Association Between Swimming Activity and Academic Performance Among Higher Education Students: A Cross-Sectional Study

Abstract

Swimming offers unique physiological and psychological benefits that may enhance academic performance in higher education students. However, empirical evidence specifically linking swimming activity to academic outcomes remains limited. A primary, cross-sectional, quantitative correlative study was conducted with 250 higher education students (mean age 20.0 years, SD = 2.21; 53.2% female) in Greece. Swimming activity was measured using the Global Physical Activity Questionnaire (GPAQ). Academic performance was assessed via the Academic Achievement Questionnaire (AAQ), comprising three factors: Study Performance, Handling Study Demands, and Social Adaptation. Data were analysed using Spearman correlations, Kruskal–Wallis tests, and multiple linear regression. Students reported swimming 1.94 days per week (SD = 1.83) for 0.98 h per session (SD = 0.82). Mean swimming intensity was 1446.53 MET·min/week (SD = 1569.65). Swimming activity varied significantly by monthly family income (H(3) = 31.37, p < 0.001) and parental education level (H(4) = 14.22, p = 0.007). However, swimming activity did not emerge as a significant independent predictor of academic performance in multivariate models when controlling for health and commitment covariates. While swimming activity demonstrates socioeconomic patterning among higher education students, its direct association with academic performance is concurrently associated with other factors including lifestyle and commitment. Interventions promoting swimming should address accessibility barriers to maximise potential academic benefits.

1. Introduction

Academic performance in higher education is a complex, multidimensional construct that encompasses study outcomes, the capacity to manage academic demands, and social integration into university life [1]. Students are exposed to a constellation of pressures—including time constraints, financial insecurity, and escalating academic obligations—that negatively impact psychosomatic health and, in turn, learning efficacy [2]. Within this demanding context [3], identifying protective and enhancing factors for academic functioning constitutes a priority for educational research and institutional policy.

While neurobiological mechanisms such as brain-derived neurotrophic factor (BDNF) release and neuroplasticity are frequently cited in the exercise literature, the cross-sectional nature of the present study precludes the empirical verification of such molecular pathways. Instead, the present study is grounded in psychological and behavioral theories of stress regulation, specifically the Effort–Recovery Model [4] and Social-Cognitive Theory [5]. From a behavioral perspective, regular physical activity serves as a structured stress-regulation strategy. By providing a dedicated time for physical exertion and psychological detachment from academic demands, exercise facilitates emotional regulation and cognitive recovery [6,7]. Engaging in a disciplined routine such as swimming inherently promotes the development of structured daily habits, time management, and self-regulatory capacity [8]. These behavioral and psychological adaptations are highly relevant to survey-based research, as they directly support the students’ capacity to handle academic demands and adapt to the university environment [5,9], providing a robust theoretical framework for the observed associations without relying on unmeasured physiological mechanisms.

Whereas land-based aerobic exercise has been consistently linked to improved executive function and academic outcomes [10], and neurobiological research has demonstrated that physical exercise promotes neuroplasticity and cognitive function [11], swimming occupies a distinct position. Unlike general water aerobics, swimming requires rhythmic breath control and bilateral cross-lateral motor coordination, which have been hypothesized in the prior literature to offer distinct cognitive and regulatory demands. However, the present cross-sectional study does not measure neurophysiological parameters, and therefore cannot verify claims about neural efficiency or hypoxic-like conditions. Instead, isolating swimming allows us to examine behavioral patterns associated with these specific motor and respiratory demands, without making causal claims about underlying physiological mechanisms. As a low-impact, full-body aerobic activity, it provides comprehensive musculoskeletal training while minimising joint stress, rendering it suitable for individuals across a wide range of fitness levels and physical conditions [12]. The aquatic environment confers additional benefits beyond those attributable to exercise alone: hydrostatic pressure facilitates venous return and reduces peripheral oedema; buoyancy decreases effective body weight and allows a wider range of motion; and the rhythmic, patterned nature of breathing in swimming has been linked to meditative and stress-attenuating effects [13,14]. Swimming has been associated with reductions in anxiety, depressive symptomatology, and perceived stress, alongside improvements in self-esteem, emotional regulation, and subjective well-being [15,16].

Aquatic environments—conceptualised as ‘blue spaces’—have additionally been linked in environmental psychology to psychological detachment from everyday stressors [17]. While the present study does not measure cognitive restoration directly, these psychological benefits may plausibly translate into academic benefits through enhanced concentration, reduced burnout, and improved emotional regulation in academic settings.

Despite this theoretical plausibility and the growing body of evidence linking general physical activity to academic outcomes, the specific contribution of swimming to academic performance among university students has received remarkably little empirical attention. While much of the foundational literature focuses on the developmental benefits of physical activity in children and adolescents, the transition to university represents a critical period of heightened psychosocial stress and autonomous demand management. Thus, the self-regulatory benefits of exercise are particularly vital for young adults navigating the unstructured environment of higher education. Most of the existing literature focuses on land-based exercise or paediatric samples, leaving a significant gap in understanding how aquatic exercise interacts with the unique demands of higher education, particularly when individual motivational, socioeconomic, and environmental factors are considered as moderators. Specifically, we hypothesize that: (H1) the relationship between swimming and academic adaptation is stronger for students with high intrinsic motivation; and (H2) environmental access to facilities positively moderates this relationship, whereas financial strain acts as a negative moderator.

