Assessing Collaborative Learning Profiles in Nursing and Medical Students: Preliminary Development and Exploratory Validation of the Attitudes Toward Collaborative Learning Scale Authors: Shih‐Tzu Huang, Mei‐Chen Kuo, Yu‐Min Huang, Chiou‐Fang Liou Published: 2026 ABSTRACT ABSTRACTAimsTo develop and preliminarily validate the Attitudes Toward Collaborative Learning Scale (ATCLS) and characterize profiles of collaborative learning attitudes among nursing and medical students.DesignAn exploratory cross‐sectional quantitative study.Review MethodsParticipants completed the ATCLS. Reliability and construct validity were evaluated using internal consistency and confirmatory factor analysis. Separate hierarchical and K‐means cluster analyses were conducted for nursing and medical students. One‐way analysis of variance with Games–Howell post hoc comparisons characterized differences among clusters identified separately within each discipline.Data SourcesData were collected from 199 nursing and 50 medical students at two universities in Central Taiwan between October 2018 and August 2020.ResultsDistinct patterns of collaborative learning attitudes were observed across the separately identified nursing and medical student profiles. The Learning from Peer Interaction dimension demonstrated stronger measurement properties than the Task Engagement dimension. Cluster differences further characterized heterogeneity in collaborative learning attitudes within each discipline.ConclusionThe findings provide preliminary evidence supporting the development and exploratory validation of the scale and suggest heterogeneity in collaborative learning attitudes. The profiles require further validation in larger and more diverse samples.Implications for the Profession and/or Patient CareRecognizing variability in collaborative learning attitudes may help nurse and medical educators tailor strategies to support collaborative and interprofessional learning.ImpactCollaborative learning attitudes were heterogeneous within nursing and medical student groups, supporting a profile‐based perspective rather than assuming uniform attitudes toward collaborative learning.Reporting MethodThis study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines.No Patient or Public ContributionPatients or members of the public were not involved because participants were university students.Keywords: cluster analysis, collaborative learning, confirmatory factor analysis, interprofessional education, nursing education Abbreviations analysis of variance Attitudes Toward Collaborative Learning Scale average variance extracted confirmatory factor analysis interprofessional education Learning from Peer Interaction Task Engagement Introduction Collaborative learning is a pedagogical approach grounded in sociocultural learning theory, which conceptualizes learning as the co‐construction of knowledge through social interaction, dialogue, and peer engagement (Arvaja et al. 2007; Hesse et al. 2015; Laal and Laal 2011; Vygotsky 1978). Within healthcare education, collaborative learning provides an important foundation for developing communication, teamwork, reflective practice, and other competencies required for effective interprofessional collaboration (Bereiter and Scardamalia 2014; Slavin 2014). Evidence from both nursing and medical education supports the educational value of collaborative learning. In nursing education, collaborative learning has been associated with improved academic performance, professional identity formation, teamwork, and shared responsibility (Baghcheghi et al. 2011; Leyva‐Moral and Riu Camps 2016; Roqueta‐Vall‐Llosera et al. 2024). Similarly, studies in medical education have shown that collaborative learning can enhance problem‐solving, critical thinking, and collaborative skills through structured team‐based educational activities and research‐oriented learning experiences (Sahoo and Mohammed 2018; Umesh et al. 2023). Despite these reported benefits, previous research has primarily evaluated the educational outcomes of collaborative learning following instructional interventions (Zhang and Cui 2018). Comparatively little attention has been paid to students' underlying attitudes toward collaborative learning before participation in formal interprofessional education (IPE) (Panel IPE 2011; Thistlethwaite 2012), although such attitudes may influence their engagement in collaborative learning activities. This gap provides the rationale for the present study. This study builds on prior work examining students' learning trajectories in a simulation‐based educational program designed to foster systems thinking. Findings from that study indicated that students demonstrated different learning orientations (e.g., more fixed versus more adaptive approaches) when engaging with complex clinical scenarios, peer discussion, and reflective learning. These observations suggest that students may enter collaborative learning environments with varying levels of openness to and engagement in peer interaction. However, existing instruments primarily assess interprofessional competencies or readiness in the context of structured IPE programmes. Few tools specifically capture students' foundational attitudes toward collaborative learning before such experiences. Addressing this gap is important for understanding how students approach and engage in collaborative learning at an early stage. Therefore, this study aimed to develop the ATCLS, provide preliminary exploratory evidence of its psychometric properties, and characterize collaborative learning profiles among nursing and medical students. Given the exploratory nature of the study and the relatively small medical student sample, the inclusion of medical students was intended to provide preliminary insights into potential variation across professional education contexts rather than to establish definitive comparisons between nursing and medical students. Literature Review Collaborative learning activities, including problem‐based learning, peer feedback, and simulation‐based teamwork, are important