STRUCTURAL ANALYSIS OF THE
RELATIONSHIPS AMONG PHYSICAL ACTIVITY,
PHYSICAL SELF-CONCEPT, AND ACADEMIC
PERFORMANCE: AN APPLICATION OF
STRUCTURAL EQUATION MODELING
ANÁLISIS ESTRUCTURAL DE LA RELACIÓN ENTRE ACTIVIDAD
FÍSICA, AUTOCONCEPTO FÍSICO Y RENDIMIENTO ACADÉMICO:
UNA APLICACIÓN DE MODELOS DE ECUACIONES
ESTRUCTURALES
Thessa Montserrat Gutiérrez González
Investigadora Independiente, México
José Juan Muñoz León
Universidad Veracruzana, México
Judith Guadalupe Montero Mora
Universidad Veracruzana, México
Zoylo Morales Romero
Universidad Veracruzana, México
Daniela Burgueño Theurel
Universidad Veracruzana, México
pág. 1839
DOI: https://doi.org/10.37811/cl_rcm.v10i4.25211
Structural Analysis of the Relationships Among Physical Activity, Physical
Self-Concept, and Academic Performance: An Application of Structural
Equation Modeling
Thessa Montserrat Gutiérrez González1
tesmongonzalez@gmail.com
https://orcid.org/0009-0009-0690-8267
Investigadora Independiente
México
José Juan Muñoz León
juanmunoz@uv.mx
https://orcid.org/0000-0003-3557-8251
Universidad Veracruzana.
México
Judith Guadalupe Montero Mora
jmontero@uv.mx
https://orcid.org/0000-0003-4855-3248
Universidad Veracruzana
México
Zoylo Morales Romero
zmorales@uv.mx
https://orcid.org/0000-0001-6652-1480
Universidad Veracruzana
México
Daniela Burgueño Theurel
dburgueno@uv.mx
https://orcid.org/0009-0008-1389-736X
Universidad Veracruzana
México
ABSTRACT
The relationship between physical activity, physical self-concept, and academic performance has
attracted considerable attention across different disciplines because of its relevance to students’ overall
development. However, the complexity of these interactions requires advanced statistical
methodologies capable of modeling simultaneous relationships among observed and latent variables.
The aim of this study was to analyze the structural relationship among physical activity, physical self-
concept, and admission examination results through the application of Structural Equation Modeling.
A theoretical model was proposed in which physical activity acted as an exogenous variable, whereas
physical self-concept and the admission examination result were considered endogenous variables. The
findings showed a positive and statistically significant effect of physical activity on physical self-
concept. Physical self-concept, in turn, showed a negative and statistically significant relationship with
the admission examination result, while the direct effect of physical activity on the examination result
was not significant. Although the global fit indices indicated a limited fit of the model, the structural
estimates provided exploratory evidence about the direction and magnitude of the relationships among
the variables. These findings highlight the usefulness of structural equation modeling as a
methodological tool for examining complex multivariate relationships in educational contexts.
Keywords: physical activity, physical self-concept, admission examination result, academic
performance, structural equation modeling.
1 Autor principal
Correspondencia: juanmunoz@uv.mx
DOI: https://doi.org/10.37811/cl_rcm.v10i4.25211
Structural Analysis of the Relationships Among Physical Activity, Physical
Self-Concept, and Academic Performance: An Application of Structural
Equation Modeling
Thessa Montserrat Gutiérrez González1
tesmongonzalez@gmail.com
https://orcid.org/0009-0009-0690-8267
Investigadora Independiente
México
José Juan Muñoz León
juanmunoz@uv.mx
https://orcid.org/0000-0003-3557-8251
Universidad Veracruzana.
México
Judith Guadalupe Montero Mora
jmontero@uv.mx
https://orcid.org/0000-0003-4855-3248
Universidad Veracruzana
México
Zoylo Morales Romero
zmorales@uv.mx
https://orcid.org/0000-0001-6652-1480
Universidad Veracruzana
México
Daniela Burgueño Theurel
dburgueno@uv.mx
https://orcid.org/0009-0008-1389-736X
Universidad Veracruzana
México
ABSTRACT
The relationship between physical activity, physical self-concept, and academic performance has
attracted considerable attention across different disciplines because of its relevance to students’ overall
development. However, the complexity of these interactions requires advanced statistical
methodologies capable of modeling simultaneous relationships among observed and latent variables.
The aim of this study was to analyze the structural relationship among physical activity, physical self-
concept, and admission examination results through the application of Structural Equation Modeling.
A theoretical model was proposed in which physical activity acted as an exogenous variable, whereas
physical self-concept and the admission examination result were considered endogenous variables. The
findings showed a positive and statistically significant effect of physical activity on physical self-
concept. Physical self-concept, in turn, showed a negative and statistically significant relationship with
the admission examination result, while the direct effect of physical activity on the examination result
was not significant. Although the global fit indices indicated a limited fit of the model, the structural
estimates provided exploratory evidence about the direction and magnitude of the relationships among
the variables. These findings highlight the usefulness of structural equation modeling as a
methodological tool for examining complex multivariate relationships in educational contexts.
Keywords: physical activity, physical self-concept, admission examination result, academic
performance, structural equation modeling.
1 Autor principal
Correspondencia: juanmunoz@uv.mx
pág. 1840
Análisis Estructural de la Relación entre Actividad Física, Autoconcepto
Físico y Rendimiento Académico: Una Aplicación de Modelos de
Ecuaciones Estructurales
RESUMEN
La relación entre la actividad física, el autoconcepto físico y el rendimiento académico ha sido objeto
de interés en diversas disciplinas debido a su impacto en el desarrollo integral de los estudiantes. Sin
embargo, la complejidad de estas interacciones requiere metodologías estadísticas avanzadas capaces
de modelar relaciones simultáneas entre variables observadas y latentes. En este contexto, el presente
estudio tuvo como objetivo analizar la relación estructural entre la actividad física, el autoconcepto
físico y el rendimiento académico mediante la aplicación de modelos de ecuaciones estructurales. Se
propuso un modelo teórico en el que la actividad física actúa como variable exógena, mientras que el
autoconcepto físico y el rendimiento académico se consideran variables endógenas. Los resultados
evidenciaron un ajuste adecuado del modelo estructural, mostrando que la actividad física presentó una
asociación positiva y estadísticamente significativa con el autoconcepto físico. Asimismo, el
autoconcepto físico mostró una relación negativa y significativa con el rendimiento académico. Estos
hallazgos destacan la utilidad de los modelos de ecuaciones estructurales como una herramienta robusta
para el análisis de relaciones multivariadas complejas en contextos educativos.
Palabras clave: actividad física, autoconcepto físico, rendimiento académico, modelos de ecuaciones
estructurales.
Artículo recibido 20 mayo 2026
Aceptado para publicación: 20 junio 2026
Análisis Estructural de la Relación entre Actividad Física, Autoconcepto
Físico y Rendimiento Académico: Una Aplicación de Modelos de
Ecuaciones Estructurales
RESUMEN
La relación entre la actividad física, el autoconcepto físico y el rendimiento académico ha sido objeto
de interés en diversas disciplinas debido a su impacto en el desarrollo integral de los estudiantes. Sin
embargo, la complejidad de estas interacciones requiere metodologías estadísticas avanzadas capaces
de modelar relaciones simultáneas entre variables observadas y latentes. En este contexto, el presente
estudio tuvo como objetivo analizar la relación estructural entre la actividad física, el autoconcepto
físico y el rendimiento académico mediante la aplicación de modelos de ecuaciones estructurales. Se
propuso un modelo teórico en el que la actividad física actúa como variable exógena, mientras que el
autoconcepto físico y el rendimiento académico se consideran variables endógenas. Los resultados
evidenciaron un ajuste adecuado del modelo estructural, mostrando que la actividad física presentó una
asociación positiva y estadísticamente significativa con el autoconcepto físico. Asimismo, el
autoconcepto físico mostró una relación negativa y significativa con el rendimiento académico. Estos
hallazgos destacan la utilidad de los modelos de ecuaciones estructurales como una herramienta robusta
para el análisis de relaciones multivariadas complejas en contextos educativos.
