College women: eating behaviors and help-seeking preferences

Adolescence, Summer, 2002 by Anne M. Prouty, Howard O. Protinsky, Donna Canady

More and more women are struggling with eating disorders, and they are doing so at younger and younger ages, often starting at puberty (Bruch, 1981; Garner & Garfinkel, 1980; Mitchell & Eckert, 1987; Shisslak, Crago, Neal, & Swain, 1987; Striegel-Moore, 1995). Nattiv and Lynch (1994) estimated that one to three percent of the general Western female population meet formal criteria for disordered eating, with a higher prevalence among adolescent and young adult women. It has been documented that a majority of American college women exhibit at least a few of the symptoms of disordered eating (Hesse-Biber, 1989, Protinsky & Marek, 1997), and prevalence studies have shown that it is common for fifteen percent or more of college campus women to meet diagnostic criteria for anorexia nervosa or bulimia nervosa (e.g., Borgen & Corbin, 1987; Heatherton, Nichols, Mahamedi, & Keel, 1995; Hesse-Biber, 1989; Ratcliff, 1986). Other researchers have documented that the most common onset of eating disorders is around age 18 (Thelen, Mann, Pruitt, & Smith, 1987). In addition, some research has shown that belonging to certain groups (for example, participating in sports or being a member of a particular religious faith) can increase a woman's likelihood of disordered eating (Rosen, McKeag, Hough, & Curley, 1986; Sykes, Gross, & Subishin, 1986; Sykes, Leuser, Melia, & Gross, 1988). The present study was conducted to answer three questions: (1) What percentage of college women would be diagnosed with disordered eating at a large, mid-Atlantic university? (2) Would the women who were diagnosable differ from those who were not on a variety of demographic variables? (3) What types of support services for disordered eating do these college women want?

By doing this research, we hoped to gain insight not only into the numbers and demographic characteristics of college women with and without disordered eating, but also to learn what kinds of services students thought would be most helpful. Thereby, we would provide an up-to-date picture of both the state of the problem and its potential solution from the students' perspectives.

METHOD

Sample and Data Collection

A random sample of 10% (1,066) of the female students enrolled in a large, public, mid-Atlantic, rural university were chosen from university registrar enrollment data. The stratified sample was proportional to the actual number of women in each year of study, undergraduate freshmen through graduate students. Surveys were distributed via the mail, and due to the anonymous protocol, no follow-up contact was possible. Participation was voluntary, and the study was approved by the university's human subjects review committee.

Instruments

The demographic questionnaire asked women to identify their age, race, college class level, whether they were in a sorority or participated in organized sports, their religious background, their current level of religiosity and which religion (if any), their relationship status, and their parents' relationship statuses. This questionnaire also asked the following: if they were concerned about their weight or eating habits to whom would they most likely go for help; if they had an eating problem from whom would they seek help; and who would be their first and second choices for support if they chose to go to therapy.

Although no self-report instrument alone can diagnose an eating disorder, the abbreviated Eating Attitudes Test (EAT-26; Garner et al., 1982) is a widely used self-report screening measure for the symptoms of anorexia nervosa and bulimia nervosa and has been used in multiple studies in North America and Europe (e.g., Nelson, Hughes, Katz, & Searight, 1999; Ratchliff, 1986; Williams, Schaefer, Shisslak, Gronwaldt, & Comerci, 1986). The EAT-26 contains twenty-six items with six possible answers ranging from never (0) to always (3).

Reliability estimates were run to reassess the EAT-26 for its appropriateness for use with this more modern population of young women. Alpha coefficients for the present study's total sample were as follows: Factor I (diet) .88, Factor II (bulimia) = .78, Factor III (oral) = .60, which compared favorably against Garner et al.'s (1982) university control group (Factor I = .86, Factor II = .61, Factor III .46), and therefore support the use of the EAT-26 with the current sample of female college students.

The possible scores on the EAT-26 range from 0 to 53. A score of 20 on the EAT-26 (Garner et al., 1982) was used as the cutoff. Women in this study who scored 20 or above were placed in the "probable diagnosis" (Dx) category, and those below were placed in the "non-probable diagnosis" (Ndx) category.

RESULTS

Sample Characteristics

In all, 589 women responded to our survey out of 1,066 women who were contacted. Of the 589, 578 respondents' data were completed and used in the analyses, giving a return rate of 54.2%. Ten of the 578 respondents chose not to answer three or fewer of the demographic form questions. We chose to use their cases, but these omitted data account for the changing Ns in the results.