Motives, constraints and market profiles for Olympic sport participation: a latent profile analysis with the 2018 Winter Olympics and skiing

Sport participation legacy delivery became a moral obligation for organizing committees and local governments, as a major justification for public spending on the Olympics (Weed et al., 2015). However, questionable outcomes regarding the actual delivery are no secret. In a systematic review of the 2012 London Olympics' legacy, no evidence of increased participation numbers was found but only limited changes in participation patterns among those already participating (e.g. switching sport, increasing participation frequency; Potwarka and Wicker, 2021; Weed et al., 2015). Reis et al. (2017) looked into multiple Olympics and failed to find meaningful evidence for the legacy, reporting limited support (e.g. participation numbers increased only in non-Olympic sports). Recently, Xing et al. (2024) found that volunteers' sport participation during the 2022 Beijing Olympics increased three months post-Olympics but declined within the next nine months.

Researchers investigated such largely null or limited findings, consistently concluding that hosting the Olympics does have the potential to boost sport participation, but that this potential often remains unrealized due to mismanagement of opportunity (e.g. Misener et al., 2015; Shi and Bairner, 2022). The potential can come from two tracks of legacy delivery: (1) the direct process grounded in the trickle-down effect, explained by demonstration (i.e. showcasing Olympic sport), role-model (i.e. athletes as role models) and festival (i.e. the Olympics as a wide-reaching event) effects and (2) the indirect process based on enhanced sport support (e.g. sport culture, infrastructure, media). Regarding its mismanagement, researchers (e.g. Potwarka and Wicker, 2021; Weed et al., 2015) pointed out lack of leveraging strategies as the primary cause, which are necessary for removing obstacles to sports participation (e.g. limited access, low self-efficacy) and facilitating it through measures such as educational campaigns and organizational collaboration, and are thus emphasized as prerequisites for delivering the legacy.

Chalip et al. (2017) particularly attributed the mismanagement to stem from supply-side failure and lack of demand-focused strategies to increase participation, calling for an investigation into key factors shaping relevant decision-making. However, researchers (e.g. Kokolakakis et al., 2019; Reis et al., 2017) raised concerns about the limited knowledge on motives for sport participation in the legacy context. Others (e.g. Potwarka et al., 2020) argued that the legacy cannot be achieved without addressing potential constraints. Weed et al. (2015) explained that decision-making for sport participation to be influenced more by perceived benefits (relevant to motives) than by perceived disadvantages (relevant to constraints); they also highlighted the need to view sports participation decision-making as a staged process (e.g. pre-contemplation, contemplation, preparation, action and maintenance stages) because determinants are known to vary by stage, suggesting progression-based tailored strategies for better relevance/effectiveness of marketing efforts (also supported by Xing et al., 2024). Investigating key determinants (i.e. motives/constraints) of sport participation decision-making across and within progression-based profiles/segments can address these limitations and inform strategies for delivering the legacy.

To understand key factors shaping sport participation decision-making as a staged process, the “balance between motives and constraints” framework can serve as a useful framework. Introduced as part of Jackson et al. (1993) “negotiation of leisure constraints” paper (1993), the balance framework explains motives and constraints and their interplay/balance in shaping leisure behavior and intention. The balance framework is theoretically grounded in Crawford et al.’s (1991) “hierarchical model of leisure constraints” (1991), which depicts leisure participation as occurring through the two stages of leisure preference formation and transfer of preference into behavior. The model further explains the staged process as being sequentially interrupted by three types of constraints, that are intrapersonal (related to personal attributes/states; e.g. lack of interest, lack of knowledge), interpersonal (related to relationships; e.g. lack of friends) and structural (related to situational factors; e.g. time, cost). Intrapersonal constraints are described as impeding preference formation, structural constraints as disrupting the translation of preferences into behavior, and interpersonal constraints as hindering both. Building on the model, Jackson et al. (1993) pointed out that leisure decision cannot be fully understood through constraints alone and that attention is needed to how people “negotiate” to overcome constraints (i.e. the negotiation proposition). Through the balance framework (Figure 1) and as an extension of the negotiation proposition, Jackson et al. (1993) proposed motive as a key component of leisure decision-making and constraint negotiation. Motives are the “energy that initiates, directs, and sustains leisure involvement” (Losier et al., 1993, p. 154) that forms leisure preference/behavior as well as prompts and maintains efforts to overcome/negotiate constraints (Jackson et al., 1993). Motives can be categorized into internal (formed by needs, goals and values; e.g. enjoyment, social bonding) and external (formed by environmental/social factors; e.g. media, promotion) motives (Kim and Trail, 2010).

