Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript addresses an important topic in sport psychology: the associations among basic psychological need satisfaction, self-determined motivation, athlete burnout, and dropout intention in Korean water polo athletes. The focus on a small and relatively understudied sport population is a strength, and the authors appropriately acknowledge that their cross-sectional data cannot establish temporal or causal mediation. The sample of 173 currently active athletes is also substantial relative to the reported national registered water-polo population.

However, the manuscript has important conceptual and psychometric limitations that affect the interpretation of the structural model. The central concern is the very strong overlap between athlete burnout and dropout intention. The authors report a latent correlation of .839 and acknowledge that the Fornell–Larcker criterion was not met for this construct pair. The separate two-factor model also demonstrated poor absolute fit, with CFI = .888 and RMSEA = .171. Therefore, the conclusion that burnout is the “strongest proximal correlate” of dropout intention should be presented much more cautiously. Major revision is recommended.healthcare-4597455-peer-review-v1.pdf

Overall assessment

The manuscript has a coherent theoretical basis in Self-Determination Theory and provides a useful attempt to examine need satisfaction, motivational quality, burnout, and dropout intention within a single model. The authors have made a commendable effort to report reliability estimates, confirmatory factor analysis, indirect effects, model comparison, and limitations related to the cross-sectional design.

Nevertheless, the study is based entirely on self-report data collected at one time point over a very short online survey period. The proposed sequential pathway is theoretically plausible, but the design cannot test whether need satisfaction precedes motivation, whether motivation precedes burnout, or whether burnout predicts later dropout intention or actual withdrawal. The results should be described consistently as cross-sectional associations within a theory-informed structural model.

The manuscript should also avoid treating SEM path direction as proof of a developmental process. Alternative models are plausible. For example, athletes with high dropout intention may report greater sport devaluation and exhaustion; burnout and dropout intention may be mutually reinforcing; or a common negative sport-evaluation factor may underlie both constructs.

Title and abstract

The title is generally suitable, but “Indirect Associations through Motivation and Burnout” could be interpreted as a demonstrated mediation process. Because all variables were measured concurrently, the title should be softened to indicate a cross-sectional structural model or theory-informed indirect associations.

The abstract is well organised but should include the cross-sectional design explicitly. It should also state that all variables were measured using self-report questionnaires and that dropout intention, rather than observed dropout, was the outcome.

The abstract highlights burnout as the strongest proximal correlate of dropout intention. This conclusion requires stronger qualification because burnout and dropout intention showed substantial overlap. The reported latent correlation of .839, the failure of the Fornell–Larcker criterion, and the poor fit of the focused two-factor model indicate that the constructs were not cleanly distinguished in the present data. The abstract should acknowledge this overlap or avoid presenting burnout as a clearly independent predictor.

The phrase “basic psychological need satisfaction was linked to dropout intention mainly through a sequential indirect pathway” should be softened. The model included not only the sequential path through self-determined motivation and burnout but also the direct path from need satisfaction to burnout and then to dropout intention. Therefore, the sequential path was one significant indirect association rather than the sole or necessarily dominant mechanism.

Introduction

The Introduction provides a clear rationale for studying athlete retention in Korean water polo. The discussion of the limited competitive base, constrained pathways, and potential consequences of athlete withdrawal is relevant. However, some contextual claims about the structure and vulnerability of Korean water polo should be supported more carefully. The argument relies partly on news sources and a small number of local references. The authors should ensure that all claims regarding athlete numbers, women’s programme development, team concentration, and structural constraints are accurate, current, and adequately referenced.

The theoretical rationale is generally appropriate, but the manuscript should distinguish more clearly between need satisfaction and need frustration. The study measures need satisfaction only. Yet some interpretations imply that participants may be experiencing need frustration, lack of autonomy, insecurity, or adverse coach–athlete relationships. These constructs were not assessed and should not be inferred directly from low need-satisfaction scores.

The manuscript also uses sport commitment theory primarily to explain the association between burnout and dropout intention. This is reasonable, but the model does not include core sport-commitment variables such as enjoyment, involvement opportunities, alternatives, social constraints, personal investments, or perceived costs. The sport-commitment perspective should therefore be described as a complementary interpretive lens rather than as a directly tested theoretical framework.

The hypotheses should be framed as associations rather than directional causal propositions. For example, instead of stating that “self-determined motivation will be negatively associated with athlete burnout,” the language should clarify that the study tests whether these variables are statistically associated in the hypothesised direction.

Participants and recruitment

The manuscript reports that 173 athletes represented approximately 58.2% of the national registered water-polo population. This is potentially impressive, but it requires careful treatment. The registered population figure is cited for 2026, whereas survey collection took place in August 2026. The authors should confirm that the registry figure reflects the eligible population during the recruitment period and that it includes the same categories of athletes as the study.

The recruitment procedure may introduce selection bias. Survey links were distributed through teams, coaches, officials, organisations, and athlete networks. Athletes who were more connected to their teams, more motivated, or more willing to discuss their sport experience may have been more likely to participate. Conversely, athletes with the strongest burnout or dropout intention may have declined participation or may already have left the sport. The authors should acknowledge that the survey included currently active athletes only and may underestimate distress or dropout risk among athletes who were less engaged or already close to leaving.

The sample was overwhelmingly male, with only 11 female athletes. This should be discussed more explicitly. The study cannot meaningfully examine sex-related patterns, and the conclusions should not be generalised to female water-polo athletes without caution.

The sample combines high-school, collegiate, and general-level athletes, who may differ substantially in age, training load, educational demands, autonomy, career stage, financial conditions, injury exposure, and likelihood of leaving the sport. These differences were not incorporated into the structural model. The authors should consider whether competitive level, age, playing experience, injury status, or national-team experience should be treated as covariates or explored in sensitivity analyses. At minimum, the lack of adjustment for these potentially relevant factors should be acknowledged.

Measures

The measurement strategy requires substantial clarification.

The Basic Needs Satisfaction in Sport Scale is an appropriate instrument in principle. However, the manuscript should state whether the Korean version used had previously undergone formal translation, cultural adaptation, and psychometric validation. The phrase “adapted for the water polo context” is not sufficient. The authors should describe the adaptation process, whether items were translated and back-translated, whether expert review or pilot testing was conducted, and whether any wording was modified.

The same concern applies to the Athlete Burnout Questionnaire and dropout-intention measure. The manuscript states that the dropout scale was translated and adapted from studies of youth handball and soccer players. The authors should provide details of translation, cultural adaptation, content validity, and pilot testing. It is particularly important to clarify whether the four dropout-intention items refer to leaving a current team, leaving competitive water polo, reducing participation, or not participating in the next season. These are not necessarily the same behavioural decision in a small sport with limited team alternatives.

The treatment of self-determined motivation is a major issue. The authors created an “adapted Relative Autonomy Index” using intrinsic motivation, integrated regulation, identified regulation, introjected regulation, and amotivation, while omitting external regulation. This deviates from the full SMS-II structure and may alter the construct represented by the score.

The authors should explain why external regulation was not measured. External regulation is theoretically important in Self-Determination Theory and may be particularly relevant in highly structured sport systems. Its omission may distort the motivational continuum and the RAI score.

The weighting scheme should also be justified more rigorously. Although the manuscript cites prior research on scoring methods, an adapted RAI is still a researcher-created composite. The authors should report the distribution, reliability, and validity of this composite in the present sample. More importantly, the use of a single manifest RAI score in SEM ignores measurement error and obscures potentially different associations of intrinsic, identified, integrated, introjected, and amotivated regulation with burnout and dropout intention.