The present study addresses this gap with three primary objectives: (1) to characterise patterns and determinants of swimming activity among university students in Greece; (2) to examine the bivariate associations between swimming activity and the three dimensions of academic performance—Study Performance, Handling Study Demands, and Social Adaptation; and (3) to identify significant predictors and moderators of the swimming–academic performance relationship in multivariate models. The study is theoretically informed by Social-Cognitive Theory [5] and ecological models of health behaviour [18], which emphasise the interplay between individual self-regulatory capacities, motivational orientations, and environmental affordances in shaping health-related behaviour and its consequences.

2. Materials and Methods

2.1. Research Design

A primary, cross-sectional, quantitative correlative study was conducted in October 2025. The design enabled the simultaneous examination of relationships between swimming activity and academic performance at a defined time point, and allowed for moderation testing across individual, motivational, and environmental variables. The study followed a predefined analytical framework aligned with the stated research questions and hypotheses, thereby minimising the risk of post hoc hypothesis generation. The cross-sectional nature of the design permits the identification of statistical associations and concurrent relationships but does not allow for causal inference or the determination of temporal precedence between variables.

2.2. Participants

The sample comprised 250 higher education students enrolled in Greek universities (133 females, 53.2%; 117 males, 46.8%), with a mean age of 20.00 years (SD = 2.21, range 18–28). A priori power analysis using G*Power 3.1 indicated a minimum required sample of 184 participants for multiple linear regression with 12 independent variables (f² = 0.15, α = 0.05, 1 − β = 0.80); the final sample of 250 provided adequate statistical power for both main effect and moderation analyses [19].

Sociodemographic composition was as follows: 65.2% were first-year students. Monthly family income distribution was: 14.4% below €1000; 46.0% €1000–2000; 20.0% €2000–3000; and 19.6% above €3000. Parental educational attainment was: primary education 18.0%, secondary education 24.0%, bachelor’s degree 20.8%, master’s degree 24.0%, and doctoral degree 13.2%. Regarding environmental access, 79.6% of students lived within 5 km of swimming facilities, while 89.2% reported that their institutions did not provide swimming programmes. These characteristics are summarised in Table 1.

Table 1. Sociodemographic and environmental characteristics of the sample (N = 250).

Variable	Category	N	%
Gender	Female	133	53.2
	Male	117	46.8
Year of study	First year	163	65.2
	Other years	87	34.8
Monthly family income	<€1000	36	14.4
	€1000–2000	115	46.0
	€2000–3000	50	20.0
	>€3000	49	19.6
Parental education	Primary	45	18.0
	Secondary	60	24.0
	Bachelor’s	52	20.8
	Master’s	60	24.0
	Doctoral	33	13.2
Proximity to facilities	≤5 km	199	79.6
	>5 km	51	20.4
Institutional swimming programmes	Available	27	10.8
	Not available	223	89.2

A notable characteristic of the sample is the overrepresentation of first-year students (65.2%), which may influence the assessment of academic adaptation, as these students are actively navigating the initial transition to university life. Additionally, the sampling strategy based on convenience and snowball recruitment limits the representativeness of the findings, and results should be generalised with caution.

2.3. Instruments

2.3.1. Swimming Activity

Swimming activity was assessed using the swimming-specific section of the Global Physical Activity Questionnaire (GPAQ) [20], a validated instrument developed by the World Health Organization to measure physical activity across domains. The GPAQ has demonstrated satisfactory concurrent [21], convergent [22], criterion [23], and face validity [24] in multiple international studies. For the present study, three items pertaining to recreational swimming were extracted: (1) whether the participant engages in swimming for at least 10 continuous minutes (dichotomous: Yes/No); (2) the number of days per week on which swimming is performed; and (3) the duration per session in hours.

Swimming intensity was computed by multiplying frequency (days/week) by session duration (minutes) by 8 METs, corresponding to the metabolic equivalent for moderate-intensity swimming as specified by the Compendium of Physical Activities [25]. The questionnaire was administered in Greek, using a validated translation by Pateraki [26], which demonstrated satisfactory content validity. The swimming activity items indicated high factor loadings (range 0.90–0.95), satisfactory reliability (α = 0.85), and adequate convergent validity (AVE = 0.87).

To avoid statistical redundancy and multicollinearity, the continuous “Swimming Intensity (MET·min/week)” variable was used as the sole, non-redundant predictor of aquatic engagement in all subsequent multiple regression and moderation analyses. An exploratory Swimming Activity Composite Score (ranging from 0 to 13.5) was initially calculated by summing standardised values of frequency, duration, and MET·min/week; however, because MET·min/week is arithmetically derived from frequency × duration × 8 METs, including it alongside its own components would introduce perfect multicollinearity and artificially inflate the contribution of these variables. Therefore, the composite score was excluded from all inferential modelling and is reported here solely for descriptive context.

Additionally, three supplementary items were assessed: (1) swimming motivation (mastery-oriented vs. performance-oriented); (2) predominant type of swimming practice (recreational vs. competitive); and (3) swimming experience in years.

2.3.2. Academic Performance

Academic performance was assessed using the Academic Achievement Questionnaire (AAQ) [27,28], a multidimensional instrument validated among university populations. Academic performance data were self-reported by participants; this represents a methodological limitation, as participants may have been inclined to provide inflated grades due to social desirability bias. Future studies should aim to collect objective academic records directly from Higher Education Institutions (HEIs) to mitigate this bias.

