In the context of stochastic choice, we introduce a model which admits a cardinal notion of interactive influence among individuals. The model presumes that individual choice is not only determined by idiosyncratic evaluations of the alternatives but also by the influence from the observed behavior of others. We establish that the equilibrium defined by the model is unique, stable and falsifiable. Moreover the underlying preference and influence parameters are uniquely identified from, arguably, limited data. The baseline model includes two individuals with conformity motives. Generalizations to multi-individual settings and negative interactions are also introduced and analyzed. Our analysis can be interpreted as providing an empirically meaningful foundation for network learning models.