Sport team identification: a social identity perspective comparing local and distant fans

Purpose – Supportingdistantteamsisafrequentphenomenon.Throughthelensofthe socialidentitytheory, this research aims to examine differences between local and distant fans regarding drivers of team identification. Design/methodology/approach – A multigroup structural equation model was employed. The data were collected through an online survey with 1,285 sports fans. Findings – Team distinctiveness constitutes an important aspect fueling identification for all fans, whereas congruence between own and team personality is important for local and displaced fans only. Team prestige does not impact identification for either group. Practical implications – To build up a base of highly identified supporters, clubs should emphasize those aspects of team brands that fans consider distinctive. When targeting local fans, clubs should also focus on communicating the brand ’ s unique personality aspects. Originality/value – This is the first study that assesses the potential differences behind fans ’ social identification with local and distant teams.

Research emphasizes the role of sport clubs as carriers of local identity (G omez-Bantel, 2016) and establishes a strong connection of geographic identity and identification with the local team (Collins et al., 2016;Uhlman and Trail, 2012). However, developments in sports broadcasting have reduced the dependency of teams on local communities (Andreff, 2008;Hutchins et al., 2019). Also, new media have de-spatialized fans' opportunities to interact with teams and other fans (Pegoraro, 2013). Consequently, identification with nonlocal sports teams has become a growing phenomenon. Several clubs from European and American leagues have been able to extend their fan bases beyond their own region, often to other countries. The trend of identifying with a nonlocal team alters the marketing landscape in team sports. Managers face geographic expansion opportunities and must at the same time defend their home market positions.
While it has been demonstrated that geographic identity explains attachment to the hometown club (Lewis, 2001;Reifurth et al., 2019), we lack a systematic understanding regarding the dynamics of identification formation with a distant club. Previous explanations for distant club or league involvement have focused entirely on the product itself either its attributes (e.g. skills of the players) or its functional benefits (i.e. social connections and escape from everyday stress) (Pu and James, 2017;Al Ganideh and Good, 2015). What remains poorly understood, however, is how a club brand's ability to fulfill fans' identity based needs drives identification with distant (or local) sports club. We understand which features of the product are important for consumers and which benefits they derive from following clubs, but we lack knowledge on how clubs become a part of the person's social self. It is crucial to understand this process from theoretical and managerial perspectives alike, as social identity guides people's behavior to act in line with the communities they perceive to belong to (Champniss et al., 2015). Moreover, no previous study has addressed and tested potential differences regarding the importance of antecedents for identifying with distant or local teams. Drawing on the social identity theory (SIT), the aim of this research is to identify antecedents of team identification (TI), specifically focusing on comparisons between local and distant fans. In summary, we address this research question: RQ. Which self-definitional needs drive and reinforce fans' identification with local and distant sports teams?