approaches for developing clinical reasoning, empathy, and interdisciplinary understanding (Mägi et al. 2024). These approaches have expanded through innovations such as virtual simulation and Collaborative Online International Learning (COIL), particularly following the COVID‐19 pandemic, enabling students to engage in diverse and intercultural learning contexts (Roqueta‐Vall‐Llosera et al. 2024; Shields et al. 2022). Despite their widespread implementation, limited attention has been given to how students' attitudes toward collaborative learning are formed before formal collaborative or interprofessional training. Foundational attitudes, including openness to peer perspectives and willingness to engage in shared learning, may influence students' participation in and responses to collaborative learning environments. Existing instruments, such as the Interprofessional Collaborative Competency Attainment Survey (ICCAS) (Archibald et al. 2014), focus on professional roles and collaborative behaviours within structured team settings. These tools do not specifically capture students' baseline attitudes toward collaborative learning, which may influence their subsequent engagement in IPE. Table 1 summarizes instruments used to assess collaborative learning and related constructs. TABLE 1Current measures for collaborative learning.Instrument (author[s])Descriptions of the scaleDimensions/subscaleValidationsSampleReadiness of health care students for interprofessional learning (RIPLS) (Parsell and Bligh 1999)The scale was developed to assess the readiness for interprofessional education in students and professionals in the field of healthcare. The tool was specifically developed to assess their attitudes and perceptions in understanding their readiness for interprofessional learning and change.The scale has a total of 19 items within three subscales: Teamwork and collaborationProfessional identityRole and responsibility Good construct validity and reliabilityStudents in healthcare professionsMultiple‐group Measurement Scale for Interprofessional Collaboration (Kenaszchuk et al. 2010)Rubrics of interprofessional collaboration education (IPE). The focus of the scale is to rate each other's rubric during the IPE practice. Labelled communicationAccommodationIsolation Reliability: Raykov's statistic; p‐values were 0.76 (Communication), 0.85 (Accommodation), and 0.76 (Isolation) Validity: Construct validity with a predictive approach and current validity were established. All clinicians working in acute care wards: physicians, nurses, and other healthcare providers rating each otherMeasuring at the team level Interprofessional collaborative competency attainment survey (ICCAS) (Archibald et al. 2014) An interprofessional collaboration program evaluation. Self‐assessment for interprofessional collaborative practice. Items of the scale focused on the competencies of interprofessional collaboration education (IPE)Adapted and modified items from the RIPLS.Single dimension Excellent Cronbach's α for internal consistency. The exploratory factor analysis explained a total of 71.4% of variance. Including both healthcare professions and non‐health care professions: clinicians and studentsInterprofessional Attitude Scale (IPAS) (Norris et al. 2015)This scale relates to the core competencies for Interprofessional Collaborative PracticeIPAS consists of 27 items in five subscales: TeamworkRoles and responsibilitiesPatient‐centrednessInterprofessional biases, diversity, and ethicsCommunity‐centredness Inadequate validity and reliability (Cronbach's α of 0.61–0.92).Validated by a total of 701 students from a healthcare program Current measures for collaborative learning. Teamwork and collaboration Teamwork and collaboration Professional identity Professional identity Role and responsibility Role and responsibility Labelled communication Labelled communication Accommodation Accommodation Isolation Isolation Reliability: Raykov's statistic; p‐values were 0.76 (Communication), 0.85 (Accommodation), and 0.76 (Isolation) Validity: Construct validity with a predictive approach and current validity were established. All clinicians working in acute care wards: physicians, nurses, and other healthcare providers rating each other All clinicians working in acute care wards: physicians, nurses, and other healthcare providers rating each other Measuring at the team level Measuring at the team level Interprofessional collaborative competency attainment survey (ICCAS) (Archibald et al. 2014) Excellent Cronbach's α for internal consistency. The exploratory factor analysis explained a total of 71.4% of variance. Teamwork Teamwork Roles and responsibilities Roles and responsibilities Patient‐centredness Patient‐centredness Interprofessional biases, diversity, and ethics Interprofessional biases, diversity, and ethics Community‐centredness Community‐centredness Collaborative learning attitudes may also vary according to professional background, gender (Viking et al. 2023), and cultural context (Nguyen et al. 2005). Nursing students, who are often educated within team‐oriented learning environments, may demonstrate stronger preferences for collaboration, whereas medical students may exhibit greater variability associated with differing educational norms (Kenaszchuk et al. 2010; Li and Cui 2022; Feng et al. 2023). Cultural influences, particularly in Confucian‐influenced contexts, may further shape communication patterns and attitudes toward collaboration through an emphasis on harmony and indirect communication (Huang and Chang 2017; Lynn‐Sze 2021; Yum 1988). Understanding such variation may inform the design of collaborative learning strategies that are responsive to diverse learner characteristics. Theoretical Frame Theories and Prior Empirical Works The theoretical foundation of this study is grounded in sociocultural theory (Vygotsky 1978) and situated learning theory (Lave and Wenger 1991), together with prior empirical work examining the development of systems thinking competence in nursing students. Sociocultural theory conceptualizes learning