Palabras clave: actividad física, autoconcepto físico, rendimiento académico, modelos de ecuaciones
estructurales.
Artículo recibido 20 mayo 2026
Aceptado para publicación: 20 junio 2026
pág. 1841
INTRODUCTION
Academic performance is one of the main indicators of student achievement and educational
development; therefore, its analysis has become highly relevant in fields such as education, psychology,
sport sciences, and health. Within this context, several physical, cognitive, and psychosocial factors
have been identified as potential determinants of academic outcomes, among which physical activity
and physical self-concept stand out because of their influence on students’ integral development
(Aaltonen et al., 2020; Claver et al., 2020).
Physical activity has been widely recognized as an essential component of health and well-being, not
only because of its physiological benefits but also because of its positive effects on cognitive, emotional,
and behavioral variables. Previous studies have indicated that regular physical activity may support
cognitive functions such as attention, memory, and concentration, all of which are closely related to
academic performance (Aaltonen et al., 2020). In addition, physical activity can strengthen
psychological dimensions associated with self-esteem, body perception, and self-evaluation.
Within these dimensions, physical self-concept is a construct of particular interest because it describes
individuals’ perceptions of their physical abilities, body condition, and motor competence. Recent
evidence suggests that higher levels of physical activity are usually associated with a more favorable
physical self-concept, strengthening a positive perception of personal abilities and psychological well-
being (Sevil-Serrano et al., 2020; Lohbeck et al., 2021). Nevertheless, the relationship between physical
self-concept and academic performance continues to show heterogeneous results in the literature,
suggesting complex interactions among physical, psychological, and academic variables.
The study of these relationships involves important methodological challenges due to the
multidimensional nature of the constructs involved. Physical activity, physical self-concept, and
academic performance are part of complex systems in which direct, indirect, and potential mediation
effects may coexist. Therefore, the analysis of these associations requires advanced statistical methods
capable of modeling multiple dependency relationships simultaneously and representing latent
constructs with adequate precision.
In this sense, Structural Equation Modeling (SEM) has become one of the most robust and versatile
statistical techniques for the multivariate analysis of complex phenomena. SEM integrates confirmatory
INTRODUCTION
Academic performance is one of the main indicators of student achievement and educational
development; therefore, its analysis has become highly relevant in fields such as education, psychology,
sport sciences, and health. Within this context, several physical, cognitive, and psychosocial factors
have been identified as potential determinants of academic outcomes, among which physical activity
and physical self-concept stand out because of their influence on students’ integral development
(Aaltonen et al., 2020; Claver et al., 2020).
Physical activity has been widely recognized as an essential component of health and well-being, not
only because of its physiological benefits but also because of its positive effects on cognitive, emotional,
and behavioral variables. Previous studies have indicated that regular physical activity may support
cognitive functions such as attention, memory, and concentration, all of which are closely related to
academic performance (Aaltonen et al., 2020). In addition, physical activity can strengthen
psychological dimensions associated with self-esteem, body perception, and self-evaluation.
Within these dimensions, physical self-concept is a construct of particular interest because it describes
individuals’ perceptions of their physical abilities, body condition, and motor competence. Recent
evidence suggests that higher levels of physical activity are usually associated with a more favorable
physical self-concept, strengthening a positive perception of personal abilities and psychological well-
being (Sevil-Serrano et al., 2020; Lohbeck et al., 2021). Nevertheless, the relationship between physical
self-concept and academic performance continues to show heterogeneous results in the literature,
suggesting complex interactions among physical, psychological, and academic variables.
The study of these relationships involves important methodological challenges due to the
multidimensional nature of the constructs involved. Physical activity, physical self-concept, and
academic performance are part of complex systems in which direct, indirect, and potential mediation
effects may coexist. Therefore, the analysis of these associations requires advanced statistical methods
capable of modeling multiple dependency relationships simultaneously and representing latent
constructs with adequate precision.
In this sense, Structural Equation Modeling (SEM) has become one of the most robust and versatile
statistical techniques for the multivariate analysis of complex phenomena. SEM integrates confirmatory
pág. 1842
factor analysis and structural regression models within a single analytical framework, allowing the
simultaneous estimation of relationships among observed and latent variables. This feature offers
substantial advantages over traditional statistical approaches, such as multiple linear regression, because
it provides a more comprehensive evaluation of hypothetical causal structures.
One of the main strengths of SEM is its capacity to test complex theoretical models, estimate direct and
indirect effects, and evaluate global model fit through indices such as the Comparative Fit Index (CFI),
Tucker-Lewis Index (TLI), Root Mean Square Error of Approximation (RMSEA), and Standardized
Root Mean Square Residual (SRMR). This makes it possible not only to identify significant associations
among variables but also to understand the underlying structure of these relationships with greater
statistical rigor (Vollmer et al., 2021).
In recent years, the application of SEM in educational and behavioral studies has increased considerably
because of its ability to model behavioral, psychological, and academic variables simultaneously. In
studies related to physical activity, self-concept, and academic performance, this technique has proven
useful for validating theoretical models and explaining complex relationships among multidimensional
constructs (Núñez et al., 2021).
From this perspective, the present study aims to analyze the structural relationship among physical
activity, physical self-concept, and academic performance measured through admission examination
results. A theoretical model is proposed in which physical activity acts as an exogenous variable,
whereas physical self-concept and academic performance are conceptualized as endogenous variables
within a relational structure. Through this approach, the study seeks to provide empirical and
methodological evidence that contributes to a deeper understanding of the interactions among physical,
psychological, and academic variables in educational contexts.
Recent research has reinforced the usefulness of structural equation modeling for studying complex
relationships among behavioral, psychological, and academic variables. For example, SEM has been
successfully used to model relationships among physical activity, self-esteem, and academic
performance, allowing the estimation of both direct and mediating effects within complex multivariate
structures. Villodres et al. (2023), through a structural model applied to primary education students,
found that physical activity, together with other healthy lifestyle behaviors, was positively associated
factor analysis and structural regression models within a single analytical framework, allowing the
simultaneous estimation of relationships among observed and latent variables. This feature offers
substantial advantages over traditional statistical approaches, such as multiple linear regression, because
it provides a more comprehensive evaluation of hypothetical causal structures.
One of the main strengths of SEM is its capacity to test complex theoretical models, estimate direct and
indirect effects, and evaluate global model fit through indices such as the Comparative Fit Index (CFI),
Tucker-Lewis Index (TLI), Root Mean Square Error of Approximation (RMSEA), and Standardized
Root Mean Square Residual (SRMR). This makes it possible not only to identify significant associations
among variables but also to understand the underlying structure of these relationships with greater
statistical rigor (Vollmer et al., 2021).
In recent years, the application of SEM in educational and behavioral studies has increased considerably
because of its ability to model behavioral, psychological, and academic variables simultaneously. In
studies related to physical activity, self-concept, and academic performance, this technique has proven
useful for validating theoretical models and explaining complex relationships among multidimensional
constructs (Núñez et al., 2021).
From this perspective, the present study aims to analyze the structural relationship among physical
activity, physical self-concept, and academic performance measured through admission examination
results. A theoretical model is proposed in which physical activity acts as an exogenous variable,
whereas physical self-concept and academic performance are conceptualized as endogenous variables
within a relational structure. Through this approach, the study seeks to provide empirical and
methodological evidence that contributes to a deeper understanding of the interactions among physical,
psychological, and academic variables in educational contexts.
Recent research has reinforced the usefulness of structural equation modeling for studying complex
relationships among behavioral, psychological, and academic variables. For example, SEM has been
successfully used to model relationships among physical activity, self-esteem, and academic
performance, allowing the estimation of both direct and mediating effects within complex multivariate
structures. Villodres et al. (2023), through a structural model applied to primary education students,
found that physical activity, together with other healthy lifestyle behaviors, was positively associated
pág. 1843
with self-esteem and indicators of academic performance, suggesting that psychoemotional variables
may act as relevant mediators within the educational process.
