Patterns of change in athletic identity and mental health in elite Polish athletes with disabilities during the COVID-19 pandemic: a 3-month longitudinal study Authors: Piotr K. Urbański, Allen E. Cornelius, Katarzyna Antosiak-Cyrak, Tomasz Tasiemski, Britton W. Brewer Published: 2026 ABSTRACT BackgroundThis study examined patterns of change in athletic identity and their associations with mental health indicators in elite Polish athletes with disabilities during the COVID-19 pandemic. Methods: In a 3-month longitudinal study conducted from April to July 2021, 91 athletes enrolled in the Polish Paralympic Preparation Program twice completed the Athletic Identity Measurement Scale–Third Generation (AIMS-3G), the Hospital Anxiety and Depression Scale (HADS), and the Revised UCLA Loneliness Scale. Using Bonanno’s pooled-SD approach, athletes were classified as showing decreasing, stable, or increasing athletic identity. Group differences in change scores were tested with multivariate analysis of variance and Bonferroni post hoc tests.ResultsMost athletes showed stable athletic identity (77%), whereas decreasing (13%) and increasing (10%) patterns were less common. The multivariate effect was significant, F(6, 172) = 4.30, p < .001, Wilks’ Λ = .76, ηp2 = .13. Athletes with decreasing athletic identity showed greater increases in depression and loneliness than athletes with stable or increasing athletic identity (p ≤ .01).ConclusionsAthletic identity remained predominantly stable, but decreases in athletic identity were associated with worsening depression and loneliness. Brief AIMS monitoring alongside mental health screening may be useful for identifying athletes who may require additional support during prolonged disruptions to sport participation. Background Self-identity pertains to the ways in which people define themselves with regard to the beliefs, goals, and values to which they are strongly committed and that they find personally expressive [1]. In the realm of sport, self-identity has been examined extensively in the form of athletic identity [2], which refers to the extent of identification with the athlete role [3]. Just as people’s conceptions of themselves in general have been demonstrated to be both stable and malleable [4], athletic identity has been conceptualized as consistent over time but susceptible to the influence of contextual factors [5]. Changes in athletic identity Evidence in support of the hypothesized influence of contextual factors on athletic identity has come from studies in which changes in athletic identity following situations where sport participation was threatened or compromised have been documented. For example, sample-level decrements in athletic identity have been reported after sport career termination [6–8], a poor competitive season [9], sport team deselection [10], and anterior cruciate ligament (ACL) reconstruction [11, 12]. When individual trajectories in athletic identity after ACL surgery were examined over periods from 6 to 24 months [13], the most common pattern was stability (i.e., no change, found in 48–68% of participants), followed by a decreasing pattern (found in 19–45% of participants) and an increasing pattern (found in 7–14% of participants). In terms of situational contexts in which sport participation was threatened or compromised, the COVID-19 pandemic was a prime example. Encompassing the SARS-CoV-2 novel coronavirus (COVID-19) and efforts toward mitigating its effects (e.g., travel restrictions, social distancing, quarantines, lockdowns), the pandemic had sport-specific effects such as reduced interaction with coaches and teammates, canceled or delayed competitions, decreased access to training facilities [14], diminished training volume and physical fitness [14, 15], and elevated risk of injury [15] that threatened the sport involvement of many athletes worldwide. Posing a potential threat to what was, for some athletes, a self-defining pursuit [16–18], the pandemic created a context in which changes in athletic identity could occur. Although relatively little research on changes in athletic identity during the pandemic has been conducted, findings from several qualitative studies suggest that the pandemic had a disruptive impact on the athletic identity of competitive athletes [16–18] and resulted in reduced athletic identity for some individuals, stable athletic identity for other individuals, and, infrequently, increased athletic identity [16]. In a quantitative, longitudinal study, Graupensperger [19] reported that 43% of a sample of intercollegiate athletes in the United States had lower levels of athletic identity during the pandemic than their prepandemic (i.e., two months earlier) levels, 28% had unchanged levels, and 29% had higher levels. In another quantitative, longitudinal study, Örencik et al. [20] found that the vast majority (i.e., 92%) of elite Swiss athletes in their sample had athletic