Who Performs Best Under Pressure? The Role of Sleep, Anxiety, and Attention in Exam Performance Across Medical, Law, and Economics Students

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript is dealing with a very important and very hot area of education. However, However, I have several comments and I recommend several major improvements.

The manuscript, describes research on humans, but does not include information regarding Ethics Committee or Institutional Review Board approval.

The sample examined is 51 students and is separated in 14 students in Medicine, 20 students in Law and 17 in Economics. For my opinion the sample separated in three scientific areas is quite limited for each group.

In statistical analysis, I recommend to apply Cronbach’s alpha for internal consistency to every questionnaire.

The authors, stated that “Group comparisons: One-way ANOVAs were used for normally distributed 191 data, and Kruskal-Wallis tests were used for non-normally distributed variables. 192 Post hoc analyses were conducted using Tukey’s HSD.” (191 – 193). Please provide more information about, where One-way ANOVA was used and where Kruskal-Wallis tests were used.

In introduction the authors present 2 research questions, but in the results the authors are dealing with 3 research questions.

The authors state that “Although they often possess more theoretical knowledge about health-promoting behaviours, this awareness did not seem to translate into healthier sleep practices—a paradox also discussed in the literature.” (267 – 270) Please provide relevant references.

Author Response

We sincerely thank Reviewer 1, as well as the editors, for their time, effort, and insightful feedback. The comments provided greatly contributed to improving the clarity, methodological precision, and overall coherence of the manuscript. All points were carefully considered and addressed through targeted revisions. The revised version now explicitly includes ethical approval information, clarifies statistical procedures, expands on sample-related limitations, aligns research questions across sections, integrates Cronbach’s alpha values for all instruments, and adds relevant supporting references in the Discussion. All changes are marked in the manuscript using Track Changes and labelled by reviewer comment.

Comment 1: “The manuscript… does not include information regarding Ethics Committee or Institutional Review Board approval.”

Response: We thank the reviewer for this observation. The manuscript included this information in Section 2.3 – Procedures. However, to enhance clarity and ensure visibility, we have slightly rephrased the sentence to explicitly state that the study received formal approval from the Ethics Committee, including the institutional name and number reference. Change implemented in manuscript – page 5, lines 316–317 [Change suggested by Reviewer 1 – Comment 1]

Updated sentence in the manuscript: This study received ethical approval from the Ethics Committee of Universidade Católica Portuguesa (CES-UCP), under reference number 40/2024.

Comment 2: “The sample examined is 51 students and is separated in 14 students in Medicine, 20 in Law and 17 in Economics. For my opinion the sample separated in three scientific areas is quite limited for each group.”

Response: We acknowledge that the small number of participants per academic program represents a significant limitation, especially when conducting group comparisons across distinct scientific domains. As explained in the manuscript, recruitment was constrained by students’ availability during the final exam period and by the differing enrollment sizes and attendance patterns across programs. In particular, the reduced number of medical students reflects the low registration rate for final exams in this course, as most students passed through continuous assessment. We have now expanded the limitations section to clarify both the logistical challenges and the interpretative caution required when analysing such a small and diverse sample. We agree that future studies should aim for larger and more balanced groups to ensure statistical power and generalizability. Change implemented in manuscript – page 11, lines 593–599 (Change suggested by Reviewer 1 – Comment 2).

Updated sentence in the manuscript: Moreover, the small sample size, conditioned by participant availability and the number of enrolled students in each academic program, also constituted a limitation. In particular, the reduced number of medical students reflects the low registration rate for final exams in this course, as most students passed through continuous assessment. Moreover, the small size and disciplinary separation of the three groups limits the generalizability of between-group comparisons. Future studies should therefore aim to include broader and more balanced samples to ensure statistical robustness and external validity

 

Comment 3: “In statistical analysis, I recommend to apply Cronbach’s alpha for internal consistency to every questionnaire.”

Response: We appreciate this valuable contribution. While internal consistency analyses (e.g., Cronbach’s alpha) were not recalculated in our sample, we fully acknowledge the importance of reporting such psychometric properties. Therefore, we have now added to the Instruments section the Cronbach’s alpha values reported in the Portuguese validation studies of each instrument used in this research. These coefficients reflect the internal consistency of the versions applied to our participants and ensure the psychometric soundness of the tools employed. Change implemented in manuscript – page 4, lines 258–259, 267, 273–274, 282–283, 292–293, 300 (Change suggested by Reviewer 1 – Comment 3).

Updated sentence in the manuscript: This version presented a Cronbach’s alpha of 0.75, indicating a good internal consistency.” (PSQI); “…which reported a Cronbach’s alpha of 0.83.” (ESS)

 “Internal consistency values for the Portuguese version were excellent, with Cronbach’s alpha of 0.91 for state anxiety and 0.89 for trait anxiety.” (STAI);  “…and presented a Cronbach’s alpha of 0.89.” (BDI-II); “The Portuguese version reported Cronbach’s alpha values ranging from 0.53 to 0.87, depending on the condition.” (Stroop); “…which demonstrated an internal consistency (Cronbach’s alpha) of 0.69.” (Go/No-Go).

 

Comment 4: “The authors stated that ‘Group comparisons: One-way ANOVAs were used for normally distributed 191 data, and Kruskal-Wallis tests were used for non-normally distributed variables. 192 Post hoc analyses were conducted using Tukey’s HSD.’ Please provide more information about where One-way ANOVA was used and where Kruskal-Wallis tests were used.”

Response: As requested, the manuscript was revised to clarify which variables were analyzed using One-way ANOVA and which required the use of Kruskal–Wallis tests. This clarification is grounded in the assumption testing procedures described in Section 2.4. Specifically, One-way ANOVAs were applied to the variables that met the assumptions of normality and homogeneity — including sleep quality (PSQI), state anxiety, trait anxiety, and Stroop Interference scores. In contrast, the Go/No-Go variable violated these assumptions and was therefore analyzed using the non-parametric Kruskal-Wallis test. These group comparisons addressed the first research question, which examined differences between academic programs regarding these psychological and cognitive dimensions. Section 2.4. Data Analysis was updated on page 5, after line 343, to explicitly specify which statistical test was used for each variable, and to clarify the link to the first research question. The sentence was added as a new paragraph following the existing bullet point on group comparisons.