Similarly, Zurita-Ortega et al. (2023) developed a structural equation model to evaluate the relationship
among levels of physical activity, family functioning, and self-concept in elementary and high school
students. Their results showed that higher levels of physical activity were associated with higher scores
in self-concept dimensions, supporting the hypothesis that regular physical activity favors psychosocial
development and a positive perception of the physical self.
Regarding academic performance, recent studies have noted that the association between physical
activity and school achievement is not always direct but is often mediated by cognitive and
psychological variables. Muntaner-Mas et al. (2022), using SEM, reported that physical fitness and
executive function significantly mediated the relationship between physical activity and academic
achievement, suggesting that the effects of physical activity on academic performance operate through
more complex intermediate mechanisms than a simple linear association.
Likewise, research in university contexts has shown that structural models can capture complex
relationships among self-concept, emotional variables, and academic performance. Ubago-Jiménez et
al. (2023) found that self-concept is significantly related to academic performance, although the
magnitude and direction of this association may vary depending on the population context and the
psychosocial characteristics analyzed. These findings support the need to continue exploring these
relationships through advanced statistical approaches that allow a more precise understanding of their
interactions.
METHODOLOGY
This study was conducted under a quantitative approach, with a cross-sectional, non-experimental
design and a correlational-explanatory scope. The main objective was to analyze the relationships
among physical activity, physical self-concept, and academic performance in university students
through the application of structural equation modeling (SEM). This technique was selected because of
its capacity to model complex relationships among observed and latent variables simultaneously,
making it possible to identify both direct and indirect effects within the same analytical framework.
with self-esteem and indicators of academic performance, suggesting that psychoemotional variables
may act as relevant mediators within the educational process.
Similarly, Zurita-Ortega et al. (2023) developed a structural equation model to evaluate the relationship
among levels of physical activity, family functioning, and self-concept in elementary and high school
students. Their results showed that higher levels of physical activity were associated with higher scores
in self-concept dimensions, supporting the hypothesis that regular physical activity favors psychosocial
development and a positive perception of the physical self.
Regarding academic performance, recent studies have noted that the association between physical
activity and school achievement is not always direct but is often mediated by cognitive and
psychological variables. Muntaner-Mas et al. (2022), using SEM, reported that physical fitness and
executive function significantly mediated the relationship between physical activity and academic
achievement, suggesting that the effects of physical activity on academic performance operate through
more complex intermediate mechanisms than a simple linear association.
Likewise, research in university contexts has shown that structural models can capture complex
relationships among self-concept, emotional variables, and academic performance. Ubago-Jiménez et
al. (2023) found that self-concept is significantly related to academic performance, although the
magnitude and direction of this association may vary depending on the population context and the
psychosocial characteristics analyzed. These findings support the need to continue exploring these
relationships through advanced statistical approaches that allow a more precise understanding of their
interactions.
METHODOLOGY
This study was conducted under a quantitative approach, with a cross-sectional, non-experimental
design and a correlational-explanatory scope. The main objective was to analyze the relationships
among physical activity, physical self-concept, and academic performance in university students
through the application of structural equation modeling (SEM). This technique was selected because of
its capacity to model complex relationships among observed and latent variables simultaneously,
making it possible to identify both direct and indirect effects within the same analytical framework.
pág. 1844
Data were obtained through an online survey administered to university students from a Mexican
institution. After the data cleaning and validation process, a total of 208 complete records were available
for the analysis. The database included sociodemographic variables such as age, sex, and degree
program, as well as information related to physical activity, physical self-concept, and academic
performance. For analyses that included the admission examination result, complete cases were used
according to the availability of the PCE, PMA, PRIN, PCL, and PING scores.
The sample consisted of university students of different sexes (117 women, 85 men, and 6 who preferred
not to report this information) and from different disciplinary areas, which provided a heterogeneous
population for analysis. The average age of participants was 19.18 years (SD = 3.01). Regarding
academic performance, the global score on the admission examination had a mean of 67.47 points (SD
= 11.53). Students also reported devoting an average of 2.6 hours per day to physical activity.
The physical activity variable was defined as the number of daily hours devoted to physical activity,
obtained through self-report. Physical self-concept was measured using the Physical Self-Concept
Questionnaire (CAF) by Goñi, Ruiz de Azúa, and Rodríguez (2006), one of the reference instruments
for Spanish-speaking populations. This questionnaire consists of 36 items grouped into dimensions
related to physical ability, physical condition, physical attractiveness, and strength, allowing physical
self-concept to be modeled as a latent construct. Academic performance was operationalized through
the global score obtained by participants on the university admission examination, used as a quantitative
indicator of academic performance in this study.
The statistical analysis was performed in Python using the semopy library for estimating the structural
equation model. A clear and sufficient theoretical description of these models can be found in Tarka
(2018). Initially, a descriptive analysis was conducted to characterize the sample and explore the
behavior of the variables of interest. Subsequently, a structural model was specified in which physical
activity was considered an exogenous variable, while physical self-concept and academic performance
were modeled as endogenous variables. The model included direct relationships between physical
activity and physical self-concept, between physical self-concept and academic performance, and
between physical activity and academic performance.
Data were obtained through an online survey administered to university students from a Mexican
institution. After the data cleaning and validation process, a total of 208 complete records were available
for the analysis. The database included sociodemographic variables such as age, sex, and degree
program, as well as information related to physical activity, physical self-concept, and academic
performance. For analyses that included the admission examination result, complete cases were used
according to the availability of the PCE, PMA, PRIN, PCL, and PING scores.
The sample consisted of university students of different sexes (117 women, 85 men, and 6 who preferred
not to report this information) and from different disciplinary areas, which provided a heterogeneous
population for analysis. The average age of participants was 19.18 years (SD = 3.01). Regarding
academic performance, the global score on the admission examination had a mean of 67.47 points (SD
= 11.53). Students also reported devoting an average of 2.6 hours per day to physical activity.
The physical activity variable was defined as the number of daily hours devoted to physical activity,
obtained through self-report. Physical self-concept was measured using the Physical Self-Concept
Questionnaire (CAF) by Goñi, Ruiz de Azúa, and Rodríguez (2006), one of the reference instruments
for Spanish-speaking populations. This questionnaire consists of 36 items grouped into dimensions
related to physical ability, physical condition, physical attractiveness, and strength, allowing physical
self-concept to be modeled as a latent construct. Academic performance was operationalized through
the global score obtained by participants on the university admission examination, used as a quantitative
indicator of academic performance in this study.
The statistical analysis was performed in Python using the semopy library for estimating the structural
equation model. A clear and sufficient theoretical description of these models can be found in Tarka
(2018). Initially, a descriptive analysis was conducted to characterize the sample and explore the
behavior of the variables of interest. Subsequently, a structural model was specified in which physical
activity was considered an exogenous variable, while physical self-concept and academic performance
were modeled as endogenous variables. The model included direct relationships between physical
activity and physical self-concept, between physical self-concept and academic performance, and
between physical activity and academic performance.
pág. 1845
The global fit of the model was evaluated using several goodness-of-fit indices commonly employed in
SEM, including χ², CFI, GFI, AGFI, NFI, TLI, RMSEA, AIC, and BIC. The obtained values were χ² =
113.613, CFI = 0.649, GFI = 0.638, AGFI = 0.357, NFI = 0.638, TLI = 0.376, RMSEA = 0.239, AIC =
22.892, and BIC = 62.768.
Overall, the obtained indices suggest a limited fit of the structural model to the observed data. This
behavior may be associated with several factors, including the sample size. Although the sample
allowed the model to be estimated and the proposed structural relationships to be explored, relatively
small sample sizes in SEM may affect the stability of estimates and limit the quality of the global model
fit. Despite this limitation, the analysis made it possible to identify relevant structural relationships
among the variables studied, providing useful evidence for understanding the interactions among
physical activity, physical self-concept, and academic performance.