Initial dimensionality was explored using Exploratory Factor Analysis (EFA) via Principal Component Analysis with Varimax rotation. The EFA yielded three stable factors, which were retained for the present study:

–

Study Performance: average and semester-specific examination grades, rated on a 0–10 scale. This factor explained 32.79% of total variance, with high factor loadings (0.71–0.90), high reliability (α = 0.87), and satisfactory convergent validity (AVE = 0.74).

–

Handling Study Demands: self-assessed ability to organise study time, plan for examinations, and manage academic requirements, rated on a 6-point Likert scale (1 = Not at all to 6 = Without problem). This factor explained 30.44% of variance, with high reliability (α = 0.93) and strong convergent validity (AVE = 0.92).

–

Social Adaptation: perceived fit with peers and adaptation to university life, rated on a 6-point Likert scale. This factor explained 20.27% of variance with high reliability (α = 0.82; AVE = 0.85).

The complete factor structure and psychometric properties of the AAQ are presented in Table 2.

Table 2. Factor structure and psychometric properties of the Academic Achievement Questionnaire (AAQ).

Item	Factor 1 Study Performance	Factor 2 Handling Study Demands	Factor 3 Social Adaptation
Average exam mark (last semester)	0.90
Worst exam mark (last semester)	0.87
Average mark (university study)	0.83
Best exam mark (last semester)	0.71
Plan time to prepare for exams		0.93
Organise time for study		0.93
Handle study requirements		0.87
Fit among classmates			0.88
Adapt to university life			0.85
Eigenvalue	2.95	2.74	1.82
Variance explained (%)	32.79	30.44	20.27
Cronbach’s α	0.87	0.93	0.82
AVE	0.74	0.92	0.85

Note. KMO = 0.79; CFI = 0.96; TLI = 0.93; SRMR = 0.05; RMSEA = 0.124. Extraction method: Principal Component Analysis; Rotation: Varimax. The RMSEA value of 0.124 indicates suboptimal CFA model fit; results should be interpreted with this limitation in mind.

Subsequently, a Confirmatory Factor Analysis (CFA) was conducted using maximum likelihood estimation to evaluate the structural validity of the three-factor model identified in the EFA. The CFA yielded mixed fit indices: CFI = 0.96, TLI = 0.93, SRMR = 0.05, but RMSEA = 0.124. While the CFI and TLI values exceed conventional thresholds for acceptable fit, the RMSEA value of 0.124 indicates poor model fit (values > 0.10 are generally considered unacceptable). Therefore, the three-factor structure cannot be considered fully confirmed by the CFA. The factor structure was retained for analytical purposes given its theoretical coherence and comparability with previous validation studies [27,28]; however, results involving the AAQ subscales should be interpreted with caution in light of this suboptimal fit.

Inter-factor correlations were all positive and statistically significant (Study Performance–Handling Study Demands: ρ = 0.50; Study Performance–Social Adaptation: ρ = 0.37; Handling Study Demands–Social Adaptation: ρ = 0.54; all p < 0.01), supporting convergent relationships among the subscales, but not compensating for the inadequate overall model fit indicated by the RMSEA.

2.3.3. Swimming Commitment and Motivation

Swimming commitment was assessed using the Sport Commitment Questionnaire-2 (SCQ-2) [29], comprising six subscales: Enthusiastic Commitment (α = 0.91), Swimming Enjoyment (α = 0.88), Social Constraints (α = 0.84), Desire to Excel–Mastery Achievement (α = 0.86), Academic Commitment (α = 0.89), and Behavioural Engagement (α = 0.72). Items were rated on 5-point or 6-point Likert scales. Emotional Engagement was assessed using a validated 5-item subscale (α = 0.81).

2.3.4. Health and Lifestyle

Physical and psychological health were assessed using the WHOQOL-BREF [30], comprising items on physical pain, energy, sleep satisfaction, daily activities, and psychological well-being. Lifestyle factors including physical activity (beyond swimming), nutrition, and overall health satisfaction were assessed using the Health-Promoting Lifestyle Profile II (HPLP-II) [31], which satisfies content, factorial, concurrent, and convergent validity.

2.3.5. Environmental and Socioeconomic Variables

Three dichotomous items assessed the availability of swimming infrastructure: (1) proximity to swimming facilities within 30 min’s travel; (2) provision of swimming programmes by the participant’s educational institution; and (3) provision of swimming programmes or facilities by the local community. Monthly family income and parental education level were collected as categorical sociodemographic variables.

2.4. Procedure

Participants were recruited via a combination of convenience and snowball sampling: the researcher initially distributed the questionnaire to known student contacts and relevant social media groups, from which it was further circulated peer-to-peer [32]. Participants were fully informed of the study’s purposes, anonymity, voluntary nature, and their right to withdraw at any time, through an introductory information sheet. Informed consent was obtained from all participants prior to data collection, in accordance with APA ethical guidelines [33]. The study received approval from the institutional ethics committee and was conducted in accordance with the Declaration of Helsinki.

2.5. Data Analysis

Data were analysed using IBM SPSS 26 and Jamovi 2.6.26.0. Descriptive statistics (mean, standard deviation, frequency, and percentage) were computed for all variables. All continuous variables were tested for normality using the Shapiro–Wilk test [34]; normality was rejected for all factors (all p < 0.001), justifying the use of non-parametric inferential procedures for bivariate analyses. Although raw variable distributions violated normality assumptions (necessitating Spearman’s rho for bivariate correlations), diagnostic tests (Q-Q plots, Kolmogorov–Smirnov test on standardised residuals, and plots of residuals versus fitted values) confirmed that the residuals of the multiple linear regression models were approximately normally distributed and exhibited homoscedasticity, thereby satisfying the assumptions for Ordinary Least Squares (OLS) regression.