Understanding leisure decision through the lens of the balance/interplay between motives and constraints has been actively applied to sport participation (e.g. Filo et al., 2021; Ntovoli et al., 2025). In review studies (Crossman et al., 2024; Wang et al., 2020), among the most frequently researched constraints for sport participation were injury, commitment, peer influence, time, cost, motivation, knowledge and access, while motivators/facilitators were enjoyment, personal goals, connection, competition, well-being and fitness. Others identified constraints (e.g. weather, danger, and crowd) and motives (e.g. nature, cause, training opportunity) that are more specific to a particular sport or population (e.g. winter sport, non-participants; Filo et al., 2021; Malasevska et al., 2024; Wang et al., 2020). These findings demonstrated the need to incorporate both general and context-specific factors, as relevant motives/constraints varied by context. Also, intrapersonally, interpersonally and structurally constraining as well as internally and externally motivating factors were reconfirmed to each respectively play a meaningful and unique role in forming sport participation decisions (Crossman et al., 2024), where participation patterns (e.g. motives, selection factors) varied by constraint-type/level (Aicher et al., 2018). When sport participation decision-making was assessed as a staged process (combining the framework with psychological continuum or transtheoretical models; e.g. Alexandris et al., 2017; Balaska et al., 2012), intra-/inter-personal constraints imposed stronger effects in the early stages of behavioral changes and structural constraints in the latter stages, evidencing heterogeneity among stage-based sub-populations and the need for segmentation strategies when tackling constraints. The strengths of the balance framework were exhibited through these studies (e.g. Aicher et al., 2018; Crossman et al., 2024), which lie in providing a more holistic understanding of diverse determinant-types and their interplay and offering in-depth insights into the process/sequence of participation decision-making.

While some studies employed the framework for Olympic spectatorship or viewership (e.g. Funk et al., 2009; Kim and Kim, 2021), application of the balance framework to Olympic sport participation is rare (with one relevant exception: Yang et al., 2024). The application can offer a more comprehensive understanding about key determinants of decision-making but necessitates efforts to compile an exhaustive list that incorporates general and context-specific factors; sport- and Olympic-specific factors warrant attention in this study. Regarding skiing (i.e. our sport of interest), many studies have examined ski-specific factors, identifying constraints such as snow/weather, venue distance, equipment and injury risk (e.g. Malasevska et al., 2024; Toogood et al., 2014) and motives such as nature and speed/thrill (e.g. Wang et al., 2020) as sport-specific. Olympic-specific factors (e.g. Olympic media exposure, sport promotional campaigns, the Olympics' social popularity, new venues; Kim and Pu, 2021) are critical for explaining the mechanism of sport participation legacy delivery (e.g. role model effect, festival effect, easier access; Reis et al., 2017; Shi and Bairner, 2022). However, these factors have largely not been examined in sport/ski participation studies. A study guided by the balance framework and conducted in the Olympic context can fill the gap, exhaustively examining relevant motives/constraints and their interplay and informing Olympic-based strategies. In particular, assessment right after hosting the Olympics can be beneficial for capturing perceptions of motives/constraints after exposure to the Olympic hype. The period is when the Olympic impact is prominent (Kim and Kim, 2021), an impact that will be short-lived if not properly managed and thus must be exploited to deliver sport participation legacy (Chalip et al., 2017).

The balance framework's application to Olympic sport/ski participation can also add to our knowledge of the stage-based process and heterogeneity, based on the diverse and sequential effects that motive-/constraint-types exert on decision-making processes (Aicher et al., 2018); in the Olympic context, this is an area with a paucity of research. A motive-/constraint-based market profiling study can help. When market profiles are identified based on motive-/constraint-based patterns, these patterns represent unique characteristics of each profile (i.e. profile-based heterogeneity; Spurk et al., 2020) and can inform the profile's placement in the decision-making process. Referring to the balance framework, placement in the early vs. later stages implies different progression (preference formation vs. conversion into behavior), influential motives/constraints (e.g. intrapersonal vs. structural), response to constraints (block vs. modify participation), participation likelihood (low vs. high), etc. (Crawford et al., 1991). Such differences (i.e. stage-based heterogeneity) are critical information that should be reflected in marketing strategies (Balaska et al., 2012; Weed et al., 2015). As shown, motive-/constraint-based market profiling and the interpretation of its findings based on the balance framework can disclose profile-/stage-based heterogeneity and inform diversified strategies. However, knowledge is limited in the Olympic context and the profiling criteria adopted in prior studies were either motives (e.g. Ntovoli et al., 2025) or constraints (e.g. Priporas et al., 2015; Vassiliadis et al., 2018). This study combines both criteria and applies them to Olympic sport/ski participation. A motive-/constraint-based market profiling study is conducted to answer the following research questions, focusing on the 2018 PyeongChang Winter Olympics and skiing.

