Review Reports - Anthropometric and Psychological Variables Associated with Body Mass Index: A Sex-Stratified Evaluation of Students from a Romanian University

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Summary of the manuscript and key contributions

This manuscript examines anthropometric, developmental, and psychosocial variables associated with body mass index (BMI) in 445 university students from the University of Oradea, with analyses stratified by biological sex. The study evaluates neck circumference, Body Adiposity Index (BAI), right-hand 2D:4D ratio, life satisfaction, and perceived social support using correlation analyses, multivariable linear regression, ROC curves, and pairwise AUC comparisons.

The study addresses a relevant topic and has several strengths. Anthropometric measurements are described in considerable detail, repeated measurements were averaged, validated Romanian versions of the psychological scales were used, and the authors attempt to distinguish association from classification performance. The inclusion of sex-stratified analyses is also potentially informative.

However, several methodological and interpretive issues need to be addressed before the main conclusions, particularly those concerning screening utility and sex-specific cut-off values, can be adequately supported.

Title

Strengths

The title is clear and accurately identifies the population, principal outcome, and sex-stratified analytical approach.

Suggestions

The term “associated with BMI” is appropriate and preferable to predictive terminology given the cross-sectional design. This cautious formulation should also be maintained consistently throughout the manuscript.

Abstract

Strengths

The Abstract clearly reports the sample, major variables, analytical procedures, and principal regression and ROC findings.

Suggestions

The practical interpretation of the ROC-derived cut-offs should be substantially tempered. The reported thresholds are derived and evaluated within the same convenience sample and use BMI-defined obesity as the reference criterion. They should therefore be described as exploratory, internally derived thresholds, not as established screening alternatives suitable for implementation.

The statement that psychological variables operate on “separate structural layers” should also be revised. Non-significant regression coefficients do not demonstrate statistical independence

Introduction

Strengths

The Introduction provides a broad rationale for examining alternative anthropometric markers and acknowledges well-known limitations of BMI. It also gives a rationale for considering psychosocial variables and sex-specific anthropometric patterns.

Suggestions

The theoretical positioning of R-2D:4D as a marker of prenatal hormonal exposure should be more cautious. The manuscript frequently presents this interpretation as established biological fact. It would be more appropriate to describe 2D:4D as a putative or proposed proxy marker, particularly because its validity as a direct marker of prenatal androgen exposure remains debated.

The hypotheses also appear unusually specific. H2 predicts that R-2D:4D will be significant in standalone ROC analyses but lose significance in the multivariable model, which closely mirrors the reported result. The authors should clarify whether these hypotheses were genuinely formulated before data analysis. If they were not prospectively specified, these statements would be better framed as exploratory research questions rather than confirmatory hypotheses.

Methods

Strengths

The anthropometric procedures are described in detail, including positioning, measurement instruments and repeated measurement. Averaging three determinations is a methodological strength.

The use of validated Romanian versions of the SWLS and MSPSS is also appropriate.

Suggestions

1. The principal conceptual issue concerns the reference outcome used to evaluate “screening performance.” Obesity is defined entirely by BMI ≥ 30 kg/m², and the study then evaluates whether anthropometric measures such as BAI and neck circumference can classify this BMI-derived category.

This does not demonstrate that these variables identify adiposity, metabolic risk, or obesity more accurately than BMI. It only demonstrates agreement with a category that is itself defined by BMI.

The problem is particularly important for BAI because BMI and BAI are both mathematically derived anthropometric indices and both contain height in their formulas. Consequently, part of the very high association and variance explained may arise from mathematical coupling/shared anthropometric components rather than independent biological information.

The authors should explicitly acknowledge this limitation and substantially revise statements describing BAI or neck circumference as “alternatives” to BMI. Validation against an independent criterion such as DXA-derived body fat, percentage body fat, visceral adiposity or cardiometabolic outcomes would be required to support stronger claims of screening validity.

2. The sample-size justification is not aligned with the primary analyses. The Cochran finite-population formula is appropriate for estimating a population proportion under specified precision assumptions, but the study uses convenience sampling and the principal analyses are sex-stratified multiple regressions and ROC analyses.

Therefore, exceeding a Cochran-derived threshold of 378 does not establish adequate statistical power or precision for the sex-specific ROC analyses, particularly for estimation of optimal cut-offs.

The authors should either provide a sample-size/precision justification appropriate to the main regression and classification analyses or clearly state that the sample-size calculation was only a general recruitment target and not an inferential power calculation.

The reported post hoc observed power of 1.000 should preferably be removed, as observed power is a deterministic function of the observed effect and provides little additional information beyond the effect estimate and its confidence interval.

3. The prevalence of obesity and the number of cases used in each ROC analysis must be reported. BMI itself is surprisingly absent from the descriptive table, despite being the main dependent variable and the criterion used to define obesity. The manuscript should report, separately for males and females:

mean and SD of BMI;
BMI range;
number and percentage classified as obese;
number and percentage classified as non-obese.

Without these data, readers cannot evaluate the stability of the AUC estimates, sensitivities, specificities, and data-derived cut-offs.

4. The cut-offs require internal validation and substantially more cautious interpretation. The “optimal” thresholds are derived by maximizing the Youden index and their performance is then reported in the same sample.

This produces optimistically biased estimates of classification performance. At minimum, the authors should consider bootstrap validation or cross-validation and report uncertainty around sensitivity, specificity, and ideally the cut-off itself. Without such validation, values such as 100% sensitivity should not be interpreted as stable screening performance.

5. Sex stratification alone does not demonstrate sex differences in associations. The authors estimate separate models for males and females, but differences in statistical significance between two subgroup models do not establish that the corresponding regression coefficients differ significantly.

If sex-specific associations are an important conclusion, pooled models including sex × predictor interaction terms should be tested, or another formal coefficient-comparison procedure should be used.

Sex-specific descriptive analyses may remain useful, but statements that a relationship is stronger, weaker, or structurally different by sex should be supported by formal interaction tests.

6. The regression diagnostics are incomplete and partly misinterpreted. The manuscript relies heavily on Durbin–Watson and states that values around 2 demonstrate that residuals are “completely free from autocorrelation.”

For independent cross-sectional observations, serial autocorrelation is not normally the principal concern, and Durbin–Watson does not establish general residual independence.

More relevant diagnostics should be reported for:

linearity,
homoscedasticity,
normality of residuals,
influential observations/leverage,
potentially heteroskedasticity-robust standard errors if needed.

Normality of the raw predictor variables is not itself an assumption of linear regression.

7. Reliability results promised in the Methods are missing from the Results. The authors state that Cronbach's alpha and McDonald's omega were calculated for SWLS and MSPSS, but these coefficients do not appear to be reported subsequently.

Please provide the reliability coefficients for the present sample, preferably separately by sex if the primary analyses are sex-stratified.

8. The rationale for the 50:50 sex goodness-of-fit test is unclear. Table 1 tests whether the observed sex distribution differs from an assumed 50:50 distribution. Unless an equal distribution is theoretically or population-based expected, this test has limited interpretive value.

If representativeness is the concern, the appropriate comparison would be with the actual sex distribution of the eligible University of Oradea student population. Moreover, stratifying the subsequent analyses does not “correct” selection bias arising from convenience recruitment.

Results

Strengths

Regression coefficients, standardized effects, confidence-related diagnostics and ROC parameters are reported in considerable detail, and the stratified presentation is generally easy to follow.

Suggestions

1. The bivariate results do not fully correspond to the stated analytical plan. The Methods state that Pearson correlations were calculated between the somatic, digital and psychosocial markers and continuous BMI scores.

However, Table 4 contains only NC, R-2D:4D, BAI, SWLS and MSPSS; BMI is absent from the matrix.

Because association with BMI is Objective 1 and a central part of the study, BMI should be added to the correlation matrix.

2. The very high R² values require more cautious interpretation. The models explain approximately 85% of BMI variance in males and 91% in females, driven primarily by BAI and neck circumference.

Given that BMI and BAI are themselves derived anthropometric quantities and share height as a mathematical component, these values should not be interpreted as evidence that these measures explain 85–91% of the biological determinants of body mass. They primarily reflect strong concurrent anthropometric overlap.

3. The pairwise ROC comparisons appear excessive relative to the research question. A large number of pairwise AUC comparisons are performed separately in males and females without discussion of multiplicity. The authors should either justify the full comparison matrix, consider multiplicity control, or focus on a smaller number of pre-specified comparisons of genuine clinical/research interest.

There is also limited interpretive value in statistically comparing AUCs for anthropometric screening variables against psychological scales that were not designed as diagnostic obesity classifiers.

4. AUC values below .50 require careful interpretation. For female participants, SWLS has an AUC of .357. This reflects an inverse relationship with the outcome rather than simply “poor classification” in a direction-neutral sense. The direction of the ROC test should be specified clearly.

Discussion

Strengths

The Discussion connects the anthropometric findings to previous literature and generally acknowledges that the design is cross-sectional

Suggestions

The manuscript repeatedly interprets the disappearance of the R-2D:4D coefficient after including BAI and neck circumference as “shared variance absorption” and implies that the anthropometric variables “mask” a meaningful prenatal biological effect. This interpretation is more speculative than the results allow.

A non-significant adjusted coefficient can simply indicate that the variable does not contribute unique information once the other variables are included. The data cannot determine that a genuine prenatal effect exists but is being “masked.”

More importantly, absence of significant associations for SWLS and MSPSS does not establish independence or separate structural systems. The Discussion states that body mass classifications and psychosocial dimensions “operate as separate variables” and the Conclusions go even further, describing them as “entirely separate structural layers.”

This language should be replaced by something such as: “No independent associations between SWLS/MSPSS and BMI were detected in the present multivariable models.”

If the authors wish to demonstrate equivalence or absence of a meaningful association, an equivalence-testing framework with a pre-specified smallest effect size of interest would be needed.

The negative correlations between SWLS and MSPSS are also unexpected given the usual positive relationship between perceived social support and well-being. The authors speculate about compensatory mechanisms and overprotective family ties, but these mechanisms were not measured. This section should be shortened and framed explicitly as speculation. Given the unusual direction, confirming scoring procedures and reporting the reliability coefficients is particularly important.

Limitations

The current limitations section is too brief relative to the methodological constraints of the study.

In addition to cross-sectional design, single-site recruitment and self-report psychometrics, it should explicitly acknowledge:

convenience sampling;
lack of an independent criterion of adiposity or metabolic health;
mathematical overlap between BMI and BAI;
internally derived, non-validated ROC cut-offs;
potentially limited numbers of obesity cases within each sex subgroup;
absence of formal tests comparing sex-specific associations;
possible optimism in ROC sensitivity/specificity estimates;
limited generalizability of thresholds beyond this sample.

Conclusions

Strengths

The Conclusions correctly emphasize that the findings apply to the investigated cohort.

Suggestions

The statements concerning “practical alternatives” and “screening options” remain too strong given that obesity was defined using BMI itself and the thresholds were derived and tested within the same sample.

A more appropriate conclusion would state that neck circumference and BAI showed strong concurrent discrimination of BMI-defined obesity within this sample, and that the proposed thresholds require independent validation against both BMI categories and, preferably, direct adiposity or metabolic-health measures before clinical or institutional implementation.

Minor/editorial points

There are also several wording and editorial issues that should be corrected. For example, expressions such as “structural layers,” “area tracking configurations,” “variance absorption,” and “statistical hierarchy” make the manuscript more difficult to read than necessary. Simpler statistical language would improve clarity.

The reference/discussion consistency should also be checked carefully. For example, the Discussion refers to Ononamadu et al. (2025), whereas reference 47 is listed as 2017.

Overall evaluation

The manuscript addresses a potentially useful applied question and contains several methodological strengths, including detailed anthropometric measurement, sex-stratified reporting, validated psychosocial measures, multivariable analyses and ROC analyses.

However, the manuscript currently overinterprets the ability of anthropometric variables to classify an outcome defined by BMI itself, particularly given the mathematical overlap between BMI and BAI. The proposed sex-specific screening cut-offs are also derived and evaluated in the same convenience sample without internal or external validation, and the number of obesity cases underlying these analyses is not reported.

Additional concerns involve the sample-size justification, lack of formal tests of sex differences, incomplete regression diagnostics, missing psychometric reliability results, omission of BMI from the correlation matrix, and interpretation of non-significant psychological effects as evidence of statistical independence.

These issues are important but appear addressable through reanalysis, fuller reporting, and substantial moderation of the interpretation.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript presents a cross-sectional study of 445 university students from the University of Oradea, examining associations between anthropometric indices, a prenatal proxy marker, psychosocial variables, and BMI status. From a pre-editorial perspective, the study has several strengths: ethical approval is reported, informed consent is declared, conflicts of interest and funding are stated, and the sample size calculation is described.

The provided PDF does not include a visible title; BMI descriptive statistics and obese/non-obese counts are absent; the MSPSS scoring appears inconsistent; multiple comparison corrections are not applied; ROC cut-off values are not internally validated.

The title could not be evaluated because it is not visible in the supplied file. This is itself an editorial concern, because a Q1 submission must present a clear, informative title that reflects the study design, population, and primary variables. The abstract is structured and includes sample size, sex distribution, regression coefficients, R² values, and ROC results. However, the research gap is weakly justified, effect sizes and confidence intervals are largely missing, and the conclusion overstates the practical utility of the findings. Phrases such as “strong screening capacity” and “low-cost institutional screening” are not supported by a cross-sectional design without external validation. The keywords are relevant but somewhat generic, and they overlap heavily with the title. The Introduction follows a broad obesity narrative but lacks a focused funnel structure. It does not clearly articulate why neck circumference, BAI, R-2D:4D, life satisfaction, and social support should be examined together, nor does it present testable hypotheses. The mechanistic rationale linking prenatal androgen exposure to adult BMI is speculative, and the psychosocial variables are not adequately justified.

The Methods section is detailed in some areas, particularly anthropometric measurement procedures, but it contains several critical methodological weaknesses. The sample size calculation uses the Cochran formula, which is appropriate for estimating a population proportion but not for powering multivariable regression or ROC analyses. No a priori power analysis is provided for the regression models, and no event-per-variable or power calculation is given for the ROC analyses. The number of obese and non-obese participants is not reported, which is a fundamental omission because ROC curves, sensitivity, specificity, and cut-off values cannot be properly interpreted without knowing the outcome prevalence and group sizes. The sampling procedure relies on voluntary enrollment, raising concerns about selection bias. Important confounders such as physical activity, diet, sleep, smoking, medication use, and socioeconomic status are not controlled. Missing data handling is not described, and no STROBE-style participant flow diagram is provided. Although the study is cross-sectional, the manuscript repeatedly implies screening and classification utility, which is not justified without prospective validation.