The following analytical strategy was applied:

(1)

Spearman rank correlations to examine bivariate relationships between swimming activity and academic performance factors;

(2)

Mann–Whitney U tests to examine gender differences in swimming activity and academic performance;

(3)

Kruskal–Wallis H tests to examine group differences across monthly family income and parental education level, with post hoc Bonferroni correction for multiple comparisons;

(4)

Multiple linear regression with bootstrapping (B = 1000 samples) to identify independent predictors of each academic performance factor, using 95% bias-corrected confidence intervals. All regression models were screened for multicollinearity using the Variance Inflation Factor (VIF; acceptable threshold: VIF < 3) and for independence of residuals using the Durbin–Watson statistic;

(5)

Moderation analysis using Hayes’ PROCESS macro (Model 1) [35] to test whether selected contextual, motivational, and lifestyle variables significantly moderated the relationship between swimming intensity and each academic performance outcome. To control for the increased risk of Type I errors due to the large number of interaction tests conducted (12 moderators × 3 outcomes = 36 tests), the Benjamini–Hochberg False Discovery Rate (FDR) correction was applied to all interaction p-values. Only interaction effects that remained significant at the adjusted threshold (p_adj < 0.05) are interpreted as statistically robust moderators. Both nominal and FDR-adjusted p-values are reported alongside the moderation analysis results to ensure full transparency and allow independent verification of the correction procedure.

It is important to note that, due to the cross-sectional design, all analyses identify concurrent statistical associations rather than causal effects. The term “predictor” is used in the statistical sense (i.e., a variable that accounts for variance in the outcome within the regression model) and does not imply temporal precedence or causal direction.

3. Results

3.1. Descriptive Statistics: Swimming Activity

Overall, 65.2% of students (n = 163) reported engaging in swimming at an intensity sufficient to increase breathing or heart rate for at least 10 continuous minutes. Among these 163 students who reported swimming, the predominant motivation was mastery-oriented (99.4%; n = 162), with 77.3% (n = 126) engaging in recreational swimming and 58.9% (n = 96) in competitive swimming (categories were not mutually exclusive). It should be noted that intrinsic motivation was not directly assessed via a dedicated instrument; therefore, the high prevalence of recreational swimming is interpreted descriptively rather than as direct evidence of intrinsic motivational drivers. Mean swimming experience was 8.71 years (SD = 5.48, range 1–22).

In a typical week, students reported swimming on an average of 1.94 days (SD = 1.83, range 0–6), with sessions averaging 0.98 h (SD = 0.82, range 0–2.25 h). Mean swimming intensity, computed as MET·min/week, was 1446.53 (SD = 1569.65, range 0–6480). Descriptive statistics for all swimming activity variables are presented in Table 3.

Table 3. Descriptive statistics for swimming activity variables (N = 250).

Variable	Category/Min–Max	N or M	% or SD
Engages in structured swimming	No	87	34.8
	Yes	163	65.2
Swimming motivation	Mastery motivation	162	99.4
	Performance motivation	31	19.0
Type of swimming	Recreational	126	77.3
	Competitive	96	58.9
Swimming experience (years)	1–22	8.71	5.48
Frequency (days/week)	0–6	1.94	1.83
Session duration (hours/day)	0–2.25	0.98	0.82
Swimming intensity (MET·min/week)	0–6480	1446.53	1569.65
Swimming activity composite score	0–13.5	3.01	3.27

Note. For swimming-specific variables (motivation and type), the denominator is the subsample of students who reported swimming activity (N = 163). Motivational categories are not mutually exclusive; participants could endorse multiple achievement-goal orientations simultaneously, explaining the overlapping percentages between mastery and performance profiles.

3.2. Descriptive Statistics: Academic Performance

Mean Study Performance was M = 7.52 (SD = 1.20) on the 0–10 scale. Students rated Handling Study Demands at M = 4.52 (SD = 1.21; range 1–6), indicating moderate-to-high perceived capacity for demand management. Social Adaptation was rated at M = 4.41 (SD = 1.18). Taken together, these descriptive findings suggest an overall moderate-to-good academic functioning profile across the sample.

3.3. Sociodemographic and Environmental Predictors of Swimming Activity

3.3.1. Gender

Male students reported significantly higher swimming activity (M = 4.24, SD = 3.46) than female students (M = 1.93, SD = 2.81; Z = −4.62, p < 0.001, d = 0.69). Males also demonstrated higher swimming intensity scores, consistent with gender differences in sport participation reported in the broader literature. Additionally, males showed significantly higher emotional engagement (M = 3.11 vs. M = 2.96; Z = −2.63, p = 0.008, d = 0.17), though no significant gender differences were observed in academic commitment or behavioural engagement.

3.3.2. Monthly Family Income

Kruskal–Wallis analysis revealed significant differences in swimming activity across income brackets (H(3) = 31.37, p < 0.001). Post hoc Bonferroni comparisons indicated that students from families earning above €3000 per month reported significantly higher swimming activity (M = 5.93, SD = 3.42) than those earning €1000–2000 (M = 2.36, SD = 2.85; p < 0.001) and below €1000 (M = 2.03, SD = 3.07; p < 0.001). This socioeconomic gradient in swimming participation likely reflects differential access to private swimming facilities, membership costs, and transportation.