as a socially mediated process in which knowledge is constructed through interaction, shared meaning‐making, and participation in collaborative activities. The concept of the Zone of Proximal Development further emphasizes that learning can be facilitated through engagement with others, particularly in contexts that require negotiation and the co‐construction of understanding (Vygotsky 1978). The present study was also informed by previous qualitative work examining students' learning trajectories in a simulation‐based systems thinking educational program. In that work, students described learning processes involving the interpretation of complex clinical situations, coordination of multiple perspectives, and reflection on the interdependence of actions within healthcare systems. These observations suggested that students may differ in how they engage in peer interaction, collaborative problem‐solving, and shared learning processes, providing an empirical rationale for examining collaborative learning attitudes in the present study. Existing research in IPE has primarily focused on collaborative competencies and educational outcomes following structured interprofessional learning experiences (Reeves et al. 2012; Saragih et al. 2023; Spaulding et al. 2021). In contrast, relatively little attention has been given to students' underlying attitudes toward collaborative learning as a foundational construct that may influence their engagement in collaborative educational environments. Taken together, these theoretical perspectives and prior empirical observations provide a rationale for conceptualizing collaborative learning attitudes as a multidimensional construct. Accordingly, this study aimed to develop the ATCLS, provide preliminary exploratory evidence of its psychometric properties, and characterize heterogeneity in collaborative learning attitudes among nursing and medical students. Conceptual Framework Guided by sociocultural theory and situated learning, this study conceptualizes collaborative learning attitudes as a foundational construct that may shape students' engagement in socially mediated learning environments. In simulation‐based systems thinking education, students participate in collaborative processes that require peer interaction, negotiation of meaning, and task‐oriented engagement. From this perspective, collaborative learning attitudes are conceptualized as comprising two interrelated dimensions: Learning from Peer Interaction (LEARN) and Task Engagement (TASK). These dimensions reflect how students approach shared learning processes and participate in collaborative tasks. These attitudes may vary across individuals according to prior learning experiences and contextual factors. Therefore, a person‐centred analytical approach was used to characterize potential heterogeneity in collaborative learning attitudes by identifying profiles among nursing and medical students. This framework positions collaborative learning attitudes as a potential precursor to engagement in simulation‐based and interprofessional learning environments and provides a conceptual basis for examining individual differences in attitudes toward collaboration. Figure 1 presents the conceptual framework of the study. Conceptual framework guiding the development and exploratory testing of the ATCLS. This framework illustrates how sociocultural theory (Vygotsky 1978), situated learning (Lave and Wenger 1991), the Confucian‐influenced educational context (Yum 1988; Huang and Chang 2017; Lynn‐Sze 2021; Feng et al. 2023), and prior empirical observations from simulation‐based systems thinking education informed the development of the ATCLS. Collaborative learning attitudes are conceptualized as comprising two interrelated dimensions: Learning from Peer Interaction (LEARN) and Task Engagement (TASK). A person‐centred analytical approach was subsequently used to provide an exploratory characterization of collaborative learning profiles among nursing and medical students. Methods and Materials Study Design This study employed an exploratory cross‐sectional design. A person‐centred analytical approach was used to characterize profiles of collaborative learning attitudes. Hierarchical cluster analysis was first conducted to determine the number of clusters, followed by K‐means cluster analysis to refine cluster membership. Participants were clustered based on their scores on the ATCLS. Ethical Considerations This study received ethical approval from the institutional review boards of both participating institutions (CSMUH No. CS18119; CRREC‐108‐105). Informed consent was obtained using an implied consent approach. A written explanation of the study, including its purpose, voluntary nature, confidentiality, and the right to withdraw at any time without penalty, was provided on the first page of the questionnaire. Completion and return of the questionnaire were considered to indicate informed consent. Participation was entirely voluntary, and students were assured that their decision to participate or decline would not affect their academic evaluation or standing. All data were collected anonymously and analysed in aggregate form. No personally identifiable information was collected, thereby protecting participants' confidentiality and privacy. Sample and Setting Participants were purposively recruited from nursing and medical programmes at two universities in central Taiwan between October 2018 and August 2020. The sampling strategy intentionally included a larger proportion of nursing students than medical students to broadly reflect the greater representation of nurses relative to physicians in healthcare teams. The final sample consisted of 249 students, including 199 nursing students and 50 medical students. Of the participants, 179 (72%) were female, and 201 (81%) were aged 18–23 years. Instrument Development To assess students' attitudes toward collaborative learning before formal IPE, we developed the ATCLS. The instrument was specifically designed to capture students' perceptions, beliefs, and engagement related to