This cross-sectional research study was approved by the Institutional Review Board at the authors’ institution. All subjects provided consent to participate in the online survey for this study.
Sample
Subjects for this study were undergraduate women between the ages of 18-24 years old, enrolled at a university located in the northeast of the United States. Women were recruited via email and class announcements (during Fall 2019) from a variety of introductory classes. As an incentive to participate, subjects in select courses were given extra credit for completing the online survey for this study. The online survey was administered via Qualtrics.
Measures
The anonymous online survey included demographic and health-related questions such as self-reported age, gender, ethnicity, height, weight, and whether participants received prior diagnosis or treatment for an eating disorder. These questions were followed by reliable and valid instruments, as described in more detail below. It should be noted that questions were structured to be inclusive of those who were not raised by their biological parents. For example, a question set about mothers would be prefaced with the following statement: “Please note that ‘mother’ refers to any female guardian figure.” Additionally, all items on the survey, except for the demographic and health-related questions, were asked on a 5-point scale with responses of never, rarely, sometimes, usually, and always, and computed for average scale scores for each instrument.
Disordered Eating Symptoms
Disordered eating symptoms were measured using the 26-item Eating Attitudes Test (EAT-26) [20]. Although items are typically rated on a 6-point scale, the response categories were altered to a 5-point scale in order to maintain consistency among all items. This modification prevented scoring the instrument for clinical purposes; however, it made the survey user friendly for the subjects who participated [8]. Mean scores were calculated for each subscale (dieting, bulimia and food preoccupation, and oral control) and total score for the questionnaire with higher mean scores indicating greater disordered eating behaviors.
The Mother-Daughter Relationship
Two factors, regard (5 items) and responsibility (8 items), from the Parent Adult-Child Relationship Questionnaire (PACQ) was used to assess the adult daughter’s perception of the relationship she has with her mother [21]. In short, the “regard” factor relates to the positive emotions that are typically associated with attachment or care. Alternatively, the “responsibility” factor corresponds to negative emotions associated with the adult-child’s feelings of being responsible for their mother’s happiness. Items were averaged with higher scores indicating greater maternal regard and responsibility.
Weight and Appearance Related Conversations
Two of the seven subscales, emphasis on maternal weight (3 items) and appearance weight control (6 items), from the Childhood Family Mealtime Questionnaire (CFMQ) were used to assess conversations and behaviors surrounding appearance and weight loss in the household [22]. The questions were formatted retrospectively to gain a better understanding of the family environment when the participant was most likely living with her caregivers. Items were averaged on each scale with higher scale scores indicating greater emphasis on maternal weight and appearance weight control.
The final set of questions in the online survey were adapted from a theater-based obesity prevention program [23]. In contrast to the CFMQ, this set of questions was framed to examine family weight-related behaviors within six months prior to taking the survey. The Parent Weight Talk Scale included 7 items such as “In the last 6 months, how often have your parents: encouraged you to diet or lose weight, gone on a diet, talked about wanting to lose weight, or made comments about other people’s weight.” Items were averaged with higher mean scores indicating greater parent weight talk in the last six months.
Data Analysis
Survey data from Qualtrics was exported into a Microsoft Excel document (version 2020, Microsoft Corp., Redmond, WA) and reviewed for adequacy. Subjects who did not meet the age criteria were excluded, as were subjects who failed to answer questions from any of the four measures. Body Mass Index (BMI) was calculated based on self-reported heights and weights and then categorized into weight status groups as recommended by the Centers for Disease Control and Prevention [24]. Additionally, the scoring of scales was conducted in Microsoft Excel. Following preliminary analysis in Microsoft Excel, the data was exported into the Statistical Package for the Social Sciences (version 26, IBM Corp., Armonk, NY) for further analysis. Descriptive statistics were performed for all variables. Lastly, four separate multiple regression analyses were performed to determine predictors of unhealthy dieting, bulimic behaviors, oral control, and disordered eating behaviors. The independent variables included in each of the models were as follows: Mother Regard, Mother Responsibility, Emphasis on Maternal Weight, Appearance Weight Control, Parent Weight Talk, BMI, and History of an Eating Disorder Diagnosis or Treatment. The dependent variables in each model were the subscales and total scale score from the EAT-26 instrument. To assess for multicollinearity of independent variables in the regression models, Pearson correlation analysis and variance inflation factors were examined. Standardized and unstandardized beta coefficients, standard errors, adjusted R-squared, and 95% confidence intervals were reported in each of the four models. Significant level was set at p < 0.05.