In the investigation, late adolescents (ages 18–24) will be of primary interest as a key target population of the Olympic movement and the 2018 Olympics' organizing committee (IOC, 2012; PBCOG, 2011; as further discussed in “Methods”). Late adolescents are in the developmental stage characterized by a shift from parental dependency to financial and social independence, enabling leisure decision-making on one’s own (Trail and James, 2019). Prior studies describe this group as being prone to social influence and media exposure, expressive, upbeat, confident, liberal and open to change (e.g. Kim et al., 2017) and more receptive to the Olympic exposure (Kokolakakis et al., 2019; Potwarka et al., 2020). Health, stress management, peer influence and university/school programs are motives and time, cost and alternative leisure options are constraints commonly affecting sport participation in this group (e.g. Kim et al., 2017). Sporting habits formed during late adolescents can translate into long-term participation (Trail and James, 2019), making this group strategically valuable for delivering sustainable sport participation legacy outcomes.

Based on the balance between motives and constraints framework, key market profiles among current and potential late adolescent ski (including snowboard) participants were identified using the 2018 PyeongChang Winter Olympics as the case study. Further, motives and constraints that influence ski participation intention were examined and compared among profiles.

LPA was used for this, which is a probabilistic and finite mixture model-based clustering analysis. In this person-centered approach, patterns of observed continuous variables (i.e. the indicators; in this case, motives/constraints) are detected to determine profile classification. LPA assumes unobserved heterogeneity and treats classification as a latent categorical variable, estimating classification based on the probability of assignment to a particular profile; the approach differs from traditional approaches (e.g. k-means, hierarchical cluster analysis) that treat classification as a non-latent categorical variables and take a distance approach estimating classification based on nearness (in a plot; Weller et al., 2020). Researchers (e.g. Spurk et al., 2020; Vermunt and Magdison, 2004; Weller et al., 2020) argue that LPA is superior for its use of probability in profile classification, its ability to sophisticatedly treat configurations with many variables and its provision of tests for model specification. LPA was adopted in prior leisure/sport studies guided by the balance framework (e.g. Agans et al., 2025; Kim, 2024; Ntovoli et al., 2025), demonstrating its compatibility as well as usefulness in developing diversified marketing strategies.

LPA was further combined with multi-group regression. Treating profile classification as a grouping variable (recommended if entropy as an indicator of classification certainty is above 0.80; Clark and Muthen, 2009), post hoc analysis can test if and how the relations among determinants and outcomes vary by profile (Ferguson et al., 2020). The approach was found effective in revealing profile-based heterogeneity and informing profile-specific strategies in previous studies on sport participation motives/constraints (e.g. Ntovoli et al., 2025).

For this study, the 2018 PyeongChang Winter Olympics presents a compelling case because of its strong emphasis on the legacy; moreover, valuable insights can be offered as research on the Winter Olympics is scarce. The vision of the 2018 Olympics was “New Horizons”, aiming to expand the winter sport market in the Asian region and among younger generations (PBCOG, 2011). With winter sport not among the most popular participation sports in the host country (Kim et al., 2012), promoting winter sport and positioning the host city as a “winter sport hub” were key objectives/justifications for the 2018 Olympics (PBCOG, 2011). In particular, skiing was a once fast-growing winter sport that had been experiencing gradual declines in participation numbers since its peak in 2013 (Lee, 2022). Expectation for a ski participation legacy were high with significant investments made in the ski industry (e.g. venue renovation/construction, high-speed train connection) and promotional programs (e.g. Suhorang ski camp, school ski club initiatives). The host country won its first-ever Olympic medal in a snow sport at the 2018 Olympics, setting up an advantageous stage for building a ski participation legacy.