identity and psychosocial resource (i.e., perceived social support, resilience, and self-esteem) profiles that remained stable over a 9-month period during the pandemic. Similarly, in a quantitative, longitudinal study of Olympic athletes, sample-level athletic identity did not change over a two-month period during the pandemic [21]. Given the dearth of longitudinal research on differences in athletic identity over the course of the pandemic, the primary purpose of the current study was to examine longitudinally individual patterns of change in athletic identity during the pandemic. Associations between changes in athletic identity and psychological distress A secondary purpose of the current study was to investigate the extent to which changes in athletic identity were associated with changes in selected mental health indicators (i.e., anxiety, depression, and loneliness) during the pandemic. In addition to disruption of self-identity [16–18], psychological effects of the COVID-19 pandemic in athletes included elevations in mental health factors such as depression and anxiety [15, 22–24]. Research on the relationship between athletic identity and mental health in sport participants during the pandemic has yielded inconsistent results. In line with research indicating that athletes whose sense of who they are as individuals is derived strongly from their involvement in sport (i.e., high athletic identity) tend to experience the highest levels of psychological distress when facing transitions that threaten their participation in sport such as injury [3] and sport career termination [25], positive associations between athletic identity and indicators of psychological distress during the COVID-19 pandemic have been documented in four studies [26–28]. In contrast, negative associations between athletic identity and indicators of psychological distress during the COVID-19 pandemic have also been documented in four studies [19, 29, 30]. Athletic identity was unrelated to psychological distress during the COVID-19 pandemic in one additional study [31]. The inconsistency across studies in the associations between athletic identity and psychological distress during the COVID-19 pandemic may be at least partially attributable to the research designs used in the studies. Cross-sectional research designs were used in 7 of the 9 studies in which the relationship between athletic identity and psychological distress during the COVID-19 pandemic was examined, and in neither of the 2 remaining studies in which longitudinal research designs were used were measures of both athletic identity and psychological distress administered longitudinally. In the lone study in which longitudinal athletic identity data were collected during the pandemic, changes in athletic identity were positively related to psychological well-being and negatively related to depression symptoms [19]. The current study To fulfill the aims of the study—examining changes in athletic identity and assessing the extent to which those changes were related to changes in psychological distress during the pandemic—a longitudinal investigation was conducted with a sample of elite athletes with disabilities in which measures of athletic identity and mental health indicators were administered twice over a three-month period in 2021. A method of determining patterns of change after threatening events developed by Bonanno et al. [32] was used to identify individual trajectories in athletic identity. Based on previous findings [16, 19], a full array of athletic identity change patterns (i.e., increasing, stable, decreasing) was expected. Similarly, the results of Graupensperger et al. [19] suggest that a decreasing pattern of athletic identity during the pandemic may be associated with less favorable psychological functioning, particularly lower well-being and greater depressive symptoms. Methods This study aimed to examine patterns of change in athletic identity and their associations with changes in anxiety, depression, and loneliness in elite Polish athletes with disabilities during the COVID-19 pandemic. This was a two-wave longitudinal online study conducted among all athletes enrolled in the Polish Paralympic Preparation Program. Participants Participants were 91 enrollees (60 men and 31 women) in the Polish Paralympic Preparation Program prior to the 2020 Tokyo Summer Paralympic Games (held in 2021) and 2022 Beijing Winter Paralympic Games (held in March 2022). A variety of disabilities were reported by participants: amputation (29%, n = 26), spinal cord injury (22%, n = 20), visual impairment (18%, n = 16), cerebral palsy (7%, n = 6), muscular dystrophy (1%, n = 1), and other (24%, n = 22). The most frequently reported sports in which participants were involved were swimming (17%, n = 15), athletics (15%, n = 14), goalball (13%, n = 12), cycling (12%, n = 11), and fencing (11%, n = 10). Participants reported that: (a) they were an average of 31.02 (SD = 12.04) years of age; (b) an average of 17.85 (SD = 14.48) years had elapsed since their injury occurred or their disease had been diagnosed; and (c) they had an average of 9.56 (SD = 7.92) years of experience as a Paralympian. Measures As part of a broader study of the psychological responses of elite athletes with disabilities during the COVID-19 pandemic, participants provided information on demographic and sport-related variables and completed measures of athletic identity, anxiety, depression, and loneliness. A self-report questionnaire was used to assess demographic and sport-related variables such as age, gender, nature and duration of disability, sport, intended hours of training, actual hours of training, and Paralympic experience. Athletic identity was assessed with the 4-item athletic identity scale from the Athletic Identity Measurement Scale—Third Generation (AIMS-3G) [33]. Responses to the athletic identity scale are given on a Likert-type scale from 1 (strongly disagree) to 7 (strongly agree), with total scores ranging from 4 to 28. Brewer et al. [33] presented evidence in support of the scale’s reliability and validity, including a Cronbach’s alpha coefficient of 0.86 and an intraclass (test–retest) correlation coefficient of 0.85. Cronbach’s alpha values of 0.92 (Time 1) and 0.93 (Time 2) were obtained for the athletic identity scale in the current study. The tool was used among athletes with disabilities [34]. The Revised UCLA Loneliness Scale [35] was used to assess loneliness. The scale features 20 items (e.g., “I feel left out,” “I feel isolated from others,” “I lack companionship”) with response options of 0 (never), 1 (rarely), 2 (sometimes), and 3 (often). Russell et al. presented support for the internal consistency (Cronbach’s alpha = 0.94), concurrent validity, and discriminant validity of the loneliness scale. Cronbach’s alpha values of 0.93 (Time 1) and 0.92 (Time 2) were obtained for the loneliness scale in the current study. The Hospital Anxiety and Depression Scale (HADS) [36] was used to measure anxiety and depression. Used frequently in research on persons with disabilities [37], the HADS features two 7-item subscales that assess anxiety (HADS-A) and depression (HADS-D) using a scale with endpoints of 0 (very rarely) and 3 (often). Thus, HADS-A and HADS-D total scores range from 0 to 21, with higher scores corresponding to greater anxiety and depression, respectively. Cronbach’s alpha values in the current study were 0.81 and 0.72 for the HADS-A and the HADS-D, respectively. Procedure This study was approved by the Ethical Committee of Poznan University of Medical Sciences (KB-742/21) and conducted in accordance with the Declaration of Helsinki. Data were collected during a period of high COVID-19 infection rates in Poland (April 2021; Time 1) and a period of low COVID-19 infection rates in Poland (July 2021; Time 2). In April 2021, all athletes in the Polish Paralympic Preparation Program were recruited for participation electronically and were emailed an online survey link via the Google Forms platform. Reminders to complete the survey were emailed to prospective participants over a two-week period. Participants provided informed consent prior to completing the survey. In April 2021, participants completed demographic/sport items, the AIMS-3G, HADS, and the Revised UCLA Loneliness Scale; in July 2021, the identical battery was re-administered. In light of previous research findings [9, 10], three months was deemed sufficient to detect changes in athletic identity given the turbulence of the COVID-19 pandemic. Data analysis Descriptive statistics were computed for the demographic and sport-related variables, the AIMS-3G athletic identity scale, the Revised UCLA Loneliness Scale, and the HADS. Changes in athletic identity were determined through methods that Bonanno et al. [32] used to study changes in depression after bereavement and adapted to examine changes in emotional responses [38] and athletic identity [13] after knee surgery. Specifically, participants’ Time 2 athletic identity scores were compared with their Time 1 athletic identity scores. Participants whose Time 1 athletic identity scores exceeded their Time 2 athletic identity scores by an amount larger than the pooled standard deviation for the full sample’s (Time 2 athletic identity – Time 1 athletic identity) difference score were deemed to have shown a decrease in athletic identity. Participants whose Time 2 athletic identity scores exceeded their Time 1 athletic identity scores by an amount larger than the pooled standard deviation for the full sample’s (Time 2 athletic identity – Time 1 athletic identity) difference score were deemed to have shown an increase in athletic identity. Participants whose Time 2 athletic identity scores exceeded their Time 1 athletic identity scores by an amount smaller than the pooled standard deviation