Theoretical background
The social identity is defined as "those aspects of an individual's self-image that derive from the social categories to which she perceives herself as belonging" (Tajfel and Turner, 2001, p. 101). The SIT argues that people define and evaluate themselves in terms of the groups to which they belong and aims to explain intergroup behavior by postulating that groups compete with one another over status and prestige (Tajfel and Turner, 1979). As an extension to the SIT, the self-categorization theory (SCT) focuses on the construction of social identity through self-categorization and how this process leads to ingroup-stereotypic self-perception and group-norm-conform behavior (Abrams and Hogg, 1990;Turner et al., 1987). The SIT can, thus, be defined as "social identity theory of intergroup relations" and the SCT as "social identity theory of the group" (Abrams and Hogg, 2010, p. 179). According to the SCT, a key motive for people to affiliate with groups is to develop a positive social identity by fulfilling the needs for self-enhancement and reduction of subjective uncertainty (Hogg, 2000). People seek to develop their social identity by affiliating with groups they subjectively evaluate as (1) being high in prestige and positively distinctive compared with relevant out-groups (Tajfel and Turner, 1979) and (2) being highly entitative with clearly identifiable prototypes (Hogg, 2000). Group prototypes define attitudinal and behavioral patterns, which reduce subjective uncertainty about how to feel and act (Hogg, 2000). People can derive their social identities from diverse groups, such as professional and student communities, hobby groups or religious entities. Also brands can offer meaningful social identities especially if they satisfy consumers' needs for (1) self-consistency by matching with their own identity and (2) self-differentiation and self-enhancement by being positively distinctive and prestigious (Bhattacharya and Sen, 2003). A brand's ability to fulfill these needs enhances pro-brand behaviors such as repurchasing or positive word of mouth (Bhattacharya and Sen, 2003). Based on this framework, Stokburger-Sauer et al. (2012) and Wolter et al. (2016) confirmed self-brand similarity and brand distinctiveness, but not brand prestige, as predictors of brand identification.
Based on the SIT, a consumer's social identification with a sports team, i.e. team identification (TI) (Lock and Heere, 2017), is defined as "that part of an individual's self-concept, which derives from membership in a community anchored around a sports team, based on the emotional value attached to that membership, and the knowledge of, engagement with, and evaluation of the community itself." (Heere, 2016, p. 216). In one of the earliest studies on TI, Wann and Branscombe (1990) focused on the impact of identification on fans' tendency to tighten their ties with successful teams or loosen them with unsuccessful ones. They showed that highly identified fans maintain their association even if the team was losing. Later, several studies investigated the link between TI and loyal consumer behavior and confirmed that highly identified fans demonstrate higher levels of loyalty. For example, Laverie and Arnett (2000) and Matsuoka et al. (2003) found that TI is a more important predictor of game attendance than satisfaction. A number of studies have confirmed the link between TI and attitudinal loyalty (Stevens and Rosenberger, 2012;Karjaluoto et al., 2016), i.e. "fans' psychological commitment to a team" (Bauer et al., 2008, p. 207). A recent study by Woratschek et al. (2020) shows how TI furthers loyalty both directly and indirectly through enhancing fans' perceptions of team performance and stadium atmosphere.
Early research into the antecedents of TI found that localness of a team is one of the most important factors (Wann et al., 1996;Jones, 1997). Kolbe and James (2000) showed that supporting their hometown team is essentially important for many fans. Uhlman and Trail (2012), demonstrated that attachment to the city was the strongest influencer of TI. Wang and Tang (2018) found that the geographic community group experience enhances identification with the team. Collins et al. (2016) examined the impact of hometown identification on TI among fans who have moved away from the city of their favorite team (i.e. "displaced fans") and found that (1) distance and time displaced do not compromise hometown TI and (2) fans who remain more attached to their original hometown also display a higher level of identification with the team located in their original hometown. Cohen (2017) demonstrates that "commitment to community" is the strongest predictor of TI, and suggests that the socialization processes that take place in one's geographic community strengthen the relationship with the team.
This link between hometown attachment and identification with the hometown team cannot, however, manifest itself for fans who support a team from a city where they have never lived. The strong connection between regional identity and local TI raises the question why fans identify with distant teams. Extant research has focused on product attributes, such as style of play or player skills, product benefits such as escape from daily routines and feelings of achievement and brand attributes such as appearance and history (Kerr and Emery, 2011;Pu and James, 2017;Bodet et al., 2020;Miranda et al., 2022). To explore fans' reasons to adopt the club as a part of their self-concept, we investigate possible differences in local and distant fans' underlying social identity based needs for self-enhancement through prestige and distinctiveness and for reduction of subjective uncertainty through personality congruence.

Research model and hypotheses
Our main hypotheses focus on the differential effect between local, displaced and distant fans regarding the three main antecedents of consumer-brand identification discussed in the literature (see proposed model on Figure 1). Although not proposing a formal hypothesis, Local vs. distant fansa comparative study we also examine the role of hometown identification as a potential explanatory factor underlying these differences.

