Crossing Cultures, Gaining Weight? A Multidimensional Analysis of Health Behaviors in Chinese Students Overseas

Response to Reviewer #1 Comments

Thank you very much for taking the time to review our manuscript and for the constructive comments. We have addressed each point in a point by point response below and revised the manuscript accordingly. All revisions and corrections are visible in the resubmitted files via Track Changes; in addition, for ease of review, all modified text within the manuscript is explicitly marked in red font.

Point-by-point response to Comments and Suggestions for Authors

Comments 1: Resolve the ethical contradiction (waiver vs. informed consent) and document the consent process.

Response 1:  Thank you for pointing this out. We have corrected the inconsistency and clarified the ethics section.

This study was reviewed and approved by the Kyung Hee University Institutional Review Board (IRB: KHGIRB-25-341). Written informed consent was obtained from all participants prior to data collection; no waiver of consent was requested or granted. All data were de-identified before analysis to ensure confidentiality.

[Page3 , and line 123]

Comments 2: Describe and justify criteria for weight gain/loss (minimum percentage, time since immigration).

Response 2: Thank you for this important point. Previous weight was collected using an event-anchored item tied to the migration date. Specifically, participants were asked to report their weight immediately before departure and, to enhance accuracy, the question explicitly referenced the routine pre-departure medical examination required for university enrolment/visa procedures in many regions. Where available, participants were instructed to consult their official examination record (hospital/clinic booklet or electronic record) when reporting this value. This approach standardises recall to a uniform timepoint and reduces telescoping bias. We have clarified these procedures in the Methods and acknowledged residual recall limitations in the Discussion.

We have added this to the “Group” subsection of the Methods.

Weight change was calculated as current minus previous weight. For the primary analysis, we classified gain if Δweight > 0 and loss if Δweight < 0. Previous (“baseline”) weight was self-reported at an event-anchored timepoint: “What was your body weight (kg) during the two weeks before leaving for Korea?” To enhance accuracy, participants were reminded that a pre-departure medical examination is commonly required for enrolment visa procedures; where available, they were asked to consult the recorded weight from that examination when answering. Current weight was self-reported in kilograms.

[Page3 , and line 138]

Comments 3: Add sensitivity analyses—model continuous variation and apply clinical thresholds (≥5% or ≥3 kg).

Response 3: Thank you for this suggestion. We agree that modelling continuous Δweight and applying clinical thresholds can be informative. In the present dataset, baseline weight is pre-departure and self-reported, and with only two time points, small absolute errors (e.g., rounding on recalled weight) would make near-threshold reclassification (±1.0 kg or ≥3 kg/≥5%) unstable and create sparse strata, risking misleading inferences. We therefore kept the a priori directional outcome (gain: Δ>0; loss: Δ<0) and added an explicit limitation noting that threshold-based definitions are preferable in future longitudinal work with objective measurements. In the section on limitations, we have added:

We acknowledge that this sign-based definition may capture very small, non-clinical fluctuations; future longitudinal work should adopt clinically interpretable thresholds (e.g., ±1.0 kg or ≥5%) and include objective monitoring and dietary assessment.

[Page14 , and line 374]

Comments 4: Include (or justify absence of) dietary measures—without energy-intake control, weight-gain interpretation is incomplete.

Response 4: Thank you for this important point. We did not collect dietary variables. We have added an explicit limitation noting that diet is a major unmeasured confounder and explaining how unmeasured dietary changes could affect the interpretation of PA- and sleep-related associations. The revised text has been inserted in Discussion Limitations.

For example, we cannot distinguish a third-year student who lost weight early and later regained it from one who steadily gained weight throughout the study period. Repeated measures would be required to resolve the trajectory. Finally, despite rigorous exclusion criteria and data cleaning, unmeasured confounding (e.g., diet composition, hormonal status, cultural adaptation) may still influence the observed relationships. Future longitudinal studies utilizing wearable devices and biomarker data could provide more robust causal inferences.