Updated sentence in the manuscript: Specifically, group comparisons aimed to examine whether students from Medicine, Law, and Economics differed significantly in sleep quality (PSQI), anxiety (state and trait), and executive attention performance (Stroop Interference and Go/No-Go tasks), thereby addressing the first research question. One-way ANOVAs were conducted for variables that met the assumptions of normality and homogeneity, including sleep quality (PSQI), state anxiety, trait anxiety, and Stroop Interference scores. In contrast, the Go/No-Go task violated these assumptions and was therefore analyzed using the non-parametric Kruskal-Wallis test. This methodological choice ensured the appropriateness of statistical procedures according to the distributional characteristics of each measure.

 

Comment 5: “In introduction the authors present 2 research questions, but in the results the authors are dealing with 3 research questions.”

Response: We thank the reviewer for highlighting this inconsistency. As noted, the Results section includes a third research question (3.2: “Are medical students better positioned in terms of sleep, anxiety, and executive attention?”) which was not explicitly stated in the Introduction. This question reflects a relevant subdimension of the first research question, focusing specifically on medical students, who were hypothesized to have a more favorable psychological and cognitive profile due to their academic background. Although it was not initially listed among the formal research questions, it naturally emerged from the study’s aims, theoretical framework, and descriptive findings. Furthermore, it was consistently addressed in the statistical analyses. To ensure full alignment across the manuscript, the Introduction was revised to include this third research question alongside the original two, thereby accurately reflecting the scope and objectives of the study. The Introduction section (page 3, lines 165–167) was updated to include the third research question, ensuring consistency with the structure of the Results section.

Updated sentence in the manuscript: The results are presented in alignment with the three research questions guiding this study. Specifically, the analysis explores: (1) whether sleep quality, executive attention, and anxiety differ across academic programs; (2) whether medical students demonstrate superior cognitive and emotional functioning compared to their peers; and (3) whether these psychological variables predict academic performance under evaluative conditions.

Comment 6: “The authors state that ‘Although they often possess more theoretical knowledge about health-promoting behaviors…’ Please provide relevant references.” (267–270)

Response: We agree that this claim required appropriate support from literature. Two references have now been added to substantiate this observation:

• Alrashed et al. (2022), which highlights the paradox between medical students’ health-related knowledge and their heightened vulnerability to stress, anxiety, and poor sleep hygiene.
• Dahlin et al. (2005), which documents elevated levels of psychological distress and maladaptive behaviours among medical students, despite their background in health education.

These sources reinforce the discrepancy between theoretical awareness and behavioral practices, aligning with the argument presented in the Discussion.

Change: Addition of references (Page 9, lines 473–475) to support the paradox between medical students’ theoretical knowledge and actual behaviors.

Reviewer 2 Report

Comments and Suggestions for Authors

This study addresses a timely and important topic by examining how sleep quality, anxiety, and executive attention relate to exam performance across students in different academic disciplines. The methodology is sound and clearly described, and the manuscript is well organized and professionally written.

The introduction does a good job of setting the context and framing the need for this interdisciplinary comparison. The inclusion of both psychological (anxiety, depression) and cognitive (executive attention) measures, along with academic performance, is a notable strength.

The findings are informative, particularly in highlighting differences among student groups and identifying sleep quality and anxiety—but not executive attention—as significant predictors of performance. This nuance is valuable and should be emphasized more strongly in the conclusion.

Suggestions:

• While you acknowledge sampling and ecological validity concerns, further discussion on how these may affect interpretation of group comparisons (especially given differences in exam requirements) would strengthen the transparency of the study.

• The null findings for executive attention should be contextualized more critically—were the Stroop and Go/No-Go tasks too limited in capturing real-world academic functioning?

• The discussion could further explore how institutions might tailor support based on discipline-specific challenges—e.g., cognitive training for law students, stress management for medical students.

• A graph or figure showing exam grade distributions across groups or a regression plot might enhance accessibility and clarity.

 

Author Response

We sincerely thank Reviewer 2 for the thoughtful and constructive feedback, which offered valuable theoretical and methodological insights that helped refine the manuscript’s conceptual grounding and practical implications. All comments were carefully considered and fully addressed through targeted revisions.

The Introduction was strengthened to include a clearer theoretical justification for the selection of psychological constructs (sleep, anxiety, and executive attention) and their relationship to academic performance. Methodological transparency was enhanced by expanding the Limitations section to discuss sampling constraints, ecological validity, and potential selection bias across academic programs. The Discussion was substantially refined to contextualize null findings for executive attention, highlight the limits of task-based measures, and integrate practical, discipline-specific recommendations for academic support.

While we appreciated the suggestion to add figures, we opted to retain the current table-based format, as it already conveys the results effectively given the sample size. All changes are detailed in the point-by-point responses that follow.

 

Comment 1: “There should be a stronger theoretical justification for the chosen psychological constructs (sleep, anxiety, executive attention) and their relevance to academic performance. The authors may consider including further literature to support why these variables were selected and how they are expected to interact.”

Response: We thank the reviewer for highlighting the need to reinforce the theoretical rationale behind the selected psychological constructs. The variables examined in this study — sleep quality, anxiety (state and trait), and executive attention, were chosen based on robust empirical and theoretical foundations that link these domains to academic functioning, particularly under conditions of stress such as exam periods.

Firstly, the impact of sleep quality on academic performance is well-documented in the literature. Poor sleep is associated with impaired memory consolidation, reduced concentration, and diminished executive functioning (Gomes et al., 2009; Curcio et al., 2006). These effects are especially relevant during examination periods, where cognitive demands and stress are heightened.

Secondly, anxiety, particularly test anxiety and trait anxiety, has been consistently associated with poorer academic outcomes (Spielberger, 1983; Eysenck et al., 2007). High anxiety can disrupt attention, working memory, and information processing speed, all of which are crucial during high-stakes evaluations.

Finally, executive attention encompasses core cognitive processes such as inhibitory control, cognitive flexibility, and interference resolution, which are essential for goal-directed behavior and exam-related problem-solving (Lezak et al., 2012; Diamond, 2013). Tasks such as the Stroop and Go/No-Go provide standardized measures of these subdomains.

