Examining the Factor Structure and Subgroup Invariance of the Deliberate Denial of Disordered Eating Behaviors Scale

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript addresses a clinically and psychometrically meaningful gap in the assessment of disordered eating concealment. The effort to extend evaluation of the DDEBS-12 to a broader, more demographically diverse sample is commendable, as is the systematic application of configural, metric, and scalar invariance testing. That said, several concerns regarding reporting completeness, methodological transparency, and interpretive accuracy require substantive attention before the manuscript can be considered for publication.

MAJOR COMMENTS

1. A likely reporting error appears in Table 1, where the standard deviation for age is listed as 16.79. For a college undergraduate sample with a mean age of 21.28 years, a standard deviation of this magnitude is implausible and would imply a distribution extending well into negative values or spanning decades. The authors should verify and correct this value. If the reported SD reflects a data entry or formatting error, the corrected value should be clearly indicated and any downstream implications for participant eligibility (participants were required to be at least 18 years old) should be briefly addressed.
2. The model respecification process, while guided by theoretical rationale, lacks sufficient transparency regarding the decisions to exclude specific items. The authors note that three items were removed because they were “ambiguously worded and could be interpreted either way,” yet do not report the factor loadings or modification indices for these items, nor do they present the full pattern of misfit that motivated each analytic step. A more rigorous account of the sequence of model comparisons, including fit statistics for each alternative tested, is essential to evaluate whether the final DDEBS-restriction model was identified through principled theoretical reasoning or through post-hoc pruning driven by statistical optimization. The inclusion of a supplementary table reporting all tested models with their fit statistics would substantially strengthen this section.
3. The reliability section reports Cronbach’s alpha exclusively. Current psychometric standards recommend the reporting of McDonald’s omega (ω) as a more robust estimate of internal consistency, particularly when the tau-equivalence assumption underlying alpha is unlikely to hold. Given that the factor loadings in Table 2 range from .530 to .800, the assumption of equal loadings is clearly not met. The authors should report omega alongside alpha throughout the manuscript.
4. The handling of partial scalar invariance across levels of disordered eating engagement requires a more careful discussion. The authors report that the CFI change (Δ = .019) exceeded the .01 threshold, indicating non-invariance at the scalar level, yet the implications of this finding for scale interpretation and score comparisons are not adequately addressed. Specifically, when scalar invariance does not hold, latent mean comparisons between groups are not fully justified, and the claim that the DDEBS-restriction is equivalently appropriate “for use with” non-clinical versus subclinical/clinical individuals requires qualification. The authors should clarify whether partial scalar invariance was explored (i.e., whether any items showed invariant intercepts while others did not), which would allow a more nuanced interpretation, and should explicitly note the inferential constraints this finding places on cross-group mean comparisons.
5. The validity correlations reported in Section 3.3 contain what appears to be a typographical error: the correlation between the DDEBS-restriction and the EPSI-restricting subscale is listed as “r = 48” without a decimal point. More substantively, the authors characterize this correlation as “moderate” and as sufficient to support criterion validity while also indicating discriminant differentiation from the criterion measure. However, a correlation of this magnitude between a measure of denial of restriction and a direct measure of restrictive eating pathology warrants more careful interpretation. The discussion of what this relationship implies for the construct boundary of the DDEBS-restriction, and how it compares to the original DDEBS-12 validity evidence, should be expanded. The emerging literature on the role of emotional concealment and behavioral secrecy in the maintenance of restrictive eating disorders offers conceptual frameworks that could help the authors situate this finding more precisely (e.g., work examining the interaction between denial, ego-syntonic cognition, and illness trajectory in restrictive eating disorders would be relevant here).
6. The exclusion of 679 participants (17%) who scored above 1 SD on the Underreporting of Disordered Eating Behaviors Scale deserves more critical discussion. While this procedure is methodologically defensible, the authors do not report whether the excluded participants differed systematically from the retained sample on key demographic or clinical variables. If excluded participants were disproportionately drawn from the subclinical/clinical group, the composition of the final analytic sample may be biased in a way that affects the invariance results and the validity correlations. At a minimum, a comparison of the demographic and disordered eating profiles of excluded versus retained participants should be reported or addressed.