Perceived team prestige
Team prestige, usually conceptualized as the perceived reputation of the sports team, is closely related to a team's competitive success although one might argue that success is a more temporary phenomenon than prestige. Nonetheless, prestigious clubs usually base their reputation on competitive success. Interestingly, prior research on the relationship between club success and fan commitment shows that highly identified fans do remain loyal even in times of limited competitive success (Wann and Branscombe, 1990). This finding does not necessarily undermine the role that success plays in enhancing identification but questions its necessity. With respect to a differential role of prestige for local versus distant fans we propose that perceived team prestige is a more important source of identification for distant fans. Strong identification with the local region likely plays an important role for local and displaced fans' identification with the local team, and a lack of success (and hence prestige) may not undermine the identification. A fan opting to support a nonlocal team, however, has a choice (the lack of success by a local team maybe one reason to support a distant team) and may therefore more likely seek out teams with high levels of prestige. We therefore hypothesize the following: H1a. Perceived team prestige does not influence TI of local fans.
H1b. Perceived team prestige does not influence TI of displaced fans.
H1c. Perceived team prestige positively influences TI of distant fans.

Perceived team distinctiveness
While the often public consumption of sports and the strong social coherence among fans of specific teams applies to all types of fans, the roots of perceiving the team to be unique are likely different. Local fans may project emotional ties with their residence (Lewicka, 2011), often based on its perceived uniqueness (Ek Styven et al., 2020), onto the local team they support. The same phenomenon may impact identification of displaced fans with their previous residence. Distant fans cannot link the team to a place of their residence but are likely to perceive the team they support as distinct for other reasons (e.g. attractive style of play). Moreover, support to a nonlocal team may itself IJSMS stem from a desire for distinction (Andrijiw and Hyatt, 2009). We therefore hypothesize the following: H2a. Perceived team distinctiveness positively influences TI of local fans.
H2b. Perceived team distinctiveness positively influences TI of displaced fans.
H2c. Perceived team distinctiveness positively influences TI of distant fans.

Self-team personality congruence
Brand personality research (Aaker, 1997;Mal€ ar et al., 2011) discusses the role of congruity between one's own and the focal brand's personality. Most empirical findings show that higher match between a brand and own personality supports the development of brand identification (e.g. Grohmann, 2009). We expect this effect to be more pronounced for local versus distant fans. Because any local sports team (with maybe the exception of recently formed or relocated teams) is closely associated with its home city (Heere and James, 2007), strong identification with the team of one's hometown is unlikely if identification with the city is lacking. Based on Heider's balance theory (Woodside and Chebat, 2001) we argue that high (low) levels of identification for both the city and the team are more common than high identification for the team and low identification for the city (or vice versa). A local fan with low congruence with the team's personality would also find herself at odds with the city they live ina conflict which a distant fan would experience less. We therefore hypothesize the following: H3a. Self-team personality congruence positively influences TI of local fans.
H3b. Self-team personality congruence positively influences TI of displaced fans.
H3c. Self-team personality congruence positively influences TI of distant fans but less so than for local and displaced fans.

Methodology
Sample and data collection We tested our model of TI with online survey data of fans supporting 16 Finnish ice hockey clubs. We selected a within-country approach in order to exclude the possible impacts of globalization-related consumption dispositions, such as consumer cosmopolitanism and ethnocentrism (Cleveland et al., 2009). The survey was administered to members of a fan message board. The questionnaire included measures for model constructs, as well as questions regarding current and prior residences. The final number of usable responses was 1,285; 62.5% (N 5 803) of the respondents were categorized as local fans, 22% (N 5 283) as displaced fans and 15.5% (N 5 199) as distant fans (see demographic characteristics and fan segments by team in Supplementary Material Appendix 1 on Table A1 and Figure A1).

