Modeling the Association Between Virtual Reality-Based Sports Facility Safety Analysis and Motivational Climate, Behavioral Outcomes, and Continuance Intention to Use in Sport Science Higher Education

Abstract

This cross-sectional study aimed to examine the associations among Physical Activity and Sport Sciences students’ self-reported motivational climate, perceived usefulness, behavioral intention, and continuance intention to use Virtual Reality (VR), as assessed immediately after a single common VR session on sports facility safety analysis. It also sought to propose and validate a theoretical model predicting students’ self-reported continued intention to use VR within a single-group design. This session was conducted with 445 undergraduate students who participated in an immersive VR experience focused on evaluating the safety of sports facilities. Immediately after the activity, participants completed questionnaires assessing usability, continuance intention to use, behavioral intention, and mastery- and performance-oriented motivational climates. Results revealed high levels of perceived usefulness, behavioral intention, and intention to use VR. The mastery-oriented motivational climate scored significantly higher than the performance-oriented climate. Structural equation modeling showed that perceived usefulness and a mastery-oriented climate were positively associated with behavioral intention, while a performance-oriented climate showed a negative association. Behavioral intention was the strongest predictor of future self-reported intention to use VR. Gender differences were also observed, associated with perceptions of mastery climate and perceived usefulness. Overall, the findings support VR as a promising educational tool for sports facility safety training and innovative learning in higher education.

1. Introduction

New technologies have profoundly transformed today’s society, and the use of technological devices is ubiquitous in many everyday activities, playing a fundamental role in shaping cultural identities and practices around the world (Alsaleh, 2024). This phenomenon is mainly due to the integration of technology into development and implementation processes in several fields, such as sports (F. Wang & Wei, 2020).

Fast-paced technological advances have also permeated the field of education, where implementation has grown rapidly over the last decade (Sosa Alonso & Bethencourt Aguilar, 2019) due to technology’s ability to improve the teaching–learning process (Calabuig-Moreno et al., 2020; Qian & Luo, 2026). Furthermore, clear benefits of its use and impact have been demonstrated in various areas, such as the quality of learning, improved academic results, student participation and engagement, the development of meaningful and personalized learning experiences, increased accessibility and flexibility in education, and the development of 21st-century skills, such as critical thinking, collaboration, communication, and digital literacy (Akintayo et al., 2024; Alotaibi, 2024; Ghory & Ghafory, 2021; Mariscal et al., 2020; Ma et al., 2024; Zou et al., 2025).

One such technology is Virtual Reality (VR), understood as an immersive technology that creates simulated environments in which users can interact (Schaffernak et al., 2020). It accurately captures the five senses of the human body by constructing a 1:1 three-dimensional model (real world-virtual world), which promotes an immersive experience and user interaction with virtual scenes, enhancing the authenticity of the experience (C. Li & Li, 2020). It can be accessed through various screens, such as desktop computers, head-mounted displays, or automatic cave-type virtual environments (Makransky & Petersen, 2021).

Therefore, VR is rapidly emerging as a highly effective technology for training diverse skills and creating immersive, interactive learning environments. VR enables realistic simulations and digital competitions, supporting both cognitive and motor skill development across education, industry, and sports (Lampropoulos & Kinshuk, 2024; Molina-Carmona et al., 2018).

This potential makes it important to consider the association between new technological devices and tools in the field of education as a fundamental key to the development of a modern, novel and innovative teaching approach (Pérez-Muñoz et al., 2024). However, the successful implementation of technology in education depends on a number of factors that must be taken into account, such as teacher competence and training, and the available infrastructure and resources (Akintayo et al., 2024; Ma et al., 2024).

1.1. Virtual Reality in Education

As with technological innovations in general, we are currently witnessing a growing adoption of VR in education, especially in the field of sports (Akhiruyanto & Yudhistira, 2024; Gómez-García et al., 2019). In the contemporary social context, where technological advances and their associated benefits are omnipresent, it is imperative that educational institutions do not remain on the sidelines of this phenomenon (Gómez-García et al., 2019; Pérez-Muñoz et al., 2024).

In this regard, the association between VR and teaching methodologies has been described as positive, as it is reported to provide innovative and engaging learning experiences for students (Akhiruyanto & Yudhistira, 2024; Pérez-Muñoz et al., 2024). Some of the benefits demonstrated in the educational context are the creation of more immersive and adapted learning environments, facilitation of the integration of diverse students, and improvement of the overall classroom atmosphere (Geisen et al., 2023; Klochko & Fedorets, 2022; Kuleva, 2024). On this point, VR allows interaction with real-world content in a potentially immersive environment, improving the ability to control and interact in safe environments with objects that are not found in students’ everyday experience, a fact that has been linked to their involvement, motivation, concentration, acquisition of skills, and retention of knowledge (Abich et al., 2021; Bores-García et al., 2024; Kuleva, 2024; P. Wang et al., 2018).

However, despite the many benefits of using VR in education, it is not without its limitations. On the one hand, some critics point to its low accessibility due to the significant costs of implementing and maintaining these VR devices in an educational context (Soltani & Morice, 2020), although the situation is changing as the technology becomes increasingly affordable and easy to use, a fact that is encouraging the proliferation of studies related to its use, mainly in the field of education and training (Stefan et al., 2023). On the other hand, some teachers and students are uncomfortable using it (Kuleva, 2024) and consider it complex to use, which is why they do not adopt this technology (Shonfeld & Greenstein, 2021), encountering challenges such as a lack of understanding of science and technology on the part of older teachers, which leads to a lack of skills and confidence in the use of technology-based learning strategies (Akhiruyanto & Yudhistira, 2024).

