Social identity

Social identity, as a primarily externally oriented dimension in Cieslak II (5) framework, reflects how individuals perceive themselves as athletes through the lens of social recognition and their integration within the sporting community [11]. This dimension encompasses both personal identification with the athletic role and recognition from others [12]. Recent research has revealed that social identity in sports extends beyond mere role recognition, influencing team cohesion, leadership dynamics, and performance outcomes [13]. Athletes with strong social identity demonstrate enhanced commitment to training and competition [4], while also showing greater resilience in facing competitive challenges [14]. In team sports like basketball, social identity plays a particularly crucial role as athletes must integrate individual and collective identities while pursuing both personal and team objectives [15].

Exclusivity

The exclusivity dimension represents the degree to which an individual#39;s self-worth is predominantly determined by their athletic performance [16]. This dimension has emerged as a critical factor in understanding both the benefits and potential risks of strong athletic identification. While high levels of exclusivity often correlate with enhanced motivation and dedication to athletic excellence [4,17], they may also create vulnerabilities during career transitions or injury recovery periods [18]. Recent longitudinal studies have demonstrated that athletes with high exclusivity scores show greater dedication to training but may face significant challenges in developing alternative identities or preparing for post-athletic careers [19,20]. The relationship between exclusivity and psychological well-being appears to be curvilinear, with moderate levels associated with optimal outcomes while extreme levels may predict adjustment difficulties [21].

Self-Identity

Self-identity, representing a core internal dimension in Cieslak II (5) framework, captures the strength of athletic self-reference in an individual#39;s overall self-concept [22] and constitutes the cognitive structure through which athletes process information and experiences from a self-referential perspective [23]. Contemporary research has expanded our understanding of athletic self-identity#39;s role in psychological functioning, demonstrating its influence on goal-setting behaviors, motivation patterns, and resilience development [24]. Athletes with well-developed self-identity often exhibit higher levels of sports commitment and achievement motivation [25,26]. However, the development of self-identity must be balanced with other aspects of personal growth to prevent overidentification with the athletic role [27]. Longitudinal research has revealed that the stability of athletic self-identity varies across developmental stages, with adolescence representing a particularly critical period for its formation and consolidation [10].

Negative affectivity

Negative affectivity in athletic identity represents emotional responses to adverse sporting experiences, including performance setbacks, injuries, or career transitions [3]. Recent research has significantly expanded our understanding of this dimension#39;s role in athlete well-being and performance. Studies indicate that athletes with high negative affectivity may experience more significant psychological distress during challenging periods, particularly following injuries or during competitive slumps [28]. Longitudinal investigations have revealed that elevated negative affectivity can predict increased vulnerability to depression and anxiety, especially during major career transitions or following serious injuries [29,30]. Moreover, contemporary research has identified important mediating factors that influence the relationship between negative affectivity and psychological outcomes, including social support systems, coping strategies, and resilience mechanisms [31].

Positive affectivity

Positive affectivity reflects the enjoyment, satisfaction, and positive emotional experiences derived from athletic participation [5]. This dimension has emerged as a critical factor in understanding sustained athletic engagement and performance excellence. Recent studies have demonstrated that positive affectivity contributes significantly to athlete motivation, engagement, and performance outcomes through various psychological mechanisms [32]. Athletes with high positive affectivity typically demonstrate greater resilience in facing challenges and maintain stronger commitment to their athletic pursuits [33]. Research has also revealed important relationships between positive affectivity and flow states, optimal performance experiences, and enhanced recovery from setbacks [34]. Furthermore, longitudinal studies have shown that positive affectivity plays a crucial role in buffering against burnout and maintaining long-term athletic engagement [35].