[Page14 , and line 377]

Comments 5: Re-examine statistical models—avoid adjusting for sex in stratified analyses; consider adjustment for duration of stay and socioeconomic factors; control for multiplicity.

Response 5: Thank you for the suggestion. In our design, weight change is calculated as current minus pre-departure weight, with the baseline event-anchored to “two weeks before leaving for Korea”; participants were prompted to consult pre-departure medical-exam records where available. Because the survey captured one current weight and eligibility required ≥6 months of residence in Korea, our estimates reflect net change from pre-departure to the survey timepoint and cannot determine whether the change began in the first year or later years. We have clarified this in the Limitations and note that a prospective, multi-wave cohort from matriculation to graduation would be needed to identify onset and trajectories.

Moreover, because weight change was defined as the difference between a pre-departure (event-anchored) baseline and a single current measure after ≥6 months in Korea, our estimates represent net change and do not identify when during university (e.g., first vs. later years) the change began.

[page 14, line 361]

Comments 6: Assessment noted that while IPAQ, STQ, and ASSIS are cited, details on Chinese versions/validation were missing.

Response 6: Thank you for this helpful comment. We have added a dedicated subsection—2.3.6. Questionnaires (Chinese Versions and Adaptation)

2.3.6. Questionnaires (Chinese Versions and Adaptation)

All three instruments were administered in Chinese, as follows: (i) the IPAQ (short, self-administered) was downloaded from the official IPAQ repository and scored using the 2005 combined protocol; (ii) the Sleep Timing Questionnaire (STQ) does not have an official or published Chinese form, so sleep timing was measured using a Chinese adaptation developed via forward–back translation, expert-panel reconciliation, and cognitive debriefing with Chinese international students; and (iii) the Acculturative Stress Scale for International Students (ASSIS) Chinese version, previously validated in Chinese student samples. No wording changes were made to the IPAQ or ASSIS beyond formatting for online delivery.

[page 5, line 195 ]

Comments 7:  Definitions of PA/SB/sleep categories were unclear.

Response 7: Thank you for pointing this out. We have explicitly operationalized all categories in the Methods and table notes.

Physical activity (PA): Weekly minutes were computed as days × minutes/session for moderate (MPA) and vigorous (VPA) activity. Categories were defined as: Insufficient (<150 min/week MPA and <75 min/week VPA), Optimal (meeting either ≥150 min/week MPA or ≥75 min/week VPA), and High (exceeding ≥300 min/week MPA or ≥150 min/week VPA, or the equivalent combination).

Sedentary behaviour (SB): Average daily sitting time was derived from weekday (×5) and weekend (×2) reports. Categories: Low (<5 h/day), Normal (5-8 h/day), Excessive (>8 h/day). SB was defined as waking behaviour in a sitting, reclining, or lying posture with ≤1.5 METs; prolonged standing was not considered sedentary. Sleep: Self-reported average sleep duration was grouped as Short (<7 h/day), Normal (7–9 h/day), Long (>9 h/day).

[page 4, line 153]

Comments 8: Language sometimes overly causal.

Response 8: Thank you for this point. We have systematically moderated causal language across the manuscript.

This study highlights gender-specific associations among psychological health, PA, SB, and sleep duration in determining weight change among Chinese international students. While stress and moderate PA were associated with weight change in males, fatigue and vigorous PA were more salient correlates among females. Sleep deprivation was consistently associated with weight gain across both sexes. Given the cross-sectional design, these findings indicate associations rather than causal effects.These findings reflect the broader impact of environmental and behavioral adaptation in migrant student populations and emphasize the need for integrated, tailored health promotion strategies.

[page 14, line 385]

Additional clarifications

Regarding your point that using only age and sex as covariates is insufficient, the authors should consider including additional potential confounders such as duration of stay, living arrangements (collected but not controlled), and academic level/program type. Through this study we also identified several limitations and hope to address and supplement these in future research. Regarding your suggestion to tone down the causative connotation of “predictor,” we have made minor revisions to several sections. In the discussion on further policy and intervention implications (e.g., university health promotion programs), we added the following at the end: Practically, universities should consider sleep-hygiene programs for international students (regular sleep-wake schedules, limits on late-night screen use, and quiet hours policies). Complementary actions include sedentary break prompts on campus; accessible, supervised PA options (emphasizing moderate-intensity offerings for males and fatigue-aware pacing for females); and stress/acculturation supports (brief CBT or mindfulness workshops, peer mentoring, and language/cultural services).