The primary target market of the 2018 Olympics was younger generations, in line with the IOC's Olympic legacy goals (IOC, 2012) and because Asia was the “youngest and fastest growing winter sport market in the world with the largest aggregate youth population” (PBCOG, 2011, p. 19). Among young people (from ages 15 to 24; UN, 2023) in the host country, the most viable target appeared to be late adolescents (from ages 18 to 24), as they are capable of making their own leisure decisions, more receptive to Olympic exposure, have shown significant growth in market size and may turn into long-term consumers (Kokolakakis et al., 2019; Lee and Kim, 2016; Trail and James, 2019). More importantly, the group was deemed to have sufficient time for leisure and to be more likely to pick up new hobbies (MOGEF, 2019), as it faces fewer educational constraints than those aged 14 to 17 in the host country (who have very limited time for leisure activities, as they are tied up at public schools and private educational institutes from early morning until late night; Yonhap, 2019). Therefore, late adolescents in the host country were set as our target population.

Identification of motives and constraints. Compiling a list of motives and constraints that specifically fit and captures the uniqueness of the context was critical for this study. With no readily available scale for the Olympics sport (i.e. ski) participation legacy context, a list was developed using thought-listing, which is a procedure useful for compiling an exhaustive list of factors by eliciting thoughts (Cacioppo and Petty, 1981); the approach has been employed in sport marketing, scale development and motive/constraint studies (e.g. Kim, 2024; Ross et al., 2006).

Participants were recruited from the target population through convenience sampling (n = 97) and asked to freely list and describe any factors/thoughts that affect their decision to participate in skiing. With a total of 824 thoughts collected and coded by two experts, motives and constraints were identified and categorized into different motive-/constraint-types. Inter-rater reliability was 93.7%; for disagreements, a third expert was consulted to reach consensus. Then, key motives/constraints were selected based on the percentage of participants mentioning the factor (the cut-off was 20%), expert panel reviews to check for content and face validity and consistency with relevant studies (e.g. Fredman and Heberlein, 2005).

Seven internal motives of curiosity, enjoyment, escape, achievement, health, body image and socialization and two external motives of media coverage and program were identified (Table 1). Four intrapersonal constraints of lack of knowledge, skill, interest and energy, one interpersonal constraint of lack of friends and five structural constraints of financial cost, time constraint, accessibility, safety and weather were compiled.

Measurement. Single items were constructed for each motive and constraint, as is common practice in thought-listing and context-specific studies (e.g. Heere, 2010; Walsh et al., 2013). In prior studies, single-item measures were “just as reliable and valid” as multi-item measures when measuring needs and values in sport experiences (relevant to motives; Trail et al., 2023) and sport consumer perceptions (e.g. Ko et al., 2021; Kwon and Trail, 2005) and they are known to capture the core of marketing constructs parsimoniously (Rossiter, 2002). Single-item measures can be rigorous as long as the construct is unambiguous and content, face and criterion validities are checked (Trail et al., 2023).

Following Diamantopoulos et al.‘s guidelines (2012), single items were developed based on established conceptualization and/or scales (e.g. Alexandris et al., 2002; McDonald et al., 2002; Manfredo et al., 1996), expert reviews and statistical criteria; if necessary, items were constructed based on common descriptions in thought-listing and expert reviews. Items were either 7-point Likert-type (with anchors of 1 = strongly disagree and 7 = strongly agree) or semantic differential scales (see Table 1). Two versions (English and Korean) were prepared by bilingual speakers using back-translation procedures, ensuring reliability. Items and translations were reviewed by experts for face/content validities and by potential respondents for clarity. Additionally, a pilot test was conducted (n = 102) and inter-item correlations were within a reasonable range (less than 0.7; three exceptions). Concurrent/predictive validity was ensured; compared to existing scales (e.g. Ahn and Hwang, 2011; Alexandris et al., 2002), correlations among respective factors were high (all greater than 0.75) and the variance in participation intention explained by our scale was meaningful (28.4%).

Items to measure participation intention (using a seven-point Likert scale from Cronin et al., 2000) and demographics (i.e. gender, age, education and household income) were also prepared. An online survey questionnaire was designed with Qualtrics, including items for consent (one item), screening (two items), motives (nine items), constraints (ten items), participation intention four items) and demographics (four items). After another expert review, the questionnaire was ready for use.