In addition, some teachers express concern about the use of VR in education, believing that it may limit physical interactions between students and direct experiences (Amin & Govilkar, 2015). In this regard, it should be noted that although virtual technology can facilitate learning, it cannot replace the experience and motor skills acquired through physical activity, as effective learning requires a combination of practical experience, direct feedback from teachers and deep reflection (Bolatuly-Omarov et al., 2025; Levac et al., 2019). Therefore, the effectiveness of training does not lie in the use of virtual reality itself, but in how it is designed to encourage active interaction aligned with learning objectives. When used properly, VR has particularly been associated with the acquisition of procedural knowledge through embodied learning experiences (Abich et al., 2021).

1.1.1. Virtual Reality as a Tool for Sports Facilities Safety Education

Traditional safety training has proven difficult to transfer to the real world (Gao et al., 2019), as it has mostly been conducted in person in the classroom, with some previous studies demonstrating a low level of engagement among participants and an inability to optimally achieve the acquisition of safety knowledge and the improvement of attitudes, beliefs, behaviors, and health (X. Li et al., 2018; Ricci et al., 2016). In this context, VR technology has been proposed as capable of safely simulating dangerous situations that would otherwise be inaccessible, potentially offering advantages over other traditional training methods (Stefan et al., 2023).

In the field of sports, there are safety requirements for sports facilities and equipment that must be identified and analyzed for accident prevention (C. Li & Li, 2020). In this regard, VR has been proposed as offering an important opportunity to improve the effectiveness of safety training and related aspects, due to its higher level of presence, its ability to fail safely, and its ability to present various scenarios that are difficult to replicate in the real world due to financial constraints or safety issues (Stefan et al., 2023), which some authors associate with an improvement in safety management capabilities (Hui & Yan, 2024), although this association has not been directly tested in the present study. In this area, the use of VR for learning complex spatial arrangements has proven beneficial for students, who show greater understanding, greater enthusiasm for the subject matter, and a desire to continue using this type of technology during their learning, improving their spatial abilities (Fogarty et al., 2018; Molina-Carmona et al., 2018; Zhang et al., 2022).

However, research such as that by Leder (Leder et al., 2019) did not find significantly worse results in terms of changes in risk perception, learning outcomes, or decision-making in safety training using PowerPoint with vivid cinematic scenes versus immersive virtual reality, the latter being much more expensive. In this respect, review studies related to safety training tend to focus on a specific area, making it difficult to identify benefits that can be extrapolated to other contexts (Stefan et al., 2023), as these reviews are very specific in terms of scope and task, limiting the ability to establish meaningful connections between the effectiveness of training in different fields (Abich et al., 2021). Therefore, despite the growing interest in researching new trends in VR in the field of sport, there is still a need for research topics to improve knowledge, as there is not much research available (Akhiruyanto & Yudhistira, 2024; Calabuig-Moreno et al., 2020; Pérez-Muñoz et al., 2024).

1.1.2. Technology Evaluation Using Technology Acceptance Models

In the educational field, several theories have been proposed regarding expectations, intentions, and preferences towards new technologies. Various models and theories aim to understand the perception of information technology users and the acceptance of information systems. Among these are the Technology Acceptance Model (TAM) (Davis, 1989), the Theory of Planned Behavior (TPB) (Ajzen, 1991), the Unified Theory of Acceptance and Use of Technology (UTAUT) (Venkatesh et al., 2003), and the combined TAM–TPB model (C-TAM–TPB) (Taylor & Todd, 1995). According to TAM (Davis, 1989), technology acceptance is influenced by beliefs about the consequences of its use. One of the factors influencing behavioral intention is usefulness, a recurring element in various theories (TAM, TPB, C-TAM–TPB, and UTAUT) (Venkatesh et al., 2003). Additionally, to determine if a technology will be used optimally, it is necessary to identify different external variables that can influence these factors, such as usefulness or behavioral intention (X. Huang et al., 2023; Yong Varela et al., 2010). In the present study, the “Usefulness” label refers specifically to the five (positive) retained EUS/SUS items described in Section 2.1, which chiefly assess ease of use, integration into the task, learnability, confidence, and willingness to use the device frequently, and should therefore not be automatically equated with TAM-style perceived usefulness or with instrumental learning value. The same applies to the Behavioral Intention scale, whose items address satisfaction, perceived motivation, and perceived competence acquisition rather than a general behavioral-intention construct as defined in TAM or TPB; the interpretations offered in the Discussion have been framed to reflect what these administered items actually measure.

One of the most widely used theories to explain student motivation and academic success is the achievement goal theory, which analyses how students adopt different types of goals: mastery motivational or performance motivational (Elliot, 2005). In this regard, the motivational climate generated by the teacher in the classroom (mastery motivational climate or performance motivational climate) is a precursor to students’ goal orientations, which have a powerful influence on motivation, basic psychological needs, and performance (Brisimis et al., 2022).

This hypothetical research model has been developed to illustrate the association between the motivational climate (mastery motivational climate or performance motivational climate) and usefulness, continuance intention to use and behavioral intention, and how these factors relate to one another, also considering the influence of gender (Figure 1).

Figure 1. Research hypothetical model.

In light of the above, the objectives of this research aimed to: (1) examine the associations among the motivational climate, perceived usefulness, behavioral intention, and continuance intention to use reported by sport science degree students after completing an immersive VR-based sports facility safety-analysis activity; and (2) propose and validate a theoretical model to predict students’ self-reported continued intention to use VR.

Based on the theoretical framework described above, the following research hypotheses were formulated:

H1. 

The mastery and performance motivational climate, perceived usefulness, and behavioral intention reported by students after the VR-based safety-analysis activity are significantly associated with one another and with continuance intention to use VR.

H2. 

A theoretical model integrating motivational climate, perceived usefulness, behavioral intention and gender will significantly predict sport science degree students’ self-reported continuance intention to use VR.