Together, these constructs were selected to capture the interplay between emotional regulation (via anxiety), physiological readiness (via sleep), and cognitive control (via attention), all of which contribute meaningfully to academic performance. We have now clarified this rationale in the Introduction and expanded the theoretical background accordingly. Change implemented in manuscript – page 2, lines 81–92 Expanded theoretical justification for the inclusion of sleep, anxiety, and executive attention as predictors of academic performance. [Change suggested by Reviewer 2 – Comment 1]

Updated sentence in the manuscript: To better understand the cognitive and emotional components influencing perfor-mance under pressure, three psychological domains have emerged as particularly rel-evant. Academic performance during high-stakes examinations relies not only on prior learning but also on students’ cognitive and emotional readiness. In this context, three psychological domains have shown strong empirical links to academic outcomes: sleep quality, anxiety, and executive attention. Poor sleep can impair memory, attention, and executive functioning, especially under academic stress [20 ,21]. Anxiety, particularly trait and test-related anxiety, has been associated with diminished cognitive efficiency and lower academic achievement [22, 23]. Finally, executive attention — including inhibitory control and interference resolution — is essential for sustaining focus, managing time, and adapting to demanding academic tasks [24, 25]. These variables were therefore selected as key predictors of academic performance during exams.

 

 

Comment 2: “While you acknowledge sampling and ecological validity concerns, further discussion on how these may affect interpretation of group comparisons (especially given differences in exam requirements) would strengthen the transparency of the study.”

Response: We thank the reviewer for highlighting the need to more thoroughly address the implications of sampling differences and ecological validity on group comparisons. Although the Limitations section already acknowledged this concern, we agree that further clarification was necessary regarding its influence on the interpretation of results.

Specifically, the medical and economics students who participated in this study were only required to take final exams if they failed the continuous assessment or opted to improve their grades, while law students were mandated to take the final exam regardless of prior performance. This structural difference may have introduced selection bias into the sample, particularly by over-representing academically vulnerable or highly motivated students in Medicine and Economics. Such a configuration may skew within-group variability and compromise the comparability of academic performance across fields.

Nonetheless, this selective participation also enhances the ecological validity of the study, as it reflects real-life academic dynamics under authentic pressure. As discussed in the Introduction and Limitations, academic performance under high-stakes conditions is not solely determined by cognitive abilities, but also by emotional and behavioral regulation, which may differ according to context and motivation.

We have now revised the Limitations section to explicitly clarify how these structural and motivational factors may differentially impact each group and influence outcome interpretation.

[Change implemented in manuscript – page 10, lines 569–577. Suggested by Reviewer 2 - Comment 2]

Updated sentence in the manuscript: Related to this, the mandatory nature of final exams for Law students, compared to the optional or compensatory nature of exam participation in Medicine and Economics, may have led to selection effects that differentially shaped the cognitive and emotional profiles of each group. Specifically, Medicine and Economics subgroups may include a disproportionate number of academically vulnerable or high-achieving students seeking grade improvement, whereas the Law sample represents a broader academic spectrum. These differences should be taken into account when interpreting comparative outcomes and underscore the importance of contextualizing group comparisons within pro-gram-specific assessment frameworks.

 

 

 

Comment 3: “The null findings for executive attention should be contextualized more critically—were the Stroop and Go/No-Go tasks too limited in capturing real-world academic functioning?”

Response: We appreciate the reviewer’s suggestion to critically contextualize the null findings for executive attention. This observation is especially pertinent given the complexity of academic functioning and the challenges of translating cognitive performance in controlled settings to real-world demands. As noted in the literature, tasks such as the Stroop Interference Task and the Go/No-Go Task are well-established tools for assessing components of executive attention—namely, interference control and inhibitory control, respectively (Diamond, 2013; Lezak et al., 2012). However, despite their psychometric strengths, these paradigms may lack ecological validity when it comes to capturing the multifaceted cognitive skills required during real academic evaluations, which often involve sustained attention, planning, and task-switching in dynamic contexts.

Furthermore, academic performance is influenced by the interaction of multiple domains, including motivation, study habits, emotional regulation, and metacognitive strategies (Almeida & Soares, 2003; Lezak et al., 2012). Thus, while executive attention remains theoretically relevant, the specific instruments used may not fully reflect the complex demands of exam performance in naturalistic settings. This limitation has now been acknowledged in the Discussion section to provide a more nuanced interpretation of the null results and to highlight the need for future studies to adopt more ecologically valid measures of executive functioning.Change implemented in manuscript – page 11, lines 600–606. [ Change suggested by Reviewer 2 – Comment 3]

Updated sentence in the manuscript: Regarding the cognitive measures, it is also important to consider that, despite their robust psychometric properties, the Stroop and Go/No-Go tasks may fall short in capturing the multidimensional nature of executive demands in real academic contexts. Exam situations often require prolonged attentional control, planning, decision-making under stress, and adaptive flexibility—functions that extend beyond the inhibitory and inter-ference resolution processes measured by these tools. As such, the null findings may reflect instrument limitations rather than the absence of an executive attention effect.

Comment 4: “The discussion could further explore how institutions might tailor support based on discipline-specific challenges—e.g., cognitive training for law students, stress management for medical students. The null findings for executive attention should be contextualized more critically—were the Stroop and Go/No-Go tasks too limited in capturing real-world academic functioning?”

Response: Following the identification of distinct cognitive-emotional profiles across academic programs, we agree that institutional support could be better tailored to the demands and challenges specific to each discipline. This perspective is aligned with our findings, which highlight the differential impact of sleep quality, anxiety, and executive attention on student performance under evaluative pressure.

In the revised manuscript, we have expanded the Discussion section to include practical implications for intervention. These include sleep hygiene and time management strategies for medical students, who exhibited higher anxiety and poorer sleep; emotional regulation and consistency-building for Economics students, whose performance was marked by weaker inhibitory control despite better sleep; and executive functioning support for Law students, whose strong exam performance may be hindered by lower sustained attention under verbal and time-pressured tasks.

By integrating these differentiated recommendations, we aim to emphasize the importance of designing context-sensitive academic interventions grounded in the cognitive and emotional profiles of students from each discipline. Change implemented in manuscript – page 9, lines 495–507. [ Change suggested by Reviewer 2 – Comment 4]

Updated sentence in the manuscript: Such preliminary insights have practical implications, particularly insights for tailoring support strategies to the specific cognitive and emotional challenges faced by students in each academic discipline. Medical students, who reported higher anxiety and poorer sleep quality, may benefit from interventions focused on sleep hygiene, time management, and emotion regulation to mitigate the effects of academic overload [20,5]. Law students, who exhibited strong academic outcomes but weaker inhibitory control, could benefit from executive attention training and techniques to strengthen sustained focus under verbal and time-pressured conditions [24]. For Economics students, whose performance suggested more stable emotional profiles but lower inhibitory control, promoting self-regulatory strategies and consistent study routines [3] may enhance academic con-sistency. These differentiated approaches highlight the importance of designing field-sensitive academic support programs that address the interplay between cognitive demands, emotional readiness, and performance

Comment 5: “A graph or figure showing exam grade distributions across groups or a regression plot might enhance accessibility and clarity.”