After receiving IRB approval (#17–021), respondents were recruited via convenience sampling through social media and major universities. Data were collected online two to four weeks after the closing ceremony of the 2018 Winter Olympics to capture participants' perception about skiing after Olympic media exposure. The period was selected because the Olympic impact was prominent (Kim and Kim, 2021), an impact that must be exploited to deliver a sustainable post-event legacy (Chalip et al., 2017). A total of 545 usable responses were collected, after dropping 41 unreliable (e.g. all answers were the same) or incomplete responses. Demographics of respondents were as follows [gender: 49.6% male, age: average 20.43 years, education: 87.2% undergraduate, household income median: between 50,000,000 and 750,000,000 won]. Our pool of respondents was more educated compared to the target population, which was not seen as a major concern as skiing tends to attract those in higher socioeconomic status (Krueger, 2022).

After data screening and assumption tests, LPA was conducted using motives/constraints as profiling criteria (using Mplus7). First, the appropriate number of profiles was judged based on (1) information criteria such as AIC and BIC (to test model fit), (2) BLRT (i.e. bootstrap likelihood ratio test; to see if the k-class model is better than the k-1 class model), (3) entropy and class probability (to examine classification accuracy), (4) class proportion and size (to check profile usefulness) and (5) theoretical interpretability (for domain usefulness) (Nylund et al., 2007; Sinha et al., 2021; Spurk et al., 2020). Then, after running LPA with the appropriate number of profiles, perceived levels of motives/constraints were examined for each profile. Multi-group regression further investigated which motives/constraints had significant effects on participation intention within and across profiles. Profile classification was set as a categorical grouping variable (cf. prerequisite: entropy above 0.80; Ferguson et al., 2020), motives/constraints as predictors and ski participation intention as the distal dependent outcome (that is not used in LPA modeling).

Data were screened and assumptions (e.g. no missing data, linearity, multicollinearity and normality) were checked. Inter-item correlations were all in a reasonable range (less than 0.70), with two exceptions (Table 2). Motive-/constraint-based LPA with various numbers of profiles (from 1 to 7) was conducted and compared to find the best-fitting model. AIC and BIC were checked for model fit. While smaller AIC and BIC are considered as better fits, adding more classes/complexity to the model almost always leads to smaller values in large data sets with many indicators (Sinha et al., 2021); in such cases, searching for the “elbow” (i.e. the point of inflection in the plot, where an increase in model complexity/the number of classes does not yield a comparable decrease in AIC/BIC) is recommended (Masyn, 2017). For classification accuracy, entropy (i.e. an indicator of effective data partitioning) higher than 0.8 and class probability (i.e. the accuracy of classification per profile) higher than 90% can be retained (Muthen and Muthen, 2000). Class proportion higher than 5% or class sizes larger than n = 50 can be pursued to ensure each profile is worthy of attention. Table 3 indicates that the 5-class model was the most optimal. Additionally, a series of BLRT tests were conducted (Nylund et al., 2007) to verify that the 5-class model was statistically superior to the 1-, 2-, 3- and 4-class models (all at p < 0.05; cf. the comparison with the 6-class, which did not converge). The 5-class model made theoretical sense, supporting interpretability. Thereby, the 5-class model was selected for LPA.

LPA was conducted, identifying five market profiles in the context. Means of motives, constraints and participation intention for each profile are reported in Table 4. Each profile is named to represent its key characteristics in accordance with the balance framework. Profile 1 is labeled as the “amotivated” (7.9%) for reporting the lowest levels of internal and external motives and the lowest participation intention; the profile reported high intrapersonal, interpersonal and structural constraints. Profile 2 is named “passionate” (16.3%). The profile can be characterized by high internal and external motives, low intrapersonal and interpersonal constraints and medium structural constraints, resulting in the highest participation intention. Profile 3 is labeled “non-committal” (26.4%) for showing somewhat medium (or neutral) levels of all motive- and constraint-types as well as participation intention. Profile 4 is the “structurally constrained” (29.0%) where structural constraints seem to be the main hindrance, despite internal motives being high and intra-/inter-personal constraints being low; participation intention was the second highest here. Profile 5 can be labeled “motivated but constrained” (20.4%) for presenting high internal motives as well as high intrapersonal and structural constraints. Compared to the “structurally constrained” profile, this profile appeared to be as motivated yet more constrained and intrapersonal constraints became a concern.