3.3.3. Parental Education Level

Significant differences in swimming activity were found across parental education levels (H(4) = 14.22, p = 0.007). Students with PhD-educated parents reported higher swimming activity (M = 5.68, SD = 3.91) than those with secondary-educated parents (M = 2.26, SD = 2.90; p = 0.012). This pattern likely reflects cultural capital transmission, whereby higher parental education is associated with greater valuation of and exposure to structured physical activity from an early age.

3.3.4. Proximity to Facilities and Community Programmes

Students living within 1 km of swimming facilities reported substantially higher activity (M = 3.92, SD = 3.68) than those residing more than 5 km away (M = 1.78, SD = 2.18; H(2) = 28.13, p < 0.001). Access to community swimming programmes was also associated with significantly higher activity levels: students with such access reported M = 3.94 (SD = 3.49) compared to M = 1.78 (SD = 2.64) for those without (Z = −4.81, p < 0.001). These environmental determinants are consistent with ecological models of health behaviour emphasising the role of structural affordances in physical activity participation [18].

3.4. Bivariate Associations Between Swimming Activity and Academic Performance

Spearman correlation analysis was conducted to examine the relationships between the swimming activity composite score and the three academic performance factors. Results are presented in Table 4.

Table 4. Spearman correlations between swimming activity and academic performance factors.

Academic Performance Factor	ρ	p-Value	Interpretation
Study Performance	0.08	0.221	Non-significant
Handling Study Demands	0.31	<0.001	Significant, moderate
Social Adaptation	0.14	0.027	Significant, weak

Note. ρ = Spearman correlation coefficient. N = 250.

A significant moderate positive correlation was found between swimming activity and Handling Study Demands (ρ = 0.31, p < 0.001), indicating that students who are more active in swimming report greater perceived capacity to manage academic workload, time, and requirements. A weak but significant positive correlation was also found with Social Adaptation ($\rho$ = 0.14, p = 0.027). While statistically detectable, this small effect size indicates that swimming activity alone accounts for a minimal proportion of the variance in social adaptation, and its practical educational relevance should be interpreted with caution within a broader ecosystem of social factors. No significant relationship was found between swimming activity and Study Performance (ρ = 0.08, p = 0.221).

Supplementary bivariate analyses also confirmed positive and significant correlations between swimming activity and physical health (ρ = 0.51, p < 0.001) and academic commitment (ρ = 0.22, p < 0.01). These associations were retained as covariates in the subsequent multivariate models.

3.5. Multivariate Predictors of Academic Performance

Multiple linear regression models with bootstrapping (B = 1000) were constructed for each academic performance factor. Independent variables entered into each model were those that demonstrated statistically significant bivariate associations. All models satisfied multicollinearity criteria (VIF < 3 for all predictors) and demonstrated residual independence (Durbin–Watson ≈ 2.0; see Table 5). However, it is important to explicitly acknowledge that the explained variance of these models is relatively low (R² = 0.140–0.235). Academic functioning is a complex, multifactorial construct influenced by numerous unmeasured variables (e.g., prior attainment, cognitive abilities, institutional factors), which strictly limits the sole predictive value of physical activity metrics.

Table 5. Multicollinearity diagnostics for regression models predicting academic performance.

Moderator/Covariate	VIF	Durbin–Watson
Proximity to swimming facilities	1.08	2.09
Community provision of swimming programmes	1.09	2.09
Academic commitment	1.00	2.04
Behavioural engagement	1.05	2.18
Emotional engagement	1.07	2.22
Enthusiastic commitment	1.26	2.08
Physical activity lifestyle	1.21	2.02
Nutrition	1.05	2.08

Note. VIF = Variance Inflation Factor. Values < 3 indicate acceptable multicollinearity.

3.5.1. Study Performance

The model was statistically significant (F(5, 244) = 8.79, p < 0.001, R² = 0.153, Adjusted R² = 0.135). Significant predictors were parental education level (β = 0.21, t = 3.31, p = 0.001) and physical health (β = 0.18, t = 2.91, p = 0.004). Swimming activity was not a statistically significant independent predictor of Study Performance in the multivariate model (β = 0.03, t = 0.51, p = 0.611, 95% CI [−0.06, 0.12]).

3.5.2. Handling Study Demands

The model was statistically significant (F(6, 243) = 12.45, p < 0.001, R² = 0.235, Adjusted R² = 0.215). Swimming activity emerged as a significant independent predictor of Handling Study Demands (β = 0.18, t = 2.88, p = 0.004, 95% CI [0.06, 0.31]). Additional significant predictors were physical health (β = 0.31, t = 5.14, p < 0.001) and academic commitment (β = 0.22, t = 3.61, p < 0.001). These results suggest that swimming activity is concurrently associated with demand management capacity, reflecting an associative relationship rather than a verified causal pathway.

3.5.3. Social Adaptation

The model was statistically significant (F(5, 244) = 7.93, p < 0.001, R² = 0.140, Adjusted R² = 0.122). Significant predictors were physical health (β = 0.24, t = 3.67, p < 0.001) and emotional engagement (β = 0.19, t = 3.01, p = 0.003). Swimming activity did not independently predict Social Adaptation (β = 0.07, t = 1.12, p = 0.265, 95% CI [−0.05, 0.19]).