Cross-cultural validation of athletic identity measurement

The universality of the athletic identity construct has been progressively supported through cross-cultural validation studies employing measurement scales across diverse cultural and linguistic contexts. The original Athletic Identity Measurement Scale (AIMS; Brewer, Van Raalte [3]) was initially developed and validated among Western populations and has since served as the foundation for numerous cross-cultural adaptations. Martin, Eklund [22] provided early evidence for the scale#39;s applicability beyond its original context by validating it among athletes with disabilities. Subsequently, the scale has been adapted and validated in a range of non-Western settings: Sohrabi and Shajie [36] validated the AIMS-Plus among Iranian physical education students, supporting a three-factor structure; Li [37] confirmed the original factor structure in a Hong Kong sample; and Kamran and Rafique [38] supported a five-factor structure among Pakistani athletes. In mainland China, Zhang and Shao [39] validated the original AIMS, providing initial evidence for the construct#39;s relevance in the Chinese sporting context, though this version assessed only the cognitive dimensions without affective components.

Collectively, these cross-cultural studies demonstrate that the core dimensions of athletic identity are recognizable across cultural boundaries, supporting the construct#39;s theoretical universality. However, they also reveal culturally specific variations in factor structures, inter-dimensional relationships, and psychometric properties, underscoring the necessity of culturally adapted measurement tools. The present study extends this body of cross-cultural evidence by validating the more comprehensive AIMSP—which incorporates both cognitive and affective dimensions—among adolescent athletes in mainland China, a population and cultural context that remain underrepresented in the athletic identity measurement literature.

Research objectives

While Zhang and Shao [39] validated the AIMS in mainland China, the AIMSP#39;s five-dimensional structure, including affective components absent in AIMS, requires separate validation. The AIMSP provides a more comprehensive theoretical framework by incorporating both cognitive and emotional aspects of athletic identity [5]. The validation of AIMSP-C is particularly crucial given China#39;s unique sports system that integrates athletic training with academic education. This “sports-education integration” system creates distinctive challenges for young athletes developing their athletic identity, especially in team sports like basketball, where players must balance multiple roles.

The present study aims to validate the Chinese version of AIMSP (AIMSP-C) among adolescent basketball players through examining its structural validity using confirmatory factor analysis, evaluating its reliability and convergent validity, and assessing its cultural appropriateness within the Chinese sporting context. A successfully validated AIMSP-C would provide researchers and practitioners with a culturally appropriate tool for understanding and supporting Chinese athletes’ psychological development, particularly during critical developmental periods. Furthermore, this validation would contribute to both theoretical understanding of how cultural factors influence athletic identity formation and practical applications in athlete development programs within China#39;s systematic sports training system.

Participants

A cross-sectional study design was employed. Validation study was conducted to examine the psychometric properties of the Chinese version of the Athletic Identity Measurement Scale Plus (AIMSP-C). Participants were recruited through random sampling from secondary school basketball programs across Shandong Province, China. The sampling strategy ensured representation from diverse competitive levels and training backgrounds.

Eligibility criteria for participation included: (1) current enrollment in secondary school basketball programs; (2) active involvement in organized basketball training and competition; (3) proficiency in reading and comprehending Mandarin Chinese; and (4) voluntary participation with informed consent. Students were excluded if they reported major health conditions interfering with regular basketball participation within the previous six months or lacked competitive experience at minimum city-level competitions.

Sample Size Determination Sample size requirements for confirmatory factor analysis were calculated using Arifin (40) web-based calculator, incorporating parameters established by contemporary structural equation modeling research. The calculation considered the following specifications: anticipated CFI of 0.95, five-factor structure with 4–5 items per factor, expected factor loadings of 0.6, inter-factor correlations of 0.25, significance level (α) of 0.05 (two-tailed), and desired power (1 – β) of 90%. Accounting for potential attrition (15%), the minimum required sample size was determined to be 478 participants. The achieved sample size (N = 604) substantially exceeded this threshold, ensuring robust parameter estimation and stable factor solutions.

Measures

Ethics and procedures

The study received ethical approval from the Universiti Sains Malaysia Human Research Ethics Committee (USM/JEPeM/22050298). The adaptation of the AIMSP into Chinese (AIMSP-C) followed a rigorous translation and validation process adhering to established cross-cultural adaptation guidelines.