Measures
We used previously validated scales (slightly modified to a sports setting) for construct measurement. The items were translated to Finnish. To safeguard the accuracy of the translation, back translation was conducted. "Team prestige" and "team distinctiveness" were measured by scales validated by Stokburger-Sauer et al. (2012). For "self-team personality congruence," the respondents were asked to evaluate their own and their favorite team's personality traits by using the 17-item sports club brand personality scale developed by Schade et al. (2014). Thereafter, an overlap score was calculated by computing Euclidean distance D and transforming it to a similarity measure S with a range of 1 (lowest overlap) to 7 (highest overlap) by applying linear transformation (cf. Stokburger-Sauer et al., 2012). Finally, for the measurement of "team identification," we used a construct combining items Local vs. distant fansa comparative study from two widely used scales by Wann and Branscombe (1990) and Robinson and Trail (2005) (Appendix , Table A2). For all items, we employed a seven-point Likert scale.
The extant literature provides no established distinction between local and distant fans. We defined (1) a local fan as a person who resides, (2) a displaced fan as a person who has previously resided (but moved away) and (3) a distant fan as a person who has never resided within a 50-km radius of the home arena of her favorite team. The 50-km demarcation represents close approximation of functional areas of major Finnish cities (Nurmio et al., 2017). Within these limits, community identification can develop through frequent utilization of a city's service infrastructure and cultural ecosystems. As an additional item, we asked respondents to indicate the municipality which they mostly identify with, but leave the item unanswered if such identification is lacking.

Data analysis and findings Common method variance
We took measures to decrease the risk of common method variance (CMV) (Podsakoff et al., 2003;Chang et al., 2010). In the questionnaire, we ordered item batteries to reduce the propensity for consistent answers (for example, background items, such as distance-related items, hometown and local identification, were displayed between the items measuring focal constructs to reduce the propensity to answer team-related items in a consistent pattern). Afterward, we conducted Harman's single factor test (Harman, 1976), through which a shared variance of 43.35% was estimated. As the commonly accepted threshold is 50% (Harman, 1976), we conclude that CMV is no serious concern.
Geographic identification by fan group Our study proposes that differences in antecedents of TI may result from the alignment of a fan's regional identification and the city where the favorite team is located. If the two are identical, TI can build on regional identification. We created a crosstab (Appendix , Table A3) of the three fan groups by their primary geographic identification and tested the statistical significance of the differences via a chi-square test of independence. The association between variables was significant with X 2 (df 5 4, N 5 1,285) 5 515.55 and p 5 0, indicating that the groups differ from each other in terms of whether their favorite team is from the city of their primary identification. The post hoc analyses revealed that the local and displaced fans were clearly more likely to identify with the city in which their favorite team is playing, whereas distant fans were much more likely to identify with another city. Also, local and displaced fans' identification with their favorite team's city differs (displaced fans are more likely to identify with another city), indicating that geographic identification may change over time when moving away.

Measurement model
To assess the validity and reliability of latent constructs, we subjected them to a single-group confirmatory factor analysis (CFA) in AMOS. For model fit examination, following fit indices were used: chi-square (X 2 ) / degree of freedom (df), comparative fit index (CFI), root mean square error of approximation (RMSEA) and standardized root mean square residual (SRMR). The model fits the data reasonably well (X 2 5 120.361, df 5 24, CFI 5 0.987, RMSEA 5 0.056 and SRMR 5 0.037). Although X 2 /df (5.015) scantly exceeds a threshold of 5 (Wheaton et al., 1977), we consider the fit acceptable for a large sample, especially because the values of CFI, RMSEA and SRMR indicate a good model fit (Hu and Bentler, 1999). Cronbach's alpha for the constructs ranged from 0.86 to 0.92, all exceeding a minimum threshold of 0.7 (Hair et al., 2018), providing evidence of internal consistency. All standardized factor loadings were significant and exceeded 0.7 (Appendix , Table A2), with the minimum threshold set at 0.5 (Hair et al., 2018). Average variance extracted (AVE) for each construct exceeded a minimum threshold of 0.5. We also have evidence of discriminant validity because the square root of AVE in every latent construct exceeds individual correlations between them (Fornell and Larcker, 1981) (Table 1).

