Confirmatory factor and measurement invariance analyses of the Eating Disorder Examination Questionnaire in sexual minority men and women

Abstract

Objective The present study aimed to investigate the factor structure of the Eating Disorder Examination Questionnaire (EDE‐Q) in a large sample of cisgender sexual minority men and women, and subsequently, to evaluate measurement invariance by gender. Method The sample consisted of 962 sexual minority adult men (n = 479) and women (n = 483) who completed online self‐report surveys. Confirmatory factor analysis was conducted using two previously supported factor structures (Friborg et al.'s four‐factor model and Grilo et al.'s brief three‐factor model) as well as the original four‐factor structure of the EDE‐Q. Results Results indicated that the best fitting models were Friborg et al.'s four‐factor model (CFI = .974, RMSEA = .098, SRMR = .0 70) and Grilo et al.'s brief three‐factor model (CFI = .999, RMSEA = .049, SRMR = .017). The model fit of both factor structures were nearly identical when examined separately for men and women. The original four‐factor structure could not be supported in this sample. Measurement invariance analyses further indicated that the best fitting models were invariant by gender in sexual minority individuals. Internal consistency was adequate for all subscales of Friborg et al.'s and Grilo et al.'s models. Discussion The present study provides support for the use of the EDE‐Q in sexual minority men and women. Additionally, findings demonstrate that the EDE‐Q performs similarly in sexual minority men and women. Future research is needed to further evaluate measurement invariance of the EDE‐Q by sexual orientation, gender identity, and race.

Structural validity of the Eating Disorder Examination—Questionnaire: A systematic review

“…Aside from the study of Peterson et al (2020), which recruited a sample of transgender youth, there were three studies which explicitly stated genders other than male or female, with .1% (Rand‐Giovannetti et al, 2020), .25% (Jenkins & Davey, 2020), and 1.3% (Zickgraf et al, 2020) of the respective samples comprising other gender identities. Some were more evenly balanced across genders (e.g., Klimek et al, 2021), although predominantly non‐female samples tended to be purposively sampled (e.g., Peterson et al, 2020; Scharmer et al, 2020). Only a minority of studies (around one‐quarter) recruited from clinical settings, with nearly half of these including both clinical and non‐clinical participants.…”

“…The Weight Concern and Shape Concern subscales have been found to be highly correlated, and several studies in the current review generated latent structures through EFA whereby items of these subscales were considered under a “Weight and Shape Concern” subscale (e.g., Carey et al, 2019; Darcy et al, 2013; White et al, 2014). There was mixed evidence for the presence of a “Global” index of eating pathology, with several studies (Friborg et al, 2013; Klimek et al, 2021; Rand‐Giovannetti et al, 2020) generating conflicting findings regarding higher‐order models, suggesting that interpretation of the Global score might remain cautious, particularly in non‐female or ethnic minority groups (Goel, Burnette et al, 2022). Similarly, whilst some studies found that a one‐factor solution emerged from EFA (e.g., Peterson et al, 2020), others failed to find support using CFA (e.g., Calugi et al, 2017; Penelo et al, 2013).…”

Ascertaining an efficient eligibility cut-off for extended Medicare items for eating disorders

“…A 7-item version of the EDE-Q exists (EDE-Q7; 4 ), showing robust fit in confirmatory factor analyses, good internal consistency, and measurement invariance by sex, self-identified gender, and overweight status. 4–6 It is the only form of the EDE-Q to have a supported factor structure, including the full version. 7 The aim of this investigation is to identify an EDE-Q7 cut-off commensurate to 3 on the EDE-Q, offering the possibility of a less burdensome assessment for Medicare item eligibility.…”

Women of diverse sexual identities admit to eating disorder treatment with differential symptom severity but achieve similar clinical outcomes

“…Future studies should include assessments of other constructs (e.g., drive for muscularity) to ensure that symptoms are accurately captured and understood across sexual orientation groups. Relatedly, while the EDE-Q appears to be valid, reliable, and invariant across gender in SM samples (Klimek et al, 2021), it is important to evaluate factorial invariance across the spectrum of sexuality given possible conceptual distinctions in cognitive symptoms such as shape and weight concerns, particularly between SM and heterosexual women. Regarding sampling, future studies should recruit a more racially/ethnically diverse sample to allow for examination of unique ED risk profiles informed by intersecting identities.…”

“…While the EDE‐Q is valid and reliable, there is limited support for its original factor structure (Berg, Peterson, Frazier, & Crow, 2012; Rose, Vaewsorn, Rosselli‐Navarra, Wilson, & Weissman, 2013). Consistent with recent factor analytic and invariance testing in SM populations, this study included 11‐item shape and weight concern, 5‐item preoccupation and restriction, and 3‐item cognitive restraint subscales (Klimek et al, 2021). A fourth three‐item eating concern subscale was not used due to poor reliability ( α 's = 0.58–0.73) across subsamples and timepoints.…”