A summary of the three multiple regression models is presented in Table 6.

Table 6. Summary of multiple regression models predicting academic performance factors.

Dependent Variable	Model F	R2	Adj. R2	Significant Predictors (β, p)
Study Performance	F(5244) = 8.79 ***	0.153	0.135	Parental education (0.21, p = 0.001); Physical health (0.18, p = 0.004)
Handling Study Demands	F(6243) = 12.45 ***	0.235	0.215	Swimming activity (0.18, p = 0.004); Physical health (0.31, p < 0.001); Academic commitment (0.22, p < 0.001)
Social Adaptation	F(5244) = 7.93 ***	0.140	0.122	Physical health (0.24, p < 0.001); Emotional engagement (0.19, p = 0.003)

Note. *** p < 0.001. β = standardised regression coefficient. Models used bootstrapping with B = 1000 samples.

3.6. Moderation Analysis: Swimming Intensity and Academic Performance

Moderation analyses were conducted to assess whether selected contextual, motivational, and lifestyle variables significantly moderated the relationship between swimming intensity and each academic performance factor. Results for Study Performance are presented in Table 7, and findings for Handling Study Demands are presented in Table 8.

Table 7. Moderation effects on the relationship between swimming intensity and Study Performance.

Moderator	F(1, 246)	t	p-Value	p_adj (FDR)
Proximity to swimming facilities	1.84	1.36	0.176	0.234
School provision of swimming programmes	0.001	0.03	0.978	0.978
Community provision of swimming programmes	14.73	3.84	0.002	0.004
Academic commitment	4.95	−2.23	0.027	0.045
Behavioural engagement	13.03	−3.61	0.004	0.008
Emotional engagement	30.50	−5.52	<0.001	<0.001
Enthusiastic commitment	5.79	2.40	0.017	0.028
Swimming enjoyment	2.50	−1.58	0.115	0.153
Social constraints	0.16	−0.41	0.687	0.763
Desire to excel/mastery achievement	3.64	1.91	0.058	0.087
Physical activity (lifestyle)	0.69	−0.83	0.407	0.452
Nutrition	9.87	−3.14	0.002	0.004

Note. Bold values indicate statistical significance after Benjamini–Hochberg FDR correction (p_adj < 0.05).

Table 8. Moderation effects on the relationship between swimming intensity and Handling Study Demands.

Moderator	F(1, 246)	t	p-Value	p_adj (FDR)
Proximity to swimming facilities	2.47	1.57	0.117	0.175
School provision of swimming programmes	2.46	−1.57	0.118	0.157
Community provision of swimming programmes	19.32	4.40	<0.001	<0.001
Academic commitment	0.42	0.65	0.519	0.566
Behavioural engagement	11.30	−3.36	<0.001	<0.001
Emotional engagement	1.59	−1.26	0.201	0.241
Enthusiastic commitment	7.31	2.70	0.007	0.017
Swimming enjoyment	0.28	−0.53	0.595	0.595
Social constraints	6.63	−2.57	0.011	0.022
Desire to excel/mastery achievement	2.55	1.60	0.112	0.192
Physical activity (lifestyle)	7.85	−2.80	0.006	0.018
Nutrition	32.31	−5.68	<0.001	<0.001

Note. Bold values indicate statistical significance after Benjamini–Hochberg FDR correction (p_adj < 0.05).

To control for the increased risk of Type I errors due to multiple interaction tests, the Benjamini–Hochberg False Discovery Rate (FDR) correction was applied. Only interaction effects that remained significant at the adjusted threshold (p_adj < 0.05) are interpreted as robust moderators.

A statistically significant moderation effect was observed for community provision of swimming programmes (F(1, 246) = 14.73, p = 0.002, p_adj = 0.004), indicating that the availability of community swimming programmes is associated with a different pattern in the relationship between swimming intensity and Study Performance. Specifically, the conditional effect of swimming intensity on Study Performance was statistically significant when community programmes were available (b = 0.0004, 95% CI [0.0003, 0.0006]), but not when they were absent (b = 0.0000, 95% CI [−0.0002, 0.0002]). The unstandardized coefficient (b = 0.0004) is numerically small due to the large measurement scale of the ‘Swimming intensity’ variable (measured in MET·min/week, M = 1446.53, SD = 1569.65). Therefore, this interaction indicates that the presence of community facilities is associated with a statistically significant positive association that is otherwise absent, though the cross-sectional design precludes causal claims about “enabling” or “fundamentally altering” trajectories.

Emotional engagement was the strongest individual moderator (F(1, 246) = 30.50, p < 0.001, p_adj < 0.001), indicating a significant interaction between swimming intensity and students’ emotional investment in their studies. The negative direction of the interaction term suggests that at higher levels of emotional engagement, the marginal academic return of swimming intensity was attenuated, possibly because highly engaged students already perform well regardless of physical activity levels.

For Handling Study Demands, full model parameters and confidence intervals are presented to ensure transparency. Statistically significant moderation effects (all p_adj < 0.05) were identified for: community provision of swimming programmes (F(1, 246) = 19.32, p < 0.001, p_adj < 0.001), behavioural engagement (F(1, 246) = 11.30, p < 0.001, p_adj < 0.001), enthusiastic commitment (F(1, 246) = 7.31, p = 0.007, p_adj = 0.012), social constraints (F(1, 246) = 6.63, p = 0.011, p_adj = 0.018), physical activity lifestyle (F(1, 246) = 7.85, p = 0.006, p_adj = 0.010), and nutrition (F(1, 246) = 32.31, p < 0.001, p_adj < 0.001).