 

Response: We agree that graphical representations can enhance the accessibility and interpretability of quantitative results. However, we decided not to include an additional regression figure or exam grade distribution plot, given the limited sample size and the fact that key descriptive data (means, standard deviations, ranges, and significant differences across fields) are already clearly presented in Table 1. Furthermore, the regression findings are straightforward, with only two significant predictors (sleep quality and state anxiety), which are already described in the Results and Discussion sections. For these reasons, we believe that an additional figure would not add meaningful value to the current version of the manuscript, but we remain open to including one should the editorial team deem it necessary.

Reviewer 3 Report

Comments and Suggestions for Authors

Overall, I thought the manuscript was well-written with a logical flow that touched on all the required areas. However, there are several issues that I find to be concerning:

The gaps listed (lines 57-75) do not reflect the research questions. Specifically, lines 59-60 addresses course-specific demands which is not examined empirically in this study. Also lines 68-69, which discuss attendance patterns are never addressed again. Similarly, lines 81-84 list the study's questions, but they differ from those listed on lines 199-202.

There are several concerns regarding research limitations as well, some of which are addressed by the authors, and others that are not. This includes the small sample size, using a convenience sample, differences in course assessment structure (lines 103-109), no explanation of how scores are assessed if there is no clinical cut-off score for 3 of the instruments (lines 140, 157, and 163). Students provide self-reported exam scores, which may be inaccurate. 

Additionally, I find several assumptions throughout the manuscript concerning, particularly in lines 267-285. This section discusses the type of study and work that different students perform, but there are no references listed to validate that assumption. I also am wary of the assumption that certain students perform differently and have different sleep quality, executive attention, and anxiety due to the type of work a student performs within their major rather than the type of personality they have. This study assumes the former, an assumption that is too large for me to be comfortable with. Lines 283-285 state, ..."distinct academic programs elicit specific cognitive and behavioral profiles, shaped by the curricular structure, assess demands, and learning strategies typical of each course," but those things were not tested in this study. While there were connections between "distinct academic programs" and "specific cognitive and behavioral profiles", we cannot claim causation, and we cannot claim that those connections are due to "the curricular structure, assess demands, and learning strategies typical of each course." These assumptions are too large to justify from this limited study.

Regarding references: This study examined university students aged 19-25, with an average age of 20. Half (~10) of the references are for studies on adolescence (ages 10-19) and half (~10) of the references are for studies on university students. The studies on adolescence should be removed as that research does not adequately reflect the population being studied in this manuscript.

Another concern regarding references is the age of the studies referenced. Not including references to the scales/index/questionnaires used in the methods section, 18 of the 29 other references were more than 10 years old.

Finally, two references within the discussion section reference a test manual, when I was expecting them to reference a study or literature review (reference 41 and 6 on lines 279 and 298 respectively). I don't see how manuals from 1993 and 1983 could make the points the authors are trying to make in those instances.

My suggestion regarding the literature used is to tighten the scope of references to include only university students within the last 10 years, if possible, and increase the number of references made. There are too many assumptions made without any reference to research for my comfort. 

Author Response

We sincerely thank Reviewer 3 for the thoughtful and constructive feedback. All comments were carefully considered and led to substantial improvements in the manuscript. The issues raised regarding conceptual assumptions, methodological limitations, and reference selection were addressed through targeted revisions, as detailed in the responses to Comments 1–6.

 

Comment 1: “The gaps listed (lines 57–75) do not reflect the research questions. Specifically, lines 59-60 address course-specific demands which is not examined empirically in this study. Also, lines 68-69, which discuss attendance patterns are never addressed again. Similarly, lines 81-84 list the study’s questions, but they differ from those listed on lines 199–202.”

Response: We thank the reviewer for this thoughtful observation. In response, we removed the paragraph previously placed between lines 72–75, which included research gaps unrelated to the study’s design (e.g., curricular structure, attendance patterns, or national scope). These elements had already been identified and adjusted following feedback from previous reviewers. Consequently, the currently retained paragraph at the end of the introduction now aligns directly with the research questions presented on lines 199–202 and focuses solely on psychological variables (sleep, anxiety, executive attention), without unsupported assumptions. No new additions were made at this stage, as the issue had already been addressed. Change implemented in manuscript – page 2 [Deletion of previous paragraph on contextual gaps (e.g., attendance, institutional scope).]

Updated sentence in the manuscript: None added for this revision. The updated paragraph was previously introduced and already reflects the reviewer’s concern.

 

Comment 2: “There are several concerns regarding research limitations as well, some of which are addressed by the authors, and others that are not. This includes the small sample size, using a convenience sample, differences in course assessment structure (lines 103–109), no explanation of how scores are assessed if there is no clinical cut-off score for 3 of the instruments (lines 140, 157, and 163). Students provide self-reported exam scores, which may be inaccurate.”

Response: We appreciate the reviewer’s observation regarding the need for greater methodological transparency. In response, we revised the “Limitations and Future Directions” section to explicitly include two points raised:

(1) the absence of clinical cut-off scores for three of the cognitive and emotional measures (STAI, Stroop Test, and Go/No-Go Task), and

(2) the reliance on self-reported final exam grades, which may be influenced by social desirability or memory bias.

These additions aim to clarify that the study’s conclusions are based on relative comparisons within the sample, rather than clinical thresholds, and that exam performance data should be interpreted with caution due to the self-report format. This ensures full alignment with the empirical scope and methodological constraints of the study.Change implemented in manuscript – page 11, lines 586–592. [Change suggested by Reviewer 3 – Comment 2]

Updated sentence in the manuscript: Third, limitations were also found at the level of instrumentation. Three of the in-struments used (STAI, Stroop Test, and Go/No-Go) do not have established clinical cut-off scores, which limits the interpretation of individual scores in clinical terms. As such, results should be viewed as relative comparisons within the sample, not diagnostic in-dicators. Moreover, the use of self-reported final exam grades may introduce bias due to memory inaccuracies or social desirability, which should be considered when interpreting academic performance outcomes.

 

Comment 3: “The manuscript assumes that academic programs shape distinct cognitive and behavioral profiles (lines 267–285), but this was not empirically tested and lacks sufficient references. While the discussion connects observed group differences to curricular demands, the study did not measure those demands or establish causal links. These assumptions are too large to justify from this limited study.”