for the full sample’s (Time 2 athletic identity – Time 1 athletic identity) difference score were deemed to have a stable athletic identity pattern. To examine the extent to which changes in athletic identity from Time 1 to Time 2 were associated with changes in anxiety, depression, and loneliness over the same period, a one-way multivariate analysis of variance (MANOVA) was conducted. Time 2-Time 1 difference scores for the HADS-A, HADS-D, and loneliness scale were compared across the increased athletic identity, decreased athletic identity, and stable athletic identity groups. Significant effects were followed up with Bonferroni post hoc tests. Results Means and standard deviations of HADS-A, HADS-D, and loneliness scale scores are presented in Table 1. The mean athletic identity scale change score from Time 1 to Time 2 was −0.47 (SD = 3.46). Consequently, participants whose athletic identity scale scores increased by 4 or more were considered to have had an increasing pattern of athletic identity, participants whose athletic identity scale scores increased or decreased by less than 4 were considered to have had a stable pattern of athletic identity, and participants whose athletic identity scale scores decreased by 4 or more were considered to have had a decreasing pattern of athletic identity.Table 1Means (and Standard Deviations) of HADS-A, HADS-D, and loneliness scale scores for patterns of change in athletic identity categoriesPattern of Change in Athletic Identity CategoryVariableDecreased athletic identity (n = 12)Stable athletic identity (n = 70)Increased athletic identity (n = 9)HADS-AApril 20216.58 (4.62)5.66 (3.66)5.78 (3.27)July 20216.58 (4.62)4.69 (3.61)3.89 (2.98)HADS-DApril 20215.17 (4.15)3.76 (3.01)3.22 (3.53)July 20218.08 (4.38)3.66 (3.19)2.89 (3.37)Loneliness ScaleApril 202128.00 (6.85)35.04 (11.35)40.11 (14.96)July 202132.58 (9.26)33.51 (10.89)33.47 (10.74)HADS-A Hospital Anxiety and Depression Scale–Anxiety, HADS-D Hospital Anxiety and Depression Scale–Depression Table 1Means (and Standard Deviations) of HADS-A, HADS-D, and loneliness scale scores for patterns of change in athletic identity categoriesPattern of Change in Athletic Identity CategoryVariableDecreased athletic identity (n = 12)Stable athletic identity (n = 70)Increased athletic identity (n = 9)HADS-AApril 20216.58 (4.62)5.66 (3.66)5.78 (3.27)July 20216.58 (4.62)4.69 (3.61)3.89 (2.98)HADS-DApril 20215.17 (4.15)3.76 (3.01)3.22 (3.53)July 20218.08 (4.38)3.66 (3.19)2.89 (3.37)Loneliness ScaleApril 202128.00 (6.85)35.04 (11.35)40.11 (14.96)July 202132.58 (9.26)33.51 (10.89)33.47 (10.74)HADS-A Hospital Anxiety and Depression Scale–Anxiety, HADS-D Hospital Anxiety and Depression Scale–Depression Means (and Standard Deviations) of HADS-A, HADS-D, and loneliness scale scores for patterns of change in athletic identity categories HADS-A Hospital Anxiety and Depression Scale–Anxiety, HADS-D Hospital Anxiety and Depression Scale–Depression As shown in Fig. 1, the majority of the participants (n = 70, 77%) displayed a stable pattern of athletic identity, with smaller portions of the sample displaying decreasing (n = 12, 13%) and increasing (n = 9, 10%) patterns of athletic identity.Fig. 1Individual AIMS-3G score changes from Time 1 to Time 2 for athletes classified as decreasing, stable, or increasing in athletic identity Individual AIMS-3G score changes from Time 1 to Time 2 for athletes classified as decreasing, stable, or increasing in athletic identity Results of the MANOVA examining Time 2-Time 1 difference scores for the HADS-A, HADS-D, and loneliness scale across the three patterns of change in athletic identity revealed a significant multivariate effect, F(6, 172) = 4.30, p < 0.001, Wilks’ Λ = 0.76, partial eta-squared = 0.13. Significant univariate effects were found for the HADS-D, F(2, 88) = 7.82, p < 0.001, partial eta-squared = 0.15, and loneliness scale, F(2, 88) = 7.20, p = 0.001, partial eta-squared = 0.14, difference scores, but not for the HADS-A, F(2, 88) = 1.23, p = 0.30, partial eta-squared = 0.03, difference scores. Bonferroni post hoc tests indicated that the HADS-D scores of participants with a decreasing pattern of athletic identity increased from Time 1 to Time 2 to a significantly greater extent than those of participants with stable (p < 0.001) and increasing (p = 0.01) patterns of athletic identity. Similar findings were obtained for loneliness scale scores, which increased from Time 1 to Time 2 to a significantly greater extent for participants with a decreasing pattern of athletic identity than for participants with stable (p = 0.01) and increasing (p = 0.001) patterns of athletic identity. Discussion In the present investigation, changes in athletic identity during the COVID-19 pandemic were examined through a longitudinal research design in which athletic identity and mental health data were collected twice from elite athletes with disabilities over a 3-month