Evaluation of multigroup invariance
To secure the equivalence of the measurement model between groups, measurement invariance was tested by a multi-sample CFA. The first step was to assess configural invariancethat is, whether the same factorial organization applies to all groups (Putnick and Bornstein, 2016). The multi-sample CFA indicates a good fit (X 2 5 192.07, df 5 72, CFI 5 0.984 and RMSEA 5 0.036), providing evidence of configural invariance.
Second, we test for metric invariance, i.e. whether the items contribute to the underlying latent constructs in a similar way across groups (Putnick and Bornstein, 2016). Toward this end, factor loadings between groups were constrained to be equal. The metric model likewise fits the data well (X 2 5 220.03, df 5 84, CFI 5 0.982 and RMSEA 5 0.036). Because the X 2 difference test is sensitive to sample sizes, which in our data differ substantially between groups, the comparison between unconstrained and metric models relied on the difference between CFI (Cheung and Rensvold, 2002) and RMSEA (Chen, 2007). ΔCFI between the unconstrained model and the metric model was À0.002 and ΔRMSEA was 0. As the recommended thresholds for ΔCFI and ΔRMSEA are 0.005 and 0.01, respectively (Chen, 2007), we conclude that metric invariance is present.
Finally, we test for scalar invariance to ensure that mean differences in the latent constructs catch all the differences in the variances of the items (Putnick and Bornstein, 2016). Scalar invariance is tested by constraining the intercepts of latent variables to be equal. Also the scalar model fits the data well (X 2 5 247.61, df 5 96, CFI 5 0.98 and RMSEA 5 0.035) and a deterioration compared to the metric model is very low (ΔCFI and ΔRMSEA being À0.002 and 0.001), providing evidence for scalar invariance.
Hypotheses testing The hypotheses were tested by a multigroup structural equation model. The structural model fits the data well (X 2 5 218.11, df 5 90, CFI 5 0.983 and RMSEA 5 0.033). The path coefficients ( Figure 1) indicate some between-groups differences whereas the link between team prestige and TI is insignificant and the link between team distinctiveness and TI is strong and positive across all groups (β local 5 0.359 and p 5 0; β displaced 5 0.388 and p 5 0; β distant 5 0.329 and p 5 0), we find a differing impact of self-team personality congruence on TI: this relationship is positive for local and displaced fans (β local 5 0.188 and β displaced 5 0.179), but no relationship for distant fans is present. To test the statistical significance of the group-specific differences, pairwise models (local fans vs displaced fans, local fans vs distant fans and displaced fans vs distant fans) were constructed, X 2 difference tests between an unconstrained model and a Local vs. distant fansa comparative study model with constrained path coefficients were conducted and if X 2 difference test revealed differences between groups, differences were investigated path by path. The results of pairwise models revealed no differences between local and displaced fans (ΔX 2 5 1.19 and p 5 0.756) indicating a similar formation of TI between these two groups and statistically significant differences between distant fans and the other two fan groups (vs local fans ΔX 2 5 9.37 and p 5 0.025 and vs displaced fans ΔX 2 5 7.86 and p 5 0.049) (Appendix , Table A4). Second, because of the similarity between local and displaced fans, a final pairwise model was constructed by comparing the combined sample of local and displaced fans with the sample of distant fans. Path-by-path X 2 difference tests (see Appendix 1, Table A5) within the final pairwise model show that the impact of team prestige on TI remains negligible among both groups (ΔX 2 5 0.199 and p 5 0.656). Hence, as team prestige does not impact TI among local (and displaced) fans, we can confirm H1a and H1b. However, against our expectations, prestige has no impact on TI among distant fans, either, leading to a rejection of H1c. Second, no differences can be identified between the groups (ΔX 2 5 0.302 and p 5 0.583) in the relationship between team distinctiveness and TI. The link is positive and equally strong between groups, allowing us to confirm H2a-c. Finally, a statistically significant betweengroup difference in the link between self-team personality congruence and TI was found (ΔX 2 5 9.033 and p 5 0.003). Hence, as the path coefficients in the model indicate, self-team personality congruence predicts TI among local (and displaced) fans, whereas this link is missing for distant fans. Therefore, we can confirm H3a and H3b, whereas H3c must be rejected as expected, self-team personality congruence is less important for distant than for local (and displaced) fans, but is not related to TI for distant fans.