Response: In response, we revised the discussion section (lines 267–285) to remove assertions suggesting causal links between curricular structures and cognitive profiles. Instead, we now interpret the observed group differences in sleep quality, anxiety, and executive attention as potential reflections of broader psychosocial dynamics that may vary by academic environment, without attributing them directly to course structure.

To address the concern regarding unsupported assumptions, we eliminated the statement that “distinct academic programs elicit specific cognitive and behavioural profiles,” and reframed the interpretation as a theoretical possibility, not an empirically established fact. Additionally, while Dyrbye et al. (2005) is cited in this section, we retained this reference because it is a widely acknowledged and extensively cited study on mental health in medical students, aligning with our sample and research scope. Its inclusion helps contextualize the findings regarding medical students’ higher levels of anxiety and poorer sleep quality, while avoiding overstated claims of causality.

These adjustments ensure consistency with the empirical data collected, respect the limitations of the study design, and maintain alignment with our theoretical framework.Change implemented in manuscript – page 9, lines 489–494. [Change suggested by Reviewer 3 – Comment 3]

Updated sentence in the manuscript: Taken Together, these findings may offer preliminary insights into how emotional and cognitive patterns vary across academic contexts, potentially shaped by intersecting factors such as learning demands, stress exposure, and self-regulation habits. However, given that the curricular structures were not directly assessed, future research should explore whether and how academic environments contribute to differentiated psycho-logical functioning among students.

 

Comment 4-6: “Regarding references: This study examined university students aged 19–25, with an average age of 20. Half (~10) of the references are for studies on adolescence (ages 10–19) and half (~10) of the references are for studies on university students. The studies on adolescence should be removed as that research does not adequately reflect the population being studied in this manuscript.

Another concern regarding references is the age of the studies referenced. Not including references to the scales/index/questionnaires used in the methods section, 18 of the 29 other references were more than 10 years old.

Finally, two references within the discussion section reference a test manual, when I was expecting them to reference a study or literature review (reference 41 and 6 on lines 279 and 298 respectively). I don’t see how manuals from 1993 and 1983 could make the points the authors are trying to make in those instances.

My suggestion regarding the literature used is to tighten the scope of references to include only university students within the last 10 years, if possible, and increase the number of references made. There are too many assumptions made without any reference to research for my comfort.”

Response: We sincerely thank the reviewer for the thoughtful and detailed comments regarding the scope of assumptions, methodological limitations, and appropriateness of references. These suggestions significantly contributed to improving the rigour and clarity of the manuscript.

In response to Comment 4, we carefully revised the Introduction and Discussion to clarify that course-specific cognitive and behavioural profiles are presented as exploratory interpretations, not tested hypotheses. We made efforts to avoid overgeneralisation and speculative causal claims, adjusting the language to reflect correlations rather than assumptions. While we recognise the limitations of our sample and design, we attempted — to the extent possible — to address the reviewer’s concerns and ensure that all interpretations remain grounded in the data and literature.

Regarding Comment 5, we conducted a comprehensive review of the references and replaced most adolescent-focused citations with studies involving university-aged populations. We also updated the citation base with more recent and context-appropriate literature (from the last 10 years), particularly in the Introduction and Discussion.

In response to Comment 6, the Limitations section was expanded to explicitly mention all relevant methodological constraints, including the use of a small, non-random sample, differences in course assessment structure, absence of clinical cut-offs for some instruments, and reliance on self-reported exam grades. These adjustments were implemented across the manuscript to improve transparency and scientific integrity.

 

Change Implemented: • Adjusted interpretative language in the Introduction and Discussion to avoid overstatements and unsupported assumptions; • Replaced outdated or adolescent-focused references with university-level, recent literature wherever possible; • Expanded the Limitations section to acknowledge all methodological constraints noted by the reviewer.

Reviewer 4 Report

Comments and Suggestions for Authors

This article takes a look at the intersection of exams and sleep quality in an interesting and much-needed study,  the introduction offers a broad overview but could have connected this to students a little more in-depth,  particularly as the study explores different domains of study this could be expanded to examine the specific literature in each of these separately as well as the literature on these topics more generally. 

Methods are well described  and would allow replication with minimal extra information. it would be helpful to include a little more data on how large the cohorts were and whether the students taking this study were representative.  it's also not clear about whether there is parity between these exams which potentially could account for some of the differences seen. in particular  stronger justification for inclusion of Stroop test  and go/no  tasks is needed. 

 the results raise a number of concerns, nonsignificant findings are reported for sleep quality as significant and some of the information around the multiple regression suggests errors in analysis reporting.  On addition  the grade ranges  are very small and it may be that an an ordinal regression and use of Z scores would work better here.  it's necessary to report model fit data when reporting regression and not just the p values.  the table is overly large and would work better as three columns plus a sample column.

 the discussion section would benefit from some expansion and consideration of factors that may have affected findings ( such as whether sleep quality was different for these individuals at other times)  advice for educators and students would also benefit from a little more depth relating to current literature as well as findings from this study. 

Author Response

We thank Reviewer 4 for the constructive and thoughtful feedback, which significantly contributed to improving the clarity and methodological rigor of our manuscript. Several of the suggestions made, namely, the expansion of the theoretical framing, clarification of sample composition and exam comparability, justification for task selection, and refinement of regression reporting, were already addressed through prior revisions in response to earlier reviewers. Nonetheless, your comments helped us refine these elements further. Specifically, we revised the Introduction to better connect the study to discipline-specific literature, clarified methodological details in the Participants and Instruments sections, corrected and expanded the regression analyses (including model fit indices and standardized coefficients), and refined the Discussion to incorporate contextual factors and tailored recommendations for each academic group. We appreciate your engagement with the manuscript and believe these changes have strengthened its contribution.

 

Comment 1: “…could have connected this to students a little more in-depth, particularly as the study explores different domains of study… examine the specific literature in each of these separately as well as the literature on these topics more generally.”

Response: We thank the reviewer for this thoughtful comment. In response, we revised the introduction to better connect the research questions to the realities and literature concerning university students. Specifically, we expanded the theoretical background by:

• Adding distinct cognitive and emotional characteristics associated with each academic field (Medicine, Law, and Economics), including typical stressors and assessment demands;
• Reinforcing empirical evidence regarding sleep quality, anxiety, and executive functioning in university populations, with updated references and reduced reliance on adolescent-focused literature;
• Enhancing the rationale for choosing these domains as key predictors during exam contexts.

These changes aim to provide a more comprehensive and discipline-sensitive framing of the study, improving conceptual depth and alignment with student realities.