RESULTS AND DISCUSSION
Descriptive and Exploratory Analysis
After importing the database into the Python working environment, the integrity of the information and
the consistency of the variables included in the study were verified. The database contained
sociodemographic information on the participants, as well as variables related to physical activity,
physical self-concept, and results obtained on the admission examination.
As part of the data preparation process, variable names were standardized to ensure proper handling
during the different stages of the statistical analysis. Subsequently, an academic performance indicator
corresponding to the global admission examination score was constructed as the average of the scores
obtained in Critical Thinking (PCE), Mathematical Thinking (PMA), Research Principles (PRIN),
Reading Comprehension (PCL), and English (PING). The integration of these areas made it possible to
obtain a synthetic measure of the academic performance achieved by applicants during the selection
process.
The global admission examination score was considered the main response variable of the study, since
it constitutes an overall indicator of students’ academic performance. In addition, the scores
corresponding to each of the areas that make up the examination were analyzed individually in order to
identify specific performance patterns that might not be evident when only the global indicator was
The global fit of the model was evaluated using several goodness-of-fit indices commonly employed in
SEM, including χ², CFI, GFI, AGFI, NFI, TLI, RMSEA, AIC, and BIC. The obtained values were χ² =
113.613, CFI = 0.649, GFI = 0.638, AGFI = 0.357, NFI = 0.638, TLI = 0.376, RMSEA = 0.239, AIC =
22.892, and BIC = 62.768.
Overall, the obtained indices suggest a limited fit of the structural model to the observed data. This
behavior may be associated with several factors, including the sample size. Although the sample
allowed the model to be estimated and the proposed structural relationships to be explored, relatively
small sample sizes in SEM may affect the stability of estimates and limit the quality of the global model
fit. Despite this limitation, the analysis made it possible to identify relevant structural relationships
among the variables studied, providing useful evidence for understanding the interactions among
physical activity, physical self-concept, and academic performance.
RESULTS AND DISCUSSION
Descriptive and Exploratory Analysis
After importing the database into the Python working environment, the integrity of the information and
the consistency of the variables included in the study were verified. The database contained
sociodemographic information on the participants, as well as variables related to physical activity,
physical self-concept, and results obtained on the admission examination.
As part of the data preparation process, variable names were standardized to ensure proper handling
during the different stages of the statistical analysis. Subsequently, an academic performance indicator
corresponding to the global admission examination score was constructed as the average of the scores
obtained in Critical Thinking (PCE), Mathematical Thinking (PMA), Research Principles (PRIN),
Reading Comprehension (PCL), and English (PING). The integration of these areas made it possible to
obtain a synthetic measure of the academic performance achieved by applicants during the selection
process.
The global admission examination score was considered the main response variable of the study, since
it constitutes an overall indicator of students’ academic performance. In addition, the scores
corresponding to each of the areas that make up the examination were analyzed individually in order to
identify specific performance patterns that might not be evident when only the global indicator was
pág. 1846
considered. This complementary analysis provided a more detailed characterization of participants’
academic behavior and contributed additional elements for interpreting the structural model results.
Table 1 presents the distribution of the main descriptive variables considered in the study, including
sex, physical activity level, physical self-concept category, and the number of daily hours devoted to
physical activity.
The descriptive statistics made it possible to characterize the initial behavior of the variables included
in the analysis. Overall, the distribution of the measures of central tendency and dispersion showed
sufficient variability among participants, which is favorable for estimating the structural equation
model. In addition, the graphical exploration of the data and the descriptive analysis did not show
patterns that compromised the continuation of subsequent analyses; therefore, the associations among
variables and the specification of the structural model were evaluated.
Table 1. Distribution of the main descriptive variables.
Variable Category Frequency Percentage
Sex Women 117 56.25
Sex Men 85 40.87
Sex Preferred not to say 6 2.88
Physical activity level Low 33 15.87
Physical activity level Medium 125 60.10
Physical activity level High 50 24.04
Physical self-concept category Low 34 16.35
Physical self-concept category Medium 138 66.35
Physical self-concept category High 36 17.31
Daily hours of physical activity 1 hour 23 11.06
Daily hours of physical activity 2 hours 59 28.37
Daily hours of physical activity 3 hours 104 50.00
Daily hours of physical activity 4 hours 22 10.58
considered. This complementary analysis provided a more detailed characterization of participants’
academic behavior and contributed additional elements for interpreting the structural model results.
Table 1 presents the distribution of the main descriptive variables considered in the study, including
sex, physical activity level, physical self-concept category, and the number of daily hours devoted to
physical activity.
The descriptive statistics made it possible to characterize the initial behavior of the variables included
in the analysis. Overall, the distribution of the measures of central tendency and dispersion showed
sufficient variability among participants, which is favorable for estimating the structural equation
model. In addition, the graphical exploration of the data and the descriptive analysis did not show
patterns that compromised the continuation of subsequent analyses; therefore, the associations among
variables and the specification of the structural model were evaluated.
Table 1. Distribution of the main descriptive variables.
Variable Category Frequency Percentage
Sex Women 117 56.25
Sex Men 85 40.87
Sex Preferred not to say 6 2.88
Physical activity level Low 33 15.87
Physical activity level Medium 125 60.10
Physical activity level High 50 24.04
Physical self-concept category Low 34 16.35
Physical self-concept category Medium 138 66.35
Physical self-concept category High 36 17.31
Daily hours of physical activity 1 hour 23 11.06
Daily hours of physical activity 2 hours 59 28.37
Daily hours of physical activity 3 hours 104 50.00
Daily hours of physical activity 4 hours 22 10.58
pág. 1847
Figure 1 shows the distribution of students by sex. The sample was mainly composed of women,
followed by men and, to a lesser extent, students who preferred not to report this information.
Figure 1. Distribution of students by sex.
Figure 2 shows the distribution of the physical activity level reported by students. The highest frequency
was concentrated at the medium level, followed by the high level and, finally, the low level.
Figure 2. Distribution by level of physical activity.
Figure 1 shows the distribution of students by sex. The sample was mainly composed of women,
followed by men and, to a lesser extent, students who preferred not to report this information.
Figure 1. Distribution of students by sex.
Figure 2 shows the distribution of the physical activity level reported by students. The highest frequency
was concentrated at the medium level, followed by the high level and, finally, the low level.
Figure 2. Distribution by level of physical activity.
pág. 1848
Figure 3 presents the distribution by physical self-concept category. Most students were located in the
medium category, suggesting a general tendency toward moderate physical perceptions within the
sample.
Figure 3. Distribution by physical self-concept category.
Regarding the time devoted to physical activity, Figure 4 shows that the most frequent category was
three hours per day, followed by two hours, one hour, and four hours.
Figure 4. Daily hours devoted to physical activity.
Figure 3 presents the distribution by physical self-concept category. Most students were located in the
medium category, suggesting a general tendency toward moderate physical perceptions within the
sample.
Figure 3. Distribution by physical self-concept category.
Regarding the time devoted to physical activity, Figure 4 shows that the most frequent category was
three hours per day, followed by two hours, one hour, and four hours.
Figure 4. Daily hours devoted to physical activity.
pág. 1849
Figure 5 compares the average admission examination result according to physical self-concept
category. Although the descriptive differences should not be interpreted as causal evidence, this
comparison made it possible to observe initial variations between groups, which justified further
analysis of the association between these variables.
Figure 5. Average admission examination result by physical self-concept category.
Table 2. Average admission examination result by physical self-concept category.
Physical self-concept category Average EXAMEN_GLOBAL
Low 67.78
Medium 68.75
High 62.41
Spearman Correlational Analysis
Figure 6 presents the Spearman correlation matrix among the variables considered in the study. In
general terms, the correlations showed low to moderate associations among the dimensions of physical
self-concept, physical activity, and academic performance. As expected, the dimensions that make up
physical self-concept showed positive correlations with one another, reflecting the conceptual
consistency of the construct.