Multi-group regression was conducted to identify which motives and constraints were influential in shaping ski participation intention within each profile (Table 5). For “amotivated”, curiosity (γ = 0.48, p < 0.01), socialization (γ = 0.36, p = 0.03), program (γ = 0.20, p = 0.05), lack of skill (γ = −0.23, p = 0.04), lack of interest (γ = −0.35, p = 0.05) and weather (γ = −0.18, p = 0.07; marginal) were found to have significant effects on intention. For “passionate”, curiosity (γ = 0.38, p < 0.01), body image (γ = 0.19, p = 0.04), lack of skill (γ = 0.24, p < 0.01), lack of friends (γ = −0.16, p = 0.06; marginal), accessibility (γ = 0.21, p < 0.01) and weather (γ = −0.16, p = 0.07; marginal) were influential. For “non-committal”, curiosity (γ = 0.20, p < 0.01), achievement (γ = 0.27, p < 0.01), program (γ = 0.18, p = 0.07; marginal), time (γ = −0.28, p < 0.01) and safety (γ = −0.18, p = 0.03) were reported to have significant impacts. For “structurally constrained”, curiosity (γ = 0.26, p < 0.01), achievement (γ = 0.14, p = 0.07; marginal), socialization (γ = 0.14, p = 0.05), program (γ = 0.22, p < 0.01), lack of knowledge (γ = −0.16, p = 0.04) and time (γ = −0.27, p < 0.01) had significant influences. For “motivated but constrained”, curiosity (γ = 0.36, p < 0.01), program (γ = 0.26, p < 0.01), lack of knowledge (γ = −0.27, p < 0.01), lack of interest (γ = −0.21, p = 0.03) and time (γ = −0.23, p = 0.02) were significant factors in ski participation intention formation.

The significance of this study lies in (1) delving into the demand-side of sport participation legacy, (2) applying the balance between motives and constraints framework to the Olympic legacy context, (3) approaching participation decision-making as a staged-process, (4) compiling an exhaustive list of motives/constraints for Olympic sport (i.e. ski) participation, (5) identifying key market profiles with the 2018 PyeongChang Winter Olympics case and late adolescents and (6) investigating key determinants of participation decision-making across/within profiles.

Grounded on the balance between motives and constraints framework, an exhaustive list of motives and constraints was compiled for our context via thought-listing. Motives/constraints known to generally apply to sporting activities (e.g. escape, enjoyment, lack of knowledge) were identified, adding support to existing literature (e.g. Crossman et al., 2024; McDonald et al., 2002). Demonstrating the need for recognizing context-specific aspects, motives/constraints considered more relevant to the Olympics (e.g. curiosity, media exposure; e.g. Kim and Pu, 2021), skiing (e.g. safety, weather; e.g. Wang et al., 2020) or late adolescents (e.g. programs) contexts were also identified. Particularly, curiosity was discovered as the most influential factor in forming intention, hinting signs of the trickle-down effect. The compiled list explained a meaningful 37.3% variance (on average per profile) of ski participation intention. As reported in prior studies (e.g. Kim and Pu, 2021; Kim and Trail, 2010), internal factors (i.e. internal motives, intrapersonal constraints) had stronger effects than external ones (i.e. extrinsic motives, intrapersonal/structural constraints).

Balance between motives and constraints was applied as a framework for learning about ski participation decision-making as a staged process and for market profile analysis. Five market profiles were identified through LPA. Earlier, leisure behavior/intention was discussed as a two-step process that is (1) leisure preference formation and (2) transferring the preference to behavior (Crawford et al., 1991); intrapersonal constraints interrupt preference formation, structural constraints impede the transfer, interpersonal constraints hinder both and motives fuel the efforts to overcome these constraints (Jackson et al., 1993). Distinctive characteristics in the five profiles corresponded to different stages/issues in the process (based on Table 4). “Amotivated”, “non-committal” and “motivated but constrained” were mainly blocked by intrapersonal constraints, “structurally constrained” by structural constraints and “passionate” by none; “amotivated” and “non-committal” were also faced with low motives. Applying the process (Figure 2), “amotivated” can be considered to be in the pre-stage of preference formation. “Non-committal” and “motivated but constrained” would be undergoing preference formation. “Structurally constrained” would have formed preference but faced hindrance in turning it into behavior. “Passionate” likely completed the process. The findings are theoretically meaningful for offering evidence consistent with the behavior/intention formation process and the sequential order of three constraint-types (Jackson et al., 1993) and for verifying the applicability of the process/sequence to Olympic sport participation. Further, stage-based heterogeneity was demonstrated in participation decision-making, as advocated by researchers (e.g. Weed et al., 2015; Xing et al., 2024).