Change implemented in manuscript – Section: Introduction - Pages: Page 2 to 4 approximately

Description of Edits: Added expanded theoretical rationale for each academic domain (e.g., clinical rotations in Medicine, verbal demands in Law, quantitative pressure in Economics), and included new empirical references (e.g., Goel et al., 2009; Mota Albuquerque et al., 2023; Kumari & Jain, 2014) to clarify the relevance of sleep, anxiety, and attention in university students.

 

Comment 2: “…include a little more data on how large the cohorts were and whether the students were representative…”; “…not clear about whether there is parity between these exams…”; “…stronger justification for inclusion of Stroop test and Go/No-Go tasks is needed.”

Response: In response, we implemented several clarifications to strengthen the transparency and rationale of our methodology:

• Regarding sample size and representativeness, we explicitly stated the distribution of participants by academic program (Medicine, Law, Economics), their year of enrollment, and the non-random nature of the sample. We clarified that recruitment was limited to one private university and does not represent the broader university student population.
• Regarding parity across exams, we explained that academic performance was measured using the final exam of the curricular unit with the highest number of ECTS credits in each program, ensuring comparable academic weight across fields. We also acknowledged structural differences in exam requirements (mandatory vs. optional), which may have introduced selection bias—this is now addressed both in the Methods and Limitations sections.
• Regarding the use of the Stroop and Go/No-Go tasks, we added a concise justification in the Instruments section, highlighting their validation for assessing executive attention, their brevity, and feasibility in academic settings. We also clarified the specific executive functions each test targets (interference resolution and impulse inhibition), while acknowledging their limited ecological scope.

These additions directly address the reviewer’s concerns and enhance the clarity and methodological robustness of the article.Change implemented in manuscript – Section: 2.1 Participants, 2.2 Instruments – Pages 3–4 [Change suggested by Reviewer 4 – Comment 2]

 

Comment 3: “the results raise a number of concerns, nonsignificant findings are reported for sleep quality as significant and some of the information around the multiple regression suggests errors in analysis reporting.  On addition  the grade ranges are very small and it may be that an an ordinal regression and use of Z scores would work better here.  it's necessary to report model fit data when reporting regression and not just the p values.  the table is overly large and would work better as three columns plus a sample column.”

Response: We thank the reviewer for this constructive and attentive comment, which prompted a careful reassessment of the statistical reporting and regression analysis. Upon review, we noticed that the initial version of the manuscript included a misaligned description of the regression results — namely, p-values were reported from the refined model (with two predictors), while the sentence implied they reflected the full model. We are grateful for this observation, which allowed us to clarify the analytical steps and ensure transparency and alignment between results and text.

We have now:

• Distinguished clearly between the full model (five predictors) and the refined model (two predictors).
• Added all relevant model fit indices, including adjusted R², F-statistics, and standardized beta coefficients (β).
• Confirmed and explicitly stated that all regression variables were standardized (Z-scores), ensuring comparability of effects across variables.

In regard to the presentation of descriptive statistics, we appreciate the reviewer’s suggestion to condense the table format. While we acknowledge that the table is extensive, we chose to retain its structure because it provides a compact yet complete overview of all core variables, organized by academic program. We believe this format enhances clarity and enables the reader to grasp key differences across groups without fragmenting the information across multiple tables.We thank the reviewer for this observation. Change implemented in manuscript – Section: 3.3 Do sleep quality, executive attention, and anxiety predict academic performance? – Pages 7 [Change suggested by Reviewer 4 – Comment 3]

Description of Edits:

• Section 3.3 was fully revised to: Correct the interpretation of the regression results; Present both full and refined models with model fit indices and standardized coefficients; Include an explicit note on the use of Z-scores in all regression analyses.
• The descriptive table was retained in its original layout for transparency and clarity.

 

Comment 4:  “The discussion section would benefit from some expansion and consideration of factors that may have affected findings (such as whether sleep quality was different for these individuals at other times). Advice for educators and students would also benefit from a little more depth relating to current literature as well as findings from this study.”

Response: In revising the manuscript in response to the comments of the other reviewers, we have already expanded and refined the Discussion section in ways that directly address these points.

Specifically, the revised version now:

• Incorporates a broader discussion of contextual factors that may have influenced the findings, including assessment structure differences across academic programs and the potential impact of compensatory strategies, self-regulation, and emotional management under evaluative stress.
• Integrates updated references to strengthen the theoretical alignment between our results and current literature on sleep, anxiety, and cognitive performance in higher education.
• Provides targeted, evidence-based recommendations for educators and students, differentiated by academic field (e.g., interventions for sleep hygiene, executive attention, and emotion regulation).
• Highlights, in the limitations section, that longitudinal designs would be necessary to examine potential temporal variations in sleep quality and academic functioning — thereby indirectly addressing the reviewer’s observation about changes over time.

We are grateful for this feedback, which aligns closely with the improvements already integrated during prior revisions and helped reinforce the overall clarity and depth of the Discussion.

Change implemented in manuscript  – No additional changes were made specifically in response to this comment, as the reviewer’s suggestions were fully addressed through previous revisions to the Discussion section following the feedback of other reviewers. The current version already includes the recommended expansions, contextual considerations, and practical implications grounded in current literature.

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

I applaud your efforts to improve the article. I can tell that much work has been done. It is a significant improvement, but foundational concerns remain:

Lines 20-22 state, "Institutional initiatives, [...], may play a equipping students with the tools..." The edit is not right. It should be, "may equip students with the tools..."

Table 1 should use periods between numbers rather than commas, as that is how the rest of the manuscript is written.

The references remain a concern as only 12 of the 40 references are less than 10 years old, and they don't seem specific enough to cover all areas of the study.

The limitations are so numerous as to be problematic. While I applaud the transparency, I find them to be too numerous. The lack of generalizability, the small sample size for quantitative research, the differences in student exams, limitations with the instruments themselves, and the lack of a strong link between the results and discussion is too great to draw solid conclusions. For instance, line 396-398 states, "Despite these cognitive vulnerabilities, Law students obtained the highest exam grades, which may reflect the presence of compensatory strategies, such as academic resilience, structured learning environments, or greater verbal fluency under evaluative conditions." yet lines 139-141 state, "Law students were all required to take the final exam regardless of prior performance, reflecting the general population of that course. Medicine and Economics students took the final exam only if they failed continuous assessment or aimed to improve their grades." Due to the difference in exam-takers (those who were succeeding and failing versus those who were failing), it is not possible to compare Law students' exam scores to Medical and Economics students exam scores. Structural issues such as this one abounds in the study. 