Response to Reviewer #2 Comments

Thank you very much for taking the time to review our manuscript and for the constructive comments. We have addressed each point in a point by point response below and revised the manuscript accordingly. All revisions and corrections are visible in the resubmitted files via Track Changes; in addition, for ease of review, all modified text within the manuscript is explicitly marked in red font.

Point-by-point response to Comments and Suggestions for Authors

Comments 1: How did the authors determine the previous weight? How long before the participants moving abroad was considered valid (e.g. body weight recall from 1 year ago, 3 years ago)?

Response 1: Thank you for this important point. Previous weight was collected using an event-anchored item tied to the migration date. Specifically, participants were asked to report their weight immediately before departure and, to enhance accuracy, the question explicitly referenced the routine pre-departure medical examination required for university enrolment/visa procedures in many regions. Where available, participants were instructed to consult their official examination record (hospital/clinic booklet or electronic record) when reporting this value. This approach standardises recall to a uniform timepoint and reduces telescoping bias. We have clarified these procedures in the Methods and acknowledged residual recall limitations in the Discussion.

We have added this to the “Group” subsection of the Methods.

Weight change was calculated as current minus previous weight. For the primary analysis, we classified gain if Δweight > 0 and loss if Δweight < 0. Previous (“baseline”) weight was self-reported at an event-anchored timepoint: “What was your body weight (kg) during the two weeks before leaving for Korea?” To enhance accuracy, participants were reminded that a pre-departure medical examination is commonly required for enrolment visa procedures; where available, they were asked to consult the recorded weight from that examination when answering. Current weight was self-reported in kilograms.

[Page 3, and line 138]

Comments 2: I suggest describing physical activity, sleep, and sedentary behaviour in more detail: what are the cutoffs for high, optimal, and insufficient, or excessive, normal, and low?

Response 2: Thank you for this suggestion. We agree and have added explicit operational definitions and cutoffs for PA, SB, and sleep in the Methods.

Physical activity (PA): Weekly minutes were computed as days × minutes/session for moderate (MPA) and vigorous (VPA) activity. Categories were defined as: Insufficient (<150 min/week MPA and <75 min/week VPA), Optimal (meeting either ≥150 min/week MPA or ≥75 min/week VPA), and High (exceeding ≥300 min/week MPA or ≥150 min/week VPA, or the equivalent combination).

Sedentary behaviour (SB): Average daily sitting time was derived from weekday (×5) and weekend (×2) reports. Categories: Low (<5 h/day), Normal (5-8 h/day), Excessive (>8 h/day). SB was defined as waking behaviour in a sitting, reclining, or lying posture with ≤1.5 METs; prolonged standing was not considered sedentary.

Sleep: Self-reported average sleep duration was grouped as Short (<7 h/day), Normal (7–9 h/day), Long (>9 h/day).

[Page 4, and line 153]

Comments 3: Lines 209-238 and 239-269 are hard to follow. I suggest describing only the main or significant results.

Response 3: Thank you for this suggestion. We streamlined the Results, removing lengthy, non-significant details and keeping.

Comments 4: Tables 3 and 4, and Figure 1 are redundant. I suggest moving Tables 3 and 4 to the supplement.

Response 4: Thank you for the suggestion. We respectfully prefer to retain Tables 3–4 and Figure 1 in the main text. The tables present the primary sex-stratified multinomial models with exact estimates (ORs, 95% CIs, p) necessary for transparency and appraisal, while Figure 1 provides a complementary at-a-glance visual summary to aid interpretation across categories and sexes. These serve distinct, non-redundant purposes. To further ease interpretation, we have added a concise Notes section to the figure caption and table footnotes.