The anthropometric procedures are described with commendable detail, including device models, measurement landmarks, and triplicate measurements. However, reliability statistics such as ICC, CV, SEM, or MDC are not reported. Inter-rater reliability is also absent, and it is unclear whether the same investigator performed all measurements or whether blinding was used. The BAI formula is correctly presented, but the mathematical coupling between BMI and BAI is not discussed. Because both indices incorporate height, and because BAI uses hip circumference while BMI uses body weight, their shared variance may artificially inflate the regression R² values. This is a major interpretive issue that must be addressed. The psychosocial measures are described, but the scoring of the MSPSS is unclear. The text refers to “total social support,” yet the descriptive table reports values around 2.23 with a maximum of 3.33, which cannot represent a 12-item, 7-point Likert total score. This inconsistency undermines confidence in the psychosocial data. In addition, the psychometric properties of the SWLS and MSPSS in the present sample, such as Cronbach’s alpha or confirmatory factor analysis, are not reported.

The statistical analysis section has several shortcomings. Multiple comparison corrections are not applied to the correlation matrix, t-tests, ROC pairwise comparisons, or DeLong tests. The ROC cut-off values are derived and tested in the same dataset without bootstrap, cross-validation, or split-sample validation. Regression diagnostics are limited to Durbin-Watson and VIF/Tolerance; residual plots, homoscedasticity, linearity, and influential points are not reported. Effect sizes such as Cohen’s f² and confidence intervals for regression coefficients are missing. The manuscript relies heavily on p-values and does not adequately discuss practical or clinical significance. Given the large number of statistical tests and the absence of correction, there is a risk of false-positive findings. Overall, the Methods section does not meet Q1 methodological standards, and the combination of unvalidated cut-offs, missing outcome counts, and uncontrolled confounders constitutes a serious threat to internal validity.

The Results section is weakened by the absence of descriptive data for BMI, the primary outcome variable. Mean BMI, standard deviation, range, and obesity prevalence are not reported. The number of obese and non-obese participants in each sex subgroup is also missing, which makes the ROC results difficult to interpret. The MSPSS scoring inconsistency is visible in Table 2 and is not resolved. The ROC analysis for SWLS in female participants yields an AUC of .357, with a cut-off of 34.50, sensitivity of 2.9%, and specificity of 100%. This is clinically meaningless and should not be presented as a meaningful classification result. Similarly, the MSPSS AUC of .482 in males and .563 in females is non-significant or trivial, and these analyses add unnecessary clutter. Table 6 contains a mislabeled note referring to “Table 1,” and several tables are dense, difficult to read, and lack confidence intervals for key estimates. The Results section also contains interpretive language such as “statistically superior” and “systematically out-tracked,” which belongs in the Discussion. No multiple comparison correction is applied, so the reported significance of some pairwise ROC contrasts may be spurious. The high R² values in the regression models are impressive at face value, but they are likely inflated by the mathematical overlap between BMI and BAI. This issue is not acknowledged in the Results.

The Discussion overinterprets the findings and implies a level of practical utility that the design cannot support. The manuscript repeatedly describes the cut-off values as offering “practical, low-cost screening options,” but these thresholds were not externally validated, and no decision-curve or net-benefit analysis is provided. The suggestion that R-2D:4D has “independent classification utility” in ROC analysis but loses significance in multivariable regression is explained as variance absorption. This explanation is plausible, but it is presented too confidently without formal mediation or commonality analysis. The psychosocial findings are particularly problematic. The negative correlation between life satisfaction and social support is interpreted through speculative mechanisms such as compensatory social support or overprotective family ties, none of which can be tested with the current data. The negative association between BAI and life satisfaction in females is interesting, but it is not sufficiently developed and may be confounded by body image, cultural factors, or unmeasured psychological variables. The literature review within the Discussion is broad but superficial; it lists many studies without critically synthesizing their methodological limitations or contradictory findings. The limitations paragraph is inadequate and does not honestly address the cross-sectional design, absence of cut-off validation, lack of confounder control, mathematical coupling between BMI and BAI, or the possibility of AI-assisted writing.

The writing quality throughout the Discussion is a major concern. The prose is overly polished, repetitive, and filled with generic academic phrasing. Terms such as “structural layers,” “mathematical artifacts,” “variance absorption,” “concurrent regression spaces,” and “somatic classifications” feel artificial and do not reflect natural scientific writing. The Discussion is also too long and repetitive, often restating the Results rather than interpreting them. The reference list contains inconsistencies: some author names and years do not match the in-text citations, some DOIs are incomplete or incorrect, and at least one page range is erroneous. The use of sources such as Cureus may be acceptable in some contexts but is not ideal for a Q1 journal. The self-citation rate is not excessive, but it should be monitored. The tables and figures are not fully Q1-compliant: Table 2 is poorly formatted, Table 6 is mislabeled, and Figure 1 could not be evaluated from the extracted text. Overall, the Discussion does not meet the standards of a high-impact sports science or health sciences journal.

The manuscript does not contain a separate Conclusion section, which is a structural omission. The final paragraphs of the Discussion serve as a conclusion, but they do not clearly state the main findings, limitations, and future directions in a concise and balanced manner. The practical implications are overstated, and the screening recommendations are premature. On a 1–10 scale, I would assign the following scores: Novelty 4, Methodological Quality 4, Statistical Quality 5, Writing Quality 3, Practical Relevance 5, Editorial Priority 3, and Acceptance Probability 10–15%. Under the requested Q1 sports science framework, the decision is Reject or Desk Reject because the manuscript is outside the scope of sports science journals. If evaluated in a health sciences, obesity, or public health journal, the decision would be Major Revision. The authors should report BMI descriptives and obese/non-obese counts, clarify and correct the MSPSS scoring, apply multiple comparison corrections, validate the ROC cut-offs using bootstrap or cross-validation, address the mathematical coupling between BMI and BAI, control for key confounders, add a STROBE flow diagram, include a separate Conclusion, and substantially revise the manuscript to remove AI-generated writing patterns. Without these revisions, the manuscript is not suitable for publication in a Q1 journal.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Dear authors,

The manuscript addresses a relevant topic concerning the relationships between anthropometric indicators, the R-2D:4D digit ratio, and psychosocial components in relation to body mass index and obesity status in young adults. The topic is of interest in the context of identifying simple, non-invasive, and accessible methods for assessing weight status in university populations.

However, the manuscript presents several conceptual, methodological, statistical, and interpretative issues that require clarification and substantial revision, as detailed below:

In the Abstract and Highlights sections, some conclusions are stated too categorically considering the cross-sectional design of the study. In particular, the claims regarding the “statistical independence” of psychosocial components and the use of anthropometric cut-offs as screening alternatives for university programs go beyond what can be demonstrated by the current analyses. The absence of a significant association in a statistical model does not demonstrate structural independence, and thresholds derived from the same sample in which their performance was evaluated cannot be regarded as validated screening thresholds. I recommend reformulating these statements in more exploratory terms.
The Introduction requires a clearer conceptual focus. In its current form, neck circumference, BAI, R-2D:4D, life satisfaction, and social support are integrated within the same framework. However, I consider that the rationale for simultaneously including these variables within the same analytical framework is insufficiently developed, particularly as the transition between anthropometric indicators, the proposed prenatal marker, and psychosocial components is relatively abrupt. I recommend restructuring this section around a clearly defined gap in the existing literature and explaining the independent contribution of each category of variables to the main research question.
In my opinion, the interpretation of the R-2D:4D ratio as a marker of prenatal hormonal exposure requires greater caution. The manuscript measures the digit-length ratio in adults rather than prenatal hormonal exposure itself. Nevertheless, several sections, including the Discussion, move from the statistical association of R-2D:4D to interpretations involving “prenatal androgen patterning” or an “underlying prenatal biological signature.” I recommend clearly distinguishing between the measurement actually performed and its hypothetical biological interpretation.
Hypothesis H2 is unusually specific, as it anticipates both significant discriminatory ability of R-2D:4D in the ROC analysis and the loss of its statistical significance when anthropometric indicators are introduced into the multivariable regression model. Such a hypothesis requires a strong theoretical justification and raises the question of whether it was specified before the data were analyzed. I recommend clarifying whether the objectives and hypotheses were defined a priori and whether the study protocol or statistical analysis plan was preregistered. Otherwise, the exploratory nature of these analyses should be explicitly acknowledged.
The justification of the sample size using Cochran’s formula does not correspond to the principal analyses performed in the study. The formula presented estimates the sample size required to determine a proportion in a finite population, using p = 0.50 and a 5% margin of error. However, the main objectives of the study involve multivariable regression models and sex-stratified ROC analyses. Therefore, exceeding the threshold of 378 participants does not automatically demonstrate that the sample size is adequate for stable estimation of these models, particularly the ROC-derived cut-offs. I recommend reconsidering the sample size justification in relation to the primary analyses.
The characterization of the sample is insufficient in relation to the main objectives of the study. Although BMI represents the central dependent variable in the regression models and is used to define obesity status in the ROC analyses (BMI ≥30 kg/m²), Table 2 does not provide descriptive statistics for BMI. I recommend reporting the mean, standard deviation, and range of BMI separately for male and female participants. More importantly, given the ROC analyses, the number and percentage of participants classified as obese and non-obese should be explicitly reported for each sex subgroup. I consider this information essential for assessing the robustness of the reported AUC estimates, sensitivity, specificity, and proposed cut-off values.
The relationship between BMI and BAI requires more careful methodological discussion. BMI is calculated using body weight and height, whereas BAI is calculated using hip circumference and height. Thus, the two variables share a mathematical component based on height. Under these circumstances, the very high adjusted R² values of 0.853 in males and 0.908 in females should not be interpreted solely as indicating a very strong biological ability of the predictors to explain BMI. I recommend discussing the structural dependence between these anthropometric indicators and tempering the interpretation of these values.
There is also a conceptual issue with presenting BAI as a potential alternative screening measure for obesity. The study begins by emphasizing the limitations of BMI as an indicator of adiposity, yet subsequently uses BMI ≥30 kg/m² as the reference standard to demonstrate the ability of BAI to identify “obesity.” Consequently, the analysis demonstrates how well BAI discriminates a category defined by BMI rather than how accurately it identifies adiposity determined using an independent reference method, such as direct body composition assessment. This distinction should be clearly acknowledged when interpreting the results.
The ROC analyses represent one of the central components of the manuscript; however, their reporting is insufficient to validate the proposed cut-off values. For example, Table 6 reports an AUC of 0.982 for BAI in females, with 100% sensitivity and 91% specificity, while neck circumference shows AUC values of 0.921 in males and 0.930 in females. Given these very high levels of performance, it is particularly important to report the number of obesity cases, the classification matrices, or at least the numbers of true positives, false positives, true negatives, and false negatives corresponding to the selected thresholds.
The cut-off values identified by maximizing the Youden index appear to have been both derived and evaluated in the same sample. This approach may produce optimistic estimates of performance and provides no information regarding the stability of these thresholds in other samples. I recommend either applying an appropriate internal validation procedure, if the available data permit, or presenting these values strictly as exploratory thresholds requiring independent validation. In the absence of such validation, I recommend removing statements regarding their direct applicability to university screening programs.
Some ROC results require more careful interpretation. In females, SWLS shows an AUC of 0.357, with p = 0.003, a sensitivity of only 2.9%, and a Youden index of 0.029. Although the inverse direction of discrimination is acknowledged, statistical significance relative to an AUC of 0.50 should not be confused with useful discriminatory performance. I recommend focusing the interpretation on the direction of the effect and the actual discriminatory performance rather than on the p-value in isolation.
The Methods section states that the internal consistency of the SWLS and MSPSS was assessed using Cronbach’s alpha and McDonald’s omega; however, the resulting values are not reported in the Results section. These results should be included, preferably separately by sex if the main analyses are sex-stratified. In addition, I consider that the exact method used to calculate the MSPSS score should be specified, as the mean values of approximately 2.2–2.3 reported in Table 2 require clarification in relation to the response scale described in the Methods.
The negative relationship between life satisfaction and perceived social support requires verification and discussion. Significant negative correlations between SWLS and MSPSS are reported in both males (r = −0.266) and females (r = −0.378). Given the nature of the constructs assessed, I recommend carefully checking item coding, score calculation, and scale direction. If the finding is correct, it warrants explicit discussion rather than being subsequently used only to support the proposed separation between psychosocial and anthropometric dimensions.
The assessment of linear regression assumptions is incompletely reported. The manuscript places considerable emphasis on Durbin–Watson statistics, Tolerance, and VIF and concludes that the residuals are “completely free from autocorrelation” and that coefficient stability is “completely uncompromised.” These statements are overly categorical. I recommend reporting the assessment of linearity, homoscedasticity, residual distribution, and influential or high-leverage observations. In a cross-sectional study involving independent participants, the Durbin–Watson statistic should not be presented as a general validation of the regression model.
The interpretation of the loss of statistical significance of R-2D:4D in the multivariable regression as “variance absorption” by anthropometric phenotypes is overly mechanistic. The fact that a variable is significant in a univariable or discriminatory analysis but no longer significant in a multivariable model does not demonstrate that the other variables have “captured” a prenatal biological mechanism. Nevertheless, the manuscript uses this explanation to argue that the anthropometric indicators “capture the majority of the shared variance.” I recommend describing the finding strictly in statistical terms: R-2D:4D did not show an independent association with BMI after inclusion of the other variables in the model.
More importantly, the linear regression and ROC analyses do not test the same hypothesis and should not be presented as though one explains the findings of the other. The regression analysis examines the association between R-2D:4D and continuous BMI after adjustment for the other predictors, whereas the ROC analysis assesses the unadjusted ability of R-2D:4D to discriminate between BMI <30 and BMI ≥30. The difference between these findings does not demonstrate that the regression model “masks” a genuine biological association. I recommend removing the statement that this divergence justifies the use of both approaches to prevent distal markers from being masked.
The Chi-square test assessing deviation of the sex distribution from a hypothetical 50:50 ratio appears to have limited relevance to the objectives of the study. It is unclear why an exactly equal distribution of males and females should be expected in a convenience sample. Furthermore, the statement that this asymmetry was “addressed” by stratifying the analyses is conceptually problematic: sex stratification may be biologically justified, but it does not correct potential selection bias resulting from voluntary recruitment. I recommend removing this analysis or clearly justifying its relevance.
Comparisons between male and female participants using t-tests are reported almost exclusively in terms of p-values. Given the sample size, I recommend reporting mean differences with 95% confidence intervals and standardized measures of effect size. This would allow readers to assess the magnitude of the observed differences rather than only their statistical significance.
The statement that normal distributions are “confirmed” because skewness and kurtosis values fall between −1 and +1 is too categorical. These indices may support an approximate characterization of the distributions but do not demonstrate normality. Furthermore, in the context of regression analysis, the distribution of residuals is more relevant than the individual normality of the predictors. I recommend revising this section accordingly.
The manuscript performs a large number of statistical tests, including sex comparisons, multiple correlations, separate regression models, several ROC curves, and pairwise AUC comparisons. Nevertheless, all analyses use a p < 0.05 threshold without discussion of multiple testing. I recommend clearly distinguishing between primary and exploratory analyses and discussing the risk of Type I error or, where appropriate, applying a suitable method to control for multiple testing.
The anthropometric measurements are described in detail, and it is stated that each measurement was performed three times and subsequently averaged. However, the reproducibility of these measurements is not reported. This issue is particularly relevant for R-2D:4D, where very small differences in digit length may affect the final ratio. If these data are available, I recommend reporting a measure of repeatability/reliability, such as the intraclass correlation coefficient (ICC) or technical error of measurement. It should also be clarified whether all measurements were performed by the same investigator and whether the investigator was aware of the participants’ weight status.
The Discussion contains several biological interpretations that extend beyond the data generated by the study. For example, the findings are used to support the existence of a prenatal signature and to conclude that adult physical dimensions reflect body mass variation more directly than markers of early development. Given the cross-sectional design and the absence of direct measurements of prenatal exposures, these explanations should be presented exclusively as possible hypotheses supported by the existing literature.
The language regarding screening should be tempered throughout the manuscript. It is suggested that neck circumference thresholds of 33.90 cm in males and 34.35 cm in females may serve as “operational boundaries” for health and wellness programs. However, these thresholds were derived from a single convenience sample recruited from one university, within a restricted age range, and without external validation. I consider that the current data support, at most, the formulation of hypotheses regarding potential candidate thresholds for future studies.
The Limitations section is too restricted in relation to the methodological issues of the study. At present, it primarily discusses the cross-sectional design, the specificity of the study population, and the self-reported nature of the psychological measures. I recommend explicitly adding convenience sampling, single-center recruitment, the absence of internal or external validation of the cut-offs, the potentially optimistic ROC performance estimates, the mathematical dependence between anthropometric indicators, the absence of an independent reference standard for adiposity, and the inability to extrapolate the proposed thresholds to other populations.
The conclusion regarding the psychosocial variables is too strong. It is stated that life satisfaction and social support “operate on entirely separate structural layers” and “do not co-vary with body weight categories.” However, Table 4 shows a significant association between BAI and SWLS in females, and SWLS also shows a statistically significant AUC for obesity status in females, although in the inverse direction and without practical discriminatory utility. Furthermore, lack of statistical significance in a regression model does not demonstrate independence between constructs. I recommend reformulating the conclusion strictly in terms of the observed findings: the psychosocial variables did not show significant independent associations with BMI in the multivariable models analyzed.
The manuscript should distinguish more rigorously between association, discrimination, and prediction. Cross-sectional regression models assess concurrent associations, whereas ROC analyses assess discrimination between categories defined at the same time point. These analyses do not demonstrate prospective predictive ability and do not validate a screening tool. I recommend removing terminology that implies prediction of future risk when this was not evaluated.
I recommend revising the presentation of the post-hoc statistical power analysis. Table 5 reports very large Cohen’s f² values (5.944 and 10.111) and an “observed statistical power” of 1.000 for both models. Observed power calculated after the results have been obtained provides limited additional information beyond the reported effect estimates and p-values. It would be more informative to report confidence intervals for the principal estimates and to justify sample adequacy on the basis of estimation precision or a prospective calculation relevant to the primary analysis.
The statistical terminology used throughout the manuscript should be simplified and standardized. Expressions such as “structural layers,” “variance absorption by proximal phenotypes,” “area tracking configurations,” “structural hierarchy of classification performance,” “statistical divergence,” “classification presence,” and “risk vector” are not standard statistical terminology and, in some instances, create the impression of mechanistic conclusions that are not supported by the analyses. I recommend using direct and precise statistical language.
The manuscript requires careful editorial and linguistic revision. There are repetitions and errors such as “for this somatic index for this somatic index,” “regressions and and classification procedures,” and “target relevant to the target,” as well as unnecessarily elaborate phrasing. Minor errors are also present in the list of abbreviations, for example “ring finge.” I recommend a thorough revision of the English language and simplification of the writing style before the manuscript is considered for publication.
The Conclusions should be substantially revised to reflect the cross-sectional and exploratory nature of the study. The data support strong associations between BMI and anthropometric indicators in the analyzed sample and suggest that certain variables may discriminate BMI-defined obesity status within this dataset. However, the findings do not demonstrate that the identified thresholds represent validated screening tools, that R-2D:4D reflects a prenatal hormonal mechanism in this context, or that psychosocial variables are structurally independent of weight status. I recommend emphasizing the need to validate these findings in an independent sample before drawing implications for screening or practice.