Author Response

General Reply:  We thank Reviewer 3 for the detailed and thoughtful feedback provided in this round of revision. Your comments have significantly contributed to improving the clarity, methodological coherence, and precision of the manuscript. In this revision, we carefully addressed all points raised, including stylistic corrections, formatting adjustments, updates to the reference list, refinement of the Limitations section, and clarification of cross-programme assessment structures to ensure conceptual consistency across the manuscript.

We believe the revisions implemented have strengthened the manuscript and better situate its contribution within the appropriate empirical and methodological boundaries. We remain grateful for your constructive insights and trust that the changes made fully address the concerns raised.

 

Comment 1: “Lines 20–22 state “…may play an equipping students…”. The edit is not right. It should be “may equip students with the tools…”  & Comment 2: “Table 1 should use periods between numbers rather than commas, as that is how the rest of the manuscript is written.”

Response: We thank the reviewer for bringing this wording error and formatting inconsistency to our attention, and we followed the corrections suggested.

Change implemented in manuscript – page 1, line 22

Change implemented in manuscript – page 6, Table 1.

 

Comment 3: “The references remain a concern as only 12 of the 40 references are less than 10 years old, and they do not seem specific enough to cover all areas of the study.”

Response: We thank the reviewer for this observation regarding reference recency and specificity. In response to the first revision round, we conducted a thorough update of the reference list, replacing older and adolescent-focused studies with more recent and university-specific sources whenever appropriate. In the present revision, we incorporated two additional empirical studies published within the last three years (Alrashed et al., 2022; Mota Albuquerque et al., 2023), which directly reinforce the sections on sleep, stress, and cognitive functioning in university and medical student populations. These studies are conceptually aligned with the variables examined and meaningfully strengthen the manuscript.

We fully agree that references should be updated whenever possible. However, given the exploratory nature of the study and the interdisciplinary scope across sleep, anxiety, executive attention, and academic performance in three distinct academic programs, the availability of recent university-level literature that simultaneously addresses all these domains is inherently limited. The updated reference list therefore reflects the maximum feasible alignment with recent, population-appropriate sources without compromising the scientific accuracy or integrity of the manuscript.

Nonetheless, we remain open to including any additional recent references that the reviewer considers essential.

Change implemented in manuscript – page -9, lines 537 and 608-609, and in the reference list accordingly.

 

 

Comment 4: “The limitations are so numerous as to be problematic. While I applaud the transparency, I find them to be too numerous. The lack of generalizability, the small sample size for quantitative research, the differences in student exams, limitations with the instruments themselves, and the lack of a strong link between the results and discussion is too great to draw solid conclusions. For instance, […] Due to the difference in exam-takers (those who were succeeding and failing versus those who were failing), it is not possible to compare Law students’ exam scores to Medical and Economics students’ exam scores. Structural issues such as this one abound in the study”

Response: We appreciate the reviewer’s careful reading and the concern raised regarding the scope of the study’s limitations. Our intention in the “Limitations and future directions” section was precisely to provide a transparent account of the methodological and contextual constraints inherent to an exploratory, observational study conducted in a single institution. We fully agree that these constraints preclude strong causal or highly generalisable conclusions.

In this revised version, we have clarified this point explicitly by (i) stating at the beginning of the section that the findings are hypothesis-generating and context-specific rather than definitive, and (ii) adding a concluding sentence emphasising that the results should be interpreted as preliminary and bound to the particular assessment structures and student populations examined. We have also streamlined the text so that the limitations are organised into broader domains, sample characteristics and selection bias, assessment structure across programmes, measurement and instrumentation issues, and design-related constraints, to avoid redundancy while retaining essential information.

Regarding the specific example of differences in exam-taking conditions between Law, Medicine, and Economics students, we agree that these structural discrepancies limit the comparability of exam grades across programmes. This issue is now explicitly acknowledged  in both the Limitations section and in the Discussion, where we have tempered the interpretation of between-group differences in academic performance and stressed that cross-programme comparisons should be treated with caution.

Overall, we believe that clearly stating these limitations strengthens rather than weakens the manuscript, as it situates the study's contribution within its appropriate empirical boundaries. The revised text makes explicit that the present findings offer preliminary, discipline-sensitive insights that can inform future, more robust research rather than firm conclusions.

Change implemented in manuscript – page 10-11, lines 636–638 (addition of introductory sentence clarifying the exploratory and context-specific nature of the findings) and lines 693–696 (addition of concluding sentence emphasising the preliminary and context-bound interpretation of the results).

 

 

Comment 5: “The excerpt in lines 396–398 conflicts with lines 139–141 in terms of Law students’ exam conditions. This undermines comparability across groups.”

Response: We appreciate the reviewer's insightful observation. To clarify, the structure of the final exam is comparable across programs; however, the conditions under which students are required or eligible to sit the exam differ substantially. In Law, all students are required to complete the compulsory final exam regardless of prior performance, resulting in a sample that reflects the full student cohort. In contrast, students in Medicine and Economics only sit the final exam if they miss the continuous assessment or voluntarily opt to improve their grade. Consequently, the exam-taking samples in these programs represent selective subgroups rather than the whole academic cohort. This issue is beyond our control and difficult to overcome.

To ensure that this important distinction is clearly reflected in the manuscript, we revised the paragraph discussing Law students’ performance to remove any implication of direct comparability across programs. The revised text now explicitly states that the higher exam grades observed among Law students must be interpreted with caution due to differences in assessment participation criteria, and that variations in sample representativeness may also contribute to observed performance differences. This clarification aligns the Discussion with the methodological description provided earlier in the manuscript and avoids unintended inconsistencies.

Change implemented in manuscript – page 9, lines 547–556 (Discussion).

Reviewer 4 Report

Comments and Suggestions for Authors

The revisions made so far have strengthened the paper, particularly the updated introduction which now presents a clearer and more coherent rationale. The comparison of three cohorts remains a valuable contribution and the applied implications for discipline sensitive support are promising. When noting that the sample is not representative, it would be helpful to include the size of each cohort that students are drawn from within that same section. This will provide useful context and will sit naturally alongside the existing caveat e.g 17 economics students drawn from a cohort of XX.

