This study was conducted to test whether the weight outcomes in an online social networking group were mediated by changes to psychological outcome measures in overweight and obese individuals, following a weight management programme delivered via Facebook. The data analysed in this study were collected during a three‐armed, randomised, controlled clinical weight management trial conducted with overweight and obese adults over 24 weeks. Two intervention groups were given the same weight management programme: one within a Facebook group, along with peer support from other group members (the Facebook Group); the other group received the same programme in a pamphlet (the Pamphlet Group). A Control Group was given standard care. The primary outcome was weight; secondary outcomes included the following domains from self‐reported questionnaires: energy intake and expenditure; psychological health, social relationships, physical health, quality of life, depression, anxiety, stress, health anxiety, happiness, as well as Facebook Group participants’ opinion of this group. The Facebook Group experienced a reduction in their baseline weight measurement by week 24, significantly compared to the Control Group (p = .016). The Facebook Group recorded a significant increase in the psychological health domain during the trial (at week 12) relative to their baseline measurement, and significant compared to the Control Group (p = .022). Mediation analysis indicated a statistical trend, but not statistical significance, for psychological health as a mediator to weight loss in the Facebook Group. While both intervention groups showed significant changes in psychological outcome measures, the Facebook Group was the only group to experience statistically significant weight loss by the end of the 24 weeks. Therefore, an examination of other psychological and/or behavioural outcome measures undertaken in larger studies in the future may help to identify significant mediators to improved weight loss outcomes in online social networking groups.