A stronger analysis would model motivational regulations as latent variables, use a validated self-determination continuum model, or conduct sensitivity analyses using alternative motivational indices. The authors acknowledge this limitation briefly, but it is central rather than peripheral because the RAI is one of the main mediators in the proposed model.

Measurement model and discriminant validity

The most serious issue concerns the measurement distinction between athlete burnout and dropout intention.

The authors report that the latent correlation between burnout and dropout intention was .839. This correlation exceeded the square roots of the AVE values for both constructs, indicating failure of the Fornell–Larcker discriminant-validity criterion. This is not a minor issue, because the central conclusion relies on burnout being an empirically distinct antecedent of dropout intention.

The subsequent comparison of a two-factor and one-factor model does not fully resolve this concern. Although the two-factor model fit statistically better than the one-factor model, its absolute fit was poor: CFI = .888, TLI = .852, RMSEA = .171, and GFI = .804. An RMSEA of .171 indicates substantial model misfit. Therefore, the results demonstrate only that a one-factor model is worse than an already poorly fitting two-factor model. They do not provide strong evidence that burnout and dropout intention are adequately distinct in this sample.

This issue should be treated as a major limitation in the Abstract, Results, Discussion, and Conclusion. The authors should not simply state that the constructs were “empirically preferable” as separate factors. A more accurate interpretation is that the constructs were theoretically distinct and the two-factor model fit better than a one-factor alternative, but their empirical separation remained inadequate.

The authors should investigate whether particular burnout items—especially sport devaluation—overlap directly with dropout-intention items. Sport devaluation may involve reduced value placed on the sport, whereas dropout intention directly concerns leaving the sport. This overlap may artificially inflate the burnout–dropout association. The authors should consider conducting additional analyses excluding the sport-devaluation subscale or testing burnout dimensions separately. For example, emotional/physical exhaustion, reduced accomplishment, and sport devaluation may show different associations with dropout intention.

The manuscript should also consider alternative discriminant-validity assessments, such as HTMT, factor correlations with confidence intervals, or exploratory bifactor models. However, these analyses should not be used merely to rescue the current conclusion. If the constructs remain highly overlapping, the interpretation must be adjusted accordingly.

Item parceling

The use of item parcels for basic psychological need satisfaction and athlete burnout requires stronger justification. Parceling can simplify a model, but it can also conceal multidimensionality, local dependence, item-level misfit, and cross-loadings. This is especially relevant because both need satisfaction and burnout are theoretically multidimensional constructs.

The manuscript states that items were divided into two parcels within each subdimension by alternating items in their original order. This is a mechanical parceling method and may not be psychometrically optimal. The authors should justify why this approach was selected and demonstrate that each subdimension was sufficiently unidimensional before parceling.

The parceling approach may also contribute to an artificially clean measurement model and may obscure the overlap between sport devaluation and dropout intention. A more transparent approach would model the three burnout dimensions separately, perhaps as first-order factors under a second-order burnout construct, and examine whether the burnout–dropout association remains as large.

The full CFA is reported as showing “generally acceptable” fit, but the GFI is .800 and the model involves 12 first-order factors in a sample of 173 participants. The authors should report all factor correlations, factor loadings, residuals, and any modification indices used. The sample size may be limited for such a complex model.

Structural equation model and model fit

The structural model is described as showing acceptable fit, but this is somewhat optimistic. The RMSEA was .081, with a 90% confidence interval extending to .095, and GFI was .866. These values indicate borderline or mediocre fit rather than clearly good fit. The language should be moderated.

The manuscript should report whether maximum likelihood estimation was appropriate for the ordinal Likert-scale items and whether multivariate normality was assessed. The authors report only univariate skewness and kurtosis for four composite variables. This does not establish multivariate normality for the latent-variable model or for the item-parcel indicators.

The authors should explain how missing data were handled. Participants with incomplete responses on key variables were excluded, but the manuscript does not state how many responses were excluded, whether missingness occurred at item level, or whether complete-case analysis introduced bias.

The model does not control for potential confounders. Age, competitive division, years of experience, injury experience, national-team experience, and sex may all plausibly relate to need satisfaction, motivation, burnout, and dropout intention. It may not be feasible to include all of these variables given the sample size, but the absence of covariate adjustment should be recognised.

The direct paths from need satisfaction and self-determined motivation to dropout intention were non-significant. The manuscript then compares a partial-mediation model with an indirect-only model and concludes that the direct paths are unnecessary. This is statistically acceptable as a model-comparison exercise, but it does not demonstrate “full mediation,” particularly in cross-sectional data. The authors should avoid mediation terminology that implies a causal process.

Indirect-effect interpretation

The authors are appropriately cautious in several places, stating that the sequential indirect effect reflects a “theoretically specified statistical association rather than evidence of temporal or causal mediation.” This language should be maintained consistently throughout the Abstract, Highlights, Results, Discussion, and Conclusion.

However, some passages still overstate the findings. For example, statements that need satisfaction is linked to dropout intention “mainly through” motivation and burnout, or that the results “identify burnout as a key proximal correlate,” can imply a stronger causal ordering than the design permits.

The paper should consistently use terms such as:

• “cross-sectional association”;

• “indirect association within the specified model”;

• “theory-consistent pattern”;

• “statistical pathway”;

• “hypothesised ordering.”

The manuscript should avoid “mediator,” “transmission,” “operating through,” or “proximal predictor” unless these are explicitly qualified as model-based statistical terms.

The indirect effect from need satisfaction to dropout intention is also not limited to the sequential pathway through self-determined motivation and burnout. The authors appropriately acknowledge that need satisfaction also had a direct association with burnout, which created another burnout-related indirect route. This point should be made clearer in the Abstract and Highlights so that readers do not interpret the sequential path as the only explanatory mechanism.

Results reporting

The manuscript provides useful descriptive statistics, reliability coefficients, factor loadings, model fit indices, structural coefficients, and bootstrapped confidence intervals. Nevertheless, several reporting improvements are needed.

The demographic table should report the continuous age mean and standard deviation, not only age categories. It should also clarify how the “general” and “other” affiliation categories were combined and why positions included an “other” category.

The manuscript reports Cronbach’s alpha values, but alpha alone is not sufficient evidence of measurement quality. Composite reliability is provided for the latent constructs retained in SEM, but the authors should report it for all relevant subscales and discuss scale validity in the current sample.

The full CFA involved 12 factors, but only selected structural-model constructs are presented in Table 4. The manuscript should provide a supplementary table with all factor loadings, factor correlations, CR, AVE, and model specifications.

The focused burnout–dropout two-factor model should not be described as adequate simply because it was superior to a one-factor model. The poor absolute fit needs much more prominence.

The authors should report confidence intervals for the key standardized path coefficients where possible. Reporting only p-values can overemphasise significance. Confidence intervals would aid interpretation of precision, particularly given the sample size and the high correlation between burnout and dropout intention.

Discussion and implications

The Discussion is generally thoughtful and includes several appropriate cautions. The authors acknowledge the cross-sectional design, self-report bias, the adapted RAI, limited generalisability, and the overlap between burnout and dropout intention. These are important strengths.

However, the practical implications need more restraint. The manuscript recommends burnout monitoring, autonomy-supportive coaching, and need-supportive practice. These recommendations are theoretically reasonable, but the current design does not show that changing coaching behaviour, need satisfaction, or motivation will reduce burnout or dropout. The paper should state that these are theory-informed practical considerations rather than evidence-based intervention recommendations.

