In categorical perception research, the category boundary effect is a well-known phenomenon: Keeping physical distance equal, differences between stimuli belonging to the same category are perceived as smaller than differences between stimuli belonging to different categories (i.e., stimulus pairs crossing the category boundary; Harnad, 1987). How can this category boundary effect be explained? It is sometimes seen as a consequence of reference points that either increase (i.e., anchor effect) or decrease (i.e., magnet effect) sensitivity in their vicinity (Medin & Barsalou, 1987). The effect can be explained by the perceptual anchor hypothesis when reference points are assumed to lie at the category boundary, and by the perceptual magnet hypothesis when reference points are around the prototypical exemplars or around the ideals. Before we can determine whether reference points increase or decrease discrimination performance in their vicinity, we need to empirically determine where exactly the reference points are. Three different methods, based on well-known consequences of reference points, are used for this purpose: (a) a categorization task; (b) a discrimination task; and (c) a similarity judgment task. Additionally, we link these ideas on reference points to a Gestalt theoretical view on prägnant percepts and their impact on sensitivity to change.