2. Materials and Methods

2.1. Study Design

The study employed a descriptive cross-sectional design, following the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines (Cuschieri, 2019). A VR session was conducted with undergraduate students enrolled in the course “Sports Equipment and Facilities” of the Degree in Physical Activity and Sport Sciences at two Spanish universities located in the regions of Murcia and Andalusia. Sampling was non-probabilistic (convenience-based), comprising all students enrolled in the course who attended the session. Potential university- and cohort-level clustering effects were not modeled. During the session, students engaged with content related to the safety of various sports facilities. Following the session, participants self-completed a questionnaire (Supplementary File S1—Annex VR alumni) that included sociodemographic questions and several scales: from the Use Behavior Scale, the Behavioral Intention dimension (Cózar-Gutiérrez et al., 2015), from the Use Behavior Scale, the Intention to Use dimension (also called Continuance Intention to Use) (Nolasco de Almeida Mello & Cabero Almenara, 2020), the Usefulness dimension of the System Usability Scale (Hedlefs-Aguilar & Garza-Villegas, 2016), and the Mastery and Performance Motivational Climate dimensions from the Motivational Climate in Education Scale (Granero-Gallegos & Carrasco-Poyatos, 2020).

For clarity, the three use-related constructs are defined and operationalized as follows: (a) Behavioral intention refers to students’ self-reported disposition to engage with the VR activity and its content, assessed with six items (e.g., perceived willingness to use the system) from Cózar-Gutiérrez et al. (2015); (b) continuance intention to use refers to students’ self-reported willingness to use VR again in future learning contexts, assessed with six items adapted from Nolasco de Almeida Mello and Cabero Almenara (2020); (c) usefulness with ten items (five positive and five negative) from the EUS (Spanish) or SUS (English) scale by Hedlefs-Aguilar and Garza-Villegas, which aims to measure the perceived usability of a system, website, or interface. In this case, it refers to the use of the VR design presented to the students (e.g., the users’ perceived experience with the proposed VR design or any usability problems identified) (Hedlefs-Aguilar & Garza-Villegas, 2016).

The label “Usefulness” is used consistently throughout this manuscript (Introduction, Methods, Results, Discussion, Conclusions) to refer specifically to the five retained positive items of the EUS/SUS scale described above, and not to “usability” in the broader sense of the construct or to any other sub-dimension of the cited scale. Likewise, the two constructs actually assessed by the “Use Behavior Scale” instrument (Cózar-Gutiérrez et al., 2015; Nolasco de Almeida Mello & Cabero Almenara, 2020)—behavioral intention and continuance intention to use (Hedlefs-Aguilar & Garza-Villegas, 2016)—are the only constructs referred to by those names in this manuscript; the instrument’s original name (“Use Behavior Scale”) is retained only when citing the source scale itself, and is not used elsewhere to describe an actual, observed use-behavior outcome, which was not measured in this study.

The study received approval from the institutional Ethics Committee (code CE072207) before it began, and all procedures complied with the principles of the Declaration of Helsinki. Before data collection, participating students signed an informed consent form that detailed the purpose of the study, how their data would be processed, and the measures taken to ensure confidentiality. The VR activity itself was a mandatory, graded component of the course (contributing 1 of 10 points to the final grade) and was completed by all enrolled students independently of their decision to participate in the research. Participation in the research (i.e., completion of the study questionnaire and inclusion of responses in the dataset) was voluntary and separate from course credit; students could complete the graded VR activity without consenting to have their questionnaire data used for research purposes, and no academic consequences resulted from declining to participate in the study. In this research, all data were collected immediately after the single VR activity.

2.2. Sampling Method and Sample Size

The sample size was determined following the methodology described in previous research based on the standard deviation (SD) (Bhalerao & Kadam, 2010). RStudio version 3.15.0 was used for the calculations. Sample size and statistical power were estimated using the standard deviation of mastery climate (SD = 0.76) reported in a sample with similar characteristics (López-García et al., 2022). Considering an estimated error (d) of 0.07, a 95% confidence interval (95% CI), a statistical power of 95%, and a significance level of 0.05, the minimum required sample size was 440 participants. This calculation was based on estimating a population mean for the mastery-climate construct and was not derived specifically from power requirements for the proposed structural equation model; no a priori power analysis specific to the proposed structural equation model was conducted; this is acknowledged as a limitation.

The study involved 445 students (mean age = 22.22 ± 3.32 years), including 372 males (83.60%) and 73 females (16.40%), from two Spanish universities located in the Regions of Murcia and Andalusia. Of these, Murcia comprised (n = 253; 56.73%) and Andalusia comprised (n = 192; 43.27%). All were students enrolled in the Sports Equipment and Facilities course taught in the Degree in Physical Activity and Sport Sciences during the 2022–2023 and 2023–2024 academic years. Inclusion criteria required students to (a) be enrolled in the Sports Equipment and Facilities subject at either university during any of those academic years, and (b) attend the project session. Of the students enrolled in the course across both universities and academic years (N = 452), 445 attended the session and completed the questionnaire, corresponding to a participation rate of 98.5%; seven students did not attend the session and were therefore not included. Sample distribution by university is reported above. The distribution by academic-year cohort within each university was: Murcia 2022–2023 (n = 151), Murcia 2023–2024 (n = 102), Andalusia 2022–2023 (n = 103), and Andalusia 2023–2024 (n = 89). Formal comparison of university- or cohort-level differences was beyond the scope of the present analysis and is suggested for future research.

2.3. Procedures

The practical case required designing sports facilities in VR (Figure 2). A meeting was held with the UCAM Multimedia Laboratory (LabM-UCAM) to explore how to adapt the project to an immersive environment for the third-year Sports Equipment and Facilities subject of the Degree in Physical Activity and Sport Sciences at both universities.

Figure 2. Sports Facilities by VR.