The discussion repeatedly interprets the Korean water-polo environment as one in which limited team alternatives intensify the association between burnout and dropout. This is plausible but untested. The manuscript should avoid suggesting that the observed findings demonstrate this mechanism. If retained, it should be framed as a contextual hypothesis for future research.

The Discussion should also acknowledge reverse causality. Athletes who already intend to leave water polo may retrospectively evaluate their sport environment more negatively, report lower need satisfaction, endorse more amotivation, and report greater sport devaluation. This alternative explanation is especially important given the cross-sectional design and common-method data.

Author Response

Comment 1: Major concern: strong overlap of burnout and dropout intention (r = .839, failed Fornell–Larcker, poor fit of the two-factor model); present “strongest proximal correlate” much more cautiously.

Response: We agree. The poorly fitting two-factor model omitted the within-subdimension residual covariances that were included in the structural model; with the same specification, it fits well (Table 5, Model 1), and a one-factor model with the same residual covariances fits clearly worse (Table 5, Model 3b; Δχ²(1) = 53.757, p < .001). However, the latent correlation remains very high (.851 [.792, .911]), and the Fornell–Larcker criterion is not met for either construct. We therefore do not claim that the overlap has been resolved. “Proximal correlate” has been removed throughout, and the overlap is the first limitation.

Revisions: Abstract; Section 3.2; Table 5; Sections 4.3, 5, 6; Table S4

 

Comment 2: Results should be described consistently as cross-sectional associations within a theory-informed model.

Response: The title, Highlights, Abstract, hypotheses, Results, figure captions, Discussion, and Conclusions now describe cross-sectional, theory-specified associations; mediation language has been removed.

Revisions: Title; Highlights; Abstract; Sections 1, 3.4–3.5, 4.5, 6; Figures 1–2

 

Comment 3: SEM path direction is not proof; alternative models (reverse direction, mutual reinforcement, common negative evaluation) are plausible.

Response: We estimated an alternative model in which dropout intention precedes burnout. It fitted somewhat worse than the hypothesized model without direct paths (χ² = 239.290 vs. 231.193, equal df). We state, however, that an equivalent model in which burnout and dropout intention merely covary fits identically, so the ordering cannot be established. Reverse causality and a common negative-evaluation factor are discussed.

Revisions: Sections 2.4, 3.5, 4.3, 4.5; Table 8

 

Comment 4: Title should be softened.

Response: The title is now “Psychological Need Satisfaction, Self-Determined Motivation, Burnout, and Dropout Intention in Korean Water Polo Athletes: A Cross-Sectional Structural Equation Model”.

Revisions: Title

 

Comment 5: Abstract should state the cross-sectional design, self-report measures, and dropout intention as the outcome; acknowledge overlap; soften “mainly through”.

Response: Done. The Abstract now states the design, measures, outcome, and sex composition, reports the latent correlation between burnout and dropout intention, notes that their association remained strong without sport devaluation, and refers to both the indirect association via burnout alone and the sequential association. To keep the Abstract within 200 words, most individual path coefficients are now reported in the Results only; the Abstract retains the burnout–dropout, sequential indirect, and overlap estimates.

Revisions: Abstract; Highlights

 

Comment 6: Contextual claims about Korean water polo should be supported more carefully.

Response: These statements were softened to what the cited sources report. Claims about the competitive structure now rest on the peer-reviewed analyses of Kim [4] and Lee [8]; news reports are cited only for the dates on which the women’s national programme was formed. We also checked the cited studies against their content: the statement attributed to Murillo et al. [9] now refers only to the association of coaches’ motivating style with motivation and sport commitment; the meta-analysis of Back et al. [17] is now described as identifying motivation-related constructs among the strongest prospective predictors of dropout, with the link between need-related experiences and dropout intentions attributed to Quested et al. [34]; and the statement about the small Korean player base is now supported by [4,8]. The registry figure has been removed (see Comment 10).

Revisions: Sections 1 and 2.1

 

Comment 7: Distinguish need satisfaction from need frustration.

Response: Added; inferences about need frustration and insecurity were removed.

Revisions: Sections 1, 2.2.1, 4.2

 

Comment 8: Sport commitment theory should be a complementary interpretive lens.

Response: Revised accordingly.

Revisions: Sections 1 and 4.3

 

Comment 9: Hypotheses should be framed as associations.

Response: All hypotheses were reworded as statistical associations in the hypothesized direction.

Revisions: Section 1

 

Comment 10: Confirm that the registry figure reflects the eligible population.

Response: Thank you for this important observation. The previously reported registration count did not define the eligible population for this sample, which included adult club and masters participants who may not be federation-registered. We removed the numerical registry claim, its original citation, and the 58.2% coverage estimate from the revised manuscript. We explicitly acknowledge that population coverage and representativeness cannot be established (Section 2.1; Section 5).

Revisions: Section 2.1; Table 1 note; Section 5

 

Comment 11: Selection bias: only currently active athletes participated.

Response: Acknowledged as a limitation.

Revisions: Section 5

 

Comment 12: Only 11 female athletes.

Response: Stated, and generalization to female athletes is explicitly limited.

Revisions: Abstract; Sections 2.1, 5

 

Comment 13: Heterogeneous competitive levels; consider covariates.

Response: Models adjusting for age (or division), sex, playing experience, national-team experience, and injury gave essentially the same estimates, and no covariate was significant. Because the general level also included adult club and masters players, we additionally excluded likely club and masters players: (a) general-level participants without National Sports Festival experience (n = 150 remaining) and (b) those with less than 3 years of playing experience (n = 148). The burnout–dropout association (β = .801, .800) and the sequential association (β = −.186, −.223) were essentially unchanged.

Revisions: Sections 2.1, 2.4, 3.6, 5; Table 9; Table S5

 

Comment 14: Describe translation, adaptation, back-translation, expert review, and pilot testing.

Response: Items were written in Korean by the research team with co-researchers familiar with competitive water polo, with reference to the original scales. No back-translation, formal content-validity rating, or pilot test was conducted; this is stated and acknowledged as a limitation. We also note the change of response format and that the need-satisfaction subscale structure differs from the original BNSSS. The motivation items were developed with reference to the SMS-II and other Self-Determination Theory-based sport motivation measures, and the need-satisfaction, burnout, and dropout-intention items with reference to the BNSSS, the ABQ, and the items of Sarrazin et al. as adapted by Quested et al.; previously validated Korean versions were not used. Full Korean wording with English translations is given in Table S1.

Revisions: Sections 2.2, 2.2.1, 5; Table S1

 

Comment 15: Clarify what the dropout-intention items refer to.

Response: The items concern discontinuing water polo participation, not leaving a particular team. They combine past thoughts, current uncertainty, a conditional preference, and a next-season expectation, so the score is interpreted as a broad withdrawal-intention measure.

Revisions: Section 2.2.4; Table S1

 

Comment 16: RAI: omission of external regulation; weights; distribution, reliability, validity; separate or latent regulations.

Response: Three of the four items labelled “introjected” (Q13–Q15) describe external regulation, so the subscale is now labelled “controlled regulation (external/introjected)”. We also note that Q04 and Q07–Q08 are closer to intrinsic and identified regulation, respectively. We report the RAI distribution and composite reliability (.964). Alternative operationalizations (latent autonomous motivation; simultaneous autonomous, controlled, and amotivation factors; a unit-weighted index) showed that the sequential association is robust to the weighting scheme but is carried mainly by amotivation, which we report and discuss.