For example, the conditional effect of swimming intensity on Handling Study Demands was significant when community facilities were available (b = 0.0003, 95% CI [0.0002, 0.0004]), but not when absent (b = 0.0000, 95% CI [−0.0003, 0.0000]). Similarly, behavioural engagement significantly moderated this relationship, with the positive association of swimming intensity being stronger at low levels of behavioural engagement (b = 0.0003, 95% CI [0.0003, 0.0004]) and non-significant at high levels (b = 0.0000, 95% CI [−0.0001, 0.0001]). These full parameters confirm the statistical robustness of the interaction effects after FDR correction; however, their practical significance remains modest, and they should be interpreted as exploratory associations requiring longitudinal confirmation.

4. Discussion

The present study examined the association between swimming activity and academic performance in a sample of 250 Greek higher education students, using a multivariate design that accounted for sociodemographic, motivational, lifestyle, and environmental moderators. The socioeconomic patterning observed in access to aquatic facilities is a critical finding that warrants deeper investigation. While beyond the scope of the current psychometric analysis, future dedicated research should explore how socioeconomic status and cultural capital moderate the accessibility and academic benefits of aquatic exercise. Three principal findings emerge from the results.

First, swimming activity demonstrates marked socioeconomic patterning. Students from higher-income families and those with more educated parents reported substantially greater swimming participation. This finding aligns with social-ecological models of health behaviour [18], which posit that physical activity is shaped not only by individual motivation but also by structural factors including economic resources, cultural capital, and access to facilities. These results also suggest that swimming activity is associated with demand management capacity above and beyond the effects of health and commitment. However, it is highly probable that other forms of regular aerobic exercise confer similar self-regulatory benefits. The extent to which swimming offers unique advantages over land-based exercises remains an open question for future comparative studies. However, the observation that 89.2% of institutions in this sample lacked organised swimming programmes reinforces the relevance of systemic barriers in the Greek higher education context, and echoes evidence from other European countries regarding inequities in sport participation [36].

Second, while general physical activity and sport have been consistently linked to academic outcomes [10,37], the present findings suggest that swimming exerts a selective and domain-specific effect on academic performance. Specifically, swimming activity was a significant independent correlate of Handling Study Demands (β = 0.18, p = 0.004), but not of Study Performance or Social Adaptation in the multivariate models. This pattern is theoretically coherent: the capacity to manage competing academic demands—to organise study time, plan systematically, and cope with academic pressure—is conceptually proximate to the self-regulatory skills that regular physical activity is known to foster [38]. However, rather than attributing the observed associations to unmeasured physiological stress-coping mechanisms or direct neurocognitive ‘transferable competencies’, we interpret these findings strictly through the lens of behavioral routines and self-regulation. It is crucial to state that these mechanisms remain hypothetical and were not empirically tested in the present study. The significant association between swimming activity and the capacity to handle study demands can be understood as a behavioral overlap. Students who successfully integrate a rigorous and time-consuming activity like swimming into their weekly schedules may concurrently exhibit organizational habits that are associated with better academic demand management. According to Social-Cognitive Theory [5], these self-regulatory skills are domain-general; however, the cross-sectional nature of our data precludes concluding that swimming causes these skills to develop or that they are subsequently applied to academic obligations. Therefore, the benefits observed in our data may reflect the behavioral structuring of the students’ daily lives and their enhanced capacity for psychological detachment from academic stressors during recovery periods, rather than stemming from direct physiological transfers.

By contrast, the absence of a direct effect on Study Performance (examination grades) is noteworthy and departs from some previous meta-analytic findings [10]. This may reflect several methodological considerations. First, the cross-sectional design cannot capture the cumulative neurobiological effects of prolonged swimming engagement—effects that may require months or years to manifest in examination performance. Second, academic grades in higher education are determined by a complex constellation of factors, including prior attainment, study strategies, course difficulty, and assessment format, which collectively exert stronger direct influence than any single lifestyle variable [39]. Third, physical health emerged as the strongest correlate of Study Performance in the multivariate model, consistent with the hypothesis that swimming benefits academic outcomes primarily through health mediation rather than direct cognitive enhancement. It is also important to acknowledge that the multivariate models presented explain a modest proportion of the variance in academic outcomes (R² = 0.140–0.235). Academic functioning is a highly complex, multifactorial construct driven by numerous unmeasured variables, including prior educational attainment, cognitive abilities, and specific institutional factors. Therefore, while swimming activity emerges as a significant independent correlate for Handling Study Demands, its predictive value should be interpreted as one important component within a broader, complex ecosystem of academic determinants, rather than a standalone driver of academic success.

Third, the moderation analyses reveal important nuances in the swimming–academic performance relationship. After correcting for multiple comparisons, several hypothesized moderation effects proved statistically unstable. This underscores the exploratory nature of these interactions and suggests that future pre-registered studies are needed to confirm which environmental factors reliably moderate the swimming–academic relationship. Community provision of swimming programmes was associated with a stronger positive relationship between swimming intensity and academic outcomes when such provision existed; however, due to the cross-sectional design, it cannot be concluded that programme availability causally enables or amplifies academic success. These variables likely co-occur within broader socio-environmental contexts, and these findings require confirmation in longitudinal or interventional studies. These results are consistent with the ecological model of health promotion [18] and with emerging evidence on the role of campus sport infrastructure in student well-being [28].