Figure 5 compares the average admission examination result according to physical self-concept
category. Although the descriptive differences should not be interpreted as causal evidence, this
comparison made it possible to observe initial variations between groups, which justified further
analysis of the association between these variables.
Figure 5. Average admission examination result by physical self-concept category.
Table 2. Average admission examination result by physical self-concept category.
Physical self-concept category Average EXAMEN_GLOBAL
Low 67.78
Medium 68.75
High 62.41
Spearman Correlational Analysis
Figure 6 presents the Spearman correlation matrix among the variables considered in the study. In
general terms, the correlations showed low to moderate associations among the dimensions of physical
self-concept, physical activity, and academic performance. As expected, the dimensions that make up
physical self-concept showed positive correlations with one another, reflecting the conceptual
consistency of the construct.
pág. 1850
Regarding physical activity, positive associations were observed with the global physical self-concept
score and with its different dimensions, suggesting that greater dedication to physical activity tends to
accompany a more favorable perception of one’s own physical abilities. In contrast, the correlations
between academic performance and the physical self-concept variables were of lower magnitude,
anticipating that the relationships between these constructs may be more complex than a simple
bivariate association.
Taken together, the results of the correlational analysis provide an initial approximation to the structure
of relationships among the variables studied and support the relevance of evaluating these associations
through a structural equation model, which allows the simultaneous analysis of the direct and indirect
effects proposed in the theoretical model.
Figure 6. Spearman correlation matrix between physical self-concept dimensions and admission
examination result.
Regarding physical activity, positive associations were observed with the global physical self-concept
score and with its different dimensions, suggesting that greater dedication to physical activity tends to
accompany a more favorable perception of one’s own physical abilities. In contrast, the correlations
between academic performance and the physical self-concept variables were of lower magnitude,
anticipating that the relationships between these constructs may be more complex than a simple
bivariate association.
Taken together, the results of the correlational analysis provide an initial approximation to the structure
of relationships among the variables studied and support the relevance of evaluating these associations
through a structural equation model, which allows the simultaneous analysis of the direct and indirect
effects proposed in the theoretical model.
Figure 6. Spearman correlation matrix between physical self-concept dimensions and admission
examination result.
pág. 1851
In general terms, the observed associations showed an intensity ranging from low to moderate,
suggesting that, although the dimensions of physical self-concept share some degree of relationship,
each one contributes differentiated information about students’ physical perceptions.
The highest relationship was identified between sports ability and physical strength (ρ = 0.63), showing
a positive association of moderately high magnitude. This result indicates that participants who
perceived higher levels of ability for sports practice also tended to perceive themselves as physically
stronger, which is consistent with the conceptual structure of the Physical Self-Concept Questionnaire
(CAF). Likewise, a positive correlation was observed between physical condition and physical
attractiveness (ρ = 0.41), as well as between sports ability and physical attractiveness (ρ = 0.35),
suggesting that a better perception of physical condition and physical performance is usually
accompanied by a more favorable evaluation of body image.
In contrast, some associations among dimensions were practically nonexistent or even negative. The
absence of a relationship between sports ability and physical condition (ρ ≈ 0.00) is noteworthy, as it
indicates independence between both perceptions within the analyzed sample. Similarly, physical
condition showed a moderate negative correlation with physical strength (ρ = -0.32), suggesting that
the perception of better physical condition is not necessarily accompanied by greater self-perceived
strength. Although this behavior differs from what might be expected conceptually, it shows that the
dimensions of physical self-concept retain some independence and may be influenced by personal
experiences, specific sports practices, or individual characteristics of the participants.
With respect to academic performance, represented by the global admission examination result, the
observed associations were generally low in magnitude. The highest negative correlation was found
with the sports ability dimension (ρ = -0.25), followed by physical strength (ρ = -0.19), whereas physical
attractiveness showed an almost null relationship (ρ = -0.04). Conversely, physical condition was the
only dimension that presented a positive association with academic performance (ρ = 0.11), although
its magnitude was small. These results indicate that the relationships between physical self-concept and
academic performance are neither linear nor homogeneous, and that the different dimensions of the
construct may behave differently with respect to academic performance.
In general terms, the observed associations showed an intensity ranging from low to moderate,
suggesting that, although the dimensions of physical self-concept share some degree of relationship,
each one contributes differentiated information about students’ physical perceptions.
The highest relationship was identified between sports ability and physical strength (ρ = 0.63), showing
a positive association of moderately high magnitude. This result indicates that participants who
perceived higher levels of ability for sports practice also tended to perceive themselves as physically
stronger, which is consistent with the conceptual structure of the Physical Self-Concept Questionnaire
(CAF). Likewise, a positive correlation was observed between physical condition and physical
attractiveness (ρ = 0.41), as well as between sports ability and physical attractiveness (ρ = 0.35),
suggesting that a better perception of physical condition and physical performance is usually
accompanied by a more favorable evaluation of body image.
In contrast, some associations among dimensions were practically nonexistent or even negative. The
absence of a relationship between sports ability and physical condition (ρ ≈ 0.00) is noteworthy, as it
indicates independence between both perceptions within the analyzed sample. Similarly, physical
condition showed a moderate negative correlation with physical strength (ρ = -0.32), suggesting that
the perception of better physical condition is not necessarily accompanied by greater self-perceived
strength. Although this behavior differs from what might be expected conceptually, it shows that the
dimensions of physical self-concept retain some independence and may be influenced by personal
experiences, specific sports practices, or individual characteristics of the participants.
With respect to academic performance, represented by the global admission examination result, the
observed associations were generally low in magnitude. The highest negative correlation was found
with the sports ability dimension (ρ = -0.25), followed by physical strength (ρ = -0.19), whereas physical
attractiveness showed an almost null relationship (ρ = -0.04). Conversely, physical condition was the
only dimension that presented a positive association with academic performance (ρ = 0.11), although
its magnitude was small. These results indicate that the relationships between physical self-concept and
academic performance are neither linear nor homogeneous, and that the different dimensions of the
construct may behave differently with respect to academic performance.
pág. 1852
Overall, the correlation matrix shows that bivariate associations between the dimensions of physical
self-concept and academic performance are relatively weak, while relationships among the physical
self-concept dimensions themselves are stronger. This pattern is a relevant finding because it suggests
that the influence of physical self-concept on academic performance can hardly be explained through
simple correlations. Consequently, it is appropriate to use a multivariate approach such as structural
equation modeling, which allows the simultaneous analysis of the structure of relationships among
observed variables and latent constructs, estimating direct and indirect effects that are not evident
through bivariate analysis.
Finally, it is important to note that Spearman correlations represent monotonic associations between
pairs of variables and, therefore, do not constitute evidence of causal relationships. Their usefulness lies
in providing a first empirical approximation to the behavior of the variables studied and in supporting
the specification of the theoretical model that was subsequently evaluated through structural equation
modeling.
Structural Equation Model
Based on the evidence obtained from the descriptive analysis and the correlation matrix, a structural
equation model was estimated in order to evaluate the relationships among physical activity, physical
self-concept, and academic performance simultaneously. The model consisted of two complementary
elements: a measurement model, designed to represent physical self-concept as a latent construct based
on its observed dimensions, and a structural model, aimed at estimating the hypothesized relationships
among the variables of interest.
Figure 7 shows the specification of the theoretical model evaluated. In this model, physical activity was
proposed as an exogenous variable with a direct effect on physical self-concept and, additionally, on
academic performance, represented by the global score obtained on the admission examination.
Physical self-concept was also modeled as a latent variable with a direct effect on academic
performance, allowing the evaluation of both the influence of physical activity on students’ physical
perceptions and the possible mediating role of physical self-concept in explaining academic
performance.
Overall, the correlation matrix shows that bivariate associations between the dimensions of physical
self-concept and academic performance are relatively weak, while relationships among the physical
self-concept dimensions themselves are stronger. This pattern is a relevant finding because it suggests
that the influence of physical self-concept on academic performance can hardly be explained through
simple correlations. Consequently, it is appropriate to use a multivariate approach such as structural
equation modeling, which allows the simultaneous analysis of the structure of relationships among
observed variables and latent constructs, estimating direct and indirect effects that are not evident
through bivariate analysis.
