Xiaoxue Gao

Who they are

Her field is exercise psychology and adherence, and the review's value is in its refusal to treat "social" as one variable. Twenty-one studies met inclusion. Where most syntheses report that social support matters or does not matter, this one reports which social measures did which, and that split is the reason the review is load-bearing on this site rather than decorative.

The paper's own reading of its null results is that they reflect poor theoretical targeting rather than the absence of a social mechanism, which is the same objection Nikos Ntoumanis makes about relatedness in the intervention evidence.

Key papers

Gao, X., Song, Y.-G., Shu, Y., Silva, A. F., & Greco, G. (2026). "Perceived Capability, Social Belonging, and Behavioral Adherence in Adult Group Fitness Classes: A Systematic Review of Psychosocial Predictors", Behavioral Sciences, 16(6), 882. First author. Published 1 June 2026. 21 studies met inclusion.

Findings, read from the paper. Self-efficacy was the most consistently supported domain, with small-to-moderate positive associations with attendance-based adherence, most reliable where the efficacy measure was behaviour-specific and prospectively matched to the attendance window. The review is explicit that this is not causal, only the domain for which the association evidence recurs most. Social support and relatedness were weaker than expected, the paper's own section heading naming them as less consistent predictors, with direct associations absent or very small in several studies, one finding guidance inversely associated with attendance and one an adjusted negative association between relatedness and attendance on a weak subscale. Specific relational indicators were a different story: group identification, social attraction and social integration, and task integration over longer follow-up, were more often associated with better adherence.

Key ideas

A gym membership, a class and a team are three different social objects, and only the third reliably supplies identification. That is her review's distinction rather than a rhetorical one.

Self-efficacy, not belonging, is the best-evidenced predictor of turning up. It is also the predictor most popular lists of adherence factors leave out.

The review reports a mechanism shift over time: early participation leans on barrier self-efficacy and conscious self-regulation, sustained participation on routinisation, contextual stability, identity and prior successful participation.

Both sides, then the line. The straightforward reading of her null results is that social factors are weak predictors of exercise attendance. Her own reading, and Ntoumanis's, is that the generic measures are too distal to detect the mechanism. We land with the split she draws: generic support is a weak predictor and should be reported as one, and group identification is a different variable that does predict.

Concepts on this site

Self-Efficacy, the domain her review puts first

Group Identification, the social measure that survived

Exercise Adherence, the outcome throughout

Belonging, the generic version that did not predict

---