collaborative learning activities. The development of the ATCLS was theoretically informed by sociocultural theory (Vygotsky 1978), which emphasizes knowledge construction through social interaction, and by Confucian educational philosophy (Romero, et al. 2022), which values harmonious group interaction and respect for authority within Taiwanese educational contexts. Item generation followed a theory‐informed and qualitatively driven process. Initial items were developed from themes identified through qualitative analysis of students' reflective journals collected during a simulation‐based educational program. These findings were complemented by a comprehensive review of the collaborative learning literature to ensure conceptual coverage. The preliminary item pool was subsequently reviewed by nursing education experts, including those with expertise in interprofessional and medical–nursing collaborative education, to establish content relevance and clarity. The initial version of the ATCLS consisted of 15 items, each rated on a five‐point Likert scale (1 = strongly disagree to 5 = strongly agree), with higher scores indicating more positive attitudes toward collaborative learning. The ATCLS was conceptualized as comprising two theoretically derived dimensions. The Learning from Peer Interaction (LEARN) subscale assesses students' perceptions of learning through communication, shared knowledge, and mutual feedback in collaborative environments (Arvaja et al. 2007; Laal and Laal 2011; Hesse et al. 2015). The TASK subscale evaluates students' cognitive and emotional engagement in group tasks, including their willingness to negotiate shared responsibilities, consider diverse perspectives, and participate actively in collaborative work (Bereiter and Scardamalia 2014; Slavin 2014). Content Validity To ensure the scale's relevance and clarity, three experts in nursing education evaluated the items for content validity. These experts assessed whether each item accurately reflected the intended constructs of peer interaction and TASK. The survey items were primarily developed from themes identified in nursing students' reflective narratives describing their experiences of peer interaction, collaboration, and TASK during simulation‐based learning. Therefore, nursing education experts were considered best suited to assess whether the items accurately reflected the language, meaning, and learning experiences of nursing students. Although all three panel members had expertise in nursing education, some also had experience in IPE and medical–nursing collaborative training. This additional expertise enabled the content validity assessment to incorporate an interprofessional perspective on collaborative learning. Their expertise ensured that the content validity assessment considered collaborative learning from an interprofessional perspective. The item‐level content validity index (I‐CVI) ranged from 0.87 to 1.00, and the scale‐level content validity index (S‐CVI) was 0.94, demonstrating strong expert agreement on the appropriateness and clarity of the items. Reliability Analysis Initial item analysis of the 15‐item ATCLS was conducted using the full sample of 249 students (199 nursing and 50 medical students). Most items demonstrated acceptable item–total correlations. However, two items within the TASK subscale showed relatively low correlations with the total scale (r = 0.12 and r = 0.15), suggesting limited contribution to the construct. Based on both statistical criteria and conceptual considerations, these items were removed, resulting in a final 13‐item scale. The internal consistency of the final 13‐item ATCLS was then evaluated using Cronbach's alpha. The resulting 13‐item ATCLS was subsequently used for CFA in the full sample and for separate cluster analyses of nursing and medical students. The final scale demonstrated high internal consistency, with Cronbach's alpha of 0.93 for the overall ATCLS, 0.92 for the LEARN subscale, and 0.77 for the TASK subscale. These findings indicate excellent reliability for the overall scale and the LEARN dimension and acceptable reliability for the TASK dimension. Overall, these results provide preliminary evidence supporting the internal consistency of the ATCLS, while suggesting that further refinement of the TASK dimension may be warranted. Construct Validity Preliminary evidence of construct validity was examined using confirmatory factor analysis (CFA) conducted in R (RStudio, version 2026.04.0). CFA of the final 13‐item ATCLS was conducted using the full sample of 249 students. A two‐factor model comprising LEARN and TASK was specified. The model demonstrated good overall fit, χ 2(44) = 59.30, p = 0.06, RMSEA = 0.04, CFI = 0.97, and SRMR = 0.03. All item loadings were statistically significant (p < 0.001). Standardized factor loadings ranged from 0.39 to 0.78 for the TASK subscale and from 0.56 to 0.90 for the LEARN subscale. Most items exceeded the recommended threshold of 0.50, although one item in the TASK subscale demonstrated a relatively low loading (0.39), indicating a potential limitation in measurement performance. Figure 2 presents the confirmatory factor analysis model for the two‐factor structure of the ATCLS. CFA Model of Attitude Toward Collaborative Learning. Additional CFA models were examined after removing items with lower factor loadings; however, these modified models demonstrated poorer overall fit. Therefore, the original two‐factor model was retained based on the balance between theoretical coherence and empirical model fit. Table 2 presents a comparison of the three CFA models. Table 3 presents the item descriptions and corresponding factor loadings for the original model. TABLE 2Comparison of the three CFA models.CFA model χ 2 df p CFITLIRMSEASRMREvaluation13‐item original theoretical scale59.3440.060.970.990.040.03Best fit; retained12‐item model263.6153< 0.0010.8910.8650.1260.067Poorer fit1 11‐item model210.0343< 0.0010.9100.8850.1250.056Improved over the 12‐item but still suboptimal Note: 1 One standardized factor loading decreased to 0.48 in