Discussion
Our findings support the distinctive formation of social identification with (1) distant and (2) local sports clubs, albeit not exactly how we anticipated. The results of this study indicate that team distinctiveness has an equally strong impact on identification with a team across all three fan segments which raises intriguing questions regarding the nature of team distinctiveness, respectively. Fans identifying with their hometown team might value the localness of the team as a distinctive element (Steenkamp and de Jong, 2010) and perceive their team to be a unique local icon representing the region and its culture ( € Ozsomer, 2012). Hence, supporting a local team may reinforce fan's regional identity (Heere and James, 2007;Lock and Funk, 2016), whereas for a distant fan the favorite team might become distinctive through product-based attributes, such as style of play (Kerr and Emery, 2011) or star players (Bodet et al., 2020). In some cases, the selection of a distant favorite team may result from a deliberate effort to dissociate from the local majority (Berger and Heath, 2007).
This research provides no support for team prestige as an antecedent for TI. Although our finding is in line with other studies on brand identification (Stokburger-Sauer et al., 2012;Wolter et al., 2016), we consider this partly surprising: because of distant fans' freedom to choose any team, we expected team prestige to matter. It is known that successful teams often attract more spectators (Sass, 2016), and fans' tendency to associate publicly with a winning team has often been corroborated since the seminal study of Cialdini et al. (1976). A possible explanation for the insignificance of team prestige as an antecedent for TI is that we studied fans who had already formed a social identity as supporters of their favorite team. Perhaps team prestige matters for building TI, but, once salient, a lack of perceived prestige is unlikely to weaken it.
The importance of self-team personality congruence differs between local (displaced) and distant fans. We expected a high level of similarity between own personality and team personality to drive identification especially among local (displaced) fans, because a clear personality incongruity between the self and team renders the adoption of a local team as an agent for local identification unlikely. Our results confirm this hypothesis. However, based on extant brand identification research (Stokburger-Sauer et al., 2012;Wolter et al., 2016), we expected but did not find a positive (but weaker) relationship between self-team congruence and identification among distant fans. A possible reason for this may lie in the concept of nostalgic consumption, which is typically characterized by high self-brand connection (Youn and Dodoo, 2021;Kessous et al., 2015) and leads to preference of local brands (Dimitriadou et al., 2019) as well as to perceiving them as local icons (Heinberg et al., 2020). While this nostalgic aspect of team brand connection cannot manifest itself among distant fans, they might seek an identification with a team which is congruent to their aspired or the ideal self rather than their actual self. Ideal brand personality congruence has been shown to positively impact identification (Mal€ ar et al., 2011;Huber et al., 2018).
From a theoretical standpoint, the current findings reaffirm that an ability of a group (or an organization) to satisfy a person's need for distinctiveness furthers identification and, hence, upholds pro-group (pro-organizational) behavior. Further, the results corroborate the idea of personality match between the self and the club as an antecedent for TI among local fans. Finally, we were unable to confirm the link between organizational prestige and identification in sport club setting. A possible explanation from the perspective of the SIT is that prestige is important in forming consumer-club relationships, but afterward organizational distinctiveness suffices for consumer's self-enhancement.

