The purpose of this study was to compare the Hebrew version of the 13-item EDE-Q-13 with that the Hebrew translation of the complete 28-item EDE-Q. The responses to the bingeing and purging items of the original questionnaire were restructured and included in the scoring of the short version. The structure of the scales was compared using CFA and the pattern of correlations between the total and subscale scores of both questionnaires was observed, as well as the pattern of correlations between EDE-Q-13 and EDE-Q total scores respectively with several scales measuring related variables.
Our results supported a five-factor model for the EDE-Q-13, with subscale scores for Eating Restraint, Body Dissatisfaction, Shape and Weight Overevaluation, Bingeing and Purging. This factor structure found for the EDE-Q-13 replicated the factor structure of the EDE-Q7 presented in Machado et al. (2020). It also replicated the structure of the 28-item EDE-Q Restraint and Eating Concern subscales, with the original Weight Concern and Shape Concern items combined into a single factor (Shape and Weight Overevaluation) as in many previous studies (Zohar et al., 2017; Hilbert et al., 2012). A major disadvantage of the full EDE-Q to date is that the open-ended structure of the response categories of the bingeing and purging items has prevented them from being included in scoring and data analyses. The recoding of these items and the inclusion of Bingeing and Purging subscales in the EDE-Q-13 score is therefore one major advantage of this short version of the questionnaire. Participants who scored above 1 on the Bingeing or Purging subscales scored higher on the EDE-Q-13 total scores excluding these two subscales, providing some validity for these subscales.
Another major advantage of the EDE-Q-13 is that it is short, user-friendly and parsimonious. Its total and subscale scores correlated strongly with those of the 28-item EDE-Q, so that significant information does not seem to be missed when it is used in lieu of the longer version, and it preserves the central features of the EDE-Q.
The EDE-Q-13 also showed convergent validity. Participants who reported higher levels of eating disorder symptoms tended to have significantly lower levels of positive body experiences, positive affect, positive eating, life satisfaction and social and emotional connection to others, and significantly higher levels of negative affect.
Our study has limitations. First, the version of the EDE-Q-13 used in this study was in Hebrew, so its psychometric properties should be verified in other languages. Second, this study was conducted with a community sample of predominantly female, single, educated community sample and may therefore not be generalizable to other populations. Third, although the use of the Likert format for the binge/purge items allows researchers and clinicians to incorporate behavioral frequency information within a continuous subscale or global scale score, it also obscures the actual frequency of binge eating/purging, such that it no longer becomes possible to determine whether participants reported “clinical” levels of these behaviors (i.e., 4x/month). It is also unclear whether adding scores for bingeing and purging behaviors may result in some respondents with these behaviors receiving higher scores on the total scale that may or may not be warranted. Further studies should investigate the validity of the EDE-Q-13 in clinical settings, its ability to accurately distinguish between cases and controls and its sensitivity to change.
The EDE-Q is widely used, but reporting on the full version is time-consuming, and presents significant participant burden. This may deter some respondents from completing the entire questionnaire. Researchers wishing to use a short version of the questionnaire have tried to decide which version is most useful (Machado et al., 2020). Although the shortest version suggested had only 7 items and had excellent psychometric properties, it omitted to ask about bingeing and purging. Thus, the EDE-Q-13 builds on the 7 items but adds the bingeing and purging items, which are very important in assessing EDs. The EDE-Q-13 makes self-report less burdensome in two distinct ways: it is more than 50% shorter, and it has a unified response scale which makes it easier on the respondent. Future research should try and validate this version of the EDE-Q in other languages and in clinical settings.
The EDE-Q is widely used, but reporting on the full version is time-consuming, and presents significant participant burden. This may deter some respondents from completing the entire questionnaire. Researchers wishing to use a short version of the questionnaire have tried to decide which version is most useful (Machado et al., 2020). Although the shortest version suggested had only 7 items and had excellent psychometric properties, it omitted to ask about bingeing and purging. Thus, the EDE-Q-13 builds on the 7 items but adds the bingeing and purging items, which are very important in assessing EDs. The EDE-Q-13 makes self-report less burdensome in two distinct ways: it is more than 50% shorter, and it has a unified response scale which makes it easier on the respondent. Future research should try and validate this version of the EDE-Q in other languages and in clinical settings. Conclusions: We found that the EDEQ-13 was reliable and showed convergent validity. It is possible to use this short and user-friendly self-report to screen for the presence of eating disorders in community samples for purposes of research and for initial clinical ascertainment.