Social network companies often carry out experiments to test product changes and new ideas. In such experiments, the outcome of one user may be influenced by the treatment assignment of other users, which necessitates specialized design and analysis methods. We introduce the general additive network effect (GANE) model, which unifies many existing outcome models under an encompassing model-based framework. The model is interpretable and flexible in modeling both the treatment effect and the network influence. We show that maximum likelihood estimators are consistent and asymptotically normal for a family of model specifications. Causal quantities of interest such as the global treatment effect are defined and expressed as functions of the GANE model parameters, and hence inference can be carried out accordingly. We further propose the “power-degree” (POW-DEG) specification of the GANE model, whose performance, together with that of other specifications, is investigated via simulations.