Comments 5: Lines 292-294, please provide citation and examples from other studies.

Response 5: Thank you for the suggestion. We have added the following sentence to the Discussion:

This observation is consistent with evidence that acute stress can suppress appetite (hypophagia) via CRF and noradrenergic pathways and with laboratory studies showing that men consume less under stress whereas women tend to increase intake of palatable foods (Zellner et al., 2006; Zellner et al., 2007).

[Page 13, and line 306]

Comments 6: Lines 295-304, please consider that the authors did not measure body composition. Thus, this study could not rule out that in people who were exercising, body weight increase was caused by increased muscle mass. I suggest that the authors explore this further and also include it as a limitation.

Response 6: Thank you for this helpful suggestion. We agree that, without body‐composition data, weight gain among more active participants could reflect lean mass rather than adiposity. We have added this as a limitation and cautioned interpretation of PA-related associations in the Discussion:

Second, reliance on self-reported data introduces the potential for recall bias, social desirability bias, and misclassification. Third, because body composition was not measured (e.g., fat mass vs. Fat free mass), we cannot rule out that observed weight gain among more active participants reflects lean mass accretion rather than increased adiposity; therefore, associations involving PA should be interpreted with caution.

[Page14 , and line 366]

Comments 7: Lines 295-304, I suggest providing suggestions for exercises for maintaining weight for both genders that could be explored further based on these findings.

Response 7: Thank you for your suggestions. Regarding specific measures, we have incorporated them into the conclusion section.

Practically, universities should consider sleep-hygiene programs for international students (regular sleep wake schedules, limits on late-night screen use, and quiet hours policies). Complementary actions include sedentary break prompts on campus; accessible, supervised PA options (emphasizing moderate-intensity offerings for males and fatigue-aware pacing for females); and stress/acculturation supports (brief CBT or mindfulness workshops, peer mentoring, and language/cultural services).

[Page14 , and line 393]

Response to Reviewer 3 Comments

Thank you very much for taking the time to review our manuscript and for the constructive comments. We have addressed each point in a point by point response below and revised the manuscript accordingly. All revisions and corrections are visible in the resubmitted files via Track Changes; in addition, for ease of review, all modified text within the manuscript is explicitly marked in red font.

Point-by-point response to Comments and Suggestions for Authors

Comments 1: The analysis does not highlight whether the weight gain/loss process occurred in students in the first year of university, in subsequent years, or in students in all university years. Such an analysis would have been useful to see when the weight change process is triggered. We believe that a complex analysis, which is truly useful and based on which appropriate measures can be taken, involves not only establishing the burden at a given moment, but also its evolution from the student's arrival until the completion of their studies.

Response 1: Thank you for the suggestion. In our design, weight change is calculated as current minus pre-departure weight, with the baseline event-anchored to “two weeks before leaving for Korea”; participants were prompted to consult pre-departure medical-exam records where available. Because the survey captured one current weight and eligibility required ≥6 months of residence in Korea, our estimates reflect net change from pre-departure to the survey timepoint and cannot determine whether the change began in the first year or later years. We have clarified this in the Limitations and note that a prospective, multi-wave cohort from matriculation to graduation would be needed to identify onset and trajectories.

Moreover, because weight change was defined as the difference between a pre-departure (event-anchored) baseline and a single current measure after ≥6 months in Korea, our estimates represent net change and do not identify when during university (e.g., first vs. later years) the change began.

[page 14, line 361]

Comments 2: How can one know if a third-year student, who declares that he weighs more than he did when he arrived in South Korea, did not initially lose weight after arriving in Korea, only to gain weight later. Or vice versa.

Response 2: Thank you for this important point. In our study, weight change is defined as current minus pre-departure weight (event-anchored to two weeks before leaving for Korea). With only one current measure, we can estimate net change but cannot determine the temporal trajectory (e.g., whether a third-year student first lost weight and later regained it, or steadily gained). We have clarified this in the Limitations and note that multi-wave longitudinal follow-up from matriculation to graduation is needed to identify onset and trajectories.