For the design of the sports facilities, information from the NIDE standard was used (including a padel court, a football field, and sports halls with markings for handball/indoor football, basketball, volleyball, and badminton). Dichotomous checklists (yes/no) were created to assess compliance with the different items (Supplementary File S2—Sport facilities) (A. Gallardo-Guerrero et al., 2023). The virtual sports facilities developed ad hoc by the LabM-UCAM were created using the Spatial platform. In the various virtual environments, the errors indicated by the teaching staff were incorporated, and students were required to identify them in the evaluation checklists. Once the environments and assessment tools were designed, the implementation of the immersive proposal was carried out.

This immersive experience of VR was conducted in pairs in a designated classroom. While one student explored the virtual environment, the other monitored on a computer screen what their partner was seeing and, using the prepared checklists, asked questions to assess the safety of the sports facilities (Figure 3). The activity lasted approximately 60 to 90 min, followed by a 10 min debriefing session. This practice accounted for 1 out of 10 points in the final course grade.

Figure 3. Students participating in an immersive VR experience.

At the end of the class, students self-completed a questionnaire in Google Forms (in Spanish) regarding their VR experience. The questionnaire included sociodemographic questions (age, gender, and home university) as well as several scales included in Supplementary File S1—Annex VR alumni: the Use Behavior Scale (Behavioral Intention dimension), consisting of six items rated on a 1–5 Likert scale with a reliability of α = 0.81; the Use Behavior Scale (Continuance Intention to Use dimension), comprising six items rated on a 1–5 Likert scale with α = 0.93; Usefulness dimension of the System Usability Scale, validated in Spanish, with ten items on a 1–5 Likert scale and α = 0.92; and finally, the Mastery (4 items, α = 0.88) and Performance (3 items, α = 0.87) motivational climate dimensions of the Motivational Climate in Education Scale. The reliability coefficients reported in this paragraph (α = 0.81–0.93) correspond to the values reported in the original validation studies of each instrument (Cózar-Gutiérrez et al., 2015; Nolasco de Almeida Mello & Cabero Almenara, 2020; Hedlefs-Aguilar & Garza-Villegas, 2016; Granero-Gallegos & Carrasco-Poyatos, 2020) and to the full original item sets prior to the CFA-driven item reduction; the reliability values recalculated in the present sample, for the final retained items (Supplementary File S3—Final Annex VR alumni), are reported in Table 1 (α = 0.74–0.87) and should be used as the reference values for this study.

2.4. Statistical Analysis

The analysis was carried out in two stages. For Objective 1, descriptive statistics (means and standard deviations) of the various constructs were calculated for data analysis.

For Objective 2, the process began with a confirmatory factor analysis (CFA) on all selected dimensions to ensure the reliability and validity of each instrument used in this study. Next, the goodness of fit for the ordinal variables was assessed using measures of absolute fit: Chi-square p-value (χ² p-value) and Root Mean Square Error of Approximation (RMSEA); incremental fit measures: Tucker–Lewis Index (TLI), Comparative Fit Index (CFI), Normed Fit Index (NFI), and Goodness of Fit Index (GFI); and measures of parsimony fit: Normalized Chi-square. An acceptable fit was indicated by a χ² p-value > 0.05, an RMSEA ≤ 0.06, an SRMR < 0.08, TLI scores ≥ 0.90, CFI scores ≥ 0.90, NFI scores ≥ 0.90, GFI scores ≥ 0.90, and a normalized Chi-square value between 2 and 3 (Hair et al., 2013).

To evaluate the internal validity of the items for each latent variable, factor loading values needed to be above 0.5 to be considered significant (Hair et al., 2013). The reliability of the questionnaire was assessed using convergent validity and discriminant validity. Convergent validity of the scale items was assessed using three criteria: Cronbach’s alpha, composite reliability (CR), and average variance extracted (AVE). An acceptable fit was indicated by Cronbach’s alpha > 0.7, CR > 0.7, and AVE > 0.5 (Hair et al., 2013). Discriminant validity was evaluated using the Fornell–Larcker criterion (Hair et al., 2016). Fornell and Larcker (Fornell & Larcker, 1981) proposed that “the square roots of the AVE for each latent variable should be higher than the correlations with other latent variables” to establish discriminant validity.

Structural Equation Modeling (SEM) was performed to provide a more precise measurement of the structural relationships among all the constructs. The estimator used was diagonally weighted least squares (DWLS) as it included a categorical variable (gender). The items were treated as ordinal variables except for gender, which was treated as a categorical variable and coded as 0 = male, 1 = female. The analysis was conducted using R software version 3.6.0 (Copyright © 2019, R Foundation for Statistical Computing) with the Lavaan package. The initial hypothesized measurement model (CFA) and the final structural model (SEM) are reported separately; the modification of the initial model (item and path removal) was performed on the same sample used for hypothesis testing, without an independent validation or cross-validation subsample. This data-driven, non-cross-validated model modification is acknowledged as a methodological limitation in the Discussion.

During the model specification and refinement process, non-significant structural paths were evaluated and iteratively removed from the initial hypothesized model (Figure 1) to achieve a parsimonious and well-fitting final structural model (Figure 4). Specifically, the following structural paths were removed: Performance Motivational Climate → Usefulness (removed; not significant, β = −0.269, p = 0.068): Considered conceptually distinct from intention constructs because achievement-goal theory links (Elliot, 2005) performance climate primarily to competitive/normative outcomes rather than to instrumental-value judgments such as perceived usefulness (Brisimis et al., 2022); Gender → Performance Motivational Climate (removed; not significant, β = −0.145, p = 0.112): Removed because achievement-goal climate perceptions were hypothesized to be driven primarily by the educational/learning environment rather than direct baseline gender differences (X. Huang et al., 2023; Yong Varela et al., 2010); Gender → Behavioral intention (removed; not significant, β = −0.169, p = 0.084): Dropped as gender did not exert a direct theoretical or statistically significant association with technical usability evaluations, operating instead indirectly through motivational climate factors (Venkatesh et al., 2003); and Usefulness → Continuance intention to Use (removed; not significant, β = 0.765, p = 0.059): Removed as a direct path to eliminate structural redundancy and conceptual overlap, allowing behavioral intention and usability to directly explain variance in final technology adoption/intended use (Ajzen, 1991; Taylor & Todd, 1995). Each path was assessed individually and removed based on goodness-of-fit mismatch and statistical non-significance while re-examining theoretical plausibility before being dropped.