It is important to clarify that the AIMSP-C was translated directly from the original English version of the AIMSP as proposed by Cieslak II (5), rather than from any existing Chinese adaptation of related athletic identity scales. This decision was made for three specific reasons: (1) to ensure fidelity to the original five-factor conceptual framework as intended by the scale developer; (2) to maintain the integrity of the complete 22-item structure, including both affective dimensions (Positive Affectivity and Negative Affectivity) that are absent in earlier AIMS versions and their Chinese adaptations; and (3) to avoid potential error accumulation that could arise from translating an already adapted version. It should be noted that the AIMS validated by Zhang and Shao (39) in mainland China assessed a narrower construct using a three-factor model without affective components, whereas the present study employs the full five-dimensional AIMSP framework. This distinction is important for interpreting both the construct validity and the cross-cultural comparability of our findings.

The translation process began with two independent forward translations by bilingual researchers proficient in both English and Mandarin Chinese. These researchers then collaborated to synthesize their translations into an initial Chinese draft through detailed comparison and discussion. A panel of experts, comprising three sports psychologists, two sports science researchers, and two physical education specialists, evaluated this draft version for content validity and cultural appropriateness. Their feedback led to necessary refinements in terminology and expression.

Following expert review, two different bilingual researchers conducted back-translation of the revised Chinese version into English. This back-translated version was compared with the original AIMSP to ensure conceptual and semantic equivalence. Discrepancies were resolved through iterative discussion and modification until achieving satisfactory concordance with the original scale. To enhance face validity, the refined translation underwent cognitive debriefing with a group of twelve native Chinese speakers, who provided feedback on clarity, comprehensibility, and cultural relevance. A subsequent pilot study with fifteen adolescent basketball players confirmed the scale#39;s readability and practical utility.

Data collection occurred between February 2023 and July 2023. The research team distributed informed consent forms to potential participants and their parents or legal guardians, explaining the study#39;s purpose, voluntary nature, and confidentiality measures. Upon obtaining written consent, participants completed the AIMSP-C questionnaire through the Sojump platform, accessing it via personal or parental electronic devices. Prior to administration, researchers provided standardized instructions emphasizing honest responses and the absence of right or wrong answers. The entire data collection process adhered to ethical guidelines for research involving adolescent participants.

Test-retest reliability was assessed using a randomly selected subset of 55 participants from the original sample, who completed the AIMSP-C twice with a 2-week interval between administrations. This sample size was determined using Arifin [40] calculator based on expected ICC of 0.80, minimum acceptable ICC of 0.60, with 80% power and 10% dropout adjustment.

Statistical analysis

Initial data screening examined univariate and multivariate normality using SPSS 28.0 (IBM Corp, Armonk, NY, USA). The Kolmogorov-Smirnov and Shapiro-Wilk tests assessed univariate normality [41], while Mardia#39;s tests evaluated multivariate normality through skewness and kurtosis coefficients. Visual inspection of histograms and Chi-square versus Mahalanobis distance plots complemented these analyses following recommendations by Tabachnick, Fidell [42].

Confirmatory Factor Analysis (CFA) was conducted using Mplus 8.7 with Maximum Likelihood Robust (MLR) estimator, selected due to the non-normal distribution of the data as recommended by Muthén and Muthén [43]. Model fit was evaluated using multiple indices with criteria established by Hu and Bentler [44]: Comparative Fit Index (CFI > 0.95), Tucker-Lewis Index (TLI > 0.95), Root Mean Square Error of Approximation (RMSEA < 0.06, with 90% confidence intervals), values below 0.10 considered acceptable following Browne and Cudeck [45], and Standardized Root Mean Square Residual (SRMR < 0.08). Factor loadings exceeding 0.40 were considered acceptable following Hair, Sarstedt [46].

Model modifications were implemented using a conservative stepwise approach. Residual correlations were added sequentially based on modification indices, with the modification process stopped when acceptable fit criteria (CFI ≥ 0.95, TLI ≥ 0.95) were achieved to maintain model parsimony and avoid overfitting.

Modification indices (MI) were examined to guide model improvements, with values exceeding 10.0 considered for potential model modifications, as recommended by Byrne [47]. Only theoretically justifiable modifications were implemented, prioritizing residual correlations between items within the same factor or conceptually related items across factors. The largest MI values were addressed first in the sequential modification process, with each modification evaluated for both statistical improvement and theoretical coherence before implementation.