MINOR COMMENTS

7. The Introduction would benefit from a more explicit developmental framing of denial as a construct. The manuscript currently treats denial of disordered eating as a relatively context-free behavioral tendency, without considering how concealment and secrecy around eating behaviors may emerge and consolidate across specific developmental periods, particularly in the transition from adolescence to emerging adulthood. Theoretical accounts that situate restrictive eating and its associated denial in the context of early relational experiences, parenting dynamics, and developmental pathways to disordered eating would considerably strengthen the conceptual grounding of the DDEBS-restriction. Engaging with this literature would also clarify why denial may function differently across developmental stages and how the scale might capture variation in concealment that is rooted in distinct developmental histories rather than merely in current symptom severity (e.g., 10.3390/children12121643; 10.3389/fpsyg.2021.802567).
8. The discussion of racial invariance, while appropriately cautious, would benefit from a more nuanced engagement with the literature on cultural variation in eating disorder presentation and self-disclosure norms. The reference to Morrison and Downey (2000) concerns suicidal ideation specifically and is not an ideal anchor for generalizing about disclosure tendencies in the context of eating behaviors. More current and construct-relevant literature on cultural factors in disordered eating concealment should be incorporated.
9. Throughout the manuscript, the term “race” is used when “racial identity” or “racial group” would be more accurate and consistent with current conventions in the field. The demographic categories used (White and Black only) should be more consistently noted as a limitation, given that no information is provided about participants who may have identified with other racial or ethnic groups, whose data were apparently removed from the invariance analyses without explicit acknowledgment.

Author Response

Please see attachment

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for the opportunity to review this compelling study that examines a measure that is much needed in the field. I appreciate the authors intent to capture what clinicians anecdotally observe and seek a way to measure this phenomenon to better understand and treat disordered eating.

Author Response

Please see attachment

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

This is a well-executed psychometric study with a clear and appropriate analytic approach. However, the manuscript would be substantially strengthened by improving conceptual clarity, increasing transparency in model development decisions, and aligning claims more closely with the data and design. Addressing the points I list in the attached review will enhance both the rigor and the contribution of the work (which is currently lacking). 

Comments for author File: Comments.pdf

Author Response

Please see attachment

Author Response File: Author Response.docx

Reviewer 4 Report

Comments and Suggestions for Authors

see attached file

Comments for author File: Comments.pdf

Author Response

Please see attachment

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript has clearly benefited from the revision process. The structure is clearer, the results are well presented, and the discussion appropriately reflects the findings. The study is now suitable for publication. I recommend acceptance in its present form.

Author Response

Comment from reviewer: The manuscript has clearly benefited from the revision process. The structure is clearer, the results are well presented, and the discussion appropriately reflects the findings. The study is now suitable for publication. I recommend acceptance in its present form.

Response: Thank you so much. We appreciate the reviews you provided and believe that our manuscript is stronger for it.

Reviewer 3 Report

Comments and Suggestions for Authors

Examining the Factor Structure and Subgroup Invariance of the Deliberate Denial of Disordered Eating Behaviors Scale

Thank you for the opportunity to review this revised manuscript. The authors have clearly invested substantial effort into addressing the prior concerns, and the manuscript has improved meaningfully in conceptual clarity, transparency, and interpretive restraint. That said, a few issues remain that should still be addressed before publication.

The manuscript continues, at times, to overstate the implications of the invariance findings and the broader applicability of the revised scale. For example, the Abstract states that the DDEBS-restriction “functioned equivalently” across groups and concludes that it is “suitable for diverse populations.” Similarly, the Conclusion states that the scale is “appropriate for use with these populations.” These statements remain somewhat stronger than warranted given:

• scalar non-invariance across disordered eating groups,
• the restriction of racial analyses to Black and White participants,
• the exclusion of nonbinary gender identities from invariance analyses, and
• the non-clinical undergraduate sample.