3. Results

3.1. Confirmatory Factor Analysis

A confirmatory factor analysis (CFA) was conducted to examine the hypothesized model shown in Figure 1 (initial model). The initial analysis of the hypothetical model did not indicate an acceptable goodness of fit, with a χ² p-value > 0.05, RMSEA = 0.082, SRMR = 0.096, TLI = 0.804, CFI = 0.825, NFI = 0.780, GFI = 0.800, and a normalized Chi-square of 2.85. The behavioral intention dimension was initially composed of six items. One item for behavioral intention and five items for usefulness showed factor loadings below 0.5, and some paths were not significant. Consequently, one item for behavioral intention (final retained: five items) and five items for usefulness (out of ten originally administered; final retained: five “positive” items) (Supplementary File S1—Annex VR alumni). The corrected “ten items” to “five items”, the number consistent with the 10-total/5-removed/5-retained figures stated later in this paragraph and in Table 1, along with the non-significant paths, were removed through an iterative process. Because this modification was performed on the same sample subsequently used to test the structural model, without cross-validation, the resulting measurement model should be considered partly exploratory; each removed path was also re-examined for theoretical plausibility before being dropped. Each path was assessed and eliminated individually based on the highest degree of mismatch until all remaining paths were significant. For clarity: the Behavioral Intention scale originally comprised six items (one removed; five retained) and the Usefulness scale originally comprised ten items (five negative removed; five positive retained), consistent with Table 1 and the final questionnaire called Supplementary File S3—Final Annex VR alumni excluding items 13 and 19–23 (which were included in Supplementary File S1). The Chi-square statistic and degrees of freedom for this initial CFA model were χ² (340) = 969.00, p < 0.001.

Table 1. Validity and reliability analysis.

Construct—Items	M	SD	Loading	α > 0.7	CR > 0.7	AVE > 0.5
Mastery motivational climate
Item 1	4.47	0.73	0.84	0.74	0.74	0.50
Item 2	4.44	0.79	0.80
Item 3	4.27	0.93	0.65
Item 4	4.50	0.71	0.73
Performance motivational climate
Item 1	2.08	1.48	0.96	0.79	0.80	0.59
Item 2	2.24	1.53	0.90
Item 3	3.06	1.50	0.56
Behavioral intention
Item 1	4.37	0.74	0.78	0.83	0.83	0.50
Item 2	4.40	0.75	0.79
Item 3	4.54	0.69	0.80
Item 4	4.28	0.77	0.78
Item 5	4.67	0.64	0.80
Usefulness
Item 1	4.23	0.91	0.88	0.81	0.81	0.50
Item 2	4.06	0.90	0.69
Item 3	4.19	0.79	0.80
Item 4	4.08	0.88	0.63
Item 5	4.22	0.92	0.72
Continuance Intention to use
Item 1	4.43	0.85	0.86	0.87	0.88	0.55
Item 2	4.22	1.01	0.87
Item 3	4.37	0.82	0.86
Item 4	4.36	0.90	0.79
Item 5	4.37	0.87	0.89
Item 6	4.52	0.80	0.81

Mastery motivational climate: male M = 4.46, SD = 0.58; female M = 4.22, SD = 0.69. Performance motivational climate: male M = 2.47, SD = 1.15; female M = 2.41, SD = 1.18. Behavioral intention: male M = 4.49, SD = 0.55; female M = 4.25, SD = 0.64. Usefulness: male M = 4.21, SD = 0.68; female M = 3.90, SD = 0.77. Continuance intention to use: male M = 4.42, SD = 0–68; female M = 4.18, SD = 0.76.

The final version of the measurement model is presented in Figure 4. All item factor loadings were above 0.56, indicating significant internal validity of the latent variables (Table 1, which reports only the items retained after the removal described above). The CFA demonstrated the measurement model’s goodness of fit with the following indices: χ² p-value > 0.05, RMSEA = 0.065, SRMR = 0.056, TLI = 0.901, CFI = 0.914, NFI = 0.874, GFI = 0.879, and a normalized Chi-square of 2.29. Although RMSEA (0.065), NFI (0.874), and GFI (0.879) did not reach the thresholds specified a priori, while TLI (0.901) and CFI (0.914) did, the overall goodness of fit of the measurement model is best described as mixed or marginal rather than unambiguously acceptable; this pattern is not uncommon in models with a relatively large number of estimated parameters and comparisons, but it is reported here as a qualified rather than a fully confirmed fit (Hair et al., 2013). These fit indices correspond specifically to the CFA measurement model; the distinct, higher fit indices correspond to the subsequent structural (SEM) model, in which the latent variables are represented by their retained indicators and connected through the hypothesized structural paths. Reporting both sets of indices separately, as intended, avoids the apparent inconsistency raised in peer review. The Chi-square statistic and degrees of freedom for this final measurement model were χ² (220) = 504.162, p < 0.001.

Figure 4. Final version of the model. *: p-value < 0.05.

3.2. Construct Reliability and Validity Analysis

Based on the convergent validity and discriminant validity, the reliability of the questionnaire was evaluated. Three criteria were used to assess the convergent validity of the scale items, namely Cronbach’s alpha, composite reliability (CR), and average variance extracted (AVE) (Table 1). All three conditions of convergent validity were met, reporting greater than the minimum loading criterion in each case: Cronbach’s alpha = 0.74–0.87, CR = 0.74–0.88; AVE = 0.5–0.59.