Construct validity assessment included both convergent and discriminant validity. Convergent validity was established through factor loadings (> 0.50), Average Variance Extracted (AVE > 0.50), and Composite Reliability (CR > 0.70) as recommended by Hair, Sarstedt [46]. Discriminant validity was evaluated using the Fornell-Larcker criterion [48], requiring the square root of AVE for each construct to exceed its correlations with other constructs.

Reliability analysis encompassed internal consistency and temporal stability measures. Internal consistency was assessed through CR, with values above 0.70 indicating good reliability [49]. Test-retest reliability was evaluated using two-way mixed effects Intraclass Correlation Coefficients (ICC) as recommended by Koo and Li [50], with a 2-week interval between test and retest administrations, with values interpreted as: poor (< 0.50), moderate (0.50–0.75), good (0.75–0.90), and excellent (> 0.90).

Descriptive statistics

A total of 604 secondary school students participated in the study, with an almost equal gender distribution of 301 males (49.8%) and 303 females (50.2%). Participants ranged in age from 12 to 19 years, with a mean age of 15.53 years (SD = 1.42). Most participants were in the senior grades: Junior 1 (5.6%), Junior 2 (4.3%), Junior 3 (6.5%), Senior 1 (28%), Senior 2 (26%), and Senior 3 (29.6%). Regarding sport levels, 28.3% had no specific sport level, 30.5% were at the junior grade level, 23.2% at level 3, 11.1% at level 2, 5% at level 1, and 2% were masters. Participants’ training years varied, with 11.9% having less than one year of training, 34.6% with one year, 31.1% with two years, 14.6% with three years, and 7.8% with more than three years, as detailed in Table 1.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Demographic Information and Frequency of Participants.

https://doi.org/10.1371/journal.pone.0345181.t001

Table 2 shows the distribution of the items’ scores for the Chinese version of the Athletic Identity Measurement Scale - Plus (AIMSP). The mean (SD) scores for all items ranged between 4.58 (1.81) for AIMSP15 and 6.92 (3.06) for AIMSP6.In the 11 options, 1 = Strongly Disagree, 6 = Neutral, and 11 = Strongly Agree. Most participants provided their responses in the higher categories of the scale. Specifically, the highest percentages of participants answered “1” for AIMSP19 (9.6%), “2” for AIMSP21 (14.6%), “3” for AIMSP15 (12.3%), “4” for AIMSP19 (26.5%), “5” for AIMSP15 (24.7%), “6” for AIMSP10 and AIMSP16 (both 24.7%), “7” for AIMSP4 (30.3%), “8” for AIMSP5 (33.1%), “9” for AIMSP7 (14.1%), “10” for AIMSP6 (16.4%), and “11” for AIMSP7 (12.9%).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Distribution of the items’ score for Chinese version of AIMSP scale.

https://doi.org/10.1371/journal.pone.0345181.t002

Table 2 presents the descriptive statistics for all AIMSP-C items. Most items showed approximately normal distribution, with skewness values ranging from −0.48 to 0.58 and kurtosis values from −1.22 to −0.12. Response patterns indicated that participants generally used the full range of the 11-point scale, with mean scores ranging from 4.58 (AIMSP15) to 6.92 (AIMSP6).

Confirmatory factor analysis

Confirmatory Factor Analysis revealed that the initial model of AIMSP-C demonstrated suboptimal fit indices (CFI = 0.929, TLI = 0.918, SRMR = 0.064, RMSEA = 0.131 [90% CI: 0.126, 0.136]), as illustrated in Fig 1. Model modifications were implemented using a conservative stepwise approach based on modification indices. Item AIMSP6 was removed due to poor factor loading (0.187), which fell well below the acceptable threshold of 0.40 recommended by Hair, Sarstedt [46]. Additionally, examination of the item content “I need to participate in sport to feel good about myself” revealed potential cross-loading issues with other dimensions and cultural adaptation concerns in the Chinese context. The item#39;s removal was further supported by modification indices suggesting poor model fit contribution and its minimal impact on the Social Identity factor#39;s overall reliability. Subsequently, residual correlations were added sequentially until acceptable fit indices were achieved. The final model included two residual correlations between theoretically related items (AIMSP14-AIMSP3, AIMSP20-AIMSP5), stopping when CFI and TLI values reached approximately 0.95 to maintain model parsimony and avoid overfitting.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. CFA Diagram of the AIMSP-C Initial model.