Finally, it is important to note that Spearman correlations represent monotonic associations between
pairs of variables and, therefore, do not constitute evidence of causal relationships. Their usefulness lies
in providing a first empirical approximation to the behavior of the variables studied and in supporting
the specification of the theoretical model that was subsequently evaluated through structural equation
modeling.
Structural Equation Model
Based on the evidence obtained from the descriptive analysis and the correlation matrix, a structural
equation model was estimated in order to evaluate the relationships among physical activity, physical
self-concept, and academic performance simultaneously. The model consisted of two complementary
elements: a measurement model, designed to represent physical self-concept as a latent construct based
on its observed dimensions, and a structural model, aimed at estimating the hypothesized relationships
among the variables of interest.
Figure 7 shows the specification of the theoretical model evaluated. In this model, physical activity was
proposed as an exogenous variable with a direct effect on physical self-concept and, additionally, on
academic performance, represented by the global score obtained on the admission examination.
Physical self-concept was also modeled as a latent variable with a direct effect on academic
performance, allowing the evaluation of both the influence of physical activity on students’ physical
perceptions and the possible mediating role of physical self-concept in explaining academic
performance.
pág. 1853
Figure 7. Proposed structural equation model.
The global fit indices indicate that the structural equation model presented a limited fit to the observed
data. Specifically, the values obtained for CFI (0.649), GFI (0.638), NFI (0.638), and TLI (0.376) were
below the reference criteria commonly accepted for satisfactory fit, while RMSEA (0.239) exceeded
the threshold recommended in the literature. Overall, these results suggest that the proposed model does
not optimally reproduce the observed covariance structure; therefore, the estimated relationships should
be interpreted with due caution.
Table 3. Global fit indices of the structural equation model.
Index Obtained value Interpretation
χ² 113.613 Limited fit
CFI 0.649 Limited fit
GFI 0.638 Limited fit
AGFI 0.357 Limited fit
NFI 0.638 Limited fit
TLI 0.376 Limited fit
RMSEA 0.239 Limited fit
AIC 22.892 Information criterion
BIC 62.768 Information criterion
Nevertheless, global fit is only one criterion for evaluating structural equation models. In exploratory
studies or studies with moderate sample sizes, such as the present one, it is possible to obtain informative
structural estimates even when global fit indices do not reach ideal values. In this sense, the model made
Figure 7. Proposed structural equation model.
The global fit indices indicate that the structural equation model presented a limited fit to the observed
data. Specifically, the values obtained for CFI (0.649), GFI (0.638), NFI (0.638), and TLI (0.376) were
below the reference criteria commonly accepted for satisfactory fit, while RMSEA (0.239) exceeded
the threshold recommended in the literature. Overall, these results suggest that the proposed model does
not optimally reproduce the observed covariance structure; therefore, the estimated relationships should
be interpreted with due caution.
Table 3. Global fit indices of the structural equation model.
Index Obtained value Interpretation
χ² 113.613 Limited fit
CFI 0.649 Limited fit
GFI 0.638 Limited fit
AGFI 0.357 Limited fit
NFI 0.638 Limited fit
TLI 0.376 Limited fit
RMSEA 0.239 Limited fit
AIC 22.892 Information criterion
BIC 62.768 Information criterion
Nevertheless, global fit is only one criterion for evaluating structural equation models. In exploratory
studies or studies with moderate sample sizes, such as the present one, it is possible to obtain informative
structural estimates even when global fit indices do not reach ideal values. In this sense, the model made
pág. 1854
it possible to examine the relationships proposed in the theoretical framework and to provide empirical
evidence on the direction and magnitude of the effects among the variables analyzed.
In particular, the model identified a positive association between physical activity and physical self-
concept, as well as a negative relationship between physical self-concept and academic performance.
These results constitute the main findings of the study and are analyzed in greater detail below.
Consequently, the structural coefficients should be interpreted as exploratory evidence limited to the
sample studied, whose confirmation will require future research with larger samples and alternative
models that improve global fit.
Table 4. Estimated structural effects.
Dependent
variable
Predictor variable Coefficient β Standard
error
z p-value
Physical self-
concept
Hours of physical
activity
10.1075 1.3295 7.6026 <0.001
Admission
examination result
Physical self-
concept
-0.1530 0.0540 -2.8320 0.0046
Admission
examination result
Hours of physical
activity
-0.1796 1.0905 -0.1647 0.8692
The estimated structural coefficients show that physical activity had a direct, positive, and statistically
significant effect on physical self-concept, indicating that greater time devoted to physical activity is
associated with a more favorable perception of one’s own physical abilities and characteristics. This
result is consistent with the theoretical approach of the model and supports the hypothesis that physical
activity is a relevant factor in the construction of physical self-concept among university students.
In contrast, physical self-concept showed a direct negative and statistically significant effect on
academic performance, represented by the global admission examination score. Although the
magnitude of this effect was small, the direction of the relationship suggests that, in the analyzed
sample, higher levels of physical self-concept were associated with slightly lower scores on the
admission examination. This finding differs from part of the specialized literature and highlights the
complexity of the interactions between physical factors and academic performance, particularly when
these relationships are evaluated through multivariate models.
it possible to examine the relationships proposed in the theoretical framework and to provide empirical
evidence on the direction and magnitude of the effects among the variables analyzed.
In particular, the model identified a positive association between physical activity and physical self-
concept, as well as a negative relationship between physical self-concept and academic performance.
These results constitute the main findings of the study and are analyzed in greater detail below.
Consequently, the structural coefficients should be interpreted as exploratory evidence limited to the
sample studied, whose confirmation will require future research with larger samples and alternative
models that improve global fit.
Table 4. Estimated structural effects.
Dependent
variable
Predictor variable Coefficient β Standard
error
z p-value
Physical self-
concept
Hours of physical
activity
10.1075 1.3295 7.6026 <0.001
Admission
examination result
Physical self-
concept
-0.1530 0.0540 -2.8320 0.0046
Admission
examination result
Hours of physical
activity
-0.1796 1.0905 -0.1647 0.8692
The estimated structural coefficients show that physical activity had a direct, positive, and statistically
significant effect on physical self-concept, indicating that greater time devoted to physical activity is
associated with a more favorable perception of one’s own physical abilities and characteristics. This
result is consistent with the theoretical approach of the model and supports the hypothesis that physical
activity is a relevant factor in the construction of physical self-concept among university students.
In contrast, physical self-concept showed a direct negative and statistically significant effect on
academic performance, represented by the global admission examination score. Although the
magnitude of this effect was small, the direction of the relationship suggests that, in the analyzed
sample, higher levels of physical self-concept were associated with slightly lower scores on the
admission examination. This finding differs from part of the specialized literature and highlights the
complexity of the interactions between physical factors and academic performance, particularly when
these relationships are evaluated through multivariate models.
pág. 1855
On the other hand, the direct effect of physical activity on academic performance did not reach statistical
significance, indicating that the time devoted to physical activity did not, by itself, explain the observed
differences in participants’ academic performance. Consequently, the results suggest that the influence
of physical activity on academic performance may be mediated by other personal, psychological, or
contextual factors that were not incorporated into the structural model.
Considering the global fit obtained for the model, these results should be interpreted as exploratory
evidence limited to the population studied. Nevertheless, the identified effects provide relevant
information on the structure of relationships among the variables analyzed and constitute a starting point
for future research that incorporates larger samples, new explanatory constructs, and alternative model
specifications to improve explanatory capacity.
Figure 8. Estimated structural effects of the SEM model.
The sports competence and participation dimension was used as the reference indicator to establish the
scale of physical self-concept; therefore, its factor loading was fixed at 1.0000 and no standard error, z-
value, or p-value was estimated for this parameter. In contrast, perceived physical strength and general
physical self-evaluation presented statistically significant loadings, providing relevant evidence for the
construction of the latent variable of physical self-concept.