Revisions: Sections 2.2.2, 3.2, 3.6, 4.2, 4.5, 5; Table 2; Tables S3, S6

 

Comment 17: Alternative discriminant-validity assessments (HTMT, CIs, bifactor); examine sport devaluation; do not merely rescue the conclusion.

Response: We report HTMT with a bootstrap CI, the CI of the latent correlation, chi-square tests against 1.0 and .90, correlations of each burnout dimension with dropout intention, an EFA of the devaluation, amotivation, and dropout items, and models excluding devaluation or treating the dimensions separately. We did not estimate a bifactor model because it is poorly identified with two parcels per dimension and N = 173. In line with the reviewer’s caution, the interpretation has been adjusted rather than rescued.

Revisions: Sections 3.2, 3.6, 4.3; Tables 5, 9; Tables S4–S5

 

Comment 18: Item parceling: justify, demonstrate unidimensionality; consider second-order burnout; report loadings, correlations, residuals, and modification indices used.

Response: Single-factor CFAs provided mixed evidence of unidimensionality: CFI and SRMR were generally favorable, but RMSEA was elevated for several subscales, and the three-item scales could not be tested (Table S2). We acknowledge this as a limitation of the parceling approach. We now disclose that the six within-subdimension residual covariances were added after inspection of modification indices, which our initial submission did not report. A hierarchical representation was considered, but exact equivalence requires matching constraints and an admissible solution; we do not assume equivalence under an additional zero-disturbance constraint. We also correct the parcel description (reduced accomplishment: Q46/Q48 vs. Q47/Q49/Q50). An item-level model with second-order need-satisfaction and burnout factors gave very similar structural estimates. Item-level CFAs fitted less well, which we report. Full loadings, CR, AVE, latent correlations, and HTMT for all 12 subscales are given in Tables S2–S3. Because composite reliability computed from loadings alone ignores the parcel error covariances, we also report CR including them (need satisfaction .784; burnout .833).

Revisions: Sections 2.4, 3.2, 3.6, 5; Table 9; Tables S1–S3

 

Comment 19: Structural fit is described too optimistically.

Response: Fit is now described as marginally acceptable. We report SRMR, the Bollen–Stine test (p = .036), and robust indices (CFI = .967, RMSEA = .068).

Revisions: Section 3.3

 

Comment 20: Was ML appropriate for ordinal items; assess multivariate normality.

Response: Mardia’s coefficient indicated multivariate non-normality (442.26, z = 30.86). The primary analysis treated the five-category responses, parcel means, and composite index as continuous indicators; MLR estimates were checked for robustness to non-normality, while the Bollen–Stine procedure assessed model fit. MLR is not an ordinal-data estimator. An item-level sensitivity model without parcels was also estimated, and its fit limitations are acknowledged. The revised manuscript does not report a separate ordinal WLSMV analysis. (Sections 2.4, 3.1, 3.3, and 3.6; Table 9).

Revisions: Sections 2.4, 3.1, 3.3, 3.6; Table 9

 

Comment 21: How were missing data handled?

Response: All items were required in the online form, so there were no missing item responses and no exclusions; this is now stated.

Revisions: Section 2.1

 

Comment 22: Avoid “full mediation” and causal mediation terminology.

Response: Section 3.4 is now titled “Indirect Associations”, and the model comparison is described as models with and without direct paths, explicitly not establishing full mediation.

Revisions: Sections 3.4–3.5; Table 8

 

Comment 23: The sequential path is not the only indirect route; clarify this in the Abstract and Highlights.

Response: Table 7 now reports all three components of the total indirect association: via burnout alone (B = −.568, β = −.313; about 60%), the sequential path (about 33%), and via self-determined motivation alone (B = −.058, 95% BC CI [−.234, .141], not significant). The Abstract and Highlights refer to the association via burnout alone and the sequential association.

Revisions: Abstract; Highlights; Section 3.4; Table 7

 

Comment 24: Demographic table: age M (SD); general/other; position “other”.

Response: Added.

Revisions: Section 2.1; Table 1 note

 

Comment 25: Report CR for all subscales.

Response: CR and AVE for all 12 subscales are reported.

Revisions: Table S2

 

Comment 26: Report confidence intervals for standardized paths.

Response: Bias-corrected bootstrap CIs (5000 resamples) were added for all standardized paths. The original AMOS B, SE, C.R., and p values are unchanged.

Revisions: Table 6; Figure 2

 

Comment 27: Practical implications need more restraint.

Response: These are now presented as theory-informed considerations, not evidence-based interventions.

Revisions: Highlights; Sections 4.5, 4.6, 6

 

Comment 28: Korean context interpretations are untested.

Response: Condensed and framed as hypotheses for future research.

Revisions: Section 4.4 (consolidated; the corresponding paragraph in Section 4.1 was removed); Section 4.3

 

Comment 29: Acknowledge reverse causality.

Response: Added.

Revisions: Sections 4.5, 5

Reviewer 2 Report

Comments and Suggestions for Authors

Comments in the attachment.

Comments for author File: Comments.pdf

Author Response

Comment 1: Overlap between burnout and dropout intention; associational interpretation; adapted RAI; specific, predominantly male sample.

Response: Thank you for the positive evaluation. We conducted additional discriminant-validity and sensitivity analyses, and the overlap is now the first limitation. Associational language is used throughout. The composition of the RAI has been clarified and examined with alternative operationalizations, and generalization to female athletes and other contexts is explicitly limited.

Revisions: Sections 3.2, 3.6, 4, 5

Reviewer 3 Report

Comments and Suggestions for Authors

General Assessment

This manuscript investigates the structural relationshipsbetween basic psychological need satisfaction, self-determinedmotivation, athlete burnout, and dropout intention in 173 active Korean water polo players. The manuscript usesStructural Equation Modeling (SEM), item parceling, andnested model comparisons to examine the proposedrelationships and discriminant validity. The sample represents58.2% of the registered national athlete population, which is a strength of the study. A minor revision is required to addresstheoretical limitations regarding the Relative Autonomy Index (RAI) and the precise use of causal terminology in a cross-sectional framework.

Specific Critiques

1. Limitation of the Relative Autonomy Index (RAI):

The authors combined distinct motivational regulations into a single RAI score to maintain model parsimony. Whilestatistically justified given the sample size (), this approach obscures the independent effects of autonomous versus controlled motivation on burnout and dropout intention. The authors must add a paragraph to the Discussion section addressing this limitation. They should briefly theorize how specific sub-dimensions (e.g., amotivation versus intrinsic regulation) might independently influence the observed sequential pathway if analyzed separately.

2. Adjustment of Causal Terminology:

The study utilizes cross-sectional data, yet the manuscriptoccasionally employs terminology that implies causality orlongitudinal pathways. While the authors appropriatelyacknowledge the cross-sectional limitation in the Discussion, all causal implications (such as "effect" or "impact") should be softened to associative terms (e.g., "association," "relationship," or "sequential indirect association") throughoutthe Abstract, Results, and main text.

3. Discriminant Validity Contextualization:

The proactive use of nested competing models to confirm thedistinction between burnout and dropout intention—despite a high correlation of .839—is mathematically sound. Tostrengthen this section, the authors should explicitlycontextualize this high correlation in the Discussion. It shouldbe framed as a natural reflection of the late-stage withdrawalprocess in elite sports, which theoretically supports thecontinuum from severe burnout to actual dropout intention.

 

Author Response

Comment 1: Add a Discussion paragraph on the RAI limitation, theorizing how amotivation vs. intrinsic regulation might influence the sequential pathway.

