We investigated the effect of family identification (one's subjective sense of belonging to and commonality with the family) on self‐reported ill‐health in 206 Valencian undergraduates, with eight months between Time 1 (T1) and Time 2 (T2). While greater family identification T1 predicted lower ill‐health T2, ill‐health T1 did not predict family identification T2. Family contact T1 (one's intensity of interaction with family) was unrelated to ill‐health T2. This shows that family identification impacts positively on health over time (rather than health impacting positively on family identification over time), and this is not reducible to effects exerted by family contact. These findings indicate that encouraging patients to engage in group activities might produce negligible health gains unless the patient identifies with the group in question.