Table 1 would benefit from restructuring and should be placed at the start of the results section rather than towards the end. Presenting each discipline in its own column, with the sample total alongside, and reporting mean and standard deviation together, followed by minimum and maximum values as secondary information, would improve readability. It is also important to clarify why law appears twice in the table. If these entries represent two distinct samples, details should be provided in the methods along with a justification for combining them. If they represent two exam grades from the same individuals, these can be combined into an overall score with appropriate weighting, but the rationale and procedure need to be clearly described.

The analysis on sleep quality is not marginally significant, this is approaching significance and therefore post hoc comparisons should not be provided, these are also only approaching significance. Rather here you could simply note that descriptives appear to suggest that medical students have slightly worse sleep quality.

The multiple regression section requires further clarification and reanalysis. The first step should be to convert exam grade to z scores, since grades are drawn from different marking scales across modules and the distribution of the dependent variable is unlikely to be comparable without standardisation. Once this transformation is applied, the regression should be rerun with all predictors included. The current results show that although two predictors reach significance, the overall model is not significant and the adjusted R² value of .133 indicates that the model explains only a small proportion of the variance in exam grade. This means that the combined predictive value of the model is limited. It is important that no secondary or reduced model is conducted if the full model remains non significant. Removing non significant predictors will make the remaining variables appear more influential by default and does not address the original problem of model fit. The appropriate course is to present the full model, after converting to z scores, and interpret it in light of whether it achieves significance.

The paper’s applied value is developing well and you may wish to expand slightly on how discipline specific pressures shape sleep and anxiety, and how institutions can support student well-being around assessment periods. The study has the potential to contribute meaningfully to discussions about tailored academic support and the role of sleep in student performance.

Author Response

General Reply: We thank Reviewer 4 for the detailed and constructive feedback, which significantly contributed to strengthening the clarity, methodological precision, and applied value of the manuscript. All points raised were carefully considered, and revisions were implemented where appropriate. Specifically, the descriptive table was entirely restructured and relocated to the beginning of the Results section as recommended, the interpretation of sleep quality findings was revised to reflect descriptive rather than inferential patterns, and all regression-related concerns were addressed through a clarified methodological rationale. We also acknowledge the reviewer’s positive comments regarding the study's applied relevance and note that this aspect has already been expanded in earlier revisions of the Discussion. We are grateful for the reviewer’s engagement with the manuscript and believe that the changes implemented have strengthened its coherence, transparency, and disciplinary relevance.

 

Comment 1: “When noting that the sample is not representative, it would be helpful to include the size of each cohort that students are drawn from e.g. 17 economics students drawn from a cohort of XX.”

Response: Thank you for this helpful observation. In addition to reporting the number of participants recruited from each program, we have now added contextual information regarding the approximate size of the second-year cohorts from which these sub-samples were drawn.

According to information provided by the collaborating departments, the corresponding cohorts in the year of data collection included approximately 30–40 students in Economics, around 200 students in Law, and approximately 50 students in Medicine.

Importantly, final-exam attendance does not mirror full cohort size across programs. In Medicine and Economics, only students who failed continuous assessment or voluntarily opted to improve their grades sit the final exam, whereas in Law all students are required to attend the exam regardless of prior performance. Consequently, the exam-taking sample represents only a subset of the full cohort and reflects participation patterns that fall outside researcher control. This contextual clarification is now explicitly included to strengthen the interpretation of representativeness within the Participants Section.

Change implemented in manuscript – Section 2.1 Participants – Page 5, lines 201–208 – Addition of cohort-size contextualisation immediately following the paragraph on sample representativeness.

 

Comment 2: “Table 1 would benefit from restructuring… should be placed at the start of the results section… report mean + SD together, then min/max… clarify why Law appears twice….”

Response: We appreciate the reviewer's constructive suggestion. Table 1 has now been fully restructured in accordance with the recommended format:

• Each academic program is presented in its own column
• Mean ± SD values are provided together, followed by minimum–maximum ranges
• All variables are consolidated in a single, coherent table
• The table is now placed at the start of the Results section, immediately after the introductory paragraph

Additionally, the duplicated Law entry has been removed, ensuring clarity and readability.

Change implemented in manuscript – Table 1 restructured and relocated to the beginning of the Results section.

 

Comment 3: The analysis on sleep quality is not marginally significant… post hoc comparisons should not be provided… report descriptives instead.”

Response: We appreciate the opportunity to proceed with this important clarification. The sleep quality subsection has been revised to avoid interpreting post hoc tests as inferentially meaningful. The results are now described as “approaching significance,” and the text focuses on descriptive tendencies rather than formal group differences. This ensures that the interpretation remains statistically appropriate and aligned with the reviewer’s guidance.

Change implemented in manuscript – Revised wording in Section 3.1.1 (Sleep Quality) to remove inferential post hoc interpretation and emphasise descriptive patterns only.

 

 

Comment 4: “The multiple regression section requires further clarification… exam grades should be converted to z-scores… the regression should be rerun…”

Response: We understand the methodological suggestion, but, after careful consideration, we concluded that no changes to the regression analyses were necessary. While we fully agree that score standardisation can be useful in studies involving multiple grading systems or heterogeneous samples, several methodological characteristics of the present study justify retaining raw exam scores and the original modelling strategy.

First, the sample is intentionally stratified by academic program, and exam grades in each field follow discipline-specific evaluation structures. Standardising grades across programs—although statistically possible—would artificially homogenise differences arising from the curricula themselves rather than from the psychological variables under investigation. Because our objective was to examine whether sleep, anxiety, and executive attention predict academic performance within the natural grading context of each academic program, preserving raw grade metrics maintains interpretability and ecological validity.

Second, given the exploratory nature of the study and the modest sample size (N = 51), re-running regression models with transformed variables would not meaningfully increase model robustness and could instead reduce interpretability. The current analyses already report all required indicators (adjusted R², F-statistics, and standardised coefficients), providing a clear and transparent account of the model’s limited but informative predictive capacity.

Finally, in line with best practices for exploratory research, we opted to present both the full model (five predictors) and the reduced model (two significant predictors) without applying additional transformations that would exceed the intended scope of the study.

For these reasons, the methodological rationale for retaining raw exam scores and the original regression structure remains appropriate, and no modifications to the analyses were required.

 

 

Comment 5: “The paper’s applied value is developing well and you may wish to expand slightly…”

Response: We thank the reviewer for this encouraging comment. The study's applied implications had already been expanded in earlier revisions, particularly in the Discussion section, where discipline-specific considerations and directions for institutional support were elaborated. Balancing the comments of other reviewers, we believe the current level of detail appropriately reflects the exploratory nature of the study while providing meaningful applied insights.