Response: We added this paragraph and also tested it. When the regulations were modeled separately, amotivation was strongly associated with burnout (β = .561), whereas autonomous motivation was not (β = .095, p = .230). We discuss why amotivation is theoretically closest to devaluation and withdrawal cognitions, why autonomous motivation may overlap with need satisfaction, and how controlled regulation may relate to exhaustion.

Revisions: Section 4.2 (third paragraph); Section 3.6; Table S6

 

Comment 2: Soften causal terminology (“effect”, “impact”).

Response: Done throughout. Table 7 now refers to direct, indirect, and total associations.

Revisions: Throughout

 

Comment 3: Contextualize the high correlation as a reflection of the late-stage withdrawal process.

Response: Added in Section 4.3 as one possible interpretation, alongside the alternative that a common negative evaluation of the sport underlies both measures. Because Reviewers 1 and 4 cautioned against using this interpretation to resolve the discriminant-validity concern, we present it as a hypothesis and retain the overlap as a limitation.

Revisions: Section 4.3

Reviewer 4 Report

Comments and Suggestions for Authors

Dear authors,

Your manuscript addresses a relevant research topic concerning the relationships between basic psychological need satisfaction, self-determined motivation, athlete burnout, and dropout intention among water polo players in South Korea. The topic is of interest from both a theoretical perspective, particularly in relation to Self-Determination Theory, and a practical perspective, considering the relatively small athlete population and the structural characteristics of Korean water polo. I appreciate your attempt to integrate these four constructs into a single structural model and to examine not only their direct associations but also the sequential indirect association between psychological need satisfaction and dropout intention through self-determined motivation and burnout.

Nevertheless, the manuscript presents several methodological and interpretative issues that require further clarification and substantial revision, as outlined below:

1. The main methodological concern relates to the discriminant validity between athlete burnout and dropout intention. A very high latent correlation is reported between these two constructs (r = .839), and the Fornell–Larcker criterion is not satisfied. More importantly, the two-factor model used to support the distinction between these constructs demonstrates poor absolute model fit: χ²/df = 6.01, CFI = .888, TLI = .852, and RMSEA = .171. The fact that this model fits the data significantly better than the one-factor model does not, in itself, establish adequate discriminant validity. Rather, it indicates that a problematic two-factor model performs relatively better than an even poorer one-factor model. This distinction is particularly important because the study's main conclusion, namely the very strong association between burnout and dropout intention, depends on demonstrating that these variables represent empirically distinct constructs. I recommend a more comprehensive assessment of discriminant validity, including appropriate additional analyses and an explicit examination of the contribution of the sport devaluation dimension to its overlap with dropout intention. Without more convincing evidence, the conclusion that burnout represents "the strongest proximal correlate" should be expressed with considerably greater caution.
2. Related to the previous concern, I consider a sensitivity analysis focusing on the sport devaluation dimension of the Athlete Burnout Questionnaire Conceptually, statements reflecting the devaluation of sport participation may closely resemble statements expressing an intention to discontinue participation. Consequently, the observed association (β = .794) may partly reflect overlapping item content rather than exclusively a strong psychological relationship between two distinct constructs. This possibility is acknowledged in the Discussion section, which is appreciated; however, merely recognizing the issue is insufficient. I recommend testing the robustness of the burnout–dropout intention association when the sport devaluation component is either examined separately or excluded from the overall burnout construct. If the association remains strong, the central conclusion of the study would become considerably more convincing.
3. The operationalization of self-determined motivation requires stronger methodological justification. An adapted version of the SMS-II was used, but external regulation was not assessed. Subsequently, a modified Relative Autonomy Index (RAI) was calculated by assigning weights to intrinsic motivation (+3), integrated regulation (+2), identified regulation (+1), introjected regulation (−1), and amotivation (−2). The concern is that this modified index subsequently becomes one of the central variables in the structural model and one of the two mediators. Omitting a component of the SDT continuum and applying predetermined weights may alter the meaning of the construct compared with the original instrument. I recommend providing a more rigorous justification for both the omission of external regulation and the psychometric validity of this modified version of the RAI. Ideally, the robustness of the findings should also be examined using an alternative operationalization of motivation, provided that the available data allow such an analysis.
4. The manuscript states that the instruments were "adapted to the context of water polo" and that the dropout intention instrument was translated and adapted. However, the actual translation and cultural adaptation procedures are not described in sufficient detail. This is important because all central constructs were assessed using instruments originally developed in other contexts, and modifications to item wording may affect their validity. I recommend describing the translation and back-translation procedures, expert evaluation, pilot testing, and any other steps undertaken to establish the equivalence of the Korean versions. If previously validated Korean versions of the instruments were used, this should be explicitly stated and supported by appropriate references.
5. The use of item parceling for basic psychological need satisfaction and athlete burnout requires a more convincing justification. Two parcels were created for each subdimension by alternating items, and these parcels were subsequently used as indicators of the latent constructs. Item parceling may artificially improve global model fit indices and conceal potential item-level problems or the multidimensionality of the constructs. This issue is particularly relevant for burnout, given the already evident difficulties concerning its discriminant validity relative to dropout intention. I recommend justifying the choice of this strategy, explaining why alternating items was considered appropriate, and, if possible, presenting a sensitivity analysis without parceling or an alternative latent-variable specification.
6. The sample size requires statistical justification. Although the 173 participants represent an impressive proportion of the nationally registered athlete population, population coverage and sample adequacy for structural equation modeling are two distinct considerations. The manuscript does not report a prospective sample size calculation or a model-specific statistical power assessment. Considering the complexity of the CFA/SEM analyses, the number of estimated parameters, and the use of bootstrapping for indirect effects, I recommend including an explicit justification for the adequacy of N = 173. If no prospective calculation was performed, this should be acknowledged, and a model-specific assessment, such as a Monte Carlo simulation, could be considered instead of relying exclusively on simplified participant-to-parameter ratios.
7. The recruitment procedure requires further clarification. The manuscript indicates that invitations were distributed through teams, coaches, officials, organizations, and athlete networks. However, it remains unclear how many athletes were actually invited, how many accessed the questionnaire, how many initiated the survey, and how many were excluded before obtaining the final sample of 173 participants. Although the sample represents approximately 58.2% of the registered athlete population, this does not eliminate the possibility of self-selection bias. For example, athletes experiencing particularly high levels of burnout or an imminent intention to discontinue participation may have been less likely to participate. I recommend providing a clear description of the recruitment and exclusion process and discussing the potential implications of selection bias.
8. The assessment of statistical assumptions is incomplete. Skewness and kurtosis were examined for the four composite variables, leading to the conclusion that univariate normality was adequately satisfied. However, multivariate normality is also relevant when estimating SEM using maximum likelihood, and the manuscript does not report an assessment of this assumption. Furthermore, insufficient information is provided regarding the identification of potential multivariate outliers. I recommend expanding the Data Analysis section to address these issues and, if the assumptions underlying maximum likelihood estimation are not satisfied, applying or justifying an appropriate robust estimation procedure.
9. All variables were assessed through self-report measures collected during the same session and from the same source, raising concerns regarding common method variance. Although shared-method variance is acknowledged among the study limitations, this issue is particularly important because the structural model includes very strong associations between conceptually related constructs. I believe the manuscript would benefit from a more detailed discussion of the extent to which common method variance may have contributed to the magnitude of the observed associations. If an appropriate additional analysis is feasible, it could further strengthen the interpretation of the findings.
10. The sample shows a substantial sex imbalance, with 162 of the 173 participants being male and only 11 being female. Although this distribution may reflect the actual structure of the Korean water polo population, it considerably limits the generalizability of the findings to female athletes. I recommend clarifying whether the sample distribution approximately reflects the national athlete population and avoiding broad statements referring to "Korean water polo athletes" when such wording could imply comparable representativeness across sexes.
11. The sample includes high-school, collegiate, and general-level athletes, who may differ substantially in age, sporting experience, competitive pressure, career prospects, and reasons for continuing or discontinuing sport participation. Nevertheless, these characteristics are presented primarily descriptively and are not incorporated into the main structural model. I recommend examining whether the principal associations remain robust after controlling for relevant variables such as age or competitive division, sporting experience, and potentially injury history. I do not consider it necessary to estimate separate models using such small subsamples; however, at least one sensitivity analysis would help determine whether the observed relationships are largely influenced by the heterogeneous composition of the sample.
12. The interpretation of the sequential indirect effect is generally cautious, and I appreciate the explicit acknowledgment that cross-sectional data cannot establish temporal or causal mediation. Nevertheless, the repeated use of terms such as "pathway", "through", and "operating through" may still encourage readers to interpret the model as demonstrating an underlying psychological mechanism. In a design where basic psychological need satisfaction, motivation, burnout, and dropout intention are measured simultaneously, the ordering of variables is theoretically specified rather than empirically established. I recommend consistently using the term "sequential indirect association" and avoiding language that might suggest that the data demonstrate the psychological process BPNS → motivation → burnout → dropout intention.
13. Similarly, the structural model tests only one primary theoretically specified ordering, whereas the cross-sectional nature of the data also allows for plausible alternative models. For example, athletes who already intend to discontinue sport participation may report greater sport devaluation, higher burnout, and lower motivational quality. I am not suggesting that the authors should establish directionality using data that cannot support such conclusions. However, I believe it would be useful either to test reasonable alternative models as a sensitivity analysis or to explicitly discuss the fact that the fit of the proposed model does not establish the superiority of the hypothesized causal ordering.
14. The reported R² = .731 for dropout intention is remarkably high and could be interpreted as evidence of substantial explanatory power. However, this finding must be contextualized in relation to the very high latent correlation between burnout and dropout intention and the concerns regarding discriminant validity. In its current form, there is a risk that the 73.1% of explained variance partly reflects conceptual and measurement overlap between the predictor and the outcome. I recommend avoiding the presentation of this value as strong evidence of the model's explanatory performance until the discriminant validity issue has been adequately addressed.
15. In the Discussion section, some interpretations concerning the specific characteristics of the Korean water polo system remain speculative. For example, the suggestion that the limited number of teams and restricted transfer opportunities may strengthen the association between burnout and dropout intention is plausible. However, these organizational mechanisms were not directly assessed in the study. I recommend condensing these explanations and explicitly presenting them as hypotheses for future research rather than as explanations for the magnitude of the observed associations.
16. The practical implications are reasonable, but a clear distinction must be maintained between statistical associations and intervention effectiveness. Recommendations concerning burnout monitoring and autonomy-supportive coaching practices are theoretically well grounded. However, the study did not evaluate any intervention and does not demonstrate that modifying these variables would effectively reduce dropout intention or actual dropout behavior. This limitation is partially acknowledged, but the wording should be reviewed consistently throughout the Highlights, Discussion, Practical Implications, and Conclusions sections.
17. I consider the distinction between dropout intention and actual dropout behavior particularly important. The outcome assessed in this study represents the self-reported intention to discontinue participation rather than actual withdrawal from sport. Although this limitation is acknowledged, some statements referring to "athlete retention" may extend beyond what the data can demonstrate. Dropout intention may represent a relevant indicator, but its relationship with subsequent behavior needs to be established longitudinally in this population. I recommend consistently using the term "dropout intention" when describing the study findings and reserving references to "retention" for hypothetical implications or future research directions.
18. I believe the Limitations section should be expanded to address participant selection, the absence of a model-specific sample size justification, potential common method variance, the substantial sex imbalance, heterogeneity across competitive divisions, and the implications of using item parceling. In particular, the discriminant validity issue concerning burnout and dropout intention should not be presented merely as a secondary psychometric limitation, as it directly affects the main conclusion of the study.
19. In the Conclusions section, the central claim that burnout represents the strongest proximal correlate of dropout intention should be more explicitly qualified in light of the overlap between the constructs. If the recommended sensitivity analysis confirms a strong association between burnout and dropout intention even after reducing the potential overlap attributable to sport devaluation, the conclusion would be considerably more robust. Otherwise, the manuscript should be reframed as demonstrating a strong association between burnout-related experiences, particularly sport devaluation, and dropout intention, without making such a firm distinction between these constructs.