In conclusion, the manuscript addresses a relevant topic and benefits from a relatively large sample, standardized anthropometric measurements, and a statistical approach that attempts to examine both continuous associations and the discriminatory ability of the investigated indicators. However, in its current form, the manuscript presents important concerns regarding the justification of sample size, insufficient characterization of weight status, interpretation of the relationship between BAI and BMI, reporting and validation of the ROC analyses, interpretation of R-2D:4D, and extrapolation of the findings to screening applications. In addition, some biological and statistical interpretations extend beyond what can be supported by the cross-sectional design. I consider that the manuscript requires major revision, including reassessment and transparent reporting of the main analyses and substantial moderation of the conclusions and practical implications.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The manuscript addresses a potentially relevant and insufficiently investigated topic. However, in its current form, it presents several substantial conceptual, methodological, and structural shortcomings. The Introduction does not provide a sufficiently comprehensive and theoretically grounded overview of the existing literature. The rationale for examining the selected anthropometric, biological, and psychosocial variables in relation to BMI remains underdeveloped. Furthermore, the research questions and hypotheses are not formulated with sufficient clarity or specificity. The Methods section is unnecessarily lengthy and insufficiently focused on the actual methodological procedures and analytical decisions. Similar structural and interpretative issues are evident in parts of the Results section. The presentation should be more strictly limited to the empirical findings, with interpretation and theoretical explanation reserved for the Discussion. The statistical results should also be presented in a manner that directly corresponds to the stated research questions and hypotheses. The tables do not appear to conform consistently to APA style. The Discussion is one of the most problematic sections of the manuscript. Rather than critically interpreting the findings in light of the stated hypotheses and the literature reviewed in the Introduction, the section introduces additional bibliographic sources and, in places, reads more like a second literature review. The Discussion should not primarily expand the literature base; rather, it should explain what the present findings mean, whether and to what extent they support the proposed hypotheses, and how they compare with previously established evidence. New references may certainly be introduced where necessary to interpret unexpected or novel findings, but this should be done selectively and in direct connection with the empirical results.

Specific Comments

Title: The title should specify the origin of the sample more precisely. The term “university students” is overly broad and does not adequately reflect the sampling context, given that all participants were recruited from a single university.

Abstract: The Abstract captures the main elements of the study; however, several revisions are required. In Line 36, the notation “N” is not used appropriately and should be corrected. In addition, the statistical analyses should be removed from the Abstract or substantially reduced. Finally, the number of keywords should be limited to a maximum of five.

Line 92: The citation format should be corrected from “(12-14)” to “[12-14]”, in accordance with the citation style used throughout the manuscript.

Line 94: The citation format should be corrected from “(15)” to “[15]”.

Line 102: An appropriate citation should be provided to support the statement.

Research hypotheses: The research hypotheses should be reformulated and presented in a considerably more concise and structured manner. At present, they are overly complex and somewhat disorganized, which makes it difficult to identify precisely what is being hypothesized and what relationships are expected between the variables.

The Methods section should be substantially streamlined. At present, considerable space is devoted to theoretical background and conceptual explanations that would be more appropriately placed in the Introduction.

Line 138: formatting

Section 2.2. Sample Size and Sampling Procedure:The introductory part of this subsection should be considerably shortened. It currently relies too heavily on theoretical background and methodological justification that is not necessary at this level of detail.

Section 2.3. Anthropometric Measurements: This subsection should also be presented more concisely.

Line 313: The date on which ethical approval was obtained should also be reported.

Statistical Analyses: This section should be substantially condensed. The current presentation includes an excessive amount of theoretical background concerning the statistical methods employed

Results: The Results section requires substantial restructuring. The empirical findings should be presented first, followed by the corresponding references to the relevant tables. At present, the organization sometimes makes the reader move back and forth between the narrative and the tables, which reduces clarity. The tables also require substantial revision to conform to APA style.

Lines 399–404: Particular attention should be paid to the use and placement of parentheses.

Table 4: Table 4 is difficult to read and does not present the information in a sufficiently clear manner. Given the amount and type of information included, it would be preferable to divide it into two separate tables.

Lines 434–443: This section should be removed, as it does not appear to contribute sufficiently to the presentation of the main empirical findings and may unnecessarily interrupt the flow of the Results section.

Discussion: The Discussion is currently one of the major weaknesses of the manuscript. In its present form, it reads to a considerable extent as a second literature review, rather than as a critical interpretation of the study's findings according to research hypotheses and previous Introduction.

Lines 880 & 917: The asterisk (*) appears to be unnecessary and should be removed if it does not serve a specific reporting or formatting purpose.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The revised manuscript has improved substantially, and the authors have addressed most of the major concerns raised in the previous review. In particular, the revised version now provides sex-specific BMI distributions and obesity counts, uses Enter-method multivariable regression with additional diagnostic information, reports internal consistency estimates for the psychosocial measures, and substantially moderates the interpretation of the ROC-derived cut-offs. The limitations section also now explicitly acknowledges the lack of an independent adiposity criterion, the internally derived nature of the classification thresholds and associated optimism bias, and the mathematical overlap between BMI and BAI.

The revised treatment of the female SWLS ROC result is also more appropriate, as the AUC below .50 is now interpreted as an inverse statistical pattern rather than as evidence of practical classification utility. Overall, the manuscript is considerably clearer and more cautious. A small number of issues nevertheless remain and should be addressed before publication.

1. Sample-size rationale. The manuscript continues to describe the Cochran finite-population calculation as ensuring “adequate statistical power.” This procedure provides a sample-size estimate based on population size, confidence level, expected proportion, and margin of error; it is not a power analysis specifically tailored to the sex-stratified regression or ROC analyses conducted in this study. A new retrospective power analysis is not necessary, but the wording should be revised so that the Cochran calculation is described as a sample-size or precision-based justification rather than as evidence of statistical power.

2. Interpretation of sex-stratified findings. Some statements still imply formal differences between male and female associations. For example, certain relationships are described as stronger among female participants, and the BAI–life satisfaction finding is interpreted as suggesting “sex-specific vulnerabilities.” A statistically significant or numerically larger association in one subgroup does not itself demonstrate that the association differs significantly between sexes. Such statements should either be supported by a formal sex × predictor interaction or reframed as descriptive subgroup patterns.

3. Regression diagnostics. The addition of Tolerance/VIF and Durbin–Watson statistics is useful. However, the manuscript does not clearly report whether other standard linear-regression assumptions were examined, particularly linearity, homoscedasticity, residual distribution, and influential observations. A brief clarification would be sufficient. Durbin–Watson and multicollinearity diagnostics alone do not establish that all regression assumptions have been met.

4. Interpretation of R-2D:4D. The revised manuscript appropriately refers to R-2D:4D as a proxy in several places, but some statements remain overly definitive, including references to “phenotypic expressions of prenatal hormone exposure” and its “established theoretical role as an indirect marker of prenatal testosterone exposure.” A more cautious formulation, such as a putative or proposed proxy that has been hypothesized to reflect prenatal hormonal exposure, would be preferable.

5. Two technical/editorial corrections remain. The Introduction states that BAI estimates adiposity “without requiring height metrics.” However, height is explicitly included in the BAI formula; BAI does not require body weight. This statement should therefore be corrected. In Table 6, the column labelled “95% CI for β” appears to contain confidence intervals for the unstandardized coefficient B. For example, among males, NC is reported as B = .78, SE = .06 and β = .43, whereas the CI is [.67, .89], which corresponds to B rather than β. If so, the column heading should be corrected to “95% CI for B.”

The authors have also now explicitly acknowledged the exploratory nature of the extensive pairwise ROC comparisons and the associated risk of Type I error inflation, which adequately addresses the previous concern regarding multiplicity provided that this cautious interpretation is maintained.

Overall, the authors have responded constructively to the previous review and have substantially improved the methodological transparency and interpretation of the manuscript. The remaining points are relatively limited and can be addressed through targeted clarification and minor revisions.

Author Response

Dear Editor and Reviewer 1,

We would like to express our sincere gratitude to the reviewer for the exceptionally constructive, rigorous, and valuable feedback provided during the evaluation of our manuscript. These insights have significantly contributed to improving the statistical robustness, conceptual clarity, and overall academic quality of our study.

To ensure full transparency and a thorough evaluation, our response documentation has been structured into two distinct sections:

Section A (Point-by-Point Response): Provides a direct, detailed response to each specific comment raised by the reviewer, detailing how and where the respective issues have been addressed in the revised text.
Section B (Comprehensive Amendments Layout): Contains a unified grid of all textual modifications, data updates, and structural adjustments implemented across the entire manuscript. Because some recommendations from the review panel required overlapping updates (such as expanded regression diagnostics, Youden cut-off corrections, or sex-specific qualification boundaries), Section B is identical across all response reports. This allows the reviewer to track the exact context of every modification, even if it was initiated by a request from another member of the evaluation committee.

For efficient editorial tracking, the changes in the revised manuscript have been color-coded as follows:

Text implemented during the first round of revisions remains highlighted in red.
New modifications and additions implemented during this current round of revisions are highlighted in light blue.
All deleted text has been removed entirely from the manuscript layout to ensure clarity and text fluidity.

Part A: Point-by-Point Evaluation Matrices

For your immediate reference, the complete unedited comments from your original review report are provided directly below, followed by our unified, section-by-section response highlighting all implemented modifications.

Open Review

( ) I would not like to sign my review report

(x) I would like to sign my review report

Quality of English Language

( ) The English could be improved to more clearly express the research.

(x) The English is fine and does not require any improvement.