For example, we cannot distinguish a third-year student who lost weight early and later regained it from one who steadily gained weight throughout the study period. Repeated measures would be required to resolve the trajectory.

[page 14, line 377 ]

Comments 3: Of the 445 students participating in the study, did not even one maintain their initial weight?

Response 3: Thank you for the question. We excluded the five students with Δweight = 0 (≈1.1%) from the analytic sample, because our primary outcome was defined directionally (gain: Δ>0; loss: Δ<0), the stable category provides no directional information, and the very small cell would yield unstable estimates in regression (risk of quasi-separation/large SEs). This change is now clarified in the Methods.

Weight change was calculated as current minus previous weight. For the primary analysis, we classified gain if Δweight > 0 and loss if Δweight < 0.

[page 3, line 138]

Response to Reviewer 4 Comments

Thank you very much for taking the time to review our manuscript and for the constructive comments. We have addressed each point in a point by point response below and revised the manuscript accordingly. All revisions and corrections are visible in the resubmitted files via Track Changes; in addition, for ease of review, all modified text within the manuscript is explicitly marked in red font.

Point-by-point response to Comments and Suggestions for Authors

Comments 1: The binary classification into "weight gain" vs. "weight loss" groups is problematic. What proportion of students maintained stable weight? This potentially substantial group appears to be excluded or misclassified. Using weight loss as the reference category is counterintuitive since weight stability would be more appropriate. The authors should clarify the classification criteria and justify the exclusion/classification of students with minimal weight changes.

Response 1: Thank you for raising this. Our outcome was defined directionally as Δweight = current − pre-departure: gain (Δ>0) vs loss (Δ<0). After screening, 5/445 students (1.1%) reported Δ=0 (exact stability). Because this cell is extremely small and provides no directional information, we excluded these cases from the analytic sample to avoid unstable estimates/quasi-separation and we note in the Limitations that a threshold-based definition (e.g., ±1.0 kg or ≥5%) would be preferable in future longitudinal work.

We acknowledge that this sign-based definition may capture very small, non-clinical fluctuations; future longitudinal work should adopt clinically interpretable thresholds (e.g., ±1.0 kg or ≥5%) and include objective monitoring and dietary assessment. For example, we cannot distinguish a third-year student who lost weight early and later regained it from one who steadily gained weight throughout the study period. Repeated measures would be required to resolve the trajectory.

[Page 14, and line 374]

Comments 2: Essential temporal information is absent from the manuscript. How long have participants been in South Korea? When did the weight changes occur? Do the behavioral assessments reflect current or pre-immigration habits? Without knowing whether students have been abroad for six months or several years, readers cannot properly interpret the associations reported. The authors should provide descriptive statistics on duration of stay, clarify the timing of all measurements, and consider including duration as a covariate or stratification variable in the analyses.

Response 2: Thank you for the suggestion. In our design, weight change is calculated as current minus pre-departure weight, with the baseline event-anchored to “two weeks before leaving for Korea”; participants were prompted to consult pre-departure medical-exam records where available. Because the survey captured one current weight and eligibility required ≥6 months of residence in Korea, our estimates reflect net change from pre-departure to the survey timepoint and cannot determine whether the change began in the first year or later years. We have clarified this in the Limitations and note that a prospective, multi-wave cohort from matriculation to graduation would be needed to identify onset and trajectories.

Moreover, because weight change was defined as the difference between a pre-departure (event-anchored) baseline and a single current measure after ≥6 months in Korea, our estimates represent net change and do not identify when during university (e.g., first vs. later years) the change began.

[page 14, line 361]

Comments 3: The categorization of physical activity levels ("optimal," "high," "insufficient") and sleep duration ("excessive," "low," "normal") lack clear definitions or cutoff values. These are central to the study findings, yet readers cannot determine what these categories represent. The authors should provide specific MET-minutes/week or minutes/day thresholds for PA categories, hours/day cutoffs for sleep categories, and citations supporting these categorization schemes.

Response 3: Thank you for this suggestion. We agree and have added explicit operational definitions and cutoffs for PA, SB, and sleep in the Methods.

