Intervention effects on physical activity identity: a systematic review and meta-analysis

Physical activity (PA) identity (i.e., categorisation of oneself in a particular role) has been linked to PA behaviour in observational research, yet experimental research has seen less attention. The purpose of this meta-analysis was to examine the effectiveness of interventions to change identity and subsequent PA. Eligible studies were published in a peer-reviewed journal in English, included an experimental or quasi-experimental design in the PA domain with a measure of identity as the dependent variable, among an adult (>18 yrs.) sample. A literature search was completed in March 2024 using five common databases. The search yielded 40 independent effect sizes, representing 4939 participants. Random-effects meta-analysis showed positive changes in identity favouring the intervention over the control group g = 0.18 (95% CI = 0.11-0.24) and positive changes in a sub-sample (k = 30) of these studies that also measured PA g = 0.61 (95% CI = 0.41-0.81). Changes in identity did not have significant (Q = 43.08, p = 0.30) heterogeneity, yet changes in PA showed heterogeneity (Q = 204.62, p < .001) and follow-up moderator analyses found potential publication bias, and differences by methods (comparison group, length of intervention) and theoretical approach. Overall, PA identity can change as a result of interventions, but the effect may be smaller than changes in behaviour in these interventions.