	Yes	Can be improved	Must be improved	Not applicable
Does the introduction provide sufficient background and include all relevant references?	(x)	( )	( )	( )
Is the research design appropriate?	(x)	( )	( )	( )
Are the methods adequately described?	( )	(x)	( )	( )
Are the results clearly presented?	( )	(x)	( )	( )
Are the conclusions supported by the results?	(x)	( )	( )	( )
Are all figures and tables clear and well-presented?	( )	(x)	( )	( )

Comments and Suggestions for Authors

Sample-size rationale. The manuscript continues to describe the Cochran finite-population calculation as ensuring “adequate statistical power.” This procedure provides a sample-size estimate based on population size, confidence level, expected proportion, and margin of error; it is not a power analysis specifically tailored to the sex-stratified regression or ROC analyses conducted in this study. A new retrospective power analysis is not necessary, but the wording should be revised so that the Cochran calculation is described as a sample-size or precision-based justification rather than as evidence of statistical power.

Response regarding sample size rationale: We fully agree with the reviewer’s excellent methodological clarification regarding the distinct mathematical purposes of the Cochran formula and a formal power analysis. Following this valuable recommendation, the terminology has been revised throughout the Materials and Methods section, highlighted in light blue). The Cochran finite-population calculation is now correctly described as a precision-based justification tailored to estimate a representative sample size based on specific margin of error and confidence levels, rather than as an indicator of statistical power for the sex-stratified regression or ROC frameworks.

Furthermore, to provide a more rigorous baseline evaluation of our sample distribution, we have explicitly cross-referenced the descriptive data in Table 1 (detailing participant distribution across body mass index categories) and integrated a parallel sensitivity analysis utilizing the CUN-BAE equation in the Results section (highlighted in light blue). These additions ensure that while the Cochran formula serves as our precision justification, the empirical sample adequacy and the specific under-classification tendencies of BMI are fully transparently reported and contextualized within the text.

To ensure absolute methodological transparency and demonstrate the empirical origin of the reported distribution metrics, we have detailed below the complete, cross-tabulated prevalence data comparing the BMI and CUN-BAE structures for your immediate verification. Please note that to preserve the overall conciseness of the main manuscript and avoid oversaturating the text layout with secondary reference arrays, this descriptive baseline grid has been omitted from the final article body, while the core percentages remain clearly anchored within the text.

Table R3. Cross-Tabulation and Descriptive Distribution of the CUN-BAE Adiposity Categories

Adiposity Category (CUN-BAE)	Male Participants (N = 199) Frequency / (%)	Female Participants (N = 246) Frequency / (%)	Total cohort (N = 445) Frequency / (%)	Chi-Square Testing *(χ²)*
Normal Weight (Males < 20% / Females < 30%)	103 / (51.8)	136 / 55.3	239 / 53.7	χ² = 1.706 df = 2 p = .426
Overweight (Males: ≥ 20-24.9 / Females: 30-34.9%)	34 / 17.1	47 / 19.1	81 / 18.2
Obese (Males ≥ 25% / Females ≥ 35%)	62 / 31.2	63 / 25.6	125 / 28.1

Interpretation of sex-stratified findings. Some statements still imply formal differences between male and female associations. For example, certain relationships are described as stronger among female participants, and the BAI–life satisfaction finding is interpreted as suggesting “sex-specific vulnerabilities.” A statistically significant or numerically larger association in one subgroup does not itself demonstrate that the association differs significantly between sexes. Such statements should either be supported by a formal sex × predictor interaction or reframed as descriptive subgroup patterns.

Response regarding interpretation of sex-stratified findings: We entirely accept this valuable methodological critique regarding subgroup comparisons. Without a formal sex × predictor interaction test, drawing definitive conclusions about formal differences between the two sexes is statistically inappropriate. Following this crucial recommendation, we have completely revised our language throughout the Discussion and Conclusions sections (highlighted in light blue). All definitive and comparative assertions, such as describing relationships as "stronger" or labeling trends as "sex-specific vulnerabilities," have been removed. Instead, these findings have been strictly reframed as localized, descriptive subgroup patterns observed within this specific sample of female participants.

Regression diagnostics. The addition of Tolerance/VIF and Durbin–Watson statistics is useful. However, the manuscript does not clearly report whether other standard linear-regression assumptions were examined, particularly linearity, homoscedasticity, residual distribution, and influential observations. A brief clarification would be sufficient. Durbin–Watson and multicollinearity diagnostics alone do not establish that all regression assumptions have been met.

Response regarding regression diagnostics: We completely accept the reviewer's professional and highly rigorous methodological corrections. We agree that the Durbin–Watson statistic evaluates residual autocorrelation and does not validate normal error distribution; that the central limit theorem rule-of-thumb threshold of 30 cases does not automatically guarantee homoscedasticity or residual normality within localized sub-cohort cells; and that Cochran's equation serves strictly as a precision-based justification for a representative sample size rather than a retrospective power analysis for multi-variable regression or ROC frameworks. Following these crucial clarifications, we have thoroughly updated and corrected all such inaccurate assertions throughout the Materials and Methods, Results, and Discussion sections. All statements have been calibrated to accurately reflect correct statistical terminology, separating residual independence indicators from normal error distribution parameters.

Furthermore, in direct accordance with your request to report the explicit checks for linear framework compliance, we have formally detailed our diagnostic checking procedures within the text. Specifically, residual normality was rigorously verified through visual inspection of the normal standardized residual probability plots (P-P plots), while homoscedasticity and error variance independence were formally confirmed using standardized residual scatterplots (ZRESID vs. ZPRED). Additionally, influential cases were thoroughly monitored using Cook’s distance metrics, mathematically demonstrating that the calculated values across both sex-stratified cohorts remained strictly below the conservative critical threshold of 1.00, fully supporting the stability and robustness of our multivariable equations.

To ensure full analytical transparency and provide absolute proof that these linear regression prerequisites were comprehensively satisfied, we have attached the raw normal standardized residual probability plots (P-P plots), the residual scatterplots, and the verified Cook's distance parameter matrices for both biological sex sub-cohorts directly below for your immediate verification.

Male Sub-Cohort Diagnostics (n = 199)

Figure R1. Normal P-P Plot of Regression Standardized Residual for Male Sub-Cohort (Dependent Variable: BMI)

Figure R2. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Male Sub-Cohort (Dependent Variable: BMI)

Table R1. Influential Cases Parameter Matrix for the Male Sub-Cohort (n = 199)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.090	.005	.012	199
sex = male
Dependent Variable: BMI

Female Sub-Cohort Diagnostics (n = 246)

Figure R3. Normal P-P Plot of Regression Standardized Residual for Female Sub-Cohort (Dependent Variable: BMI)

Figure R4. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Female Sub-Cohort (Dependent Variable: BMI)

Table R2. Influential Cases Parameter Matrix for the Female Sub-Cohort (n = 246)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.130	.005	.012	246
sex = female
Dependent Variable: BMI

Interpretation of R-2D:4D. The revised manuscript appropriately refers to R-2D:4D as a proxy in several places, but some statements remain overly definitive, including references to “phenotypic expressions of prenatal hormone exposure” and its “established theoretical role as an indirect marker of prenatal testosterone exposure.” A more cautious formulation, such as a putative or proposed proxy that has been hypothesized to reflect prenatal hormonal exposure, would be preferable.

Response regarding interpretation of R-2D:4D: We completely agree with the reviewer’s excellent point regarding the need for strict scientific caution when discussing distal developmental biomarkers within cross-sectional structures. Following this valuable recommendation, we have comprehensively revised our language throughout the Discussion and Conclusions sections (highlighted in light blue).

All over-definitive assertions, such as describing the digital ratio as an "indirect marker" or discussing the "phenotypic expressions of prenatal hormone exposure," have been systematically removed. In their place, we have integrated the exact cautious formulations recommended by the reviewer, reframing the right-hand R-2D:4D ratio strictly as a "putative proxy" or a "proposed biomarker that has been hypothesized to reflect early hormonal exposure." This ensures full alignment with the exploratory nature of our study design and prevents any speculative over-interpretation of the biological mechanisms.

Two technical/editorial corrections remain. The Introduction states that BAI estimates adiposity “without requiring height metrics.” However, height is explicitly included in the BAI formula; BAI does not require body weight. This statement should therefore be corrected. In Table 6, the column labelled “95% CI for β” appears to contain confidence intervals for the unstandardized coefficient B. For example, among males, NC is reported as B = .78, SE = .06 and β = .43, whereas the CI is [.67, .89], which corresponds to B rather than β. If so, the column heading should be corrected to “95% CI for B.”

Response regarding the two technical/editorial corrections remain:

We sincerely thank the reviewer for identifying these two critical technical oversights, which have been fully corrected in the revised manuscript (highlighted in light blue):

(a) The conceptual description of the Body Adiposity Index (BAI) framework in the Introduction section has been accurately revised to state that the index estimates adiposity without requiring body weight metrics. We appreciate the reviewer pointing this out; while the formal mathematical equation was explicitly and correctly reported elsewhere in the manuscript, this specific introductory sentence was originally formulated to highlight the reliance on hip circumference rather than body mass, leading to an unintended phrasing error that has now been fully corrected.

(b) The column heading in the multivariable regression diagnostic layout (Table 6) has been corrected to "95% CI for B" to accurately reflect that the reported confidence intervals correspond to the unstandardized coefficients (B) rather than the standardized parameters (β).

The authors have also now explicitly acknowledged the exploratory nature of the extensive pairwise ROC comparisons and the associated risk of Type I error inflation, which adequately addresses the previous concern regarding multiplicity provided that this cautious interpretation is maintained.

Response: We sincerely thank the reviewer for evaluating and validating our updated methodological framing. We completely agree that maintaining a cautious, exploratory interpretation is essential for handling multiplicity boundaries. We assure the reviewer that this prudent approach and strict statistical caution have been fully maintained throughout the final narrative layout of the Discussion section to ensure complete analytical integrity.

Submission Date

19 August 2026

Date of this review

06 Oct 2026 12:54:59

Part B: Authors' Unified Comprehensive Amendments Layout

To ensure absolute review transparency and demonstrate the comprehensive structural adjustments executed across the entire evaluation panel, Part B compiles our complete, sequential responses and mathematical justifications provided directly to the other external reviewers, omitting their individual commentary texts for a more concise overview. This compiled layout is shared to offer full clarity on how concurrent recommendations—including specific analytical and formatting requests not raised within your individual report—were systematically integrated to optimize the final manuscript.

Responses to Reviewer 3:

Response to comment 1: We completely accept the reviewer's highly valid methodological critique regarding our sample size justification. We agree that the generic rule of thumb requiring at least 30 participants per subgroup does not automatically demonstrate the mathematical precision or robustness of the calculated AUC boundaries and optimal cut-off values.

Following this valuable recommendation, we have systematically removed this categorical justification from the manuscript text. Furthermore, we have explicitly integrated a formal acknowledgement of the statistical uncertainty and margin of error associated with relatively small clinical subgroup sample sizes in both our multivariable and ROC estimations directly within the Limitations of the Study section.

Additionally, to address the potential under-classification tendencies of BMI and further strengthen our sample's evaluative framework—and in direct accordance with a valuable recommendation raised by Reviewer 4—we have integrated a parallel sensitivity analysis utilizing the CUN-BAE equation in the Results section. This addition ensures full transparency regarding the prevalence layout of obesity within our lot while directly addressing the descriptive boundaries highlighted by the review panel.

Table R3. Cross-Tabulation and Descriptive Distribution of the CUN-BAE Adiposity Categories

Adiposity Category (CUN-BAE)	Male Participants (N = 199) Frequency / (%)	Female Participants (N = 246) Frequency / (%)	Total cohort (N = 445) Frequency / (%)	Chi-Square Testing *(χ²)*
Normal Weight (Males < 20% / Females < 30%)	103 / (51.8)	136 / 55.3	239 / 53.7	χ² = 1.706 df = 2 p = .426
Overweight (Males: ≥ 20-24.9 / Females: 30-34.9%)	34 / 17.1	47 / 19.1	81 / 18.2
Obese (Males ≥ 25% / Females ≥ 35%)	62 / 31.2	63 / 25.6	125 / 28.1

Response to comment 2: We completely accept the reviewer's professional and highly rigorous methodological corrections. We agree that the Durbin–Watson statistic evaluates residual autocorrelation and does not validate normal error distribution; that the central limit theorem rule-of-thumb threshold of 30 cases does not automatically guarantee homoscedasticity or residual normality within localized sub-cohort cells; and that Cochran's equation serves strictly as a precision-based justification for a representative sample size rather than a retrospective power analysis for multi-variable regression or ROC frameworks. Following these crucial clarifications, we have thoroughly updated and corrected all such inaccurate assertions throughout the Materials and Methods, Results, and Discussion sections. All statements have been calibrated to accurately reflect correct statistical terminology, separating residual independence indicators from normal error distribution parameters.

Male Sub-Cohort Diagnostics (n = 199)

Figure R1. Normal P-P Plot of Regression Standardized Residual for Male Sub-Cohort (Dependent Variable: BMI)

Figure R2. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Male Sub-Cohort (Dependent Variable: BMI)

Table R1. Influential Cases Parameter Matrix for the Male Sub-Cohort (n = 199)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.090	.005	.012	199
sex = male
Dependent Variable: BMI

Female Sub-Cohort Diagnostics (n = 246)

Figure R3. Normal P-P Plot of Regression Standardized Residual for Female Sub-Cohort (Dependent Variable: BMI)

Figure R4. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Female Sub-Cohort (Dependent Variable: BMI)

Table R2. Influential Cases Parameter Matrix for the Female Sub-Cohort (n = 246)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.130	.005	.012	246
sex = female
Dependent Variable: BMI

Response to comment 3: We completely agree with the reviewer that reporting the raw diagnostic counts significantly improves methodological transparency and allows for a more rigorous evaluation of the calculated sensitivity and specificity parameters.

Following this valuable recommendation, we have explicitly expanded the ROC classification framework (Table 6, highlighted in light blue by introducing four additional columns that comprehensively report the exact numbers of true positives (TP), false positives (FP), true negatives (TN), and false negatives (FN) corresponding to the selected optimal Youden cut-off thresholds for each anthropometric index.

Furthermore, a interpretive paragraph has been integrated into the Discussion section, highlighted in light blue) to formally discuss these raw tracking counts. This addition explicitly contextualizes the optimal sensitivity achieved by the female BAI system (correctly isolating all 34 obese individuals alongside 19 false positives) and the male neck circumference metrics (isolating 32 true positives against 4 false negatives and 25 false positives), transparently acknowledging the over-classification tendencies inherent to these somatic thresholds within university student sub-cohorts.