To determine whether the factors of one variable are different from those of another variable, discriminant construct validity was assessed. To assess discriminant validity, the Fornell–Larcker cross-loading criterion was used (Hair et al., 2016). Square roots of the AVE were greater than the absolute values of the correlation coefficients, so the discriminant validity of each construct was considered acceptable (Hair et al., 2013) (Table 2). Notably, the correlation between behavioral intention and continuance intention to use (r = 0.68) is high relative to the square roots of AVE for both constructs, which is consistent with the conceptual overlap between these two variables.

Table 2. Discriminant validity—Fornell-Larcker criterion.

Construct	1	2	3	4	5
Mastery motivational climate	0.71
Performance motivational climate	−0.169	0.77
Behavioral intention	0.60 *	−0.19	0.71
Usefulness	0.57 *	−0.02	0.68 *	0.71
Continuance intention to use	0.53 *	−0.23	0.68 *	0.66 *	0.74

Values on the diagonal (bolded) represent the square root of AVE, while off-diagonals represent correlations. *: p-value < 0.05.

3.3. Model Fit Analysis

To examine the relation between the variables, the structural model (SEM) built on the retained measurement model was tested. Structural relationships among the variables are shown in Figure 4. The fit indices indicated that the structural model provided an adequate representation of the data (χ² p-value < 0.05, RMSEA = 0.065, SRMR < 0.05, TLI = 0.989, CFI = 0.989, NFI = 0.983, GFI = 0.984, Normalized Chi-square = 2.8). These values describe the fit of the full structural equation model (measurement and structural components estimated jointly in a single Lavaan model, as detailed below), and are therefore expected to differ from the measurement-model indices reported. The Chi-square statistic and degrees of freedom for the structural model were χ² (245) = 693.202, p < 0.01. The structural model was specified and estimated as a single, full latent-variable structural equation model in Lavaan (i.e., the measurement and structural components were estimated jointly within one model syntax).

3.4. Path Analysis

Behavioral intention (β = 0.941, p < 0.001) was the strongest predictor of, and showed a strong positive association with, continuance intention to use; given the magnitude of this coefficient, the conceptual and empirical overlap between the two constructs should be considered when interpreting this result. In addition, usefulness (β = 0.586, p < 0.001), mastery motivational climate perception (β = 0.367, p < 0.001) and performance motivational climate perception (β = −0.269, p < 0.001) were significantly associated with behavioral intention (Figure 4 and Table 3). Besides, mastery motivational climate perception (β = 0.737, p < 0.001) and gender (β = −0.169, p < 0.001) were associated with usefulness positively and negatively, respectively. Finally, it is shown that gender was significantly associated with mastery motivational climate perception; the standardized coefficient is reported as β = 0.169 (z = 3.898, p < 0.001). Predictive accuracy of the structural models was determined using the coefficient of determination (R² value). From the structural model, it is observed that the R² values are 0.886, 0.858 and 0.530 for intended use, behavioral intention and usefulness, respectively. This implies that the variance of these constructs is explained by the proposed construct by 88.6%, 85.8% and 53%, which indicates moderate to high explanatory power (Hair et al., 2016). Gender explains 2.9% (R² = 0.029) of mastery motivational climate perception, which is a low level of explained variance (Hair et al., 2016).

Table 3. Proposed model evaluation results.

Relationship	Std β	Std Error	z Value	R2
Continuance intention to use
Behavioral intention	0.941 *	0.025	41.615 *	0.886
Behavioral intention
Usefulness	0.586 *	0.038	13.702 *	0.858
Mastery motivational climate	0.367 *	0.039	8.716 *
Performance motivational climate	−0.269 *	0.017	−13.015 *
Usefulness
Mastery motivational climate	0.737 *	0.032	24.053 *	0.530
Gender	−0.169 *	0.100	−4.040 *
Mastery motivational climate
Gender	−0.169 *	0.097	−3.898 *	0.029

Std error: standard error, Std β: beta value standardized, *: p-value < 0.05.

4. Discussion

The primary objective of this research was to examine the associations among the motivational climate, perceived usefulness, behavioral intention, and continuance intention to use reported by Physical Activity and Sport Sciences students after completing an immersive VR-based activity in evaluating the safety of sports facilities, where all dimensions (Mastery Motivational Climate, Behavioral Intention, Usefulness and Continuance Intention to Use) scored above 4, except for Performance Motivational Climate, which had values ranging from 2.08 to 3.06.

These results are consistent with previous findings regarding the perceived usefulness in the adoption of immersive technologies (Bower & Jong, 2020; Miguélez-Juan et al., 2019; Molina-Carmona et al., 2018; W. Huang et al., 2021; X. Huang et al., 2023). Likewise, the significant association of the mastery motivational climate with usefulness and behavioral intention suggests that a learning environment focused on personal improvement and cooperation is related to a predisposition toward the continued use of VR (Zhang et al., 2022; Zhao et al., 2023), and students reported feeling a stronger understanding of concepts after the experience (Fogarty et al., 2018), although this was not directly measured with a knowledge or performance test in the present study. In contrast, the performance motivational climate showed a negative association, in line with research linking competitive environments to lower satisfaction and engagement (Ames, 1992; Zhao et al., 2023).

