Variations in adolescents’ motivational characteristics across gender and physical activity patterns: A latent class analysis approach

Physical activity (PA) guidelines recommend that young people accumulate at least 60 min of moderate-to-vigorous physical activity (MVPA) daily. Adherence to PA guidelines is associated with a range of positive outcomes including decreased risk of excess weight, cardiovascular disease and psychological difficulties [1–3]. Despite these benefits, most adolescents are insufficiently active [4, 5]. PA is a complex behavior characterized by participation in a number of different types of activity such as sport, exercise and free play. However, previous studies have typically used continuous measures of MVPA or cut-off points to categorize participants into discrete groups reflecting whether or not recommended PA guidelines are met. Focusing on overall PA levels without taking account of the underlying context or type of PA (e.g. team or individual sports; organized i.e. that guided by a coach or instructor, or non-organized activity, such as walking, jogging, informal games) does not allow for examination of groups that differ qualitatively (e.g. active team sport vs active dance), which has further resulted in a limited understanding of the patterns of adolescents’ PA.

Efforts within the literature to investigate underlying types of PA involvement have focused on discriminating the context of participation, which is broadly defined in terms of organized (e.g. team sport, individual sport, PA lessons e.g. dance) and non-organized PA behavior (e.g. informal games of basketball, tag, or exercise such as jogging) [6, 7]. A lack of research exists regarding the relative extent to which non-organized PA contributes to overall PA levels. One longitudinal investigation into male and female adolescents’ PA throughout secondary school revealed that participation in light-to-moderate PA like walking, running, and jogging were more likely to be initiated and continued over time [8], underscoring the potential benefit of this type of PA for promoting adherence to an active lifestyle long-term.

Organized sport in contrast, has been widely researched with empirical studies indicating that participation is associated with higher levels of MVPA among young people [9, 10]. Focusing exclusively on organized sport however, may have precluded an investigation into the role of other organized non-sport PA lessons such as dance or fitness classes (e.g. aerobics, circuits), which may also contribute to total PA levels among youth. In accordance, O’Neill, Pate and Hooker [11] found that female adolescents’ objectively measured activity in structured dance lessons accounted for roughly 29% of their weekly MPVA. Failing to consider organized PA participation alongside sport may therefore result in an underestimation of overall youth PA levels whilst also overlooking an important source of MVPA among girls in particular. As such, it is evident that taking account of the different contexts of PA that underpin overall involvement can provide a greater understanding of adolescents’ PA behavior. Recent studies have attempted to address this issue by adopting person-oriented approaches, which seek to identify groups of adolescents that engage in a similar pattern of PA behavior or respond similarly on a set of observed variables, which in turn differentiates them from other subgroups [12]. Identification of different profiles of adolescents’ PA participation may be used to tailor interventions towards specific subgroups of young people promoting increased MVPA. Agans and Geldhof [13] identified five distinct clusters of organized sport participation among a sample of secondary school students that reflected participation in team sport, individual sport, a combination of team and individual sports, dance, and non-participation. However, this study failed to consider involvement in organized non-sport (e.g. gym, aerobics) and non-organized PA (e.g. walking, jogging) yielding an incomplete picture of adolescents’ PA behaviors.

Liu and colleagues [14] identified five distinct groupings of MVPA participation for male and female adolescents as part of a national US Survey. The largest female class, comprising three-quarters of the sample, was labelled ‘dancers/walkers/runners’. In contrast, for males, just less than three-quarters of the overall sample comprised the ‘basketball and runner’ class. The results indicate that preferred type of PA participation differed between boys and girls, highlighting the need to explore gender specific profiles of PA. These findings contribute to our understanding of adolescents’ PA patterns, although the context of organized versus non-organized PA was not considered.

In addition to elucidating distinct patterns of PA among adolescent boys and girls, identification of psychological factors that differ across discrete patterns of PA could prove useful for informing interventions targeting specific subgroups of adolescents. In accordance, prior investigations of latent classes of PA among youth have tended to focus solely on demographic characteristics, which have resulted in a limited understanding of modifiable factors underpinning involvement [14, 15].

Self-determination theory (SDT; [16, 17]) has been widely adopted as a framework for understanding the psychological influences and motivational processes associated with adolescents’ PA. Proponents of SDT posit that satisfaction of psychological needs for competence, relatedness and autonomy foster autonomous motivation, which underpins PA behavior. Higher perceptions of competence, relatedness and autonomy underpin intrinsic motivation to participate in sport and sustain involvement over time [18–20]. Higher perceptions of autonomy support have also been found to promote psychological need satisfaction and intrinsic motivation for organized sport participation [20, 21]. Most of these studies have focused on organized sport without discriminating team and individual sport participation. Thus, it remains unclear whether motivational profiles differ across discrete types of organized and non-organized PA participation.

