The nature of the alliance-outcome relationship is still emerging. This study examined the reciprocal influence of change in alliance to the group and change in urge to restrict in eating-disordered individuals attending a group-based day treatment. Participants (N = 238) were a transdiagnostic or mixed diagnostic sample of eating-disordered individuals consecutively admitted to a day treatment program. On a weekly basis, participants completed a measure of alliance to the group of patients with whom they attended multiple group therapies each week. After each meal, they rated the intensity of their urge to restrict food intake, and the intensity ratings were averaged per week. Latent change score analysis was used to assess the reciprocal relationship between prior change in alliance to the group with subsequent change in urge to restrict, and prior change in urge to restrict with subsequent change in alliance to the group across each participant's first 9 weeks in the program. A reciprocal causal model was a good fit to the data. Prior growth in alliance to the group was significantly associated with subsequent reduction in urge to restrict, and concurrently, prior reduction in urge to restrict was significantly associated with subsequent growth in alliance to the group. Alliance to the group and individual outcomes are dynamically related and changing constructs represented by a reciprocal causal model. Clinicians may improve group treatment by assessing alliance to the group and outcomes repeatedly, being aware of their interplay, and structuring interventions based on the mutual causal effects of change in each. (Contains 2 figures and 4 tables.)