To what extent user's stance towards a given topic could be inferred? Most of\nthe studies on stance detection have focused on analysing user's posts on a\ngiven topic to predict the stance. However, the stance in social media can be\ninferred from a mixture of signals that might reflect user's beliefs including\nposts and online interactions. This paper examines various online features of\nusers to detect their stance towards different topics. We compare multiple set\nof features, including on-topic content, network interactions, user's\npreferences, and online network connections. Our objective is to understand the\nonline signals that can reveal the users' stance. Experimentation is applied on\ntweets dataset from the SemEval stance detection task, which covers five\ntopics. Results show that stance of a user can be detected with multiple\nsignals of user's online activity, including their posts on the topic, the\nnetwork they interact with or follow, the websites they visit, and the content\nthey like. The performance of the stance modelling using different network\nfeatures are comparable with the state-of-the-art reported model that used\ntextual content only. In addition, combining network and content features leads\nto the highest reported performance to date on the SemEval dataset with\nF-measure of 72.49%. We further present an extensive analysis to show how these\ndifferent set of features can reveal stance. Our findings have distinct privacy\nimplications, where they highlight that stance is strongly embedded in user's\nonline social network that, in principle, individuals can be profiled from\ntheir interactions and connections even when they do not post about the topic.\n