“I hate it when that happens too!”: observed and perceived exchanges of social support between university student-athletes during discussions of stressors

ABSTRACT

Receiving social support is often related to adaptive outcomes for athletes, such as reduced stress and greater wellbeing. Nevertheless, in some cases, receiving social support may not reduce, or even heighten, stress reactions. Thus, research is needed to better understand the complex dynamics of social support processes among athletes. Though substantial research has examined social support in sport, the majority of these studies relied on self-report measures; there is a lack of information about the actual interactions that constitute social support exchanges among teammates. This study sought to fill this gap by observing athletes’ discussions of stressful experiences, with a focus on understanding the types of social support behaviours that are displayed. University athlete dyads (N = 46 dyads, M_age = 20.2 years, SD = 1.9) were asked to complete surveys and engage in a conversation regarding stressors in a lab setting. The conversations were video-recorded and analysed along with self-reported survey data. Results showed that in addition to emotional, esteem, and informational support, athletes commonly engaged in co-rumination – collective discussions of stressful experiences with a persistent focus on negative feelings associated with the experience. Correlational analysis revealed that observed informational support was positively associated with perceived emotional and informational support, and co-rumination was positively associated with perceived stress. Emotional, esteem, and informational support did not relate to stress perceptions. The findings were interpreted to highlight the importance of considering co-rumination and we suggest potential avenues for future research on social support in sport that might further illuminate this salient component.

KEYWORDS:

• coping
• sports
• observation
• emotion regulation
• rumination

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The anonymized data that support the findings of this study are available at: https://osf.io/nc74f/.

Notes

1 Tangible support was not assessed because the items assessing tangible support were deemed irrelevant in the context of their conversation (e.g., “ … help with travel to training and matches”, “ … help with tasks to leave you free to concentrate”; Freeman et al., 2011).

2 The value represents the two-way mixed, single-measures, absolute value agreement ICC (Hallgren, 2012). According to Cicchetti (1994), ICC values between .75 and 1.00 can be considered excellent.

3 Stress ratings were used for computing mean stress variables only if both pre- and post-conversation stress ratings were available, and if the support seeker indicated that the given stressful situation was discussed during the conversation. For example, if the support seeker indicated that the third stressful event they listed on the survey was not discussed during the conversation, or if the pre- or post-conversation stress rating for the third stressful event was unavailable, ratings for this event were not used for computing the mean pre- and post-conversation stress variables.

4 Shapiro-Wilk tests revealed statistically significant results for trust (p = .002), observed esteem support, as well as perceived emotional, esteem, and informational support, indicating they were significantly skewed (ps < .001).

5 To further test the robustness of this finding, a post-hoc linear regression model was conducted to examine whether co-rumination predicted post-conversation stress after accounting for the influence of pre-conversation stress. Pre-conversation stress was entered in step 1, and co-rumination was entered in step 2, and change in model statistics was observed. Addition of co-rumination in the second step resulted in a significantly improved fit, ΔF(1,43) = 4.85, p = .033, ΔAdjusted R² = .08. Co-rumination positively and significantly predicted post-conversation stress, b = 8.0, SE = 3.61, p = .033, after accounting for the influence of pre-conversation stress, b = .45, SE = .24, p = .068.

Additional information

Funding

This work was supported by Social Sciences and Humanities Research Council of Canada Insight Grant awarded to the second author: [Grant Number 435-2018-0869].