The moderation role of academic commitment and engagement is equally theoretically interesting. Higher levels of behavioural and emotional engagement appeared to interact with swimming intensity in predicting academic outcomes, suggesting that the benefits of swimming may be optimally realised in students who are already academically invested. This interaction may reflect a synergistic mechanism: motivated and engaged students may be more likely to use physical activity as a deliberate self-regulation strategy—scheduling swimming around study, using it to manage stress and maintain energy—thereby strengthening its academic benefits.

Limitations

Several limitations merit acknowledgement. First and foremost, the cross-sectional design precludes causal inference; the direction of association between swimming and academic outcomes cannot be established from these data. It is plausible that high-performing students have greater time and psychological resources to engage in leisure activities including swimming, introducing reverse causality. Second, self-reported swimming activity is susceptible to recall bias and social desirability effects; objective measurement via accelerometry or swimming logs would strengthen future research. Third, self-reported academic performance is similarly susceptible to recall bias and social desirability effects, as participants may be tempted to provide inflated grades. Future studies should aim to collect objective academic records directly from Higher Education Institutions (HEIs) to mitigate this bias.

Fourth, a notable limitation is the reliance on convenience and snowball sampling strategy and the high proportion of first-year students (65.2%). Because first-year students are actively navigating the initial transition to university life, their self-reported adaptation scores may differ systematically from senior students, potentially biasing the educational outcomes reported herein.

Fifth, the psychometric validation of the Academic Achievement Questionnaire (AAQ) yielded a suboptimal model fit regarding the RMSEA (RMSEA = 0.124). While the theoretical factor structure was retained for comparability with previous validation studies, this poor fit indicates that the translated instrument may not fully capture the nuances of academic adaptation in this specific demographic. This represents a methodological limitation that requires cautious interpretation of the subscale scores. Consequently, the suboptimal fit of the AAQ in this specific demographic suggests that the translated instrument may not fully capture the cultural nuances of academic adaptation in Greek higher education. Therefore, the findings related to the AAQ subscales should be considered preliminary, and a formal cross-cultural revalidation of the instrument is strongly recommended before generalizing these results.

Sixth, while the sample size was adequate for the primary multiple regression models, the statistical power to detect small interaction effects in the moderation analyses may be limited. Although the Benjamini–Hochberg FDR correction was applied to control for Type I errors, the modest sample size restricts the ability to detect complex, higher-order interactions. Future studies with larger, multi-center samples are required to robustly confirm these moderating effects.

Finally, the exclusive reliance on self-report measures introduces the potential for common method bias. Shared method variance may have artificially inflated the observed correlations between physical activity, motivation, and academic functioning, a limitation that future multi-method studies should address. Moreover, the absence of longitudinal data or dose–response information restricts conclusions about the frequency, intensity, and duration of swimming required to produce meaningful academic benefits. Also, the cross-sectional design is susceptible to self-selection bias. Students who regularly engage in swimming may differ systematically from non-swimmers in unmeasured confounding variables, such as baseline self-discipline, socioeconomic advantages, or pre-existing psychosomatic health. Therefore, the observed associations may partly reflect these underlying pre-existing differences rather than a direct effect of swimming itself.

5. Conclusions

The present study provides empirical evidence that swimming activity is a significant correlate of students’ capacity to handle academic demands. However, the cross-sectional design precludes conclusions regarding underlying mechanisms or causal pathways; thus, we cannot assert that this association operates through physical health or self-regulatory skills. Its direct effect on examination grades is not supported in multivariate models controlling for sociodemographic and health covariates. Swimming participation is strongly shaped by socioeconomic and environmental factors, highlighting the importance of systemic barriers to aquatic physical activity in the higher education context. In sum, while swimming is positively associated with academic functioning, researchers and practitioners must interpret these findings in light of the small to moderate effect sizes and the limited variance explained by the current models.

Practical Implications: The finding that community provision of swimming programmes is associated with the swimming–academic performance relationship offers cautious practical applications for university wellness coordinators and local authorities. Rather than mandating the construction of expensive new aquatic infrastructure, institutions may consider optimizing access to existing community facilities, subsidizing memberships for lower-income students to address the observed socioeconomic patterning, and promoting structured aquatic exercise as part of a broader, cost-effective portfolio of student well-being initiatives. While community swimming programmes are associated with positive academic outcomes, aquatic facilities are highly expensive to build and maintain. Given that other forms of exercise likely provide similar self-regulatory benefits, universities and local authorities should spend resources wisely by investing in minimal-infrastructure initiatives, such as organized ‘park runs’, campus walking groups, and outdoor calisthenics. A diverse, cost-effective portfolio of physical activity options will maximize population-level participation and well-being without the prohibitive costs of aquatic infrastructure. These proposed actions remain hypotheses requiring confirmation in longitudinal or intervention studies.

Also, for future research, longitudinal and experimental designs—incorporating accelerometry, ecological momentary assessment, and objective academic outcomes—are needed to establish the causal pathways and dose–response characteristics of swimming’s effects on academic performance. Particular attention should be given to the mediating roles of physical health, sleep quality, and self-regulatory capacity, as well as to potential moderating factors including swimming motivation, competence, and the social dimensions of aquatic exercise.