Physical activity (PA): Weekly minutes were computed as days × minutes/session for moderate (MPA) and vigorous (VPA) activity. Categories were defined as: Insufficient (<150 min/week MPA and <75 min/week VPA), Optimal (meeting either ≥150 min/week MPA or ≥75 min/week VPA), and High (exceeding ≥300 min/week MPA or ≥150 min/week VPA, or the equivalent combination).

Sedentary behaviour (SB): Average daily sitting time was derived from weekday (×5) and weekend (×2) reports. Categories: Low (<5 h/day), Normal (5-8 h/day), Excessive (>8 h/day). SB was defined as waking behaviour in a sitting, reclining, or lying posture with ≤1.5 METs; prolonged standing was not considered sedentary. Sleep: Self-reported average sleep duration was grouped as Short (<7 h/day), Normal (7–9 h/day), Long (>9 h/day).

[Page 4, and line 153]

Comments 4: Relying on self-reported pre-immigration weight introduces substantial recall bias, particularly given the potential for several years to have elapsed. The accuracy of this primary outcome measure needs stronger justification. The authors should consider validating these reports through university health records or passport/visa documentation, or conduct sensitivity analyses excluding participants with longer durations of stay to assess the robustness of the findings.

Response 4: Thank you for this important point. Previous weight was collected using an event-anchored item tied to the migration date. Specifically, participants were asked to report their weight immediately before departure and, to enhance accuracy, the question explicitly referenced the routine pre-departure medical examination required for university enrolment/visa procedures in many regions. Where available, participants were instructed to consult their official examination record (hospital/clinic booklet or electronic record) when reporting this value. This approach standardises recall to a uniform timepoint and reduces telescoping bias. We have clarified these procedures in the Methods and acknowledged residual recall limitations in the Discussion.

Moreover, because weight change was defined as the difference between a pre-departure (event-anchored) baseline and a single current measure after ≥6 months in Korea, our estimates represent net change and do not identify when during university (e.g., first vs. later years) the change began.

[page 14, line 361]

Comments 5: While the introduction and discussion mention dietary changes as important factors, no dietary variables were measured or controlled for. Given that diet is arguably the most direct determinant of weight change, this represents a major confounding limitation. At minimum, this should be clearly acknowledged as a limitation with discussion of how unmeasured dietary changes might affect interpretation of PA and sleep associations.

Response 5: Thank you for this important point. We did not collect dietary variables. We have added an explicit limitation noting that diet is a major unmeasured confounder and explaining how unmeasured dietary changes could affect the interpretation of PA- and sleep-related associations. The revised text has been inserted in Discussion Limitations.

For example, we cannot distinguish a third-year student who lost weight early and later regained it from one who steadily gained weight throughout the study period. Repeated measures would be required to resolve the trajectory. Finally, despite rigorous exclusion criteria and data cleaning, unmeasured confounding (e.g., diet composition, hormonal status, cultural adaptation) may still influence the observed relationships. Future longitudinal studies utilizing wearable devices and biomarker data could provide more robust causal inferences.

[Page 14, and line 377]

Comments 6: No sample size calculation or power analysis is provided, and multiple comparison corrections are not discussed despite numerous statistical tests. Additionally, the recruitment strategy ("over ten universities") is vague. The authors should specify which universities participated, what sampling strategy was used at each institution, and how this might affect the generalizability of the findings.

Response 6: Thank you for these important points. This study was designed as an exploratory, multi-campus survey with the goal of achieving maximal feasible coverage, so no a priori sample-size/power calculation was performed. Participants were recruited via international-affairs offices, programme coordinators, Chinese student associations, campus e-bulletins, and student messaging groups across universities in major regions (Yonsei University, Korea University, Ewha Womans University, Kyung Hee University, Dongguk University, Pusan National University, Chung-Ang University, Hanyang University, Kyung Hee University, etc.). Sampling was convenience with limited snowballing.

Comments 7: Regarding what you mentioned The chi-square analysis in Table 2 is difficult to interpret. What is being compared to what? 