the modified model. Comparison of the three CFA models. Note: 1 One standardized factor loading decreased to 0.48 in the modified model. TABLE 3Item content and factor loading.ItemsContentFactor loadingLearning interactionDeveloping a strategyI would love to think of strategies to overcome the obstacles of performing tasks.0.56Learning about each otherLearning and working with my peers allows me to benefit from their strengths.0.73Trial and errorI learn from mistakes that I made in the learning process.0.74AdviceI look forward to recommendations from others.0.90Other's viewpointsViewpoints from others make me think of my weakness.0.87Other's criticismCriticism from others makes me think of my weakness.0.85Learning with each otherMy peer's excellent performance helps us to smoothly complete tasks.0.84Learning from each otherMy peer's excellent performance serves as a model for my learning0.82Task engagementLearning from doingPerforming tasks often provides an opportunity to help me grow.0.78Openness to experienceI would consider it a success if I tried on tasks, making mistakes or not.0.66Task completionI would love to try new things.0.39Making mistakesIt is my chance to learn if/when I make mistakes in the learning process.0.56ProficiencyLearning how to perform the task makes me proficient.0.76 Item content and factor loading. Convergent and Discriminant Validity Convergent validity was examined using average variance extracted (AVE). The LEARN subscale demonstrated adequate convergent validity (AVE = 0.63). In contrast, the TASK subscale had an AVE of 0.42, below the commonly recommended threshold of 0.50, indicating comparatively weaker convergent validity for this dimension. Notably, the TASK subscale demonstrated acceptable internal consistency (Cronbach's α = 0.77). Taken together, these findings suggest that although the TASK items demonstrated acceptable internal consistency, the variance captured by the underlying construct was comparatively limited. Discriminant validity was evaluated using the Fornell–Larcker criterion (Fornell and Larcker 1981). The square roots of the AVE values for LEARN (0.80) and TASK (0.65) exceeded the correlation between the two constructs (r = 0.62), providing preliminary evidence that the two dimensions were empirically distinguishable. Overall, these findings provide preliminary evidence of the construct validity of the ATCLS. However, the lower AVE for the TASK subscale indicates that further refinement and independent validation of this dimension are warranted. Scoring No items required reverse coding. Subscale and total scores were calculated by summing the corresponding item scores. Higher ATCLS scores indicate more positive attitudes toward collaborative learning. Data Analysis Procedure Data were analysed using IBM SPSS Statistics for Windows, Version 30.0 (IBM Corp., Armonk, NY) and R (RStudio, version 2026.04.0). A person‐centred analytical approach was employed to characterize profiles of collaborative learning attitudes. Because nursing and medical students represented different professional groups and differed substantially in sample size, cluster analyses were conducted separately for nursing students (n = 199) and medical students (n = 50). A two‐stage clustering procedure was applied to each group. First, hierarchical cluster analysis using Ward's method and squared Euclidean distance was conducted to determine the number of clusters. K‐means cluster analysis was subsequently conducted to refine cluster membership. Participants were clustered based on their ATCLS total and subscale scores (LEARN and TASK). Cluster quality was evaluated using the silhouette measure of cohesion and separation, with values greater than 0.50 considered indicative of good cluster quality. To descriptively characterize differences among the profiles identified separately for nursing and medical students, a one‐way analysis of variance (ANOVA) was conducted. The identified cluster memberships from both professional groups were combined as the grouping variable, while ATCLS total and subscale scores (LEARN and TASK) served as dependent variables. Because the same ATCLS variables were used both to derive and subsequently characterize the clusters, the ANOVA was interpreted as a descriptive characterization of the identified profiles rather than as independent confirmatory evidence of the cluster solutions. Levene's test was used to assess homogeneity of variance. Games–Howell post hoc comparisons were conducted to further characterize differences among the identified profiles. Effect sizes (η 2) were calculated to describe the magnitude of these differences. Results Cluster Characteristics Cluster analyses identified three profiles among nursing students (n = 199) and two profiles among medical students (n = 50). Cluster sizes varied, with the smallest cluster observed among medical students (n = 8). Among nursing students, Cluster One (n = 20) demonstrated the lowest levels of collaborative learning attitudes, with lower scores across both the LEARN and TASK dimensions (mean ATCLS = 32.22). Cluster Two (n = 80) showed the highest scores across both dimensions (mean ATCLS = 59.19), whereas Cluster Three (n = 99) demonstrated intermediate levels of collaborative learning attitudes (mean ATCLS = 48.17). Among medical students, Cluster One (n = 42) demonstrated relatively higher scores on both LEARN and TASK (mean ATCLS = 55.24), whereas Cluster Two (n = 8) showed lower scores across both dimensions (mean ATCLS = 36.14). Overall, the identified profiles showed a gradient in collaborative learning attitudes from lower to higher levels within the two professional groups. Table 4 presents the characteristics of the identified profiles, and Figure 3 illustrates their ATCLS score patterns. TABLE 4Cluster profiles based on ATCLS total and subscale scores.Cluster n LEARN (Mean ± SD)TASK (Mean ± SD)ATCLS total (Mean ± SD)NC1 (Low)2018.78 ±5.1 13.44 ±2.59 32.22 ±5.40 NC2 (High)8037.21 ±2.26 21.98 ±2.3 59.19 ±3.79 N3C (Moderate)9930.93 ±2.74 17.24 ±2.73 48.17 ±3.57 MC1 (Higher)4235.17 ±3.42 