Note: soc = Social Identity, exc = Exclusivity, sel = Self-Identity, neg = Negative Affectivity, pos = Positive Affectivity..

https://doi.org/10.1371/journal.pone.0345181.g001

The modified model demonstrated acceptable fit indices (CFI = 0.959, TLI = 0.952, SRMR = 0.083, RMSEA = 0.091 [90% CI: 0.086, 0.096]), meeting established criteria for acceptable model fit while maintaining model parsimony (Table 3). The CFI, TLI, and SRMR met their respective thresholds, while the RMSEA value of 0.091 fell within the acceptable range below 0.10 [45], supporting the structural validity of the final AIMSP-C model (Table 4, Fig 2).

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. CFA Fit Indices for the AIMSP-C (Initial and Final models).

https://doi.org/10.1371/journal.pone.0345181.t003

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Factor Loadings of the AIMSP-C for Initial and Final Models.

https://doi.org/10.1371/journal.pone.0345181.t004

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. CFA Diagram of the AIMSP-C Final model.

Note: soc = Social Identity, exc = Exclusivity, sel = Self-Identity, neg = Negative Affectivity, pos = Positive Affectivity..

https://doi.org/10.1371/journal.pone.0345181.g002

Convergent and discriminant validity

The AIMSP-C demonstrated strong psychometric properties across all dimensions. Factor loadings in the final model remained robust, ranging from 0.906 to 0.997 across all items, indicating strong item-factor relationships. The Social Identity dimension showed high loadings (0.921–0.981), as did Exclusivity (0.958–0.987), Self-Identity (0.904–0.993), Negative Affectivity (0.972–0.976), and Positive Affectivity (0.906–0.997). The minor variations between initial and final models’ factor loadings suggest stability in the measurement structure.

Construct reliability analysis revealed exceptional internal consistency, with composite reliability (CR) values ranging from 0.974 to 0.989 across all factors. Internal consistency was further evaluated using Cronbach#39;s alpha and McDonald#39;s omega coefficients. For the final model, Cronbach#39;s alpha values were: Social Identity (α = 0.974), Exclusivity (α = 0.989), Self-Identity (α = 0.979), Negative Affectivity (α = 0.986), and Positive Affectivity (α = 0.982). McDonald#39;s omega coefficients showed nearly identical patterns: Social Identity (ω = 0.974), Exclusivity (ω = 0.989), Self-Identity (ω = 0.979), Negative Affectivity (ω = 0.986), and Positive Affectivity (ω = 0.983). Convergent validity was established through Average Variance Extracted (AVE) values consistently exceeding 0.50 (range: 0.904–0.949), indicating that the items effectively represent their respective constructs (Table 5). All reliability coefficients exceeded recommended thresholds, with values greater than 0.97 across all dimensions, demonstrating excellent internal consistency (Table 6).

Download:

• PNG
  larger image
• TIFF
  original image

Table 5. Correlation Matrix and Discriminant Validity of the AIMSP-C (Initial and Final models).

https://doi.org/10.1371/journal.pone.0345181.t005

Download:

• PNG
  larger image
• TIFF
  original image

Table 6. Internal Consistency Reliability Coefficients for AIMSP-C Final Model.

https://doi.org/10.1371/journal.pone.0345181.t006

Test-retest reliability

The test-retest reliability of the AIMSP-C was evaluated using the Intraclass Correlation Coefficient (ICC) values for the five subscales, indicating excellent stability over time. The ICC values were as follows: Social Identity (0.936), Exclusivity (0.989), Self-Identity (0.985), Negative Affectivity (0.988), and Positive Affectivity (0.973). These values demonstrate strong reliability, with ICC values greater than 0.70 indicating good reliability and values greater than 0.90 indicating excellent reliability (Table 7).