Response 7: Thank you for these important points. We have added the following to the Note of Table 2, to aid interpretation:

Cell entries are n (%) by sex. Percentages use the sex-specific total sample as the denominator. The Pearson χ² shown tests the independence between weight-change group (gain vs loss) and the listed categorical factor in the combined sample. Exact p-values are reported; significance markers: *** p < 0.001, ** p < 0.01, * p < 0.05. “Males/Females” denote sex-stratified counts; the χ² assesses the overall association, not a between-sex comparison.

[Page 7, and Table 2. Note]

Additional clarifications

Regarding your point that using only age and sex as covariates is insufficient, the authors should consider including additional potential confounders such as duration of stay, living arrangements (collected but not controlled), and academic level/program type. Through this study we also identified several limitations and hope to address and supplement these in future research.

Response to Reviewer 5 Comments

Thank you very much for taking the time to review our manuscript and for the constructive comments. We have addressed each point in a point by point response below and revised the manuscript accordingly. All revisions and corrections are visible in the resubmitted files via Track Changes; in addition, for ease of review, all modified text within the manuscript is explicitly marked in red font.

Point-by-point response to Comments and Suggestions for Authors

Comments 1: Design is suitable for associations; cross-sectional nature and self-report introduce bias; absence of dietary intake data weakens interpretation of “weight change.”

Response 1:  Thank you for this important point. We did not collect dietary variables. We have added an explicit limitation noting that diet is a major unmeasured confounder and explaining how unmeasured dietary changes could affect the interpretation of PA- and sleep-related associations. The revised text has been inserted in Discussion Limitations.

For example, we cannot distinguish a third-year student who lost weight early and later regained it from one who steadily gained weight throughout the study period. Repeated measures would be required to resolve the trajectory. Finally, despite rigorous exclusion criteria and data cleaning, unmeasured confounding (e.g., diet composition, hormonal status, cultural adaptation) may still influence the observed relationships. Future longitudinal studies utilizing wearable devices and biomarker data could provide more robust causal inferences.

[Page 14, and line 377]

Comments 2: Methods: specify whether pre-immigration weight/BMI was self-reported or obtained from records; clarify language validation for Chinese participants.

Response 2: Thank you for this important point. Previous weight was collected using an event-anchored item tied to the migration date. Specifically, participants were asked to report their weight immediately before departure and, to enhance accuracy, the question explicitly referenced the routine pre-departure medical examination required for university enrolment/visa procedures in many regions. Where available, participants were instructed to consult their official examination record (hospital/clinic booklet or electronic record) when reporting this value. This approach standardises recall to a uniform timepoint and reduces telescoping bias. We have clarified these procedures in the Methods and acknowledged residual recall limitations in the Discussion.

Moreover, because weight change was defined as the difference between a pre-departure (event-anchored) baseline and a single current measure after ≥6 months in Korea, our estimates represent net change and do not identify when during university (e.g., first vs. later years) the change began.

[page 14, line 361]

Comments 3: Results-some tables are dense; add brief interpretive summaries

Response 3: Thank you for this important point. We have added and modified the Note sections in Tables 2, 3, and 4.

Additional clarifications

Regarding your point that using only age and sex as covariates is insufficient, the authors should consider including additional potential confounders such as duration of stay, living arrangements (collected but not controlled), and academic level/program type. Through this study we also identified several limitations and hope to address and supplement these in future research. Regarding your suggestion to tone down the causative connotation of “predictor,” we have made minor revisions to several sections. In the discussion on further policy and intervention implications (e.g., university health promotion programs), we added the following at the end: Practically, universities should consider sleep-hygiene programs for international students (regular sleep-wake schedules, limits on late-night screen use, and quiet hours policies). Complementary actions include sedentary break prompts on campus; accessible, supervised PA options (emphasizing moderate-intensity offerings for males and fatigue-aware pacing for females); and stress/acculturation supports (brief CBT or mindfulness workshops, peer mentoring, and language/cultural services).

[page 14, line 393]