20.07 ±2.66 55.24 ±5.2 MC2 (Lower)820.86 ±7.15 15.29 ±3.5 36.14 ±7.17 Note: Values are presented as mean ± standard deviation.Abbreviations: ATCLS, Attitudes Toward Collaborative Learning Scale; LEARN, learning‐focused collaboration; MC, Medical Student Cluster; NC, Nursing Student Cluster; TASK, task‐oriented collaboration. Cluster profiles based on ATCLS total and subscale scores. Note: Values are presented as mean ± standard deviation. Abbreviations: ATCLS, Attitudes Toward Collaborative Learning Scale; LEARN, learning‐focused collaboration; MC, Medical Student Cluster; NC, Nursing Student Cluster; TASK, task‐oriented collaboration. Cluster Profiles of Collaborative Learning. Cluster Quality The clustering solutions for both nursing and medical students demonstrated good quality based on silhouette measures (> 0.5), indicating adequate cohesion and separation within the present samples. However, these indices characterize the identified solutions within the current samples and should not be interpreted as evidence of their reproducibility in other samples. Given the exploratory nature of the study and the relatively small medical student sample, particularly the limited number of participants in one cluster (n = 8), the medical student cluster solution should be considered preliminary. Differences Across Clusters Descriptive characterization using ANOVA showed differences among the identified profiles in LEARN, TASK, and total ATCLS scores. The corresponding effect sizes were η 2 = 0.74 for LEARN, η 2 = 0.51 for TASK, and η 2 = 0.77 for total ATCLS. Although these effect sizes were large, they should not be interpreted as independent confirmation of the cluster solutions because the same ATCLS variables were used both to derive and subsequently characterize the profiles, potentially inflating the observed between‐profile differences. Furthermore, because cluster solutions were derived separately for nursing and medical students, comparisons among the identified profiles should be considered exploratory and should not be interpreted as representing equivalent cluster structures across the two professional groups. Table 5 presents the ANOVA and Games–Howell post hoc comparisons used to descriptively characterize differences among the identified profiles. TABLE 5Results of one‐way ANOVA and pairwise comparison.Dependent variableSourceSSdfMSF p (η 2)Significant pairwise comparisons (Games‐Howell)1 LEARNBetween groups6657.70441664.426169.274< 0.001 (0.74) nc1 < nc3***; nc1 < mc1*** 2; n2 > others***; nc3 < mc1***, nc3 > mc2***; mc1 > mc2 nc1 = mc2 (ns3) Within groups2369.6942419.833Total9027.398245TASKBetween groups1713.0354428.25962.903< 0.001 (0.51)nc1 < nc3***; nc1 < mc1***; n2 > others***; nc3 < mc1***; mc1 > mc2***; nc1 = mc2 (ns); nc3 = mc2 (ns).Within groups1640.7912416.808Total3353.825245ATCLSBetween groups14751.05943687.765206.597< 0.001 (0.77)nc1 < others*; nc2 > others * nc3 < mc1***, nc3 > mc2***; mc1 > mc2*** Within groups4301.85624117.850Total19052.915245 Note: 1 Levene's test indicated that the assumption of homogeneity of variance was met across the five groups. However, given the unequal group sizes and small sample sizes in some clusters, the Games–Howell post hoc test was applied. 2. nc, Nursing Cluster; mc, Medical Cluster. 3. ns, non‐significance.* p < 0.05.*** p < 0.001. Results of one‐way ANOVA and pairwise comparison. nc1 < nc3***; nc1 < mc1*** 2; n2 > others***; nc3 < mc1***, nc3 > mc2***; mc1 > mc2 nc1 = mc2 (ns3) Note: 1 Levene's test indicated that the assumption of homogeneity of variance was met across the five groups. However, given the unequal group sizes and small sample sizes in some clusters, the Games–Howell post hoc test was applied. 2. nc, Nursing Cluster; mc, Medical Cluster. 3. ns, non‐significance. p < 0.05. p < 0.001. Discussion The present findings extend this perspective by suggesting that students may differ in their attitudes toward collaborative learning and potentially in how they approach socially mediated learning processes. Measurement Properties of the ATCLS Reliability and Construct Validity The present study provides preliminary evidence regarding the measurement properties of the ATCLS. The scale demonstrated strong internal consistency overall (α = 0.93) and for the LEARN subscale (α = 0.92), while the TASK subscale demonstrated acceptable internal consistency (α = 0.77). CFA supported the proposed two‐factor structure of the ATCLS within the present sample. The LEARN dimension demonstrated satisfactory convergent validity (AVE = 0.63), whereas the TASK dimension showed a lower AVE (0.42) and comparatively weaker factor loadings. This pattern suggests that task‐oriented collaboration may represent a more complex and heterogeneous construct. Although the TASK items demonstrated acceptable internal consistency, they may capture broader and more context‐dependent aspects of collaborative engagement rather than a narrowly defined homogeneous construct. Further refinement and independent validation of this dimension are therefore warranted. Theoretical Integration of LEARN and TASK Dimensions These findings can be interpreted through a sociocultural perspective, which emphasizes that learning occurs through social interaction, shared meaning‐making, and participation in collaborative activities. Within this framework, the LEARN subscale reflects students' attitudes toward peer interaction and the co‐construction of knowledge, which are central processes in sociocultural learning theory. The relatively strong measurement performance of the LEARN dimension, as indicated by higher factor loadings and adequate convergent validity (AVE = 0.63), suggests that students' attitudes toward learning through peer interaction can be captured with reasonable consistency. This finding is consistent with the theoretical assumption that collaborative learning is fundamentally grounded in socially mediated learning processes. In contrast, the TASK subscale reflects attitudes toward task‐oriented aspects of collaboration, such as responsibility, coordination, and task