Crucially, in line with the reviewer's call for analytical caution, we have explicitly qualified these high diagnostic parameters within the revised text by explaining that the high sensitivity and specificity metrics are partially driven by mathematical artifact. Because BMI, BAI, and derivative screening metrics share height as a common denominator within their formulaic layout, a high degree of variance overlap is expected by construction. This structural metric interdependence has been critically expanded directly within the main narrative of the Discussion section to ensure the reported ROC performance is interpreted with the necessary statistical prudence.

Response to comment 4: We completely agree with the reviewer’s rigorous methodological critique regarding the limitations of applying a Chi-square framework to evaluate sex distribution assumptions within a convenience sample layout. We accept that this comparison carries no prospective scientific relevance and cannot correct for underlying selection biases or recruitment-related limitations.

Following this clear recommendation, we have completely removed the Chi-square analysis along with its corresponding data matrix from the Results section of the revised manuscript. All speculative remarks suggesting that biological sex stratification resolves sampling or recruitment imbalances have been systematically deleted to ensure a strictly descriptive and transparent reporting layout.

Response to comment 5: We thank the reviewer for highlighting these specific methodological points, which have been fully clarified in the revised manuscript:

(a) The conceptual description of the Body Adiposity Index (BAI) framework in the Introduction section has been corrected to state that the index estimates adiposity without requiring body weight metrics. While the formal mathematical formula reported in the Materials and Methods was entirely correct from the baseline, this introductory phrasing error has been resolved to ensure absolute consistency throughout the paper.

(b) Regarding the calculation of Intraclass Correlation Coefficients (ICC) for measurement reproducibility, we would like to clarify that formal test-retest ICC parameter indices were not computed. Instead, to guarantee somatic measurement precision and minimize intra-observer variability across the student cohort, all anthropometric measurements were rigorously executed by the same trained evaluator. To further optimize data consistency, three consecutive independent recordings were performed for each specific metric during the evaluation, and the exact mathematical mean value of these three trials was calculated and directly logged into the final database framework.

Response to comment 6: We sincerely appreciate the reviewer’s recommendation for enhancing terminological precision. In accordance with this guidance, a thorough, final editorial revision has been systematically executed across the entire manuscript layout. Overly elaborate expressions and non-standard narrative phrasings have been replaced with direct, established statistical terminology. Furthermore, redundant explanations—particularly those regarding administrative and demographic sample breakdowns in the literature cross-references—have been eliminated, and specific ambiguous terms (such as "mirroring") have been systematically removed. These adjustments ensure that the final text strictly prioritizes conceptual clarity, concise academic expression, and accurate statistical reporting guidelines.

In conclusion, the revised version of the manuscript presents significant improvements, and a substantial proportion of the comments raised previously have been adequately addressed. Nevertheless, I consider that the remaining issues, particularly those concerning the justification of sample adequacy, the assessment of regression assumptions, and the transparency of ROC reporting, require further clarification.

Response to final comment: We are deeply grateful to the reviewer for these encouraging remarks regarding the improvements implemented in our previous revision round. In direct response to the remaining focal issues highlighted in this final summary, we hope that these newly added advancements—specifically the explicit qualification of sample uncertainty boundaries within the Limitations, the comprehensive reporting of raw linear regression diagnostics, and the enhanced transparency of the ROC parameters via the inclusion of raw counts (TP, FP, TN, FN)—have now been satisfactorily addressed within this current version. We trust that these systematic additions provide the necessary clarification and fully meet the high standards of the journal.

Responses to Reviewer 4

Response regarding the title: We sincerely thank the reviewer for this excellent recommendation regarding the accuracy of the title layout. We fully agree that broadening the geographical framing enhances both scientific precision and the international relevance of the study, rather than focusing on localized demographic details. Following this valuable suggestion, we have revised the title to incorporate the second option: "…of Students from a Romanian University." We highly appreciate this insight and, in accordance with the reviewer's note, we happily leave the final decision regarding this amendment to the discretion of the Academic Editor.

Response regarding the abstract: We completely agree with the reviewer’s valuable recommendation to streamline the Abstract for enhanced clarity and impact. Following this suggestion, the Results paragraph within the abstract has been carefully revised (highlighted in light blue). We have omitted the extensive statistical listings of R² values and detailed pairwise contrasts, replacing them with clear verbal summaries. Only the primary, key Area Under the Curve (AUC) values have been retained to preserve the core empirical findings while significantly improving the overall readability and conciseness of the introductory summary.

Response regarding the supporting citation: We completely accept the reviewer's rigorous note regarding citation precision. We agree that a general review on obesity cannot support our specific multidimensional tracking structure. Because this tridimensional integration (combining somatic physical metrics, prenatal developmental markers, and subjective psychosocial dimensions) represents our study’s unique conceptual framework, we have followed the reviewer's excellent advice and rephrased the sentence entirely as our own scientific rationale. The text now reads: "Our rationale is that combining these distinct layers allows for an exploratory evaluation of how young adult physical traits and well-being co-vary without implying diagnostic causality." (highlighted in light blue „our rationale”). Concurrently, the reference to the review [3] remains strictly placed at the end of the subsequent sentence, where it originally belongs.

Response regarding the streamlining of methods: We completely agree with the reviewer that historical or validation sample backgrounds from early literature are not required for the direct replication of our protocol. Following this excellent recommendation for structural conciseness, we have comprehensively streamlined the Materials and Methods section (highlighted in light blue):

(a) In Section 2.4, the historical narrative regarding the limitations of earlier well-being measures that led to the development of the SWLS has been removed, retaining strictly the operational parameters of the instrument.

(b) The text describing the 1988 original MSPSS validation cohort characteristics (composition, mean age, and ethnic diversity) has been systematically deleted, focusing the narrative exclusively on the scale's structure, scoring layout, and its specific Romanian validation data.

(c) In Section 2.2, the extended step-by-step mathematical breakdown of the Cochran formula alongside the narrative justification of the sample surplus has been eliminated. The section has been reduced to a concise summary statement reporting the baseline parameters and precision targets utilized to justify the final sample size layout.

Response regarding the structure of the results: We completely accept the reviewer's stylistic recommendations for enhancing the academic layout of the Results section. Following this clear structural guidance, the entire section has been thoroughly re-edited (highlighted in light blue):

(a) Every paragraph opening has been systematically restructured to lead directly with the empirical findings and scientific observations, relocating all references to the respective descriptive data grids strictly into parenthetical citations at the end of the sentences (e.g., removing any repetitive introduction entry such as "Table X presents...").

(b) All interpretive remarks, subjective justifications, and explanatory narratives have been systematically removed from the Results text. Specifically, the sentence explaining the sex enrollment asymmetry has been completely deleted from the manuscript (as the entire baseline Chi-square distribution testing layout was removed in accordance with Reviewer 3's request); the categorical justification regarding the 30-case sub-cohort threshold has been completely removed (aligning also with Reviewer 3's request); and the remaining descriptive metrics have been strictly locked to reflect dry, non-speculative statistical values.

Response regarding the APA style of the tables: We thank the reviewer for this meticulous final check, which has significantly polished the presentation quality of our data. Every identified discrepancy has been thoroughly resolved in the revised manuscript layout:

(a) Table 2 has been completely harmonized with international reporting guidelines by replacing all decimal commas with standard decimal points and modifying the abbreviation "IMC" to "BMI".

(b) The unused single asterisk definition has been systematically deleted from the explanatory footnote of the Pearson correlation matrix (Table 4), as all our statistically significant correlation tracks achieved significance directly at the .01 level (two asteriscs).

(c) The typographical error in Table 6 regarding the male SWLS parameter has been corrected to display the proper statistical metrics.

(d) Following the reviewer's excellent recommendation, Table 1 has been completely removed from the manuscript, and its basic demographic distribution data has been successfully integrated into a single, concise descriptive sentence within the main text. Consequently, all subsequent tables have been systematically renumbered throughout the entire document (e.g., the original Table 2 has now become Table 1, and so forth), ensuring full structural coherence.

Response regarding the Lines 434–443: We completely accept the reviewer's critique regarding textual redundancy. Repeating nearly every empirical coefficient within the narrative text defeats the purpose of providing an integrated data grid. Following this recommendation, the paragraph discussing the Pearson correlation layout has been thoroughly streamlined (now spanning lines 411–437 in the updated draft framework, highlighted in light blue). All repetitive statistical values and individual coefficients have been removed from the running text, retaining strictly a concise introduction that highlights only the key bivariate findings. Furthermore, the duplicated phrasing oversight identified at lines 454–456 ("alongside alongside...") has been systematically deleted to ensure a clean, precise, and professional narrative layout.

Response regarding the Discussion: We highly appreciate the reviewer's comprehensive structural and conceptual guidance for refining the Discussion section. We would like to clarify that introducing distinct bibliographic sources within the discussion matrix (refs. 40–61) was a deliberate academic choice aimed at expanding the literature comparison layout with independent frameworks, rather than duplicating the baseline citations already established in the Introduction.

Furthermore, we would like to highlight that while none of the other reviewers across the evaluation rounds raised any concerns or requested modifications regarding the original layout of the Discussion, we have fully embraced your valuable recommendations to achieve an even higher level of structural balance and thematic focus. Following this guidance, the section has been thoroughly reorganized and streamlined (highlighted in light blue):

The entire narrative has been structured directly around our three primary hypotheses (H1, H2, and H3), explicitly stating within each thematic layer whether the empirical findings supported or qualified the respective assumptions.
The extensive, study-by-study sequential summaries from earlier literature have been significantly condensed, removing unnecessary administrative or cohort-specific details from external studies while strictly retaining the core comparative benchmarks.

This restructuring substantially improves the balance of the section, ensuring that our own findings remain the central focus while preserving a highly documented, respectful literature integration.

Response regarding the Asterisks: We thank the reviewer for identifying this subtle typographical artifact. The residual asterisk in reference 61 ("Curr. Nutr. Rep. 2020, 9, 31–41") has been completely removed from the final bibliography layout. As noted during the previous evaluation phase—where a major batch of similar characters was systematically eliminated—these stray asterisks were structural artifacts automatically injected by the reference management software during the automated style conversion interface into the MDPI bibliographic framework. This final isolated entry has now been thoroughly cleaned, verified, and validated for absolute compliance.

Response regarding male neck circumference cut-off: We are exceptionally grateful to the reviewer for identifying this severe typographical discrepancy. The reviewer’s mathematical deduction is absolutely correct: a cut-off threshold of 33.90 cm is completely incompatible with the reported specificity boundaries and the baseline mean of the male sub-cohort layout.

This error was a major clerical oversight during the final data transfer phase from the statistical software grids. The optimal calculated Youden cut-off threshold for male neck circumference has been verified and fully corrected to its true empirical value across all tables (including Table 6, formerly Table 7) and throughout the narrative text of the Discussion section (highlighted in light blue). This correction restores absolute mathematical harmony and consistency across the entire document layout.

Response regarding the MSPSS score range: We would like to clarify that there is no analytical or mathematical error regarding the reported psychometric boundaries. In direct accordance with the official scoring guidelines established by the original authors of the Multidimensional Scale of Perceived Social Support (MSPSS; Zimet et al.), the final metrics can be interpreted either as a total summation score or as a mean score per item. In our study, we strictly utilized the mean score approach, which is calculated by dividing the total summation value by the 12 individual items of the instrument.

Consequently, on a 7-point Likert response layout (ranging from 1 to 7), the reported sub-cohort mean values (2.23 and 2.32) and maximum thresholds (3.33 and 3.17) are mathematically accurate, fully consistent, and reflect a lower baseline of perceived social support within this specific student cohort. This scoring structure is strictly identical to the official Romanian validation protocol of the scale, executed directly by one of the co-authors of this manuscript. This operational layout guarantees that the inverse bivariate correlation with the SWLS is statistically genuine within this sample data, providing a sound empirical foundation for our subsequent interpretive narrative in the Discussion section.

Response regarding the BMI as reference criterion and the CUN-BAE analysis: We highly appreciate this profound methodological evaluation, which directly addresses the structural core of our analytical boundaries. In accordance with these insights, the manuscript has been thoroughly updated and expanded (highlighted in light blue):

(a) Integration of the CUN-BAE Sensitivity Framework: As recommended, the complete findings derived from the CUN-BAE validation equation have been formally integrated into the Results section. This explicit reporting transparently shows that obesity rates effectively doubled compared to standard BMI tracking, reaching 31.2% in male participants and 25.6% in female participants. This addition provides a much-needed baseline validation layer regarding the under-classification vulnerabilities of BMI.

(b) Clarification on the Omron BF-511 Device Layout: Regarding the bioelectrical impedance analysis (BIA) capabilities of the Omron BF-511 system, we would like to clarify that body fat percentage tracks were not recorded or logged into the study database framework. The device was integrated into the protocol exclusively as a high-quality, verified medical grade system to guarantee maximum accuracy for body weight recordings, preventing its utilization as an independent reference metric for sub-cohort sensitivity testing.

(c) Expanded Discussion on Metric Interdependence and Overlap: In direct agreement with the reviewer’s note, the high mathematical variance overlap (R² fields ranging from .85 to .91) and the exceptionally high AUC tracks have been critically examined. We have expanded our narrative directly within the main text of the Discussion section (rather than relying strictly on the Limitations overview). The text now formally alerts the reader that these high performance metrics are partially driven by mathematical artifact, since BMI, BAI, and derivative anthropometric systems share height parameters as a common baseline scaling denominator by construction.

Table R3. Cross-Tabulation and Descriptive Distribution of the CUN-BAE Adiposity Categories

Adiposity Category (CUN-BAE)	Male Participants (N = 199) Frequency / (%)	Female Participants (N = 246) Frequency / (%)	Total cohort (N = 445) Frequency / (%)	Chi-Square Testing *(χ²)*
Normal Weight (Males < 20% / Females < 30%)	103 / (51.8)	136 / 55.3	239 / 53.7	χ² = 1.706 df = 2 p = .426
Overweight (Males: ≥ 20-24.9 / Females: 30-34.9%)	34 / 17.1	47 / 19.1	81 / 18.2
Obese (Males ≥ 25% / Females ≥ 35%)	62 / 31.2	63 / 25.6	125 / 28.1

Response regarding the Statements about statistical assumptions: We completely accept the reviewer's professional and highly rigorous methodological corrections. We agree that the Durbin–Watson statistic evaluates residual autocorrelation and does not validate normal error distribution; that the central limit theorem rule-of-thumb threshold of 30 cases does not automatically guarantee homoscedasticity or residual normality within localized sub-cohort cells; and that Cochran's equation serves strictly as a precision-based justification for a representative sample size rather than a retrospective power analysis for multi-variable regression or ROC frameworks. Following these crucial clarifications, we have thoroughly updated and corrected all such inaccurate assertions throughout the Materials and Methods, Results, and Discussion sections. All statements have been calibrated to accurately reflect correct statistical terminology, separating residual independence indicators from normal error distribution parameters.