assessment period. The most common individual pattern was one of stable athletic identity, with patterns of decreasing athletic identity and increasing athletic identity occurring less frequently. The decreasing pattern of athletic identity was associated with greater increases in depression and loneliness than the stable and increasing patterns of athletic identity during the assessment period. Large effect sizes were obtained for both depression and loneliness. The prominence of the stable pattern of athletic identity is consistent with the findings of Örencik et al. [20] and with the tenets of continuity theory [39], which holds that people interpret transitional situations such as those precipitated by the pandemic and its effects in light of their previous views of themselves and their environment, implementing coping strategies that allow them to preserve their internal psychological characteristics and external relationships with others. For the majority of participants, stability of athletic identity may have provided a sense of continuity in themselves while experiencing the challenges of the COVID-19 pandemic. Maintaining consistency in athletic identity may have had adaptive value, as increases in anxiety, depression, and/or loneliness over the assessment period were not found for the stable pattern of athletic identity group. Despite the prevalence of the stable pattern of athletic identity, nearly one-quarter of participants demonstrated either an increase or a decrease in athletic identity over a three-month period during the pandemic, corroborating research documenting pandemic-related turbulence in self-identity [16, 19]. Although the percentages of increasing and decreasing athletic identity patterns observed in the current study differed markedly from those reported previously [19], these differences may be largely attributable to the use of different criteria for defining change patterns. In the earlier study, increasing and decreasing patterns were classified on the basis of any nonzero change in athletic identity between the first and second assessments. In contrast, more stringent definitions of decreasing and increasing were used in the current study, requiring athletic identity change scores that were one standard deviation below or above the full sample’s mean change score, respectively. For participants who demonstrated a pattern of increasing athletic identity, expansion of athletic identity may have reflected an attempt to diminish the threat to athletic identity posed by the pandemic or to derive motivational benefit to offset pandemic-related restrictions in training. The inconsistency of the circumstances of the pandemic with participants’ self-defining goals may have prompted a “doubling down” on athletic identity by a tenth of the sample, a response adopted by some athletes after sustaining a severe sport injury [13]. Although such a strategy may be viable or effective for only a brief period of time [13, 40], it is consistent with a model positing that increased commitment to domains of personal or social importance is a defensive response to the uncertainty of a threat that is looming or unresolvable [41]. For participants who demonstrated a pattern of decreasing athletic identity, reduction of athletic identity may have reflected an attempt to maintain a favorable self-image by distancing themselves psychologically from an important—yet threatened—domain of self-identity [42]. Diminution of athletic identity may, therefore, have made the threat to sport participation posed by the pandemic less relevant to self-definition and self-evaluation [43]. The decreases in athletic identity reported in the current study mirror those documented both during the pandemic [16, 19] and after other events that posed a threat to sport involvement [6–8, 13]. Of the three patterns of change in athletic identity, the decreasing pattern was the only one associated with change in the mental health indicators over the assessment period. Participants who decreased their athletic identity exhibited increases in both depression and loneliness, whereas those who maintained a stable athletic identity or increased their athletic identity displayed no Time 1-Time 2 differences in anxiety, depression, and loneliness. This result is consistent with the findings that decreased athletic identity was associated with less favorable mental health during the pandemic [19], sport performance during a competitive season [33], and physical symptoms and functioning after knee surgery [13]. Because the current study employed a correlational (albeit longitudinal) design, the direction of the relationship between decreases in athletic identity and increases in depression and loneliness cannot be determined. Nevertheless, one possible interpretation is that when athletes experience poor psychological, physical, or sport-related functioning, they tend to reduce their investment