completion. Although these elements are relevant to collaborative learning, they may represent more context‐dependent or behaviourally variable aspects of collaboration, potentially contributing to the lower AVE and greater variability in factor loadings observed for this dimension. Taken together, the findings suggest that both dimensions contribute to the conceptualization of collaborative learning attitudes. The LEARN dimension appears to more directly reflect the socially mediated processes emphasized in sociocultural theory, whereas TASK captures complementary but potentially more heterogeneous aspects of collaborative engagement. Cluster Structure and Interpretation The clustering solutions demonstrated acceptable quality (silhouette > 0.50), indicating adequate cohesion and separation within the present samples. However, the identified profiles should be viewed as an exploratory characterization of heterogeneity in collaborative learning attitudes rather than as evidence of reproducible cluster structures. Differences in sample size and cluster distribution, particularly within the medical student group, may further affect the stability of the identified solutions. Heterogeneity of Collaborative Learning Profiles The present study extends the IPE and collaborative learning literature by providing preliminary evidence of heterogeneity in students' attitudes toward collaborative learning before participation in formal IPE programmes. Previous research has primarily focused on the effects of IPE interventions on collaborative competencies, teamwork, and educational outcomes (Saragih et al. 2023; Spaulding et al. 2021). In contrast, the present findings suggest that students may not enter IPE with uniform attitudes toward collaborative learning, highlighting the potential importance of considering such attitudes before formal interprofessional training. From a sociocultural perspective, such heterogeneity may reflect differences in how students internalize and engage with collaborative learning experiences, even within similar educational contexts. This interpretation is consistent with the premise that learning is shaped by participation in socially mediated environments, while individuals may differ in how they engage with these processes. This interpretation is also consistent with prior work examining students' reflective learning in systems thinking contexts, which identified patterns characterized by more fixed versus adaptive (growth‐oriented) learning orientations. Evidence from medical education similarly suggests that students with more flexible learning orientations may be more adaptable to technology‐enhanced and collaborative learning environments (Naamati‐Schneider and Alt 2023). These findings provide a complementary perspective on the observed heterogeneity and suggest that differences in learning orientations may be associated with variation in collaborative learning attitudes. However, learning orientations were not directly measured in the present study, and this potential relationship requires further investigation. Taken together, these findings suggest that collaborative learning attitudes may represent an important learner characteristic to consider when designing formal IPE curricula. Rather than assuming that all students enter IPE with similar attitudes toward collaboration, educators may benefit from recognizing potential heterogeneity in collaborative learning attitudes and considering students' differing learning needs. The educational implications of early assessment and differentiated support are discussed further below. Contextual Considerations Differences in collaborative learning attitudes may also be considered in relation to professional education contexts. Nursing education commonly emphasizes teamwork, peer interaction, shared responsibility, and collaborative practice as components of professional preparation (Archibald et al. 2014; Hesse et al. 2015; Kenaszchuk et al. 2010; Feng et al. 2023; Mägi et al. 2024; Wang et al. 2019). Such education contexts may provide students with greater opportunities for social interaction and peer learning. Nevertheless, the identification of nursing students with lower collaborative learning attitudes suggests that these attitudes may not be uniform even within a professional education context that emphasizes collaboration. The variability observed among medical students may similarly be considered in relation to literature suggesting that traditional medical education has often emphasized individual responsibility, autonomy, and hierarchical learning structures (Behrend et al. 2020). However, this interpretation should be considered preliminary given the relatively small medical student sample (n = 50), particularly the limited number of participants in one identified cluster (n = 8). Furthermore, because students' educational experiences and other contextual factors were not directly measured in the present study, professional education context should not be inferred as an explanation for the observed differences. Further research with larger and more balanced samples is needed to examine whether and how professional education contexts are associated with collaborative learning attitudes. Educational Implications Based on the findings of this preliminary study, nurse educators may consider students' differing attitudes toward collaborative learning when designing learning activities that require teamwork. For example, when assigning students to small groups, educators may consider creating groups that include students with differing attitudes toward collaborative learning rather than assuming that all students approach collaborative learning in the same way. The ATCLS may potentially serve as an assessment tool to help educators identify differences in students' collaborative learning attitudes. Combining such assessment with educators' observations and rubric‐based evaluations during learning activities may provide a more comprehensive understanding of students' collaborative learning performance. Early assessment of collaborative