Download:

• PNG
  larger image
• TIFF
  original image

Table 7. Intraclass Correlation Coefficients (ICC) for the AIMSP-C.

https://doi.org/10.1371/journal.pone.0345181.t007

Theoretical interpretation of residual correlations

The two residual correlations included in the final model (AIMSP14-AIMSP3, AIMSP20-AIMSP5) can be theoretically justified within identity theory frameworks. The correlation between AIMSP14 (“Others see me mainly as an athlete”) and AIMSP3 (“Most of my friends are athletes”) reflects the interconnected nature of social identity formation, where external recognition and peer networks mutually reinforce each other. Similarly, the correlation between AIMSP20 (“I spend more time thinking about sport than anything else”) and AIMSP5 (“Sport is the most important part of my life”) represents shared variance in exclusivity-related cognitions, reflecting the cognitive overlap in how athletes prioritize sport in their identity hierarchy.

Psychometric properties and cross-cultural validation

The psychometric properties of AIMSP-C demonstrated robust characteristics across cultures. Compared to the Iranian validation study [36], which yielded a three-factor structure with internal consistency of 0.88 among 384 physical education majors, our study achieved superior reliability indices (CR: 0.974–0.989) and validity indicators (AVE: 0.904–0.947). Similarly, while [37] Hong Kong validation maintained the original factor structure, our mainland China sample showed stronger psychometric properties, suggesting potential regional variations in athletic identity manifestation within Chinese cultural contexts. The Pakistani validation [38] also supported a five-factor structure but with relatively lower fit indices (χ²/df = 2.69, CFI = 0.96, TLI = 0.98), possibly due to their smaller sample size.

Cultural manifestations of athletic identity

A distinctive finding emerged in the relationship patterns among the five dimensions. Social identity showed strong correlations with both self-identity and exclusivity, supporting [39] findings regarding the interconnected nature of identity components in Chinese athletes. This pattern reflects the collective orientation in Chinese sports culture, where individual athletic identity development is closely tied to social recognition and team integration. The exclusivity dimension demonstrated particularly interesting characteristics in our sample. While maintaining strong psychometric properties, its relationship with other dimensions suggests a uniquely Chinese manifestation of athletic role commitment. This finding extends our understanding of how cultural context influences the balance between athletic and other life roles, a consideration particularly relevant in China#39;s sports-education integration system.

The affective components (both positive and negative) showed higher inter-correlations than typically reported in Western samples, suggesting a more integrated emotional experience among Chinese athletes. This finding aligns with research on cultural differences in emotional processing and expression, particularly in athletic contexts where performance pressure and achievement expectations are culturally mediated. Of particular significance was the exceptional test-retest reliability demonstrated by AIMSP-C, with ICC values ranging from 0.936 to 0.989 across dimensions. This temporal stability is crucial for longitudinal tracking of athletic identity development, especially during the critical adolescent period when identity formation intersects with intensive athletic training.

Practical applications and limitations

From a practical standpoint, the validated AIMSP-C provides several important applications. It offers coaches and sport psychologists a reliable instrument for evaluating athletic identity development in Chinese youth athletes, enables targeted interventions addressing specific aspects of athletic identity, and demonstrates the scale#39;s sensitivity to Chinese cultural nuances while maintaining measurement rigor. While this study establishes strong support for AIMSP-C#39;s utility, certain limitations warrant consideration. The current validation focused primarily on internal structure and reliability; future research should examine convergent and divergent validity through correlations with established measures of related constructs (e.g., self-esteem, sport commitment). Additionally, criterion validity should be assessed by examining relationships with objective performance measures and behavioral indicators of athletic identity. The sample composition, particularly the high proportion of participants without specific competitive levels (28.3%), suggests the need for additional validation with elite athletes. Furthermore, the geographic concentration in one region may limit generalizability across China#39;s diverse sporting contexts.

Future directions

Future research should include longitudinal studies tracking athletic identity development across career stages, cross-validation studies with elite athletes, investigation of regional variations within China, and examination of sport-specific identity patterns. In conclusion, the successful validation of AIMSP-C provides a robust tool for understanding athletic identity in Chinese youth athletes while contributing to the broader theoretical framework of athletic identity measurement across cultures. The findings support the universal aspects of athletic identity while highlighting culturally specific manifestations, particularly in the Chinese sporting context where athletic development is uniquely integrated with academic education.