Testing a sequence of relationships from interpersonal coaching styles to rugby performance, guided by the coach–athlete motivation model

Abstract

This study was designed to test a four stage sequence of relationships derived from the coach–athlete motivation model (CAMM; [Mageau, G. A., & Vallerand, R. J. (2003). The coach-athlete relationship: A motivational model. Journal of Sport Sciences, 21, 883–904. doi: 10.1080/0264041031000140374]) from interpersonal coaching styles to athletes' basic psychological needs, to self-determined motives, and finally rugby performance. Participants were 221 Canadian Interuniversity Sport (CIS) female rugby players who ranged from 16 to 33 years of age and reported playing rugby for an average of six years. The athletes completed a self-report questionnaire which included demographics, interpersonal coaching styles, basic psychological needs, self-determined motivation, and performance. An analysis of the structural model yielded the following fit index scores; (χ² (157) = 328.00, p = .000; Comparative Fit Index  = .95; Incremental Fit Index  = .95; Root Mean Square Error of Approximation = .07; Standardised Root Mean Square Residual = .09. The interpersonal coaching styles of autonomy support (β = .66), structure (β = .35), and involvement (β = .44) positively predicted the athletes corresponding needs. However, only autonomy significantly (β = .60) predicted self-determined motivation, while competence (β = .01) and relatedness (β = .07) did not. Finally, self-determined motivation moderately predicted rugby performance (β = .35). These findings provide initial support for propositions offered in the CAMM that the structure and involvement provided by coaches are positively associated with optimising the athletic experience. Furthermore, this study contributes to the existing literature by attempting to replicate previous studies that have examined at least one of the relationships examined in the present investigation using a varied population of female rugby players.

Keywords:

• basic psychological needs
• coaching behaviours
• self-determined motivation
• self-determination theory
• sport

Acknowledgements

Thanks are extended to the participants who gave freely of their time to invest in this research. This study was completed in partial fulfilment of a Master of Arts thesis at Brock University by the first author under the supervision of the second author.

Funding

This research was supported by a grant from the Social Sciences and Humanities Research Council of Canada (SSHRC) awarded to the second author. The first author is supported by a graduate scholarship from the SSHRC at the time of manuscript submission.

Notes

1. This study included only athletes' self-determined motives (intrinsic, integrated, and identified regulation) in order to be consistent with the CAMM. Mageau and Vallerand (2003) rationalised that the self-determined motives are necessary for optimal functioning while non self-determined motives are associated with less desirable outcomes.

2. This convenience-based sample of CIS female rugby players differed from the existing research as most studies have included samples of both genders (Amorose & Anderson-Butcher, 2007; Curran et al., 2013), and either more than one sport (Adie et al., 2008; Hollembeak & Amorose, 2005), or individual-based sports (Coatsworth & Conroy, 2009; Gagné et al., 2003).

3. Two BNSSS measurement models were analysed to determine if the basic psychological needs were better represented as five factors (competence, IPLOC, volition, choice, and relatedness) as per the BNSSS or the theoretically based three factors (competence, autonomy, and relatedness) put forth by Deci and Ryan (2002). Results from the standardised factor loading and further inspection of bootstrapped confidence interval (CI) indicated that volition and IPLOC were highly correlated (r = .93), and the upper boundary of the CI exceeded 1.00. These results therefore informed us that volition and IPLOC were not distinctly different, thus failing to support the use of the five factor model, which grounded our decision to use three factors in the path model.

4. Domain representative parcelling was selected for the multidimensional constructs as it allowed us to reflect all of the faucets/subscales of a construct within one parcel, such that each parcel was an individual representation of the overall construct (Little et al., 2002).