In conclusion, the manuscript addresses a relevant and relatively underexplored research topic within a specific athletic population, and the integration of basic psychological need satisfaction, self-determined motivation, burnout, and dropout intention into a structural model is of interest. Nevertheless, the study's main conclusion is substantially affected by the very high empirical overlap between burnout and dropout intention, while the model used to support the distinction between these constructs demonstrates unsatisfactory absolute fit indices. Additional concerns relate to the operationalization of self-determined motivation, item parceling, instrument adaptation, sample adequacy for SEM, and the potential influence of participant heterogeneity. I believe these issues are potentially addressable through additional analyses, methodological clarifications, and more cautious interpretation of the findings.

Author Response

Comment 1: Comprehensive assessment of discriminant validity; the poorly fitting two-factor model does not establish distinctiveness.

Response: See Reviewer 1, Comments 1 and 17. With the same specification as the structural model, the two-factor model fits well. The latent correlation is .851 [.792, .911]; 1.0 is rejected but .90 is not; HTMT = .836 [.749, .898]; and the Fornell–Larcker criterion is not met. We therefore retain substantial construct overlap as a major discriminant-validity limitation rather than treating better comparative fit as definitive separation.

Revisions: Section 3.2; Table 5; Table S4

 

Comment 2: Sensitivity analysis excluding or separating sport devaluation.

Response: With devaluation excluded, burnout remained strongly associated with dropout intention (β = .663, R² = .640). When the dimensions were entered separately, only devaluation had a unique association (β = .570), but reduced accomplishment was strongly associated when devaluation was omitted (β = .643).

Revisions: Section 3.6; Table 9; Table S5

 

Comment 3: Justify the omission of external regulation and the RAI; test an alternative operationalization.

Response: See Reviewer 1, Comment 16.

Revisions: Sections 2.2.2, 3.6; Table S6

 

Comment 4: Describe translation and cultural adaptation procedures.

Response: See Reviewer 1, Comment 14.

Revisions: Section 2.2; Table S1

 

Comment 5: Justify parceling; sensitivity analysis without parceling.

Response: See Reviewer 1, Comment 18. The item-level sensitivity model without parcels gave a similar burnout–dropout association (β = .806), although item-level fit was poorer and evidence for unidimensionality was mixed; these qualifications are reported (Sections 2.4, 3.2, 3.6; Table 9).

Revisions: Sections 2.4, 3.2, 3.6; Table 9

 

Comment 6: Justify the sample size, e.g., with a Monte Carlo simulation.

Response: No a priori calculation was performed; this is stated. RMSEA-based power for the test of close fit was .93. In a Monte Carlo simulation (1000 replications), power was .91–1.00 for H1–H4 and .90 for the sequential association, but only .10–.11 for the direct paths. These post hoc, model-conditional simulations do not establish prospective sample-size adequacy or bias-corrected bootstrap-test power. The non-significant direct paths are therefore not interpreted as evidence of no association. (Sections 2.4, 3.6, 4.4; Table S7).