Thus, the abstract and conclusion should reflect with greater precision that there was support for configural and metric invariance only across the tested groups, and with the caveat of limited evidence regarding scalar equivalence and broader generalizability.

In addition, although the authors appropriately acknowledge that the DDEBS-restriction reflects a narrowing of the original construct, the manuscript could more fully address the conceptual implications of this shift. At present, the paper occasionally reads as though the revised measure represents a refined version of denial of disordered eating broadly. However, the findings may instead suggest that denial of restrictive eating represents a more psychometrically coherent subtype of denial, whereas broader denial processes may be more heterogeneous and not adequately captured by a single latent construct. This distinction is theoretically important and merits brief acknowledgment, as it would strengthen the conceptual precision of the manuscript.

Another—and in my view more substantial—conceptual issue concerns the exclusion of participants scoring high on the UDEBS. The authors now provide a clearer rationale for this decision, namely that the UDEBS was intended to identify participants likely engaging in socially desirable or underreporting response patterns and thus improve the psychometric interpretability of the DDEBS factor structure. This rationale is understandable from a measurement perspective. However, the manuscript still does not fully grapple with the theoretical implications of this exclusion strategy. The construct under investigation is itself denial, concealment, and minimization of disordered eating behaviors, yet participants identified as especially prone to underreporting or “fake-good” responding were removed from the analytic sample. This creates an important conceptual tension between treating concealment as substantive psychological content versus treating it as measurement contamination. Put differently, the exclusion procedure may have disproportionately removed individuals highest on the very tendencies the scale is intended to assess. The authors appropriately note that the factor analytic pattern remained similar with and without exclusions, which is reassuring. Nevertheless, the manuscript would benefit from a more explicit acknowledgment that the resulting factor structure may reflect a psychometrically cleaner—but potentially narrower—representation of denial-related processes. This issue is especially relevant given the emergence of a restriction-focused factor and the elimination of broader denial-related content.

The manuscript would also benefit from inclusion of a full correlation table among the primary study variables. At present, only selected validity correlations are reported, which makes it difficult to evaluate the broader nomological structure of the construct and the extent to which the observed pattern aligns with existing literature on concealment and disclosure processes. For example, the positive association between the DDEBS-restriction and self-concealment is theoretically coherent and broadly consistent with prior findings indicating that individuals high in self-concealment also tend to report elevated distress and related maladaptive psychological characteristics. Similarly, the negative association with distress disclosure appears conceptually appropriate. However, without the full intercorrelation matrix—including associations among self-concealment, distress disclosure, restrictive eating pathology, and related measures—it is difficult for readers to fully evaluate the distinctiveness and interpretability of the construct. Including these correlations would substantially strengthen the psychometric and conceptual framing of the manuscript.

Minor Comments

• There are still a few places where “race” and “racial identity” appear inconsistently. A final terminology check would be helpful.
• In several locations, the manuscript refers to “subclinical/clinical” participants even though the sample appears to consist primarily of undergraduate participants categorized via EDE-Q cutoffs rather than diagnosed clinical populations. The wording should remain careful and consistent throughout.
• The paper would benefit from a brief explicit statement that the revised factor structure has not yet been cross-validated in an independent sample. Although replication is mentioned in the limitations, emphasizing this point more directly would strengthen the psychometric framing.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 3

Reviewer 3 Report

Comments and Suggestions for Authors

Thank you for the opportunity to review this revised manuscript. The authors have addressed the remaining concerns thoughtfully and thoroughly. The manuscript is now substantially clearer in its conceptual framing, more transparent in its analytic presentation, and appropriately cautious in its interpretation of the invariance findings.

I have no further substantive concerns.