An exploration of Keyes’ two-continuum model of mental health in athletes: resilience, mental illness and performance

Keyes’ (2005) two-continuum model of mental health posits two related, but distinct dimensions (1: the presence or absence of mental health – MH; 2: the presence or absence of mental illness - MI). Theoretically, athletes could experience both positive MH and symptoms of MI. Alternatively, athletes could be free from MI, but experience low levels of MH (what Keyes, 2005 terms languishing). This study presents preliminary results from an online survey examining (a) associations between resilience, MH, and MI, and (b) associations between MH, MI and performance. Participants comprised (male, n = 29; female, n = 28) athletes from a range of team (e.g., soccer, netball) and individual (e.g., triathlon, golf) sports (mean age = 23 ± 7 years). The survey comprised measures of MH (Keyes et al., 2008), MI (Connell et al., 2007), resilience (Wagnild & Young, 1987), and performance [the mean of 3 items assessing satisfaction in training, competition, and in sport generally from 0 (totally dissatisfied) to 100 (totally satisfied)]. Using proposed cut-off criteria (Connell et al, 2007; Keyes et al, 2008), some individuals (12%) reported both severe MI and high MH. In addition, a modest, negative relationship (r = -.40, p = .003) between MH and MI lends some support to Keyes’ model. Resilience (personal competence) was associated with MH (r = .50, p < .01), and MI(r = -.34, p = .01). Resilience (acceptance of self and life) was not associated with MI (r = -.24, p = .08), but was associated with MH (r = .39, p = .003). Zero-order correlations between MH and performance (r = .63, p < .001), and MI and performance (r = -.40, p = .003) are qualified by partial correlation analyses. The correlation between MH and performance remains significant when MI is controlled for (r = .59, p < .001). When MH is controlled for, the relationship between MI and performance (r = -.05, p = .76) is attenuated. Collectively, results provide some support for Keyes’ model and for considering MI and MH as separate factors influencing sport performance.