Male Sub-Cohort Diagnostics (n = 199)

Figure R1. Normal P-P Plot of Regression Standardized Residual for Male Sub-Cohort (Dependent Variable: BMI)

Figure R2. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Male Sub-Cohort (Dependent Variable: BMI)

Table R1. Influential Cases Parameter Matrix for the Male Sub-Cohort (n = 199)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.090	.005	.012	199
sex = male
Dependent Variable: BMI

Female Sub-Cohort Diagnostics (n = 246)

Figure R3. Normal P-P Plot of Regression Standardized Residual for Female Sub-Cohort (Dependent Variable: BMI)

Figure R4. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Female Sub-Cohort (Dependent Variable: BMI)

Table R2. Influential Cases Parameter Matrix for the Female Sub-Cohort (n = 246)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.130	.005	.012	246
sex = female
Dependent Variable: BMI

Response regarding the Consistency of the psychosocial conclusions: We completely agree with the reviewer that our structural conclusions regarding the psychosocial parameters must accurately and consistently reflect the sex-stratified bivariate realities. We accept that while the multivariable equations qualified these connections under simultaneous control, the standalone sex-specific bivariate vectors demonstrated a statistically significant inverse association between adiposity indicators (BMI and BAI) and life satisfaction scores strictly within the female sub-cohort, supported by significant localized AUC trajectories.

Following this excellent recommendation, we have thoroughly revised the text across both the Discussion and Conclusions sections (highlighted in light blue). All over-definitive assertions claiming a total "absence of classification capacity" or stating that higher adiposity trackers "do not necessarily track with lower satisfaction parameters" have been systematically removed. In their place, the narrative has been carefully qualified to explicitly highlight these significant descriptive sex-specific bivariate patterns, ensuring full mathematical harmony between our statistical tables and the final interpretive reporting text.

Response regarding the BAI formula: We sincerely thank the reviewer for this final rigorous check regarding the formulaic construction of the Body Adiposity Index (BAI). Every identified phrasing error and mathematical shorthand has been fully resolved in the revised manuscript layout (highlighted in light blue):

(a) The introductory description has been corrected to accurately state that the BAI framework estimates adiposity without requiring body weight metrics (rather than height parameters), removing the unintended phrasing conflict.

(b) In the Materials and Methods section, the operational description has been fully detailed to explicitly state that the Body Adiposity Index is calculated by dividing hip circumference by height raised to the power of 1.5, followed by the subtraction of the constant factor of 18, exactly matching the formal mathematical equation presented in the text.

Response regarding the Reference accuracy: We sincerely thank the reviewer for this crucial and meticulous quality check on our bibliographic integration. All identified citation discrepancies have been fully resolved in the revised manuscript layout (highlighted in light blue):

(a) The erroneous placement of reference [26] at line 254 has been corrected, removing it from the obesity-related mortality narrative where it was mistakenly injected.

(b) Reference [7] has been systematically removed from the text regarding student weight gain dynamics, as it was a stray entry from the initial drafting phase. It has been replaced with an accurate, highly specific bibliographic source, and the respective sentence has been slightly rephrased to optimize conceptual precision. The new entry has been explicitly highlighted in light blue directly within the final reference list layout.

(c) References [32] and [33], which were accidentally omitted from the main text during the previous structural editing layout, have been fully restored and anchored back into their correct positions within the narrative context where they originally belonged.

Response regarding the Quality of English Language: We highly appreciate the reviewer’s meticulous linguistic evaluation and specific line-by-line pointers, which have been instrumental in refining the professional quality of the text. A thorough final editorial check has been executed to ensure strict compliance with established scientific writing standards.

All non-standard and overly elaborate expressions have been systematically removed and replaced with conventional terminology throughout the revised manuscript (highlighted in light blue); for instance, "alpha level risk vector" has been adjusted to standard statistical significance thresholds, "Likert architecture" to response format layout, "convenience sample lot" to convenience sample or cohort, "area trace" to AUC trajectory, and "structurally mirrored" to tracked consistently.

Furthermore, all identified mechanical and typographical errors have been corrected: the duplicated conjunctions ("and and"), the missing spacing artifact ("indicatorutilized"), the repeated phrasing layout, the sentence fragment, and the misplaced punctuation marks before parenthetical citations have all been completely resolved to ensure a clean, precise, and professional presentation.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Thanks for your efforts.

Best Regards

Author Response

Dear Reviewer 2,

We sincerely thank you for your time, valuable feedback throughout the review process, and positive final evaluation of our manuscript. We greatly appreciate your encouraging comments and well wishes.

Best regards,

The Authors

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Dear authors,

I appreciate the efforts made to revise the manuscript and to address the comments previously raised. The revised version presents important improvements, but I consider that certain methodological and reporting issues have not been fully clarified and require further attention.

Specifically:

Although the distribution of participants across BMI categories is now appropriately reported, the justification of sample adequacy remains questionable. It is argued that the presence of at least 30 participants with obesity in each sex subgroup ensures the statistical stability of the ROC analyses and multivariable models. However, this criterion does not, in itself, demonstrate the precision or robustness of the AUC estimates and optimal cut-off values. I recommend removing this categorical justification and acknowledging the uncertainty associated with the relatively small number of obesity cases in each subgroup.
The assessment of linear regression assumptions remains incomplete. Although the inclusion of confidence intervals and additional effect size measures represents an improvement, these do not replace appropriate regression model diagnostics. I consider that it should be specified whether linearity, homoscedasticity, residual distribution, and the presence of influential observations were assessed. If these checks were not performed, this limitation should be explicitly acknowledged.
The interpretation of the ROC analyses is now more cautious, which represents an important improvement. However, given the very high discriminatory performance reported for certain anthropometric indicators, I continue to recommend reporting the numbers of true positives, false positives, true negatives, and false negatives corresponding to the selected cut-off values. This information would improve the transparency of the analyses and allow a more rigorous evaluation of the reported sensitivity and specificity.
The manuscript retains the Chi-square test comparing the observed sex distribution with a hypothetical 50:50 distribution. The scientific relevance of this comparison remains unclear in the context of a convenience sample. Furthermore, stratifying the analyses by sex does not correct potential selection bias. I recommend removing this analysis or providing a stronger methodological justification for its use, while avoiding the suggestion that stratification resolves recruitment-related limitations.
Some methodological descriptions require further clarification. For example, the Introduction suggests that BAI can be assessed without using height, although the formula presented in the Materials and Methods section explicitly includes this variable. This inconsistency should be corrected. In addition, I recommend clarifying how the intraclass correlation coefficients (ICC) used to assess the reproducibility of anthropometric measurements were calculated.
Although the manuscript has undergone linguistic revision, some overly elaborate expressions or formulations that do not conform to established statistical terminology remain. I recommend a final editorial revision focusing on clarity of expression, terminological precision, and the elimination of redundant wording.

Author Response

Dear Editor and Reviewer 3,

To ensure full transparency and a thorough evaluation, our response documentation has been structured into two distinct sections:

Section A (Point-by-Point Response): Provides a direct, detailed response to each specific comment raised by the reviewer, detailing how and where the respective issues have been addressed in the revised text.
Section B (Comprehensive Amendments Layout): Contains a unified grid of all textual modifications, data updates, and structural adjustments implemented across the entire manuscript. Because some recommendations from the review panel required overlapping updates (such as expanded regression diagnostics, Youden cut-off corrections, or sex-specific qualification boundaries), Section B is identical across all response reports. This allows the reviewer to track the exact context of every modification, even if it was initiated by a request from another member of the evaluation committee.

For efficient editorial tracking, the changes in the revised manuscript have been color-coded as follows:

Text implemented during the first round of revisions remains highlighted in red.
New modifications and additions implemented during this current round of revisions are highlighted in light blue.
All deleted text has been removed entirely from the manuscript layout to ensure clarity and text fluidity.

Part A: Point-by-Point Evaluation Matrices

Open Review

(x) I would not like to sign my review report

( ) I would like to sign my review report

Quality of English Language

(x) The English could be improved to more clearly express the research.

( ) The English is fine and does not require any improvement.

	Yes	Can be improved	Must be improved	Not applicable
Does the introduction provide sufficient background and include all relevant references?	( )	(x)	( )	( )
Is the research design appropriate?	( )	(x)	( )	( )
Are the methods adequately described?	( )	( )	(x)	( )
Are the results clearly presented?	( )	(x)	( )	( )
Are the conclusions supported by the results?	( )	( )	(x)	( )
Are all figures and tables clear and well-presented?	( )	(x)	( )	( )

Comments and Suggestions for Authors

Dear authors,

Specifically:

Although the distribution of participants across BMI categories is now appropriately reported, the justification of sample adequacy remains questionable. It is argued that the presence of at least 30 participants with obesity in each sex subgroup ensures the statistical stability of the ROC analyses and multivariable models. However, this criterion does not, in itself, demonstrate the precision or robustness of the AUC estimates and optimal cut-off values. I recommend removing this categorical justification and acknowledging the uncertainty associated with the relatively small number of obesity cases in each subgroup.

Response regarding the comment 1: We completely accept the reviewer's highly valid methodological critique regarding our sample size justification. We agree that the generic rule of thumb requiring at least 30 participants per subgroup does not automatically demonstrate the mathematical precision or robustness of the calculated AUC boundaries and optimal cut-off values.

Table R3. Cross-Tabulation and Descriptive Distribution of the CUN-BAE Adiposity Categories

Adiposity Category (CUN-BAE)	Male Participants (N = 199) Frequency / (%)	Female Participants (N = 246) Frequency / (%)	Total cohort (N = 445) Frequency / (%)	Chi-Square Testing *(χ²)*
Normal Weight (Males < 20% / Females < 30%)	103 / (51.8)	136 / 55.3	239 / 53.7	χ² = 1.706 df = 2 p = .426
Overweight (Males: ≥ 20-24,9 / Females: 30-34,9%)	34 / 17.1	47 / 19.1	81 / 18.2
Obese (Males ≥ 25% / Females ≥ 35%)	62 / 31.2	63 / 25.6	125 / 28.1

The assessment of linear regression assumptions remains incomplete. Although the inclusion of confidence intervals and additional effect size measures represents an improvement, these do not replace appropriate regression model diagnostics. I consider that it should be specified whether linearity, homoscedasticity, residual distribution, and the presence of influential observations were assessed. If these checks were not performed, this limitation should be explicitly acknowledged.

Response regarding the comment 2: We completely accept the reviewer's professional and highly rigorous methodological corrections. We agree that the Durbin–Watson statistic evaluates residual autocorrelation and does not validate normal error distribution; that the central limit theorem rule-of-thumb threshold of 30 cases does not automatically guarantee homoscedasticity or residual normality within localized sub-cohort cells; and that Cochran's equation serves strictly as a precision-based justification for a representative sample size rather than a retrospective power analysis for multi-variable regression or ROC frameworks. Following these crucial clarifications, we have thoroughly updated and corrected all such inaccurate assertions throughout the Materials and Methods, Results, and Discussion sections. All statements have been calibrated to accurately reflect correct statistical terminology, separating residual independence indicators from normal error distribution parameters.

Male Sub-Cohort Diagnostics (n = 199)

Figure R1. Normal P-P Plot of Regression Standardized Residual for Male Sub-Cohort (Dependent Variable: BMI)

Figure R2. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Male Sub-Cohort (Dependent Variable: BMI)

Table R1. Influential Cases Parameter Matrix for the Male Sub-Cohort (n = 199)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.090	.005	.012	199
sex = male
Dependent Variable: BMI

Female Sub-Cohort Diagnostics (n = 246)

Figure R3. Normal P-P Plot of Regression Standardized Residual for Female Sub-Cohort (Dependent Variable: BMI)

Figure R4. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Female Sub-Cohort (Dependent Variable: BMI)

Table R2. Influential Cases Parameter Matrix for the Female Sub-Cohort (n = 246)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.130	.005	.012	246
sex = female
Dependent Variable: BMI

The interpretation of the ROC analyses is now more cautious, which represents an important improvement. However, given the very high discriminatory performance reported for certain anthropometric indicators, I continue to recommend reporting the numbers of true positives, false positives, true negatives, and false negatives corresponding to the selected cut-off values. This information would improve the transparency of the analyses and allow a more rigorous evaluation of the reported sensitivity and specificity.

Response regarding the comment 3: We completely agree with the reviewer that reporting the raw diagnostic counts significantly improves methodological transparency and allows for a more rigorous evaluation of the calculated sensitivity and specificity parameters.

The manuscript retains the Chi-square test comparing the observed sex distribution with a hypothetical 50:50 distribution. The scientific relevance of this comparison remains unclear in the context of a convenience sample. Furthermore, stratifying the analyses by sex does not correct potential selection bias. I recommend removing this analysis or providing a stronger methodological justification for its use, while avoiding the suggestion that stratification resolves recruitment-related limitations.

Response regarding the comment 4: We completely agree with the reviewer’s rigorous methodological critique regarding the limitations of applying a Chi-square framework to evaluate sex distribution assumptions within a convenience sample layout. We accept that this comparison carries no prospective scientific relevance and cannot correct for underlying selection biases or recruitment-related limitations.

Some methodological descriptions require further clarification. For example, the Introduction suggests that BAI can be assessed without using height, although the formula presented in the Materials and Methods section explicitly includes this variable. This inconsistency should be corrected. In addition, I recommend clarifying how the intraclass correlation coefficients (ICC) used to assess the reproducibility of anthropometric measurements were calculated.

Response regarding the comment 5: We thank the reviewer for highlighting these specific methodological points, which have been fully clarified in the revised manuscript:

Although the manuscript has undergone linguistic revision, some overly elaborate expressions or formulations that do not conform to established statistical terminology remain. I recommend a final editorial revision focusing on clarity of expression, terminological precision, and the elimination of redundant wording.

Response regarding the comment 6: We sincerely appreciate the reviewer’s recommendation for enhancing terminological precision. In accordance with this guidance, a thorough, final editorial revision has been systematically executed across the entire manuscript layout. Overly elaborate expressions and non-standard narrative phrasings have been replaced with direct, established statistical terminology. Furthermore, redundant explanations—particularly those regarding administrative and demographic sample breakdowns in the literature cross-references—have been eliminated, and specific ambiguous terms (such as "mirroring") have been systematically removed. These adjustments ensure that the final text strictly prioritizes conceptual clarity, concise academic expression, and accurate statistical reporting guidelines.