Managerial implications
Our findings suggest that local and distant fans identify with their favorite team for partly different reasons. This difference has consequences for clubs' market segmentation and positioning. Clubs need to understand the geographic composition of their fan bases. If their distant fan segment is of sufficient size, it is advisable to compare their club perceptions with those of local fans. If reasons behind supporting the club differ, an understanding of each group's brand perception is essential to craft segment-specific activities. Because of their geographic separation from the club and local fan community, both displaced and distant fans may not experience many of the firsthand local activities and should therefore be exposed to the club in different ways (e.g. social media and online events) to remain engaged and uphold their identification with the team (Collins et al., 2016). After all, both displaced and distant fans are likely to be effective ambassadors of the club outside the primary market area and valuable resources in attracting new fans by promoting their bond with the club (Wakefield and Bennett, 2018;Lianopoulos et al., 2020).
Second, the findings emphasize fans' longing for distinctiveness as a primary component of their TI. To enhance a club brand's unique position, managers have two, ideally aligned, sources of inspiration: first, they can investigate which components of current brand perception among highly identified fans best contribute to the club's distinctiveness. Numerous marketing research techniques are available to learn about brand perceptions among important stakeholders (Koll et al., 2010) and compare these in easy-to-grasp formats (Koll et al., 2022). Second, managers should also investigate whether perceptual facets that have the potential to be leveraged (for example, a new player, the stadium and a specific team activity) would contribute to the perceived uniqueness of the club. Once such perceptions are identified, they should be accentuated in the club's brand positioning strategy. Although the results of this study indicate that distinctiveness drives identification equally across all three geography-based fan segments, these factors should be further investigated by segments (local vs distant): since team sports are fundamentally local bound, we consider it likely that a club's status as a local icon (cf. € Ozsomer, 2012) might be the most important factor of distinctiveness for local or displaced fans who identify strongly with the club's hometown. While distant fans might also show some attachment to the city of their favorite team, they likely will value other club product or brandspecific perceptions that they deem to be distinctive, such as style of play (e.g. Bodet et al., 2020).

Local vs. distant fansa comparative study
Third, our findings indicate that team personality contributes to identification among local (displaced) fans if the fans perceive it to match with their own personality. While it is possible to evaluate personality traits among fans, the results probably show strong heterogeneity within a club's fan base. It therefore seems more promising to investigate the brand's perceived personality among fans and use this information as the basis for crafting a unique and attractive team brand image. While the appeal of brands requires a certain degree of stability (Park et al., 1986), social media activities, team-related events or affiliating with local causes might be an effective means to either reinforce or change specific nuances of the team's personality. Given that personality congruence seems to drive identification only among fans who have a strong connection to the club's hometown, it is recommended to assess whether the perceived brand personality of the city or region is an integral part of the team's own personality and whether the localness can be leveraged as a team brand personality trait.

Limitations and future research
One limitation of this study is its focus on within-country distant fandom. Further research on fans identifying with foreign teams may uncover different dynamics of social identification with distant sports teams. Since this study focused on existing, mostly long-term fans (see Appendix, Table A1), our findings do not explain the initial impetus for developing a social identity with a team. Thus, a focus on the emergence of fan relationships could highlight potential differences between various stages of sports club-related social identity development. Finally, although the majority of ice hockey fans are still male, the share of female fans is quickly growing (Reedy, 2022;Hakola, 2017). Hence, we propose to investigate whether genders differ in their formation of local vs distant fandom.   0.003 Note(s): ΔX 2 5 Chi-square difference between unconstrained model and single-path constrained model. Variables: Pst 5 perceived team prestige, Dtc 5 perceived team distinctiveness, PCong 5 self-team personality congruence, TI 5 team identification and Adv 5 team advocacy. Significant differences (p < 0.05) in italic   Table A3. Crosstab of fan groups by indicated geographic identification Table A4. Summary of X 2difference tests between unconstrained and path constrained pairwise models