Regarding the first objective, the results show that students self-reported a highly positive and motivating learning experience after the VR-based safety-analysis activity (Decker & Merkt, 2025; Thomann et al., 2024). The high scores in the dimensions of continuance intention to use and behavioral intention are consistent with the idea that immersive environments can support active and meaningful learning, consistent with the principles of Self-Determination Theory (Deci & Ryan, 2000). These findings are in line with recent studies highlighting the potential of VR to increase student engagement and satisfaction through the simulation of realistic and safe contexts (Keller et al., 2025; Puiu & Udriștioiu, 2024; Radianti et al., 2020; Halabi et al., 2022; Sviridova et al., 2023), where factors such as learning value, intrinsic motivation, and personal innovativeness have also been identified as predictors of the continuance intention to use VR environments (Parong & Mayer, 2021, 2018; Solmaz et al., 2024; Tursø-Finnich et al., 2023). In this context, perceived usefulness and enjoyment have been described as key determinants for the sustained adoption of educational technologies (Fernández-Vázquez et al., 2024; Mohsen & Alangari, 2024). Continuance intention to use VR has been reported to increase when users believe that VR provides value for learning or performance, making perceived usefulness one of the most consistent predictors of intention to use (Alturkustani et al., 2025; Benar et al., 2024; W. Huang et al., 2021; Y. Huang & Soebbing, 2022).

Likewise, the mastery motivational climate was positively associated with behavioral intention, suggesting that learning environments focused on personal progress and collaboration are related to technological acceptance (Brown et al., 2002). In the present sample, students reported high scores across the perception and intention measures collected after the VR activity, which is consistent with, although not proof of, the idea that VR can successfully address the goals of digitalization in education (Rojas-Sánchez et al., 2023; Stenberdt & Makransky, 2023); this remains to be tested with objective learning or performance outcomes. Moreover, VR has been reported by other authors to enhance information retention and knowledge transfer, especially when applied in practical and experiential contexts (Makransky et al., 2021), although retention and transfer were not directly assessed in this study.

The results of the analysis indicate that the mastery motivational climate emerges as a strong and positive predictor of the perceived usefulness of VR in this sample. This suggests that learning environments that foster competence, autonomy, and a mastery-oriented approach are associated with an increase in students’ perception of the value of VR (Fern et al., 2024; Kemp et al., 2024). In contrast, the performance motivational climate was negatively related to behavioral intention, indicating that an excessively competitive environment may be associated with a lower predisposition toward continued use of VR. This pattern is consistent with the works of Ames (Ames, 1992) and Sánchez-Oliva (Sánchez-Oliva et al., 2017) and has been confirmed in recent studies on motivation in digital environments (Chen et al., 2020).

The findings also indicate that both the motivational climate and individual characteristics were associated with students’ perceptions of virtual reality’s usefulness in learning. Gender exhibited a moderate yet significant negative association, suggesting that differences in prior technological experience, interests, or confidence in using digital tools may help explain why, in this sample, male students reported perceiving VR as more useful than female students (Alturkustani et al., 2025; Parong & Mayer, 2021). Given the pronounced gender imbalance in the sample (83.6% male, 16.4% female) and the absence of measurement invariance testing across gender groups, these findings should be interpreted with caution.

The results indicate that the mastery motivational climate reported by students may be related to the immersive, sensory-rich nature of the VR experience common to all participants insofar as it appears to provide sensory experiences that reinforce the sense of competence and autonomy (Makransky & Petersen, 2021) and potentially enhance the quality of learning, especially in disciplines that benefit from practical simulation (Makransky et al., 2021), such as the analysis of compliance with various items in the safety of sports facilities; this mediating role was not directly tested in the present model and should be examined in future research.

A secondary objective was to construct and validate a theoretical model to explain and predict students’ ongoing intention to engage with VR.

The results of the structural equation modeling provide support for the proposed model, with adequate factor loadings and acceptable fit indices (CFI = 0.914; TLI = 0.901; RMSEA = 0.065 for the measurement model). The internal consistency of the constructs (α = 0.74–0.87) reinforces the internal robustness of the instrument as applied in the present sample; because item and path removal were performed on this same sample without cross-validation, generalizability to other samples should be confirmed in future studies. Although the association with gender was small in magnitude, it revealed significant differences in perceived usefulness and the mastery motivational climate. These findings are consistent with previous studies that have identified gender gaps in university-level technology adoption (Durndell & Haag, 2002; Makransky et al., 2017; Venkatesh & Morris, 2000; Wongwatkit et al., 2020) and highlight the importance of promoting inclusive strategies to reduce such disparities.

The model demonstrates acceptable internal consistency and predictive validity within this sample. In this regard, the mastery motivational climate was associated with both perceived usefulness and behavioral intention. Behavioral intention emerged as the key predictor of intention to use, although the strength of this association also reflects the conceptual overlap between the two constructs noted above. Moreover, gender differences were observed in the perception of the motivational climate. The theoretical model validated in this study also showed high predictive power, with R² values exceeding 80% for behavioral intention and intention to use, reflecting the internal consistency and robustness of the proposed framework as estimated in this sample. These findings are consistent with the idea that both motivational factors and demographic characteristics are relevant for predicting the perception of usefulness and, by extension, the behavioral intention to use VR, lending support to the applicability of models such as TAM and UTAUT in immersive educational contexts (Al-Adwan et al., 2023; Venkatesh & Davis, 2000).

Based on the proposed model, it can be observed that behavioral intention was significantly associated with perceived usefulness, the mastery motivational climate, and, negatively, with the performance climate. These findings are consistent with the notion that learning environments promoting autonomy (Arık & Erturan, 2023), personal improvement, and cooperation may foster a more positive attitude toward the use of technology (García-González et al., 2019; Kuleva, 2024; Luo & Du, 2022; Stracke et al., 2025; Tendinha et al., 2021; Umaroh et al., 2024). In contrast, settings focused on comparison and competition appear to be associated with resistance and lower willingness toward innovation (Ntoumanis & Standage, 2009).

Regarding perceived usefulness, the results indicate that the mastery climate showed a strong positive association, whereas gender showed a negative relationship. This suggests that motivational environments fostering self-efficacy and learning focused on personal improvement are associated with higher students’ perception of the technology’s usefulness. These findings are consistent with those of Baena-Extremera (Baena-Extremera et al., 2014) and other studies that analyzed self-determined motivation and technological acceptance among sports science students (Aufenanger et al., 2025; Kunz & Santomier, 2019). Conversely, the influence of gender may reflect differences in confidence or familiarity with digital tools, as previously reported in earlier research (Venkatesh & Morris, 2000; Wongwatkit et al., 2020; Ong & Lai, 2006). Therefore, it is recommended to promote inclusive strategies aimed at reducing these disparities.

