General approaches for Group Recommendation Systems start from the individual recommendations and merge them in a way to determine the best choice for the whole group. The results presented in the literature showed that traditional aggregation techniques do not seem to capture all the features of real-world scenarios. Furthermore, recent studies in Behavioral Economics evidence the necessity to define utility models that are not compatible with the self-interested utility-maximizing behavior of the traditional economic paradigm. In this work, starting from Other-Regarding Preference models that characterize the utility of an individual considering his/her own behavioral characteristics and the utility of another individual, we aim at obtaining a general model where such characteristics are described in terms of interpersonal relationships, such as, tie strength and conflict. We started by performing an analysis on opinion shifts based on a two sized groups user study, with the aim to empirically determine the extent of the considered parameters on a possible model. The results show that the opinion shifting on the evaluation of an activity to be performed in a group is related to the two considered factors.