Results from multi-group analysis offered additional support for stage-based heterogeneity, revealing that key determinants for decision-making varied by profile. For the “amotivated” profile (7.9%) that was likely in the pre-initiation stage of preference formation, internal motives and intrapersonal constraints were more influential on intention. Along with curiosity (which was an Olympic-related cross-profile determinant), socialization, lack of skill and lack of interest were noteworthy determinants in this profile. Socialization was a motive linked to social influence from friends and family, which is known as an effective driver for non-participants to start considering sport (Trail and James, 2019) and a stronger predictor in collectivist cultures and among late adolescents (Liu et al., 2025). Lack of skill was a constraint warned as a side effect of Olympic media exposure (Chalip et al., 2017; Hindson et al., 1994), occurring when viewers recognize the substantial gap in capability between athletes and themselves; this gap was perhaps perceived as especially wider among those least likely to participate in skiing, resulting in the factor functioning as a significant constraint only in this profile. Lack of interest was an influential constraint related to the reported low intrapersonal motives (e.g. enjoyment, escape) and is a more difficult challenge (Balaska et al., 2012) for not only being a barrier itself but also indicating no energy/drive to overcome such barrier.

For the “passionate” profile (16.3%), internal motives and three constraint-types were influential on intention. Notable determinants were body image, lack of skill, lack of friends and accessibility in this profile. In line with the profile likely consisting of active ski participants that successfully overcame constraints, body image was a determinant unique to the profile that can only be pursued with serious and frequent participation. Interestingly, lack of skill and accessibility had positive effects on intention, despite commonly being known as constraints (Crossman et al., 2024). The profile perhaps viewed skiing as a physically challenging (extreme) and exclusive (luxurious) leisure that is not for everyone (Williams and Fidgeon, 2000), these views may have ironically added to the sport's appeal. Consistently, Lee (2022) reported that advanced skiers in Korea tended to be committed to mastering their skills and willing to pay price premiums for access into better skiing environments. The two constraints serving as motives in this particular profile exemplified that the placement of factors in the motives/constraint continuum was dependent on the profile/context (Kim and Trail, 2010), adding support to profile-based heterogeneity.

“Non-committal”, “motivated but constrained” and “structurally constrained” were in the process of intention/behavior formation. Curiosity, program and time were determinants found in all three profiles, respectively related to the Olympics (Kim and Pu, 2021), late adolescents (Liu et al., 2025) and the Korean ski market (cf. travel time is a critical factor as day-trips are common; Kim, 2012). The “non-committal” (26.4%) was in the earlier phase of preference formation, where mediocre motives likely initiated but did not sustain efforts to overcome mediocre constraints. Achievement and safety were determinants notable in this profile. Urge to learn and fear of danger are a common mix witnessed among potential and novice skiers (e.g. Lee, 2022; Williams and Fidgeon, 2000; Yang et al., 2024), hinting the composition of this profile. Key characteristics of “motivated but constrained” (20.4%) and “structurally constrained” (29%) were somewhat similar, but the primary difference lay in intrapersonal constraints; the constraints were perceived higher (e.g. lack of skill) and more influential (e.g. lack of knowledge, lack of interest) in “motivated but constrained”, playing the more significant role in the profile than in “structurally constrained”. Crawford et al. (1991) described intrapersonal constraints to “block” preference formation and structural constraints to “modify” leisure behavior (e.g. less frequency, alternative activity), indicating fundamentally distinct challenges and the need for tailored strategies (Balaska et al., 2012). Notably, lack of interest was earlier discussed as a more difficult challenge for “amotivated”; the constraint was also a significant determinant in “motivated and constrained” but was less of a concern as perceived lack of interest was low.

Findings of this study offer practical insights into the demand-side of sport participation legacy. Diversified marketing strategies can be developed for the five market profiles identified from LPA. Based on the issue-type interrupting the process of behavior/intention formation, marketing efforts can focus on stimulating motives for “amotivation” and “non-committal” (e.g. promote benefits of skiing, use hype-videos), easing intrapersonal constraints for “amotivation”, “non-committal” and “motivated but constrained” (e.g. ski camps, tutorials) and removing structural barriers for “structurally constrained” (e.g. discounts, shuttles). Such strategic allocation of efforts can enhance marketing relevance/effectiveness (Armstrong and Kotler, 2005).