Overall, the relationship between gender and mastery climate shows that there are small variations in motivational perception according to gender (Gea-García & Martínez, 2025; Nicholls, 1984). Although limited, this result may indicate the need to further explore differences in motivational and technological perceptions between men and women within the university context, with some gender-related differences observed in motivational dimensions such as attention and satisfaction (Portuguez-Castro & Santos Garduño, 2024; Moreno-Murcia et al., 2008).

Finally, the proposed model demonstrates a notable explanatory capacity within the studied sample, revealing that motivational factors—particularly the mastery climate are associated with perceived usefulness and intention to use technology in sports science education. These findings underscore the importance of fostering positive, improvement-oriented learning environments that may enhance both intrinsic motivation and the acceptance of innovative technological tools (Carrier, 2024; Mascret et al., 2025; Poupard et al., 2025), such as virtual and augmented reality, within educational and sports contexts (A. M. Gallardo-Guerrero et al., 2025). As noted by an anonymous reviewer, reference (A. M. Gallardo-Guerrero et al., 2025) examined a closely related model (augmented reality, sports-equipment safety, motivational climate, and continuance intention to use); the present study’s differential contribution lies in applying an immersive VR modality (rather than AR), in a distinct facility-safety-auditing task, and in a larger, dual-institution sample, and this distinction is now made explicit rather than assumed.

This research highlights both theoretical and practical implications related to the use of VR for the analysis of safety in sports facilities. From a theoretical perspective, VR used within a mastery motivational climate was associated with higher students’ perceived usefulness and intention to use, in this cross-sectional sample, consistent with the idea that this design facilitates active, self-explanatory, and improvement-oriented practice. From a practical standpoint, immersive walkthroughs and inspection simulations are intended to allow users to identify risks (such as signage, evacuation routes, accessibility, lighting, surface conditions, and user flow) and to rehearse safety protocols without interfering with the actual operation of the facility. Whether this approach reduces errors, prevents incidents, and promotes the transfer of skills to real contexts by reproducing decision-making scenarios similar to those that professionals will encounter in practice remains to be empirically tested; the present study assessed students’ perceptions and intentions rather than measured safety outcomes, error rates, or skills transfer, and these claims should therefore be read as potential applications rather than demonstrated effects. Furthermore, pending such confirmatory research, VR may serve as a continuous training tool for technicians, maintenance staff, and facility managers by standardizing audit criteria, enhancing evidence-based decision-making, and contributing to the improvement of safety in sports environments.

The present study has several limitations.

First, student participation in the research (i.e., providing questionnaire data for analysis) was voluntary, and data were collected through a questionnaire, which may have introduced an inclusion bias, as more motivated students or those more familiar with immersive technologies may have been more likely to participate. Second, the absence of a control group and of pre-post or baseline measurement prevents direct comparisons between VR and other teaching/learning methods or conventional on-site safety audits; therefore, the results should be interpreted as indicative associations rather than as conclusive evidence of methodological superiority. Third, the key variables—mastery motivational climate, behavioral intention, perceived usefulness, and continuance intention to use—were measured through self-reported instruments, which are susceptible to biases related to social desirability, recall, or external factors; moreover, actual use behavior beyond the single studied session was not measured, and behavioral intention and continuance intention to use showed a very strong association (β = 0.941; r = 0.68) that suggests these constructs may not be fully empirically distinct in this dataset. Fourth, the integration of immersive technologies such as virtual reality in the analysis of sports facilities not only allows for the exploration of associations between VR use and safety-related perceptions and realistic simulation, but also enables the exploration of how different groups, based on gender, perceive and adapt to these tools; however, the pronounced gender imbalance in the present sample (83.6% male) and the lack of measurement-invariance testing limit the strength of conclusions that can be drawn about gender differences, suggesting the need for further studies with more balanced samples that examine differences in motivational and technological perceptions among university students in sport sciences. Fifth, the final structural model was derived through iterative, data-driven removal of items and non-significant paths from the same sample used for hypothesis testing, without an independent validation sample; this increases the risk of overfitting and should be addressed through replication in future work. Consequently, future studies are encouraged to incorporate controlled or quasi-experimental designs, as well as longitudinal approaches, to assess the transfer of competencies acquired through VR to real-world safety audit contexts.

In this regard, the combination of virtual environments with active learning methodologies, such as problem-based learning or gamification, could further enhance engagement and learning retention (Lampropoulos & Kinshuk, 2024; Keller et al., 2025; Hellín et al., 2023). This methodological synergy opens new lines of research on how motivation and technology interact within university-level sports education.

5. Conclusions

In conclusion, in this cross-sectional sample, students who completed the VR-based safety-analysis activity reported a mastery motivational climate, perceived usefulness, behavioral intention, and continuance intention to use that were positively associated with one another, as well as with the usefulness that the student perceived this type of content to have. Therefore, a model has been found through which the motivational climate was associated with the usefulness and behavioral intention of students in this sample. Usefulness was associated with behavioral intention, the latter showing a strong association with continuance intention to use. Because the study did not include a non-VR comparison condition, these findings describe associations among students’ post-activity perceptions rather than an effect of VR use itself. A potential practical application of using VR in sports facility management is to explore its use for training staff and students in evaluating the safety and functionality of sports equipment and spaces in a risk-free, immersive environment, with a view to the identification of potential hazards and the optimization of facility design and maintenance procedures; however, given the observational, single-session, self-report design of the present study, these applications remain to be confirmed through controlled and longitudinal research before firm causal or practical claims can be made about VR’s effect on real-world safety outcomes.