorders, 17, 403-412.

Matacin, M.L. (1994). Body image and culture: Body satisfaction, perception, and ideals in Italian women and men. Unpublished doctoral dissertation, University of Cincinnati.

MuKai, T., & McCloskey, L.A. (1996). Eating attitudes among Japanese and American elementary schoolgirls. Journal of Cross Cultural Psychology, 27, 424-435.

Raich, R., Rosen, J., Deus, J., Perez, O., Requena, A., & Gross, J. (1992). Eating disorder symptoms among adolescents in the United States and Spain: A comparative study. International Journal of Eating Disorders, 11, 63-72.

Reed, D., Thompson, J.K., Brannick, M., & Sacco, W. (1991). Development and validation of the Physical Appearance State and Trait Anxiety Scale. Journal of Anxiety Disorders, 5, 323-332.

Rosen, J.C., Srebnik, D., Saltzberg, E., & Wendt, S. (1991). Development of a Body Image Avoidance Questionnaire. Psychological Assessment: A Journal of Consulting and Clinical Psychology, 3, 32-37.

Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.

Stormer, S.M., & Thompson, J.K. (1996). Explanations of body image disturbance: A test of maturational status, negative verbal commentary, social comparison, and sociocultural hypotheses. International Journal of Eating Disorders, 19, 193-202.

Stunkard, A.J., Sorenson, T., & Schulsinger, F. (1983). Use of the Danish adoption register for the study of obesity and thinness. In S.S. Kety, L.P. Rowland, R.L. Sidman, & S.W. Matthyssee (Eds.), The genetics of neurological and psychological disorders (pp. 115-120). New York: Raven Press.

Tabachnick, B.G., & Fidell, L.S. (1996). Using multivariate statistics (3^rd ed.). New York: HarperCollins Publishers, Inc..

Thompson, J.K. (1990). Body image disturbance: Assessment and treatment. Elmsford, NJ: Pergamon Press.

Thompson, J.K., & Heinberg, L.J. (1993). A preliminary test of two hypotheses of body image disturbance. International Journal of Eating Disorders, 14, 59-64.

Thompson, J.K., Heinberg, L.J., & Tantleff, S. (1991). The Physical Appearance Comparison Scale. The Behavior Therapist, 14, 174.

Tiggeman, M., & Rothblum, E.D. (1988). Gender differences in social consequences of perceived overweight in the United States and Australia. Sex Roles, 18, 78-86.

Triandis, H.C. (1994). Culture and social behavior. New York: McGraw Hill.

Wade, T., Tiggeman, M., Heath, A.C., Abraham, S., Treoloar, S.A., & Martin, N.G. (1996). Structure of disordered eating in a twin community sample. International Journal of Eating Disorders, 19, 63-71.