learning attitudes may also help nurse educators, medical educators, and IPE leaders identify students who could benefit from additional educational support before or during team‐based learning activities. Tailored approaches, such as structured guidance, scaffolding, facilitated peer interaction, and role clarification, may be considered to support students with differing attitudes toward collaborative learning and prepare them for participation in collaborative, team‐based practice. However, these educational applications remain preliminary, and further research is needed to determine whether assessment‐informed educational strategies improve collaborative learning outcomes. Limitations Several limitations should be considered. First, the same sample was used for item refinement, CFA, and subsequent cluster analysis, which may increase the risk of sample‐dependent findings. Although post hoc CFA models involving removal of items with lower factor loadings resulted in poorer model fit and the theoretically derived two‐factor structure was retained, the psychometric findings should nevertheless be considered preliminary. Future studies should evaluate the measurement properties and cluster structure of the ATCLS using independent samples. Second, the medical student cluster solution should be considered preliminary given the relatively small medical student sample (n = 50), with only eight participants in one of the identified clusters. Third, cluster solutions were generated separately for nursing and medical students. Consequently, comparisons across professional groups should be interpreted with caution, as the cluster structures were not directly comparable. Fourth, because cluster membership was derived from the same variables used in subsequent analyses, the reported effect sizes may have been inflated. Fifth, although the silhouette values indicated acceptable cluster quality within the present samples, they do not establish the reproducibility of the identified profiles in independent samples. Finally, the cross‐sectional design precludes causal inference and limits conclusions regarding the temporal relationships among collaborative learning attitudes and related educational factors. Contribution and Future Directions This study contributes to the literature by integrating scale development with a person‐centred analytical approach to examine collaborative learning attitudes. The findings provide preliminary evidence of heterogeneity in collaborative learning attitudes and suggest that such variation may be characterized through exploratory profiles. Building on prior work examining students' learning trajectories in simulation‐based systems thinking education, the present findings raise the possibility that differences in learning orientations may be associated with variation in collaborative learning attitudes. However, learning orientations were not directly measured in the present study, and this potential relationship requires further investigation. Future research is needed to further evaluate and validate the ATCLS across diverse student populations and educational settings. In addition, longitudinal studies are recommended to examine how collaborative learning attitudes develop over time and how they relate to students' learning orientations. Further investigation is also warranted to explore potential predictors of collaborative learning profiles and their relationships with educational outcomes, including engagement in interprofessional learning and team‐based practice. Conclusion This study provides preliminary and exploratory evidence that collaborative learning attitudes are heterogeneous and may be characterized through distinct profiles. The ATCLS may offer an initial tool for assessing foundational attitudes toward collaborative learning prior to formal IPE. When supplemented by educators' observations and rubric‐based assessments during learning activities, the ATCLS may contribute to a more comprehensive understanding of students' attitudes toward collaborative learning. However, the findings should be interpreted cautiously because the same sample was used for instrument refinement, CFA, and cluster analysis, and because of the relatively small medical student sample, including the limited number of participants in one identified cluster. Future research using larger, more diverse, and independent samples is needed to further validate the ATCLS, refine the TASK dimension, examine the stability and reproducibility of the identified profiles, and investigate their relationships with educational outcomes and preparation for team‐based practice. Author Contributions Shih‐Tzu Huang: investigation, data curation, writing – review and editing, methodology, writing – original draft, validation. Mei‐Chen Kuo: writing – review and editing, resources, conceptualization, validation. Yu‐Min Huang: formal analysis, writing – review and editing, resources, software, validation, visualization. Chiou‐Fang Liou: conceptualization, investigation, funding acquisition, writing – review and editing, formal analysis, project administration, data curation, writing – original draft, software, supervision, methodology, validation, visualization, resources. Funding This work was supported by the Department of Technology, Taiwan (MOST) (106‐2511‐S‐040‐003) and Dr. Chiou‐Fang Liou, who is the principal investigator of this study. Ethics Statement This study was approved by two Institutional Review Boards (IRBs): Chung Shan Medical University Hospital (CSMUH No. CS18119) and China Medical University Research Ethics Review Committee (CRREC‐108‐105). Conflicts of Interest The authors declare no conflicts of interest. Acknowledgements This research work was supported by the Ministry of Science and Technology, Taiwan. The authors are grateful to all participants for their contribution to the research work. Data Availability Statement The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. References Associated Data Data Availability Statement The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.