Revisions: Sections 2.4, 3.6, 4.4; Table S7

 

Comment 7: Clarify recruitment (invited, accessed, initiated, excluded).

Response: The 58.2% figure has been removed because a registry count could not provide a valid sampling denominator for the adult club and masters participants (see Reviewer 1, Comment 10). An official letter was sent to coaching staff, and the Google Forms link was distributed through coaches and team networks. All items were required, so all 173 submitted questionnaires were complete. Abandoned attempts are not recorded and the number of athletes reached is unknown; we state this and discuss self-selection.

Revisions: Sections 2.1, 2.3, 5

 

Comment 8: Assess multivariate normality and outliers; use robust estimation if needed.

Response: Mardia’s test indicated non-normality, and eight multivariate outliers and three invariant responders were identified. Results were unchanged with MLR, the Bollen–Stine bootstrap, and when these cases were excluded.

Revisions: Sections 3.1, 3.3, 3.6

 

Comment 9: Common method variance.

Response: Harman’s first factor accounted for 36.2% of the item variance. The unconstrained ULMC yielded a negative method-factor variance (−0.006), an inadmissible solution, so its results were not used as evidence against common method bias. In a constrained re-estimation the variance was near its lower bound (approximately .0001), with a similar burnout–dropout path (β = .794). This boundary estimate does not rule out common method bias and is not interpreted as a conventional chi-square test. Shared method variance and common negative evaluation remain limitations (Sections 3.1, 3.6, 5).

Revisions: Sections 2.4, 3.1, 3.6, 5; Table 9

 

Comment 10: Sex imbalance and representativeness.

Response: Because the target population includes both registered and adult club/masters athletes, we could not verify sample representativeness, including the sex distribution. We explicitly limit generalization to female athletes and avoid broad population claims (Sections 2.1 and 5).

Revisions: Section 2.1; Section 5

 

Comment 11: Control for age/division, experience, and injury.

Response: Done; results unchanged. We also excluded likely adult club and masters players (see Reviewer 1, Comment 13).

Revisions: Section 3.6; Table 9; Table S5

 

Comment 12: Use “sequential indirect association” consistently.

Response: Revised throughout.

Revisions: Throughout

 

Comment 13: Test alternative models or discuss that fit does not establish causal ordering.

Response: Done; see Reviewer 1, Comment 3.

Revisions: Sections 3.5, 4.5; Table 8

 

Comment 14: Contextualize R² = .731.

Response: We now state that this value partly reflects overlap between burnout and dropout intention.

Revisions: Sections 3.3, 4.3

 

Comment 15: Korean water polo system interpretations are speculative.

Response: Condensed and framed as hypotheses.

Revisions: Section 4.4 (consolidated); Section 4.3

 

Comment 16: Distinguish associations from intervention effectiveness.

Response: Revised throughout.

Revisions: Highlights; Sections 4.5, 4.6, 6

 

Comment 17: Use “dropout intention” consistently; reserve “retention”.

Response: Revised throughout.

Revisions: Throughout

 

Comment 18: Expand Limitations; discriminant validity should not be secondary.

Response: Expanded to eleven points, with discriminant validity first. It also covers the data-driven residual covariances and the parcel allocation.

Revisions: Section 5

 

Comment 19: Qualify the conclusion.

Response: The association remained strong without devaluation, but devaluation carried the unique association. The Conclusions therefore refer to a strong association between burnout-related experiences, especially sport devaluation and amotivation, and dropout intention, rather than to a firmly established distinction or causal sequence.

Revisions: Section 6

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript is now suitable for acceptance.

Author Response

Comment 1: The manuscript is now suitable for acceptance.

Response: We are grateful to Reviewer 1 for the careful evaluation of our manuscript across both review rounds and for the positive recommendation. The constructive comments in the first round substantially improved the clarity and rigor of the paper. No further changes were required in response to this comment. The revisions addressing Reviewer 4's remaining comments have been incorporated directly into the revised manuscript.

Reviewer 4 Report

Comments and Suggestions for Authors

Dear authors,

I appreciate the substantial effort you have devoted to revising the manuscript and providing detailed responses to the comments raised during the previous review round.

Nevertheless, I believe that several aspects still require further clarification before the manuscript can be considered for publication, as outlined below:

1. I appreciate the additional analyses conducted to assess the discriminant validity between athlete burnout and dropout intention. The comparison of measurement models, the assessment of the heterotrait–monotrait ratio (HTMT), and the examination of confidence intervals represent important improvements. However, the latent correlation remains very high, and the Fornell–Larcker criterion is still not satisfied. I consider that you have adequately addressed the initial analytical request; nevertheless, I recommend consistently maintaining a cautious interpretation of the empirical distinctiveness of these constructs, including in the Abstract, Highlights, and Conclusions sections. The findings should be presented as evidence of strong associations between burnout-related experiences and dropout intention, without suggesting that the discriminant validity of these constructs has been conclusively established.
2. The sensitivity analyses concerning the operationalization of motivation represent an important contribution to the revised manuscript. However, these analyses indicate that the sequential indirect association identified in the main model depends considerably on the use of the Relative Autonomy Index (RAI) and is primarily driven by the amotivation component. Replacing the RAI with a latent autonomous motivation factor results in the sequential indirect association no longer being statistically significant. I recommend that this finding be reflected more explicitly in the interpretation of the study's theoretical contribution and in the discussion of hypothesis H6. In particular, the findings should not be interpreted as providing general support for a protective role of autonomous motivation, but rather as highlighting the specific importance of amotivation in relation to athlete burnout.
3. The clarifications regarding instrument adaptation are welcome; however, they also highlight additional measurement-related limitations. The instruments were adapted without formal translation and cultural validation procedures, and their original response formats were modified. I believe that a more explicit methodological justification for these modifications is necessary, particularly regarding the changes to the original response formats of the BNSSS, SMS-II, and ABQ. Furthermore, I recommend clearly distinguishing the evidence of internal consistency and factorial structure obtained in the present sample from evidence demonstrating psychometric equivalence with the original instruments.
4. I appreciate the inclusion of the item-level analysis without parceling, which indicates the stability of the association between burnout and dropout intention. However, the poorer fit of the item-level model and the use of residual covariances introduced on the basis of modification indices suggest that the measurement structure should be interpreted with caution. I recommend explicitly clarifying that the stability of the structural coefficients does not eliminate concerns regarding factorial validity and that the favorable fit of the main model does not constitute independent confirmation of the measurement structure of the instruments.

In conclusion, I believe that the revised manuscript represents a substantial improvement over the original submission. Most of the concerns raised during the previous review round have been addressed through relevant additional analyses, and the interpretation of the findings has become more cautious. The remaining issues primarily concern clarification of the psychometric and theoretical implications of the results and do not necessarily require the collection of additional data.

Author Response

Comment 1. I appreciate the additional analyses conducted to assess the discriminant validity between athlete burnout and dropout intention… However, the latent correlation remains very high, and the Fornell–Larcker criterion is still not satisfied… I recommend consistently maintaining a cautious interpretation of the empirical distinctiveness of these constructs, including in the Abstract, Highlights, and Conclusions sections. The findings should be presented as evidence of strong associations between burnout-related experiences and dropout intention, without suggesting that the discriminant validity of these constructs has been conclusively established.

