Physical Activity-Related Profiles of Female Sixth-Graders Regarding Motivational Psychosocial Variables: A Cluster Analysis Within the CReActivity Project.

Joachim Bachner,Yolanda Demetriou,Xavier García-Massó,Javier Molina-García,David J Sturm

doi:10.3389/fpsyg.2020.580563

Joachim Bachner, Yolanda Demetriou + Show 3 more

Open Access

https://doi.org/10.3389/fpsyg.2020.580563

Copy DOI

Abstract

IntroductionAdolescents’ physical activity (PA) behavior can be driven by several psychosocial determinants at the same time. Most analyses use a variable-based approach that examines relations between PA-related determinants and PA behavior on the between-person level. Using this approach, possible coexistences of different psychosocial determinants within one person cannot be examined. Therefore, by applying a person-oriented approach, this study examined (a) which profiles regarding PA-related psychosocial variables typically occur in female sixth-graders, (b) if these profiles deliver a self-consistent picture according to theoretical assumptions, and (c) if the profiles contribute to the explanation of PA.Materials and MethodsThe sample comprised 475 female sixth-graders. Seventeen PA-related variables were assessed: support for autonomy, competence and relatedness in PE as well as their satisfaction in PE and leisure-time; behavioral regulation of exercise (five subscales); self-efficacy and social support from friends and family (two subscales). Moderate-to-vigorous PA was measured using accelerometers. Data were analyzed using the self-organizing maps (SOM) analysis, a cluster analysis including an unsupervised algorithm for non-linear models.ResultsAccording to the respective level of psychosocial resources, a positive, a medium and a negative cluster were identified. This superordinate cluster solution represented a self-consistent picture that was in line with theoretical assumptions. The three-cluster solution contributed to the explanation of PA behavior, with the positive cluster accumulating an average of 6 min more moderate-to-vigorous PA per day than the medium cluster and 10 min more than the negative cluster. Additionally, SOM detected a subgroup within the positive cluster that benefited from a specific combination of intrinsic and external regulations with regard to PA.DiscussionThe results underline the relevance of the assessed psychosocial determinants of PA behavior in female sixth-graders. The results further indicate that the different psychosocial resources within a given person do not develop independently of one another, which supports the use of a person-oriented approach. In addition, the SOM analysis identified subgroups with specific characteristics, which would have remained undetected using variable-based approaches. Thus, this approach offers the possibility to reduce data complexity without overlooking subgroups with special demands that go beyond the superordinate cluster solution.

Highlights

Adolescents’ physical activity (PA) behavior can be driven by several psychosocial determinants at the same time
With regard to participants in cluster 3, called the positive cluster, it becomes clear that the students who felt most supported in their BPN during physical education (PE) were the ones who felt most satisfied in their BPN during PE and LT
The results support the use of a person-oriented approach, as the self-organizing maps (SOM) analysis could show that the different psychosocial resources do not develop independently of one another within a given person (e.g., Teixeira et al, 2012; Bergman and Lundh, 2015)

Summary

Introduction

Adolescents’ physical activity (PA) behavior can be driven by several psychosocial determinants at the same time. Most analyses use a variable-based approach that examines relations between PA-related determinants and PA behavior on the between-person level. Analysis of accelerometer data of over 26,000 5-17-year olds from ten countries indicated an even lower rate with 9% of boys and 1.9% of girls fulfilling the WHO guideline (Cooper et al, 2015). In addition to this gender effect, which is consistently found during adolescence, PA decreases with age (Dumith et al, 2011; Van Hecke et al, 2016). There have been ambiguous findings, a lower socioeconomic status (SES) of the household tends to have a negative effect on PA levels in children and adolescents (Sallis et al, 2000; Corder et al, 2011; Molina-García et al, 2017; Finger et al, 2018)

Methods

Results

Discussion

Conclusion