Response: We are deeply grateful to the reviewer for these encouraging remarks regarding the improvements implemented in our previous revision round. In direct response to the remaining focal issues highlighted in this final summary, we hope that these newly added advancements—specifically the explicit qualification of sample uncertainty boundaries within the Limitations, the comprehensive reporting of raw linear regression diagnostics, and the enhanced transparency of the ROC parameters via the inclusion of raw counts (TP, FP, TN, FN)—have now been satisfactorily addressed within this current version. We trust that these systematic additions provide the necessary clarification and fully meet the high standards of the journal.

Submission Date

19 August 2026

Date of this review

09 Oct 2026 21:30:37

Part B: Authors' Unified Comprehensive Amendments Layout

Responses to Reviewer 1

Table R3. Cross-Tabulation and Descriptive Distribution of the CUN-BAE Adiposity Categories

Adiposity Category (CUN-BAE)	Male Participants (N = 199) Frequency / (%)	Female Participants (N = 246) Frequency / (%)	Total cohort (N = 445) Frequency / (%)	Chi-Square Testing *(χ²)*
Normal Weight (Males < 20% / Females < 30%)	103 / (51.8)	136 / 55.3	239 / 53.7	χ² = 1.706 df = 2 p = .426
Overweight (Males: ≥ 20-24,9 / Females: 30-34,9%)	34 / 17.1	47 / 19.1	81 / 18.2
Obese (Males ≥ 25% / Females ≥ 35%)	62 / 31.2	63 / 25.6	125 / 28.1

Male Sub-Cohort Diagnostics (n = 199)

Figure R1. Normal P-P Plot of Regression Standardized Residual for Male Sub-Cohort (Dependent Variable: BMI)

Figure R2. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Male Sub-Cohort (Dependent Variable: BMI)

Table R1. Influential Cases Parameter Matrix for the Male Sub-Cohort (n = 199)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.090	.005	.012	199
sex = male
Dependent Variable: BMI

Female Sub-Cohort Diagnostics (n = 246)

Figure R3. Normal P-P Plot of Regression Standardized Residual for Female Sub-Cohort (Dependent Variable: BMI)

Figure R4. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Female Sub-Cohort (Dependent Variable: BMI)

Table R2. Influential Cases Parameter Matrix for the Female Sub-Cohort (n = 246)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.130	.005	.012	246
sex = female
Dependent Variable: BMI

Response regarding the two technical/editorial corrections remain:

We sincerely thank the reviewer for identifying these two critical technical oversights, which have been fully corrected in the revised manuscript (highlighted in light blue):

Response to final comment: We sincerely thank the reviewer for evaluating and validating our updated methodological framing. We completely agree that maintaining a cautious, exploratory interpretation is essential for handling multiplicity boundaries. We assure the reviewer that this prudent approach and strict statistical caution have been fully maintained throughout the final narrative layout of the Discussion section to ensure complete analytical integrity.

Response to Reviewer 4

(a) Table 2 has been completely harmonized with international reporting guidelines by replacing all decimal commas with standard decimal points and modifying the abbreviation "IMC" to "BMI".

(c) The typographical error in Table 6 regarding the male SWLS parameter has been corrected to display the proper statistical metrics.

The entire narrative has been structured directly around our three primary hypotheses (H1, H2, and H3), explicitly stating within each thematic layer whether the empirical findings supported or qualified the respective assumptions.
The extensive, study-by-study sequential summaries from earlier literature have been significantly condensed, removing unnecessary administrative or cohort-specific details from external studies while strictly retaining the core comparative benchmarks.

This restructuring substantially improves the balance of the section, ensuring that our own findings remain the central focus while preserving a highly documented, respectful literature integration.

Table R3. Cross-Tabulation and Descriptive Distribution of the CUN-BAE Adiposity Categories

Adiposity Category (CUN-BAE)	Male Participants (N = 199) Frequency / (%)	Female Participants (N = 246) Frequency / (%)	Total cohort (N = 445) Frequency / (%)	Chi-Square Testing *(χ²)*
Normal Weight (Males < 20% / Females < 30%)	103 / (51.8)	136 / 55.3	239 / 53.7	χ² = 1.706 df = 2 p = .426
Overweight (Males: ≥ 20-24,9 / Females: 30-34,9%)	34 / 17.1	47 / 19.1	81 / 18.2
Obese (Males ≥ 25% / Females ≥ 35%)	62 / 31.2	63 / 25.6	125 / 28.1

Male Sub-Cohort Diagnostics (n = 199)

Figure R1. Normal P-P Plot of Regression Standardized Residual for Male Sub-Cohort (Dependent Variable: BMI)

Figure R2. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Male Sub-Cohort (Dependent Variable: BMI)

Table R1. Influential Cases Parameter Matrix for the Male Sub-Cohort (n = 199)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.090	.005	.012	199
sex = male
Dependent Variable: BMI

Female Sub-Cohort Diagnostics (n = 246)

Figure R3. Normal P-P Plot of Regression Standardized Residual for Female Sub-Cohort (Dependent Variable: BMI)

Figure R4. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Female Sub-Cohort (Dependent Variable: BMI)

Table R2. Influential Cases Parameter Matrix for the Female Sub-Cohort (n = 246)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.130	.005	.012	246
sex = female
Dependent Variable: BMI

(a) The erroneous placement of reference [26] at line 254 has been corrected, removing it from the obesity-related mortality narrative where it was mistakenly injected.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

I thank the authors for their revised manuscript and for the detailed response letter. Several of my earlier comments have been addressed satisfactorily:

The sample size notation in the Abstract has been corrected and the keywords limited to five.
The citation format at lines 92 and 94 has been corrected.
The date of ethical approval has been added.
The parentheses in the statistical reporting have been corrected.
The research hypotheses have been reformulated more concisely.

I also appreciate the overall moderation of causal and clinical language throughout the manuscript. However, the main structural concerns I raised in the first round have been addressed only partially or not at all. These concern the Methods, Results and Discussion. In addition, on closer reading of the revised tables, I identified some inconsistencies in the reported data that need to be clarified. My comments follow below. The first part covers my previous points; the second covers issues arising from the revised version.

Part A: Previously raised points

Title (comment 1). I understand the authors' wish to keep the title general. However, each reviewer's comments should be weighed on their merits, and the decision belongs to the Academic Editor. My point concerns accuracy, not style. A minimal change would be sufficient, for example "…of Romanian University Students" or "…of Students from a Romanian University." I leave the final decision to the Editor.
Abstract (comment 2). Implemented in part. The Results paragraph still reports R², AUC values and pairwise contrasts. I suggest keeping only the key AUC values and replacing the remaining statistics with verbal summaries.
Supporting citation (comment 5). The citation added (ref. [3]) is a general review on the causes and consequences of obesity. It does not support the specific statement that combining anthropometric, developmental and psychological layers allows an exploratory evaluation of their co-variation. Please provide a source that supports this claim, or rephrase the sentence as the authors' own rationale.
Streamlining of Methods (comments 7, 9, 10). My request did not concern the device specifications or measurement procedures, which are indeed needed for replication. It concerned the theoretical and historical material, which has largely been retained:
- In Section 2.4, lines 279–283 describe the limitations of earlier well-being measures that led to the development of the SWLS.
- Lines 297–301 describe the composition, mean age and ethnic diversity of the original MSPSS validation sample from 1988.
- Section 2.2 still presents the full step-by-step Cochran calculation and a justification of the "deliberate surplus."
None of this is required for replication. A short description of each instrument (number of items, response format, scoring, and the Romanian validation with its reliability) would be enough. The Cochran calculation could be reduced to one sentence.
Structure of the Results (comment 13). The requested restructuring has not been carried out. Almost every Results paragraph still opens with a description of the table, for example "Table 1 presents…" (line 376), "Table 2 provides…" (line 386), "Table 3 summarizes…" (line 404), "Table 6 displays…" (line 479), and "Table 8 details…" (line 558). Each paragraph should instead begin with the empirical finding and refer to the table in parentheses. The Results also still contain interpretation that belongs in the Discussion, for example at lines 380–382, 388–390 and 547–549.
APA style of the tables (comment 13). Further harmonisation is needed:
- Table 2 uses decimal commas and the abbreviation "IMC," while the rest of the manuscript uses decimal points and "BMI."
- The note to Table 5 defines a single asterisk that is never used.
- Table 6 contains a typographical error ("t = −14" for male SWLS).
- Table 1 adds little information and could be replaced by a sentence in the text.
Lines 434–443 (comment 16). The two responses to this comment refer to different sections. Response 16 refers to the narrative of the correlation table, while point 28 of the global matrix refers to the narrative of the regression table. To clarify: my comment concerned the paragraph that repeated the correlation values already shown in the table. I accept that each table must be introduced in the text, but the introduction should be brief and highlight only the key findings. The current paragraph (lines 443–473) still repeats nearly every coefficient. It also contains a duplicated sentence at lines 454–456 ("alongside alongside a weaker positive correlation between the R-2D:4D ratio and BAI").
Discussion (comment 17). I fully agree that the Discussion should compare the findings with the existing literature, and I did not suggest removing references. My concern is the balance between the two. Lines 719–840 still consist mainly of sequential summaries of individual studies. These describe each study's sample, design and conclusions, often in detail that is not needed to interpret the present findings. Many of these sources (refs. 40–61) appear only in the Discussion and are not connected to the Introduction. Furthermore, the Discussion never states explicitly whether H1, H2 and H3 were supported. I recommend organising the Discussion around the three hypotheses. Each section should state the finding, whether it supports the hypothesis, and how it compares with two or three key studies. The remaining study-by-study summaries should be condensed.
Asterisks (comment 18). One asterisk remains in reference 61 ("Curr. Nutr. Rep. 2020, 9*, 31–41").

Part B: Issues identified in the revised version

These points concern the accuracy of the reported data and the validity of the conclusions. I raise them because they need to be resolved before the findings can be interpreted, not to expand the scope of the study.

Male neck circumference cut-off. Table 7 reports an optimal cut-off of 33.90 cm for men, with a sensitivity of 88.9% and a specificity of 84.7%. Table 3, however, gives a male mean NC of 38.46 cm (SD = 2.61, minimum = 30.20 cm). A threshold almost 1.75 SD below the mean would classify nearly all men as positive, which is incompatible with a specificity of about 85%. Please verify this value, which is also repeated in the Discussion (line 686).
MSPSS score range. I note the authors' statement that the item coding was verified. However, the method describes 12 items rated on a 7-point scale, while the reported means are 2.23 and 2.32 and the maxima are 3.33 and 3.17. These values are hard to reconcile with that response format. Please state the exact response scale and scoring of the Romanian version used (total or mean score; range). If the scoring differs from the original, please describe it. This is important because the negative SWLS–MSPSS correlation, which contradicts the literature the authors cite [35,36], forms the basis of an extended interpretation in the Discussion.
BMI as reference criterion and the CUN-BAE analysis. The response letter (point 9) reports that a CUN-BAE analysis yielded obesity rates about twice those based on BMI (31.2% in men and 25.6% in women). The authors also acknowledge that BMI underestimates adiposity in young adults. This analysis is not reported in the manuscript, yet it is directly relevant to interpreting the ROC results. In addition, the Omron BF-511 used to measure weight is a bioelectrical impedance device that provides body fat percentage. Please (a) report the CUN-BAE findings in the manuscript, and (b) state whether BIA-derived body fat was recorded. If it was, I recommend using it as an independent reference criterion, at least as a sensitivity analysis. Because BMI and BAI share height in their formulas, the very high R² values (.85–.91) and AUCs largely reflect overlap by construction. This should be discussed in the Discussion, not only mentioned in the Limitations.
Statements about statistical assumptions. Several statements in the manuscript and the response letter are inaccurate:

The Durbin–Watson statistic tests autocorrelation of residuals. It does not show that residuals are normally distributed or that the "error terms satisfy the standard multivariable linear assumptions" (lines 605–609, 668–670).
A minimum of 30 cases per cell does not guarantee "normal distribution of error terms" (response 9).
Cochran's formula estimates a proportion. It does not guarantee statistical power for regression or ROC analyses.

Please report checks of residual normality, homoscedasticity and influential cases, and correct these statements.

Consistency of the psychosocial conclusions. In women, BMI and BAI correlate significantly with SWLS (r = −.175 and −.232), and SWLS shows a significant AUC. Yet lines 688–691 state that the psychological constructs show "no statistically significant classification capacity." Lines 845–847 conclude that students with higher BMI "do not necessarily report lower life satisfaction." These statements should be qualified to reflect the sex-specific bivariate findings.
BAI formula. Lines 85–86 state that BAI requires no height measurements. This is incorrect: BAI uses height and does not use weight. Line 260 describes BAI as hip circumference divided by height. The formula is actually hip circumference divided by height to the power 1.5, minus 18.
Reference accuracy. Some citations do not match their content:

Ref. [26], a study of the SWLS, is cited for obesity-related mortality (line 254).
Ref. [7] is cited for weight gain in students compared with non-students (line 74).
Refs. [32] and [33] do not appear to be cited in the text.

Comments on the Quality of English Language

The English has improved, but it still needs careful editing. Several expressions remain unusual for scientific writing, for example "alpha level risk vector," "Likert architecture," "convenience sample lot," "area trace" and "structurally mirrored." There are also some errors: "and and" (line 246), "indicatorutilized" (line 249), a duplicated phrase at line 315, a sentence fragment at line 710, and misplaced periods before citations at lines 647 and 651.

Author Response

Dear Editor and Reviewer 4,

To ensure full transparency and a thorough evaluation, our response documentation has been structured into two distinct sections:

Section A (Point-by-Point Response): Provides a direct, detailed response to each specific comment raised by the reviewer, detailing how and where the respective issues have been addressed in the revised text.
Section B (Comprehensive Amendments Layout): Contains a unified grid of all textual modifications, data updates, and structural adjustments implemented across the entire manuscript. Because some recommendations from the review panel required overlapping updates (such as expanded regression diagnostics, Youden cut-off corrections, or sex-specific qualification boundaries), Section B is identical across all response reports. This allows the reviewer to track the exact context of every modification, even if it was initiated by a request from another member of the evaluation committee.

For efficient editorial tracking, the changes in the revised manuscript have been color-coded as follows:

Text implemented during the first round of revisions remains highlighted in red.
New modifications and additions implemented during this current round of revisions are highlighted in light blue.
All deleted text has been removed entirely from the manuscript layout to ensure clarity and text fluidity.