The current research will attempt to address these shortcomings in the literature by adopting a person-oriented approach to investigate patterns of PA involvement for male and female adolescents using multiple indicators of PA including type, context, duration and intensity. Secondly, PA profiles will be compared on dimensions derived from SDT: intrinsic motivation, competence, relatedness, autonomy and autonomy support, to determine whether psychological processes differ as a function of the pattern of PA undertaken.

The current findings provide a greater understanding of adolescents’ PA patterns and differences in psychological processes associated with PA participation. Previous studies have typically adopted variable-centred approaches (e.g. correlation, factor analysis) that look at relations among isolated variables (e.g. MVPA) based on information about the average person in the sample, which ultimately neglects intra-individual differences [36]. LCA, a person-orientated approach was therefore used in the current study to identify homogeneous subgroups of adolescents that engaged in similar patterns of PA behavior [12].

This study extends previous research [13, 14] providing additional information on the context in which PA is performed and the different ways in which male and female adolescents accumulate MVPA. Overall, boys demonstrated more active patterns of PA participation than girls, which is consistent with prior findings indicating that latent classes characterized by higher levels of MVPA derived from accelerometer and/or self-report measures of PA, comprise a higher proportion of male than female adolescents [15, 37]. In accordance, four of the five male PA profiles were considered leisure active, indicating that recommended levels of MVPA were achieved on at least three days over the past week, with two classes also achieving at least 60 mins of MVPA daily. Thus, boys can accumulate high levels of MVPA through various PA behaviors. Only two of the female classes reflecting discrete patterns of participation in team sport and individual sport, undertook sufficient MVPA to be considered leisure active, with members of the latter class also engaging in a mean of more than 60 min of MVPA daily.

The findings support prior research in highlighting the significance of organized sport participation in assisting adolescent girls to achieve and sustain higher levels of MVPA [9, 10]. The most prevalent PA profile among girls was Non-Participation while for boys it was Active Team Sport (soccer/Gaelic football/hurling/rugby), with each class representing approximately one third of the respective sample. This result supports previous findings that boys are more likely to engage in moderate-to-vigorous intensity team based activities than girls [38, 39], who are more likely to be characterized by low active patterns [15, 37].

Gender differences were also evident in relation to organized non-sport PA. A small group of male adolescent gym members comprised a leisure active fitness-oriented profile, which contrasted with a lower active pattern of dance among girls. An additional female class Run/Swim & Dance/Gym characterized by low levels of MVPA and dual participation in individual sport and organized PA emerged, that was not replicated among the male sample. The results suggest that for girls, partaking in organized dance classes once a week, alone or in addition to a weekly running or swimming training session does not necessarily mean they are classified as being leisure active. Nonetheless, the present findings indicate that organized dance classes can make an important contribution to girls’ overall PA levels with latent class members achieving an average of 25-28 min in MVPA daily. This result is in line with previous research that revealed that female adolescents’ objectively measured activity in structured dance lessons accounted for roughly 29% of their weekly MPVA [11], highlighting the significance of this mode of activity for increasing girls’ PA. Finally, a discrete profile of low active engagement in non-organized PA unique to girls emerged that was primarily characterized by participation in individual type activities like walking, jogging and running. This context of PA participation has received limited attention to date, thus this finding makes an important contribution to the existing literature by elucidating that informal exercise reflects a valid and distinct pattern of adolescent girls’ PA, representing 17.1% of the total female sample.

Conversely, for boys dual participation in non-organized PA (soccer/run/basketball) alongside team sport (soccer, Gaelic football, hurling, basketball) represented a highly active profile of mixed PA participation. Of note, roughly half of the non-organized PA bouts reported reflected engagement in informal games of soccer, contrasting girls’ inclination for more individualized types of unstructured activities. Prior research conducted on a nationally representative sample in Ireland revealed that male adolescent sport involvement is dominated by team-based invasion games (i.e. soccer, Gaelic football, basketball, hurling) [40]. This finding contributes to this literature base indicating that preferences for team games are further evident in adolescent boys’ unstructured play. Moreover, the informal nature of non-organized PA, which can be undertaken across multiple settings, alone or with friends, with little cost, commitment or organization, offers a means of accumulating MVPA that is accessible to all young people.

A secondary aim of the study was to determine whether motivational processes derived from SDT differed with respect to the profile of PA undertaken. While previous studies have found that adult sport participants demonstrate higher levels of self-determined motivation and competence than those engaged in fitness activities (e.g. aerobics, weightlifting) [41, 42], this area has received little attention among adolescent samples.