Response: We fully agree. Because the Fornell–Larcker criterion was not met and the upper limit of the HTMT confidence interval exceeded .85, discriminant validity cannot be regarded as established. We have revised the Highlights, Abstract, Results, Discussion, and Conclusions so that the findings are presented as strong associations between burnout-related experiences and dropout intention, and we have removed wording that could suggest established distinctiveness (e.g., “statistically distinguishable,” “the strongest correlates,” “the strongest association”).

Revised text (Highlights):

“…athlete burnout was strongly associated with dropout intention, although the two constructs overlapped substantially and their discriminant validity was not conclusively established.”

Revised text (Abstract):

“Burnout and dropout intention overlapped substantially (latent r = .852; Fornell–Larcker criterion not met)… Conclusions: Burnout-related experiences, particularly sport devaluation, were strongly associated with dropout intention, but discriminant validity was not conclusively established.”

Revised text (Section 3.2, end):

“…these results indicate that burnout and dropout intention were not statistically identical but strongly overlapping (a moderate discriminant-validity concern). Because the Fornell–Larcker criterion was not met, discriminant validity is not regarded as conclusively established, and the association is interpreted as a strong association between burnout-related experiences and dropout intention rather than as a relation between clearly distinct constructs.”

Revised text (Section 4.3):

“…discriminant validity was not conclusively established (the Fornell–Larcker criterion was not met, and the upper limit of the HTMT confidence interval exceeded .85)…”

Revised text (Conclusions):

“Burnout-related experiences were strongly associated with dropout intention; however, given their substantial overlap—especially with sport devaluation—and the unmet Fornell–Larcker criterion, the discriminant validity of the two constructs was not conclusively established.”

A note to Table 6 also states that H3 should be interpreted in light of this overlap.

Comment 2. …these analyses indicate that the sequential indirect association identified in the main model depends considerably on the use of the Relative Autonomy Index (RAI) and is primarily driven by the amotivation component… I recommend that this finding be reflected more explicitly in the interpretation of the study's theoretical contribution and in the discussion of hypothesis H6. In particular, the findings should not be interpreted as providing general support for a protective role of autonomous motivation, but rather as highlighting the specific importance of amotivation in relation to athlete burnout.

Response: We agree with this interpretation and thank the reviewer for stating it so clearly. We now state explicitly that H6 was supported only when motivation was operationalized as a composite index (the adapted RAI or a unit-weighted index), that the sequential association was not significant when a latent autonomous-motivation factor was used, and that the study’s theoretical contribution lies in the specific importance of amotivation rather than in a protective role of autonomous motivation. This qualification now appears where H6 is evaluated (Section 3.4, Table 6 note), in the discussion of H2 and H6 (Sections 4.2 and 4.5), and in the Highlights, Abstract, and Conclusions.

Revised text (Section 3.4):

“Moreover, this support was conditional on the operationalization of motivation: H6 was supported with the adapted RAI and with a unit-weighted index, but not when a latent autonomous-motivation factor was used (Section 3.6).”

Revised text (Section 4.2):

“Accordingly, the support for H2 does not indicate a general protective role of autonomous motivation against burnout; rather, it highlights the specific importance of amotivation.”

Revised text (Section 4.5):

“In particular, when the RAI was replaced by a latent autonomous-motivation factor, the sequential indirect association was no longer significant, whereas the corresponding association via amotivation remained significant. H6 should therefore be regarded as supported only when motivation was operationalized as a composite index (the adapted RAI or a unit-weighted index), and not when autonomous motivation was modeled as a latent factor. These findings do not provide general support for a protective role of autonomous motivation; rather, their theoretical contribution lies in highlighting the specific importance of amotivation, that is, the absence of perceived reasons for participating, in relation to athlete burnout and, through burnout, to dropout intention.”

Revised text (Conclusions):

“The sequential association was not significant when autonomous motivation was modeled as a latent factor and appeared to reflect the specific role of amotivation rather than a general protective role of autonomous motivation.”

Comment 3. …The instruments were adapted without formal translation and cultural validation procedures, and their original response formats were modified. I believe that a more explicit methodological justification for these modifications is necessary, particularly regarding the changes to the original response formats of the BNSSS, SMS-II, and ABQ. Furthermore, I recommend clearly distinguishing the evidence of internal consistency and factorial structure obtained in the present sample from evidence demonstrating psychometric equivalence with the original instruments.

Response: We thank the reviewer for this important comment and respond to it candidly. All items were administered in a single 5-point Likert agreement format so that the online questionnaire was consistent, allowing participants to respond to every item in the same format. This was a practical decision about questionnaire administration rather than one based on a psychometric rationale, and we acknowledge that it departs from the original instruments. Reducing the BNSSS and SMS-II from seven to five response options may affect score variability, and replacing the ABQ frequency format with an agreement format shifts the focus from how often burnout-related experiences occur to how strongly athletes endorse them. We have therefore (a) stated this rationale explicitly in Section 2.2, (b) acknowledged that scores are not directly comparable with those from the original instruments or with published norms, and (c) distinguished the internal consistency and factorial structure observed in this sample from psychometric equivalence with the original instruments, which was not tested. We also state explicitly in Section 2.2 and in the Limitations that the Korean items were adapted rather than formally translated and that no back-translation, formal content-validity rating, or pilot test was conducted, so cultural and linguistic equivalence with the original instruments was not established. The same points are now stated in the Limitations.

Revised text (Section 2.2):

“Evidence of internal consistency and factorial structure was therefore obtained only in the present sample (Section 3.2 and Tables S2–S3); this evidence describes how the adapted items functioned in this sample and does not demonstrate psychometric equivalence with the original instruments, as measurement invariance between the adapted and original versions was not tested. … This single response format was adopted to maintain consistency across the online questionnaire, so that participants could respond to all items in the same format. We acknowledge, however, that this decision departs from the original instruments: reducing the number of response options for the BNSSS and SMS-II may affect score variability, and replacing the frequency format of the ABQ with an agreement format shifts the focus from how often burnout-related experiences occur to the extent to which athletes endorse them. Scores obtained in this study are therefore not directly comparable with those obtained with the original instruments or with published norms, and this is acknowledged as a limitation (Section 5).”

Revised text (Section 5, Limitations):

“…The use of a uniform 5-point agreement format, which differed from the original response formats of the BNSSS, SMS-II, and ABQ, further limits direct comparability with the original instruments and does not establish psychometric equivalence.”

Comment 4. …the poorer fit of the item-level model and the use of residual covariances introduced on the basis of modification indices suggest that the measurement structure should be interpreted with caution. I recommend explicitly clarifying that the stability of the structural coefficients does not eliminate concerns regarding factorial validity and that the favorable fit of the main model does not constitute independent confirmation of the measurement structure of the instruments.

Response: We agree. Section 2.4 already noted that the data-driven residual covariances preclude treating the measurement-model fit as independent confirmation. We have now added explicit statements in the sensitivity-analysis results and in the Limitations that the stable structural coefficients do not resolve concerns about factorial validity and that the favorable fit of the main model does not independently confirm the measurement structure of the instruments.

Revised text (Section 3.6):

“The stability of the burnout–dropout coefficient in the item-level model should not, however, be taken as evidence of factorial validity: that model fitted poorly (CFI = .838, RMSEA = .088; Table 9), and stable structural coefficients do not resolve concerns about the measurement structure.”

Revised text (Section 5, Limitations):

“Item parceling and data-driven within-subdimension residual covariances may also have obscured item-level misfit; the stability of structural coefficients across specifications does not resolve concerns about factorial validity, nor does the favorable fit of the main model constitute independent confirmation of the measurement structure.”