Homophily is an intuitive behavior in opinion dynamics, although its identification is not easy when using real data. In this paper, the Stochastic Actor-Oriented Model (SAOM) is proposed to detect homophily relationships emerging from Reddit discussions on three different topics: minorities discrimination, gun control, and politics. Homophily is characterized by the similarity of opinions between actors who connect over five sequential waves, each determined by collecting data over six months. Two indices to evaluate the tendency towards homophily-based connectivity are defined by considering the deviations of the actor’s opinion from the average of all opinions and from that obtained by randomly choosing the same number of neighbors. The SAOM similarity score highlights situations in which homophily influences the variations of connections on consecutive waves and the Jaccard index is used to analyze the similarity of the graphs produced by the model with the experimental ones.