Specific motives/constraints can be targeted across or within profiles based on multi-group regression results. Curiosity and program emerged as cross-profile determinants. As the motives are linked to Olympic media exposure and school-offered programs, collaboration with Olympic media channels/commentators can be effective for boosting curiosity (e.g. depict skiing as an attractive Olympic-featured sport) and with schools to offer more programs (e.g. ski classes, camps). Within profiles, tailored approaches are required, acknowledging profile-/stage-based heterogeneity and targeting profile-specific determinants. For “amotivated”, socialization (e.g. position skiing as a social activity, instigate word-of-mouth from friends and family, offer group trips) and lack of skill (e.g. describe Olympians as relatable people rather than gifted talents) should be targeted. For “passionate”, lack of skill and accessibility can be strategically reframed (e.g. promote skiing as a challenging extreme sport, emphasize exclusivity). Notably, contrasting suggestions are made for lack of skill between “amotivated” and “passionate”, re-supporting the need for diversified strategies. Other factors worth attention are time for “non-committal”, “structurally constrained” and “motivated but constrained” (e.g. target breaks/vacations, run shuttles to reduce travel), lack of knowledge for “structurally constrained” and “motivated but constrained” (e.g. education via Olympic media, school programs, games) and achievement for “non-committal” (e.g. emphasize progression, realistic goal setting).

Certain profiles can be prioritized. “Passionate” can be a prioritized because it has the highest intention level, with a focus on participant retention; satisfying internal motives (i.e. expectations) during participation is critical (based on “expectancy-disconfirmation”). “Structurally constrained” and “motivated but constrained” are promising targets for expanding the ski market for showing high motives and medium-to-high intention; notably, “motivated but constrained” faced intrapersonal constraints but perceived lack of interest was low – i.e. the interest is there. “Amotivated” would be the least priority due to the lowest intention and smallest market size.

The 2018 PyeongChang Winter Olympics and late adolescents were our foci in this study, which were selected based on the Games’ vision emphasis on sport participation legacy and as a crucial/viable target population for the IOC and organizing committee. The study offered valuable insights, but caution is warranted when applying the findings to other contexts. Further studies on other sports (e.g. swimming, skating), events (e.g. Summer Olympics), populations (e.g. children, adults), or host countries (e.g. France, USA) are necessary to accrue knowledge; comparison across contexts can provide insights into the generalizability and context-specificness of findings.

This study's purpose was to inform leveraging strategies for building a sport participation legacy. For this, data was collected post-Olympics to capture motives/constraints after Olympic media exposure and to inform marketing strategies for sustainable post-event legacy. The collected data served our purpose well, but a longitudinal study can advance the knowledge. Investigating the changing patterns of motives/constraints before, during and after the Olympics can present in-depth insights into the Olympics as a promotional platform for sport participation, by testing for the existence of trickle-down effect and identifying the best timing for marketing/leveraging activities.

This study showcased how better knowledge of Olympic sport participation target markets can be acquired via LPA and multi-group analysis, with a focus on the 2018 PyeongChang Winter Olympics, skiing and late adolescents. Referring to the “balance between motives and constraints” framework, five motive-/constraint-based market profiles were identified: “amotivated”, “non-committal”, “motivated but constrained”, “structurally constrained” and “passionate”. Each profile faced distinct issue-types (e.g. low internal motives, high structural constraints) and occupied different stages in the participation behavior/intention formation process. Cross-profile (e.g. curiosity, program) and profile-specific determinants (e.g. lack of skill, accessibility) of ski participation intention were also discovered.

Theoretically, this study extended the balance framework to the Olympic context, advancing our knowledge on motives/constraints relevant to Olympic ski participation and their sequential interplay in forming intention. The use of LPA along with the balance framework provided a novel and rigorous approach for identifying key market profiles as well as cross-profile and profile-specific determinants of Olympic ski participation. These findings unveiled profile-/stage-based heterogeneity (in terms of perception of motives/constraints, factors shaping intention and their valence of influence, placement within and challenges faced in intention/behavior formation process, etc.). Practically, findings offered insights into demand-focused marketing strategies (e.g. cross-profile vs. profile-specific strategies, profile prioritization) that can guide the delivery of sport participation legacy.