The results revealed that adolescent girls characterized by leisure active participation in organized team or individual sport demonstrated higher levels of intrinsic motivation, competence, relatedness, autonomy and autonomy support relative to those characterized by less active profiles of organized PA (dance/gym), non-organized PA (walk/run) and non-participants.

Consistent with prior research, female adolescents characterized by leisure active sport participation demonstrated significantly higher perceptions of competence than individuals that engaged in less active PA patterns [10, 43]. In accordance, organized sports typically require a certain level of skill to gain entry, with such abilities unnecessary for taking part in unstructured PA (e.g. walking, jogging) or non-competitive pursuits such as dance lessons and gym classes. Qualitative studies further suggest that athletically competent teenagers are more likely to enjoy and pursue competitive sports as it provides a platform to demonstrate their abilities to others [43]. On the basis of these findings it is plausible to suggest that organized PA classes (e.g. spin) or informal exercise may represent a less threatening alternative to competitive sport for girls who lack actual or perceived athletic competence.

Higher perceptions of autonomy and autonomy support also differentiated girls characterized by leisure active team sport or active individual sport profiles from less active patterns of PA participation. Regular training sessions with coaches targeting improved performance may facilitate increased opportunities for athletes’ input into decision-making regarding skills that necessitate development, which supports feelings of volition over one’s sport participation [44]. Lower levels of autonomy experienced among girls characterized by participation in dance relative to team sport may be a consequence of the class structure in which limited choices are available regarding dance routine adopted. Conversely, the style of instruction adopted may impact female adolescents’ perceived autonomy. Thus, it is unclear whether the structure of the sport or coaching style differentiated participants’ perceptions of autonomy.

Leisure active team sport profiles also demonstrated higher levels of relatedness relative to PA profiles representing Organized Dance, Walk/Run/Outdoor games and Non-participants. This finding may be attributed to the team sport context that comprises individuals working interdependently as part of a group to achieve goals, which ultimately enhances feelings of connection [18, 19, 45].

The findings also revealed that girls characterized by lower active profiles of participation in Organized Dance, Organized Run/Swim & Dance/Gym, and Walk/ Run/Outdoor games demonstrated higher levels of intrinsic motivation and autonomy than non-participants indicating that even low active patterns of leisure PA are associated with more adaptive psychological processes. Prior research indicates that patterns of PA involvement characterized by lower intensity activity (e.g. walking) are more likely to be maintained across adolescence [8]. In addition, personal exercise reflecting swimming, jogging, going to the gym were the only type of activity in which participation was found to increase among Irish teenagers over the secondary school years [40]. These findings underscore the importance of identifying psychological factors associated with low active PA patterns alongside leisure active participation profiles. It should be acknowledged however, that some of the latent female PA classes contained a small number of participants. Further research comprising a bigger sample of adolescent girls characterized by participation in Organized Dance, and Organized Run/Swim & Dance/Gym is therefore needed for any statistically stronger differences relative to other PA classes to be uncovered.

Furthermore, identifying the motivational characteristics on which girls characterized by low active PA patterns differ from non-participants may also help to identify mechanisms through which inactive individuals can be helped to initiate and engage in a more active lifestyle. The current findings suggest that fostering perceptions of ownership over one’s PA participation, by structuring PA environments to facilitate decision-making opportunities such as offering different PA options within the school or community context may enable girls to identify a form of activity that they enjoy and are motivated to continue [46].

For boys, those characterized by Active Team Sport participation alone or in combination with non-organized PA (Active Mixed Type) evidenced higher scores on intrinsic motivation, competence, relatedness, autonomy, and autonomy support, relative to boys represented by profiles of Leisure Active Individual Sport, Leisure Active Gym or Non-Participation. However, additional research utilizing a larger sample of male adolescents classified within the Leisure Active Gym class in particular, may be necessary to detect any significant differences from other PA groups. While, leisure active patterns of organized sport participation irrespective of format undertaken (e.g. team vs individual) was associated with more adaptive psychological profiles relative to less active patterns of PA among girls, this finding was not replicated among boys. In accordance, boys whose profiles included team sport participation differed significantly from their individual sport counterparts, indicating that psychological processes differ dependent on the type of sport (i.e. individual v team) undertaken by males. Prior research exploring predictors of sustained participation in organized sport have typically grouped individuals that performed different types of sport together (e.g. [18]). This has precluded potential differences pertaining to team and individual sport participation to be uncovered.