Part A: Point-by-Point Evaluation Matrices

Open Review

(x) I would not like to sign my review report

( ) I would like to sign my review report

Quality of English Language

(x) The English could be improved to more clearly express the research.

( ) The English is fine and does not require any improvement.

	Yes	Can be improved	Must be improved	Not applicable
Does the introduction provide sufficient background and include all relevant references?	( )	( )	(x)	( )
Is the research design appropriate?	( )	( )	(x)	( )
Are the methods adequately described?	( )	( )	(x)	( )
Are the results clearly presented?	( )	( )	(x)	( )
Are the conclusions supported by the results?	( )	( )	(x)	( )
Are all figures and tables clear and well-presented?	( )	( )	(x)	( )

Comments and Suggestions for Authors

I thank the authors for their revised manuscript and for the detailed response letter. Several of my earlier comments have been addressed satisfactorily:

The sample size notation in the Abstract has been corrected and the keywords limited to five.
The citation format at lines 92 and 94 has been corrected.
The date of ethical approval has been added.
The parentheses in the statistical reporting have been corrected.
The research hypotheses have been reformulated more concisely.

Part A: Previously raised points

Title (comment 1). I understand the authors' wish to keep the title general. However, each reviewer's comments should be weighed on their merits, and the decision belongs to the Academic Editor. My point concerns accuracy, not style. A minimal change would be sufficient, for example "…of Romanian University Students" or "…of Students from a Romanian University." I leave the final decision to the Editor.

Abstract (comment 2). Implemented in part. The Results paragraph still reports R², AUC values and pairwise contrasts. I suggest keeping only the key AUC values and replacing the remaining statistics with verbal summaries.

Supporting citation (comment 5). The citation added (ref. [3]) is a general review on the causes and consequences of obesity. It does not support the specific statement that combining anthropometric, developmental and psychological layers allows an exploratory evaluation of their co-variation. Please provide a source that supports this claim, or rephrase the sentence as the authors' own rationale.

Streamlining of Methods (comments 7, 9, 10). My request did not concern the device specifications or measurement procedures, which are indeed needed for replication. It concerned the theoretical and historical material, which has largely been retained:

In Section 2.4, lines 279–283 describe the limitations of earlier well-being measures that led to the development of the SWLS.
Lines 297–301 describe the composition, mean age and ethnic diversity of the original MSPSS validation sample from 1988.
Section 2.2 still presents the full step-by-step Cochran calculation and a justification of the "deliberate surplus."

None of this is required for replication. A short description of each instrument (number of items, response format, scoring, and the Romanian validation with its reliability) would be enough. The Cochran calculation could be reduced to one sentence.

Structure of the Results (comment 13). The requested restructuring has not been carried out. Almost every Results paragraph still opens with a description of the table, for example "Table 1 presents…" (line 376), "Table 2 provides…" (line 386), "Table 3 summarizes…" (line 404), "Table 6 displays…" (line 479), and "Table 8 details…" (line 558). Each paragraph should instead begin with the empirical finding and refer to the table in parentheses. The Results also still contain interpretation that belongs in the Discussion, for example at lines 380–382, 388–390 and 547–549.

APA style of the tables (comment 13). Further harmonisation is needed:

Table 2 uses decimal commas and the abbreviation "IMC," while the rest of the manuscript uses decimal points and "BMI."
The note to Table 5 defines a single asterisk that is never used.
Table 6 contains a typographical error ("t = −14" for male SWLS).
Table 1 adds little information and could be replaced by a sentence in the text.

(a) Table 2 has been completely harmonized with international reporting guidelines by replacing all decimal commas with standard decimal points and modifying the abbreviation "IMC" to "BMI".

(c) The typographical error in Table 6 regarding the male SWLS parameter has been corrected to display the proper statistical metrics.

Lines 434–443 (comment 16). The two responses to this comment refer to different sections. Response 16 refers to the narrative of the correlation table, while point 28 of the global matrix refers to the narrative of the regression table. To clarify: my comment concerned the paragraph that repeated the correlation values already shown in the table. I accept that each table must be introduced in the text, but the introduction should be brief and highlight only the key findings. The current paragraph (lines 443–473) still repeats nearly every coefficient. It also contains a duplicated sentence at lines 454–456 ("alongside alongside a weaker positive correlation between the R-2D:4D ratio and BAI").

Discussion (comment 17). I fully agree that the Discussion should compare the findings with the existing literature, and I did not suggest removing references. My concern is the balance between the two. Lines 719–840 still consist mainly of sequential summaries of individual studies. These describe each study's sample, design and conclusions, often in detail that is not needed to interpret the present findings. Many of these sources (refs. 40–61) appear only in the Discussion and are not connected to the Introduction. Furthermore, the Discussion never states explicitly whether H1, H2 and H3 were supported. I recommend organising the Discussion around the three hypotheses. Each section should state the finding, whether it supports the hypothesis, and how it compares with two or three key studies. The remaining study-by-study summaries should be condensed.

The entire narrative has been structured directly around our three primary hypotheses (H1, H2, and H3), explicitly stating within each thematic layer whether the empirical findings supported or qualified the respective assumptions.
The extensive, study-by-study sequential summaries from earlier literature have been significantly condensed, removing unnecessary administrative or cohort-specific details from external studies while strictly retaining the core comparative benchmarks.

This restructuring substantially improves the balance of the section, ensuring that our own findings remain the central focus while preserving a highly documented, respectful literature integration.

Asterisks (comment 18). One asterisk remains in reference 61 ("Curr. Nutr. Rep. 2020, 9*, 31–41").

Part B: Issues identified in the revised version

Male neck circumference cut-off. Table 7 reports an optimal cut-off of 33.90 cm for men, with a sensitivity of 88.9% and a specificity of 84.7%. Table 3, however, gives a male mean NC of 38.46 cm (SD = 2.61, minimum = 30.20 cm). A threshold almost 1.75 SD below the mean would classify nearly all men as positive, which is incompatible with a specificity of about 85%. Please verify this value, which is also repeated in the Discussion (line 686).

Response regarding the male neck circumference cut-off: We are exceptionally grateful to the reviewer for identifying this severe typographical discrepancy. The reviewer’s mathematical deduction is absolutely correct: a cut-off threshold of 33.90 cm is completely incompatible with the reported specificity boundaries and the baseline mean of the male sub-cohort layout.

MSPSS score range. I note the authors' statement that the item coding was verified. However, the method describes 12 items rated on a 7-point scale, while the reported means are 2.23 and 2.32 and the maxima are 3.33 and 3.17. These values are hard to reconcile with that response format. Please state the exact response scale and scoring of the Romanian version used (total or mean score; range). If the scoring differs from the original, please describe it. This is important because the negative SWLS–MSPSS correlation, which contradicts the literature the authors cite [35,36], forms the basis of an extended interpretation in the Discussion.

BMI as reference criterion and the CUN-BAE analysis. The response letter (point 9) reports that a CUN-BAE analysis yielded obesity rates about twice those based on BMI (31.2% in men and 25.6% in women). The authors also acknowledge that BMI underestimates adiposity in young adults. This analysis is not reported in the manuscript, yet it is directly relevant to interpreting the ROC results. In addition, the Omron BF-511 used to measure weight is a bioelectrical impedance device that provides body fat percentage. Please (a) report the CUN-BAE findings in the manuscript, and (b) state whether BIA-derived body fat was recorded. If it was, I recommend using it as an independent reference criterion, at least as a sensitivity analysis. Because BMI and BAI share height in their formulas, the very high R² values (.85–.91) and AUCs largely reflect overlap by construction. This should be discussed in the Discussion, not only mentioned in the Limitations.

Table R3. Cross-Tabulation and Descriptive Distribution of the CUN-BAE Adiposity Categories

Adiposity Category (CUN-BAE)	Male Participants (N = 199) Frequency / (%)	Female Participants (N = 246) Frequency / (%)	Total cohort (N = 445) Frequency / (%)	Chi-Square Testing *(χ²)*
Normal Weight (Males < 20% / Females < 30%)	103 / (51.8)	136 / 55.3	239 / 53.7	χ² = 1.706 df = 2 p = .426
Overweight (Males: ≥ 20-24.9 / Females: 30-34.9%)	34 / 17.1	47 / 19.1	81 / 18.2
Obese (Males ≥ 25% / Females ≥ 35%)	62 / 31.2	63 / 25.6	125 / 28.1

Statements about statistical assumptions. Several statements in the manuscript and the response letter are inaccurate:

The Durbin–Watson statistic tests autocorrelation of residuals. It does not show that residuals are normally distributed or that the "error terms satisfy the standard multivariable linear assumptions" (lines 605–609, 668–670).
A minimum of 30 cases per cell does not guarantee "normal distribution of error terms" (response 9).
Cochran's formula estimates a proportion. It does not guarantee statistical power for regression or ROC analyses.

Please report checks of residual normality, homoscedasticity and influential cases, and correct these statements.

Male Sub-Cohort Diagnostics (n = 199)

Figure R1. Normal P-P Plot of Regression Standardized Residual for Male Sub-Cohort (Dependent Variable: BMI)

Figure R2. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Male Sub-Cohort (Dependent Variable: BMI)

Table R1. Influential Cases Parameter Matrix for the Male Sub-Cohort (n = 199)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.090	.005	.012	199
sex = male
Dependent Variable: BMI

Female Sub-Cohort Diagnostics (n = 246)

Figure R3. Normal P-P Plot of Regression Standardized Residual for Female Sub-Cohort (Dependent Variable: BMI)

Figure R4. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Female Sub-Cohort (Dependent Variable: BMI)

Table R2. Influential Cases Parameter Matrix for the Female Sub-Cohort (n = 246)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.130	.005	.012	246
sex = female
Dependent Variable: BMI

Consistency of the psychosocial conclusions. In women, BMI and BAI correlate significantly with SWLS (r = −.175 and −.232), and SWLS shows a significant AUC. Yet lines 688–691 state that the psychological constructs show "no statistically significant classification capacity." Lines 845–847 conclude that students with higher BMI "do not necessarily report lower life satisfaction." These statements should be qualified to reflect the sex-specific bivariate findings.

BAI formula. Lines 85–86 state that BAI requires no height measurements. This is incorrect: BAI uses height and does not use weight. Line 260 describes BAI as hip circumference divided by height. The formula is actually hip circumference divided by height to the power 1.5, minus 18.

Reference accuracy. Some citations do not match their content:

Ref. [26], a study of the SWLS, is cited for obesity-related mortality (line 254).
Ref. [7] is cited for weight gain in students compared with non-students (line 74).
Refs. [32] and [33] do not appear to be cited in the text.

(a) The erroneous placement of reference [26] at line 254 has been corrected, removing it from the obesity-related mortality narrative where it was mistakenly injected.

Comments on the Quality of English Language

Submission Date

19 August 2026

Date of this review

06 Oct 2026 08:56:59

Responses to Reviewer 1

Table R3. Cross-Tabulation and Descriptive Distribution of the CUN-BAE Adiposity Categories

Adiposity Category (CUN-BAE)	Male Participants (N = 199) Frequency / (%)	Female Participants (N = 246) Frequency / (%)	Total cohort (N = 445) Frequency / (%)	Chi-Square Testing *(χ²)*
Normal Weight (Males < 20% / Females < 30%)	103 / (51.8)	136 / 55.3	239 / 53.7	χ² = 1.706 df = 2 p = .426
Overweight (Males: ≥ 20-24.9 / Females: 30-34.9%)	34 / 17.1	47 / 19.1	81 / 18.2
Obese (Males ≥ 25% / Females ≥ 35%)	62 / 31.2	63 / 25.6	125 / 28.1

Male Sub-Cohort Diagnostics (n = 199)

Figure R1. Normal P-P Plot of Regression Standardized Residual for Male Sub-Cohort (Dependent Variable: BMI)

Figure R2. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Male Sub-Cohort (Dependent Variable: BMI)

Table R1. Influential Cases Parameter Matrix for the Male Sub-Cohort (n = 199)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.090	.005	.012	199
sex = male
Dependent Variable: BMI

Female Sub-Cohort Diagnostics (n = 246)

Figure R3. Normal P-P Plot of Regression Standardized Residual for Female Sub-Cohort (Dependent Variable: BMI)

Figure R4. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Female Sub-Cohort (Dependent Variable: BMI)

Table R2. Influential Cases Parameter Matrix for the Female Sub-Cohort (n = 246)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.130	.005	.012	246
sex = female
Dependent Variable: BMI

Response regarding the two technical/editorial corrections remain:

We sincerely thank the reviewer for identifying these two critical technical oversights, which have been fully corrected in the revised manuscript (highlighted in light blue):

Responses to Reviewer 3

Table R3. Cross-Tabulation and Descriptive Distribution of the CUN-BAE Adiposity Categories

Adiposity Category (CUN-BAE)	Male Participants (N = 199) Frequency / (%)	Female Participants (N = 246) Frequency / (%)	Total cohort (N = 445) Frequency / (%)	Chi-Square Testing *(χ²)*
Normal Weight (Males < 20% / Females < 30%)	103 / (51.8)	136 / 55.3	239 / 53.7	χ² = 1.706 df = 2 p = .426
Overweight (Males: ≥ 20-24.9 / Females: 30-34.9%)	34 / 17.1	47 / 19.1	81 / 18.2
Obese (Males ≥ 25% / Females ≥ 35%)	62 / 31.2	63 / 25.6	125 / 28.1

Male Sub-Cohort Diagnostics (n = 199)

Figure R1. Normal P-P Plot of Regression Standardized Residual for Male Sub-Cohort (Dependent Variable: BMI)

Figure R2. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Male Sub-Cohort (Dependent Variable: BMI)

Table R1. Influential Cases Parameter Matrix for the Male Sub-Cohort (n = 199)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.090	.005	.012	199
sex = male
Dependent Variable: BMI

Female Sub-Cohort Diagnostics (n = 246)

Figure R3. Normal P-P Plot of Regression Standardized Residual for Female Sub-Cohort (Dependent Variable: BMI)

Figure R4. Residual Scatterplot (Standardized Residuals vs. Standardized Predicted Values) for Female Sub-Cohort (Dependent Variable: BMI)

Table R2. Influential Cases Parameter Matrix for the Female Sub-Cohort (n = 246)


Metric	Minimum	Maximum	Mean	Std. Deviation	N
Cook's Distance	.000	.130	.005	.012	246
sex = female
Dependent Variable: BMI

Response to comment 5: We thank the reviewer for highlighting these specific methodological points, which have been fully clarified in the revised manuscript:

Author Response File: Author Response.pdf