in being (and viewing themselves as being) an athlete, potentially leaving themselves open to invest in other sources of self-definition. It is also possible that individual differences in personality may partly shape how athletes respond to threats to sport participation and identity; for example, higher neuroticism may be linked to greater vulnerability to depressive symptoms and social distress in periods of disruption [44]. However, because personality traits were not assessed in the present study, the role of neuroticism in this pattern should be treated with caution and examined directly in future research. Limitations Although the elite Paralympic sample and longitudinal design are strengths of this study, several limitations should be considered. First, the three-month interval was appropriate given the intensity of pandemic-related disruptions, but the two-wave design still limited the ability to capture more complex, non-linear, or longer-term trajectories of change. Second, the findings were based exclusively on self-report questionnaires, which are susceptible to bias, including recall bias and social desirability bias. Future studies could include additional assessment methods beyond self-report. Third, the classification into decreasing, stable, and increasing athletic identity groups was based on a pooled standard deviation criterion. Although this approach was grounded in previous research on patterns of change after threatening or disruptive events [13, 32, 38], it may have oversimplified individual variability by assigning athletes to discrete categories and may not have captured more subtle or heterogeneous patterns of change. Fourth, the correlational design precludes causal inference; therefore, the association between decreased athletic identity and worsening depression and loneliness should be interpreted cautiously. Mental health changes or unmeasured contextual factors may also have contributed to the observed pattern. Fifth, although the pandemic was global, infection rates and mitigation efforts varied locally over time. Thus, the present context differed from other pandemic studies of athletic identity change [19–21]. Nevertheless, the disruptive effects of the pandemic on psychological functioning [22] and athletic identity [16–18] are well documented. Sixth, only the sport-related aspect of self-identity was examined. Some athletes may have responded to reduced sport opportunities by shifting from an exclusive athletic identity toward a more diverse identity structure [16]. Future research should therefore assess other aspects of self-identity beyond sport. Finally, the findings may have limited generalizability beyond elite Polish athletes with disabilities enrolled in the Paralympic Preparation Program. Replication across countries, competitive levels, and sport systems is needed. Directions for future research Future research should assess aspects of self-identity beyond the sport domain, as some athletes may respond to reduced opportunities for sport involvement by reorganizing their priorities and shifting away from an exclusive athletic identity toward a more diverse sense of self [16]. This would help determine whether athletes who show decreases in athletic identity reallocate identity-related resources to other valued roles or domains. Future studies may also benefit from using multi-wave designs to examine how such identity shifts unfold over time. Practical implications These findings suggest that changes in athletic identity may be clinically and practically relevant during prolonged disruptions to sport participation. In particular, decreases in athletic identity were accompanied by less favorable changes in depression and loneliness. Monitoring athletic identity together with basic mental health indicators may therefore help practitioners identify athletes who may require additional support during periods of uncertainty, reduced training opportunities, or disrupted competition schedules. Conclusion This longitudinal study demonstrates that athletic identity remained predominantly stable among elite Polish para-athletes during the COVID-19 pandemic. Although less common, a decreasing pattern of athletic identity was associated with increases in depression and loneliness. These findings extend prior longitudinal research on athletic identity dynamics during the pandemic [19, 20] and align with evidence linking identity change to psychological adjustment following sport-related threats [13]. The results suggest that monitoring athletic identity alongside routine mental health screening may help practitioners identify athletes at greater risk of psychological distress during prolonged disruptions to sport participation. Future multi-wave studies are needed to clarify the directionality of these relationships, evaluate intervention strategies, and assess the generalizability of the observed patterns beyond the pandemic context.