On the other hand, the direct effect of physical activity on academic performance did not reach statistical
significance, indicating that the time devoted to physical activity did not, by itself, explain the observed
differences in participants’ academic performance. Consequently, the results suggest that the influence
of physical activity on academic performance may be mediated by other personal, psychological, or
contextual factors that were not incorporated into the structural model.
Considering the global fit obtained for the model, these results should be interpreted as exploratory
evidence limited to the population studied. Nevertheless, the identified effects provide relevant
information on the structure of relationships among the variables analyzed and constitute a starting point
for future research that incorporates larger samples, new explanatory constructs, and alternative model
specifications to improve explanatory capacity.
Figure 8. Estimated structural effects of the SEM model.
The sports competence and participation dimension was used as the reference indicator to establish the
scale of physical self-concept; therefore, its factor loading was fixed at 1.0000 and no standard error, z-
value, or p-value was estimated for this parameter. In contrast, perceived physical strength and general
physical self-evaluation presented statistically significant loadings, providing relevant evidence for the
construction of the latent variable of physical self-concept.
pág. 1856
Table 5. Factor loadings of physical self-concept.
Indicator Factor loading Standard error z p-value
Sports competence and
participation 1.0000 -- -- --
Physical attractiveness and
body image 0.0179 0.0449 0.399 0.690
Perceived physical strength 0.2156 0.0419 5.144 <0.001
General physical self-
evaluation 0.4189 0.0586 7.143 <0.001
Taken together, the results show that the bivariate correlations made it possible to identify preliminary
associations between the dimensions of physical self-concept and academic performance; however,
these relationships were insufficient to explain the complexity of the phenomenon analyzed. The
incorporation of a structural equation model made it possible to integrate these associations into a
broader analytical framework, representing physical self-concept as a latent construct and
simultaneously estimating the direct relationships among physical activity, physical self-concept, and
academic performance. Although the global fit of the model was limited, the structural analysis
provided evidence about the direction and magnitude of the effects among the variables, allowing the
identification of association patterns that cannot be captured through traditional correlational analyses.
CONCLUSIONS
The results obtained allow us to conclude that physical activity is positively associated with physical
self-concept among university students, confirming the main hypothesis proposed in this study. This
finding is consistent with the results reported by Sevil-Serrano et al. (2020), Núñez et al. (2021), and
Lohbeck et al. (2021), who documented that greater participation in physical activities favors a more
positive perception of students’ physical abilities and characteristics. Overall, the evidence reinforces
the importance of promoting physically active lifestyles as an element that contributes to strengthening
physical self-concept during university education.
In contrast, the relationship identified between physical self-concept and academic performance showed
a negative and statistically significant direction, a result that differs from a substantial part of recent
literature. Studies such as those by Villodres et al. (2023) and Muntaner-Mas et al. (2022) have
Table 5. Factor loadings of physical self-concept.
Indicator Factor loading Standard error z p-value
Sports competence and
participation 1.0000 -- -- --
Physical attractiveness and
body image 0.0179 0.0449 0.399 0.690
Perceived physical strength 0.2156 0.0419 5.144 <0.001
General physical self-
evaluation 0.4189 0.0586 7.143 <0.001
Taken together, the results show that the bivariate correlations made it possible to identify preliminary
associations between the dimensions of physical self-concept and academic performance; however,
these relationships were insufficient to explain the complexity of the phenomenon analyzed. The
incorporation of a structural equation model made it possible to integrate these associations into a
broader analytical framework, representing physical self-concept as a latent construct and
simultaneously estimating the direct relationships among physical activity, physical self-concept, and
academic performance. Although the global fit of the model was limited, the structural analysis
provided evidence about the direction and magnitude of the effects among the variables, allowing the
identification of association patterns that cannot be captured through traditional correlational analyses.
CONCLUSIONS
The results obtained allow us to conclude that physical activity is positively associated with physical
self-concept among university students, confirming the main hypothesis proposed in this study. This
finding is consistent with the results reported by Sevil-Serrano et al. (2020), Núñez et al. (2021), and
Lohbeck et al. (2021), who documented that greater participation in physical activities favors a more
positive perception of students’ physical abilities and characteristics. Overall, the evidence reinforces
the importance of promoting physically active lifestyles as an element that contributes to strengthening
physical self-concept during university education.
In contrast, the relationship identified between physical self-concept and academic performance showed
a negative and statistically significant direction, a result that differs from a substantial part of recent
literature. Studies such as those by Villodres et al. (2023) and Muntaner-Mas et al. (2022) have
pág. 1857
documented positive associations among variables related to physical condition, well-being, and
academic performance. This discrepancy shows that the relationship between both constructs depends
on the context, the characteristics of the population studied, and the way in which academic performance
is operationalized. In the present study, academic performance was represented by the score obtained
on an admission examination, which constitutes a specific measure of cognitive performance at a given
moment and does not necessarily reflect sustained academic performance throughout the university
trajectory.
Similarly, the structural model showed that physical activity did not have a significant direct effect on
academic performance. This result suggests that the influence of physical activity on school
performance may be mediated by psychological, cognitive, or contextual variables that were not
incorporated into the model. In this sense, the findings support the need to develop more comprehensive
models that include variables such as motivation, executive functions, psychological well-being, study
habits, or family support, which have been identified in recent research as relevant factors for
understanding academic performance.
From a methodological perspective, one of the main contributions of this study lies in the application
of structural equation modeling as a tool for analyzing relationships among observed variables and
latent constructs simultaneously. Unlike traditional correlational analyses, SEM made it possible to
represent physical self-concept as a latent variable and to integrate, within the same model, the direct
effects among physical activity, physical self-concept, and academic performance. This approach is
consistent with the methodological literature that recognizes SEM as one of the most robust techniques
for studying complex multivariate phenomena in the social and behavioral sciences (Tarka, 2018).
Although the global fit indices showed a fit below conventionally accepted criteria, the structural
coefficients provided exploratory evidence consistent with the proposed theoretical model.
Consequently, the results should be interpreted with caution and limited to the analyzed sample. Future
research should consider larger sample sizes, incorporate new explanatory constructs, and evaluate
alternative models that improve global fit and deepen the understanding of the mechanisms linking
physical activity and academic performance.
documented positive associations among variables related to physical condition, well-being, and
academic performance. This discrepancy shows that the relationship between both constructs depends
on the context, the characteristics of the population studied, and the way in which academic performance
is operationalized. In the present study, academic performance was represented by the score obtained
on an admission examination, which constitutes a specific measure of cognitive performance at a given
moment and does not necessarily reflect sustained academic performance throughout the university
trajectory.
Similarly, the structural model showed that physical activity did not have a significant direct effect on
academic performance. This result suggests that the influence of physical activity on school
performance may be mediated by psychological, cognitive, or contextual variables that were not
incorporated into the model. In this sense, the findings support the need to develop more comprehensive
models that include variables such as motivation, executive functions, psychological well-being, study
habits, or family support, which have been identified in recent research as relevant factors for
understanding academic performance.
From a methodological perspective, one of the main contributions of this study lies in the application
of structural equation modeling as a tool for analyzing relationships among observed variables and
latent constructs simultaneously. Unlike traditional correlational analyses, SEM made it possible to
represent physical self-concept as a latent variable and to integrate, within the same model, the direct
effects among physical activity, physical self-concept, and academic performance. This approach is
consistent with the methodological literature that recognizes SEM as one of the most robust techniques
for studying complex multivariate phenomena in the social and behavioral sciences (Tarka, 2018).
Although the global fit indices showed a fit below conventionally accepted criteria, the structural
coefficients provided exploratory evidence consistent with the proposed theoretical model.