The current findings suggest that team sport may play a particularly salient role among male youth, perhaps attributable to increased opportunities within the group setting to develop positive relationships that may not be supported within the context of individual activities such an individual sport or gym sessions. Performing as part of a mutually dependent group facilitates multiple opportunities to form connections, foster cooperation and develop positive relationships with teammates, which ultimately contribute to enhanced feelings of relatedness [19, 45]. The greater feelings of belonging, and affiliation arising from high quality relationships with teammates may also account for the higher levels of intrinsic motivation reported among males represented by profiles of team sport participation relative to their individual sport, gym and non-participant counterparts.

In regard to perceived competence, triumph in team sport is dependent on collaborative effort from multiple members whilst individual sport success is contingent on individual performance. Following on from this, losses may be shared among the team or blamed on errors by multiple people whereas failure in individual sport cannot be attributed to anyone but oneself. It is therefore plausible to suggest that the individual sport environment fosters greater one-on-one comparisons than team sport resulting in decreased perceptions of competence. In addition, leisure active gym members reported lower levels of competence than those characterized by active team sport participation. Boys who lack the necessary skills to gain entry into team sport may opt for this non-competitive fitness alternative that requires little athletic prowess.

Finally, males characterized by patterns of participation in individual sport and team sport significantly differed on perceived autonomy and autonomy support, with the former group performing more poorly. Prior research indicates that participation in individual sports elicits greater pressure than team sport given that performance is contingent on individual as opposed to collective performance [47]. Self-imposed or external pressure from coach or parent to act or think in certain ways undermines feelings of autonomy or sense of ownership over one’s PA participation [19], which may account for the discrepancies evidenced. Further research is necessary to determine whether male adolescents’ autonomy perceptions differ owing to coaching style quality or format of sport undertaken.

The findings offer support for SDT as a framework for exploring motivational and psychological mechanisms underpinning youth PA participation. Most studies to date that have applied SDT to youth PA have typically focused on organized sport or total MVPA, with limited attention directed towards non-organized types of activity such as walking or jogging. The current findings therefore extend the literature by providing a greater insight into the relationship between psychological processes and discrete patterns of organized and non-organized PA participation among adolescents. The results indicate that salience of psychological needs for relatedness, competence and autonomy differs dependent on pattern of PA undertaken. For example, feelings of connection are central to team sport participation but appear to be less important for attending the gym or undertaking informal PA like walking and jogging. Similarly, it appears that perceptions of autonomy may be of greater relevance than perceived competence for maintaining involvement in non-competitive organized PA like dance and fitness classes or more informal games and exercise (e.g. walking).

The cross-sectional nature of the study precludes us however, from making any conclusions regarding the direction of effects between adolescents’ PA patterns and motivational processes or whether latent classes remain stable over time. Longitudinal studies are needed to determine if psychological processes differentially predict sustained involvement, dropout or changing patterns of PA participation during adolescence.

The various profiles of PA participation that emerged have relevance for policies addressing young people’s PA engagement. A consistent finding in this study was that team sport participation was associated with higher levels of MVPA and more adaptive motivational characteristics suggesting that this pattern of activity should be promoted among adolescents. However, adopting such a strategy subsequently neglects individuals who dislike group sports and competitive activities. A simple ‘one size fits all’ approach is not sufficient for promoting participation among teenagers and consequently interventions need to be tailored to reflect different subgroups of adolescents [14]. Whilst many chances exist for young people to take part in sport at an extracurricular or extraschool level, limited opportunities are available for participation in non-sport fitness oriented activities like spinning and aerobics. In addition, non-organized PA such as walking, jogging and cycling could be fostered by increasing the ease with which such activities can be performed, for example, by developing walking and cycling friendly routes that can be safely navigated [10].

The strengths of this study include the relatively large sample size and person-orientated analysis which allowed a more nuanced insight into adolescents’ PA patterns. However, the findings of the study must be considered in view of the study limitations. PA was assessed using a self-report survey. This method has been criticized owing to susceptibility to errors in memory recall, overestimation of PA levels, and socially desirable responses [48]. However, given the relatively large sample size in this study and the need to collect information on the context of PA, the use of costly- and time-intensive objective tools, such as pedometers, was not a viable option [48]. With respect to data analysis, LCA is exploratory in nature and there is no definitive test to facilitate identification of the ‘true’ number of latent classes [49]. However, adolescents’ PA profiles were readily discriminated on motivational characteristics indicating that there were meaningful differences among classes, offering support for the validity of the latent classes [15]. Nevertheless replication of findings among additional samples of secondary school students would offer further support for the validity of these profiles. Finally, class labels were assigned to reflect the defining features of the class but may not encompass the full range of activities undertaken by class members.

This research offers a novel and nuanced insight into the underlying type of activities that constitute overall patterns of male and female adolescents’ PA and further revealed that psychological processes differ as a function of the pattern of PA undertaken.