Consequently, the results should be interpreted with caution and limited to the analyzed sample. Future
research should consider larger sample sizes, incorporate new explanatory constructs, and evaluate
alternative models that improve global fit and deepen the understanding of the mechanisms linking
physical activity and academic performance.
pág. 1858
Finally, this study shows that the incorporation of advanced analytical methodologies, such as structural
equation modeling, expands the possibilities of educational research by making it possible to understand
the simultaneous interaction among physical, psychological, and academic factors. Beyond the specific
findings obtained, the study demonstrates the potential of SEM to generate explanatory models with
greater analytical richness, contributing to the development of evidence-based research and to the
strengthening of data science applied to complex educational problems, a trend that continues to
consolidate in contemporary educational research.
REFERENCES
Aaltonen, S., Latvala, A., Jelenkovic, A., Rose, R. J., Kujala, U. M., Kaprio, J., & Silventoinen, K.
(2020). Physical activity and academic performance: Genetic and environmental associations.
Medicine & Science in Sports & Exercise, 52(2), 381–390.
https://doi.org/10.1249/MSS.0000000000002124
Claver, F., Martínez-Aranda, L. M., Conejero, M., & Gil-Arias, A. (2020). Motivation, discipline, and
academic performance in physical education. Frontiers in Psychology, 11, 1808.
https://doi.org/10.3389/fpsyg.2020.01808
Goñi, A., Ruiz de Azúa, S., & Rodríguez, A. (2006). Cuestionario de Autoconcepto Físico (CAF). EOS.
Lohbeck, A., von Keitz, P., Hohmann, A., & Daseking, M. (2021). Children’s physical self-concept,
motivation, and physical performance. Frontiers in Psychology, 12, 669936.
https://doi.org/10.3389/fpsyg.2021.669936
Muntaner-Mas, A., Mazzoli, E., Abbott, G., Mavilidi, M. F., & Galmes-Panades, A. M. (2022). Do
physical fitness and executive function mediate the relationship between physical activity and
academic achievement? An examination using structural equation modelling. Children, 9(6),
823. https://doi.org/10.3390/children9060823
Núñez, J. L., León, J., Valero-Valenzuela, A., & Conte, L. (2021). Influence of physical self-concept
and motivational processes on moderate-to-vigorous physical activity of adolescents.
Frontiers in Psychology, 12, 685612. https://doi.org/10.3389/fpsyg.2021.685612
Finally, this study shows that the incorporation of advanced analytical methodologies, such as structural
equation modeling, expands the possibilities of educational research by making it possible to understand
the simultaneous interaction among physical, psychological, and academic factors. Beyond the specific
findings obtained, the study demonstrates the potential of SEM to generate explanatory models with
greater analytical richness, contributing to the development of evidence-based research and to the
strengthening of data science applied to complex educational problems, a trend that continues to
consolidate in contemporary educational research.
REFERENCES
Aaltonen, S., Latvala, A., Jelenkovic, A., Rose, R. J., Kujala, U. M., Kaprio, J., & Silventoinen, K.
(2020). Physical activity and academic performance: Genetic and environmental associations.
Medicine & Science in Sports & Exercise, 52(2), 381–390.
https://doi.org/10.1249/MSS.0000000000002124
Claver, F., Martínez-Aranda, L. M., Conejero, M., & Gil-Arias, A. (2020). Motivation, discipline, and
academic performance in physical education. Frontiers in Psychology, 11, 1808.
https://doi.org/10.3389/fpsyg.2020.01808
Goñi, A., Ruiz de Azúa, S., & Rodríguez, A. (2006). Cuestionario de Autoconcepto Físico (CAF). EOS.
Lohbeck, A., von Keitz, P., Hohmann, A., & Daseking, M. (2021). Children’s physical self-concept,
motivation, and physical performance. Frontiers in Psychology, 12, 669936.
https://doi.org/10.3389/fpsyg.2021.669936
Muntaner-Mas, A., Mazzoli, E., Abbott, G., Mavilidi, M. F., & Galmes-Panades, A. M. (2022). Do
physical fitness and executive function mediate the relationship between physical activity and
academic achievement? An examination using structural equation modelling. Children, 9(6),
823. https://doi.org/10.3390/children9060823
Núñez, J. L., León, J., Valero-Valenzuela, A., & Conte, L. (2021). Influence of physical self-concept
and motivational processes on moderate-to-vigorous physical activity of adolescents.
Frontiers in Psychology, 12, 685612. https://doi.org/10.3389/fpsyg.2021.685612
pág. 1859
Sevil-Serrano, J., Aibar, A., Abós, Á., Generelo, E., & García-González, L. (2020). Basic psychological
needs, physical self-concept, and physical activity among adolescents. Frontiers in
Psychology, 11, 491. https://doi.org/10.3389/fpsyg.2020.00491
Tarka, P. (2018). An overview of structural equation modeling: Its beginnings, historical development,
usefulness and controversies in the social sciences. Quality & Quantity, 52(1), 313–354.
https://doi.org/10.1007/s11135-017-0469-8
Ubago-Jiménez, J. L., Corral-Robles, S., Ortega-Martín, J. L., & Melguizo-Ibáñez, E. (2023). The
interplay between academic performance, emotional intelligence, and self-concept as
predictors of violent behavior in higher education: A multi-group structural equation
modeling. Frontiers in Psychology, 14, 1124712.
https://doi.org/10.3389/fpsyg.2023.1124712
Villodres, G. C., Salvador-Pérez, F., Chacón-Cuberos, R., & Muros, J. J. (2023). Lifestyle behaviours,
self-esteem and academic performance in primary education students: A structural equation
model according to sex and school type. Children, 10(11), 1769.
https://doi.org/10.3390/children10111769
Vollmer, J., Lohmann, J., & Giess-Stüber, P. (2021). Socioeconomic status and global physical self-
concept of adolescents: A multilevel structural equation modeling approach. German Journal
of Exercise and Sport Research, 51, 160–169. https://doi.org/10.1007/s12662-020-00701-7
Zurita-Ortega, F., Alonso-Vargas, J. M., Puertas-Molero, P., González-Valero, G., Ubago-Jiménez, J.
L., & Melguizo-Ibáñez, E. (2023). Levels of physical activity, family functioning and self-
concept in elementary and high school education students: A structural equation model.
Children, 10(1), 163. https://doi.org/10.3390/children10010163
Sevil-Serrano, J., Aibar, A., Abós, Á., Generelo, E., & García-González, L. (2020). Basic psychological
needs, physical self-concept, and physical activity among adolescents. Frontiers in
Psychology, 11, 491. https://doi.org/10.3389/fpsyg.2020.00491
Tarka, P. (2018). An overview of structural equation modeling: Its beginnings, historical development,
usefulness and controversies in the social sciences. Quality & Quantity, 52(1), 313–354.
https://doi.org/10.1007/s11135-017-0469-8
Ubago-Jiménez, J. L., Corral-Robles, S., Ortega-Martín, J. L., & Melguizo-Ibáñez, E. (2023). The
interplay between academic performance, emotional intelligence, and self-concept as
predictors of violent behavior in higher education: A multi-group structural equation
modeling. Frontiers in Psychology, 14, 1124712.
https://doi.org/10.3389/fpsyg.2023.1124712
Villodres, G. C., Salvador-Pérez, F., Chacón-Cuberos, R., & Muros, J. J. (2023). Lifestyle behaviours,
self-esteem and academic performance in primary education students: A structural equation
model according to sex and school type. Children, 10(11), 1769.
https://doi.org/10.3390/children10111769
Vollmer, J., Lohmann, J., & Giess-Stüber, P. (2021). Socioeconomic status and global physical self-
concept of adolescents: A multilevel structural equation modeling approach. German Journal
of Exercise and Sport Research, 51, 160–169. https://doi.org/10.1007/s12662-020-00701-7
Zurita-Ortega, F., Alonso-Vargas, J. M., Puertas-Molero, P., González-Valero, G., Ubago-Jiménez, J.
L., & Melguizo-Ibáñez, E. (2023). Levels of physical activity, family functioning and self-
concept in elementary and high school education students: A structural equation model.
Children, 10(1), 163. https://doi.org/10.3390/children10010163