Perspective taking is typically defined as the ability to reason about others’ mental states (e.g. their beliefs, thoughts, desires, and intentions) and to understand the role of those mental states in everyday situations (I. A. Apperly, 2012). Traditional accounts of perspective taking typically analyze the ability based on three different domains: visual, affective or emotional, and cognitive perspective taking (Flavell, 2004). From a behavioral viewpoint, perspective taking skills are built upon the ability to recognize our own behavior in relation to the context. Relational Frame Theory (RFT) is a contemporary behavioral account of human language and cognition (Hayes et al., 2007). From an RFT viewpoint, perspective-taking skills involve deictic relations between individuals, spaces, and time. Instead of using the three dimensions analyzed in the other fields, RFT studies the development of complex perspective-taking skills through three types of deictic frames: interpersonal (I-YOU-OTHER), spatial (HERE-THERE), and temporal (NOW-THEN-LATER). The purpose of this dissertation was to develop a set of behavioral assessments to measure visual, emotional, and cognitive perspective-taking skills from an RFT viewpoint. This dissertation made methodological and empirical contributions to the field by proposing three behavioral computer-based protocols for evaluating the role of deictic frames on visual, emotional, and cognitive perspective taking tasks. Experiment 1 results revealed significant differences in response latency and correct response levels on interpersonal and spatial deictic frames at simple and reverse levels of complexity on a visual perspective-taking task. These findings suggested that transforming stimulus functions following a mutually entailed relationship between interpersonal and spatial frames is not equivalent to performing conditional discriminations involving both interpersonal and spatial stimuli. Experiment 2 results revealed significant differences in response latency and correct response levels on interpersonal frames with simple, reverse, and double reverse levels of complexity on an emotional perspective-taking task. The finding showed that as the complexity of the deictic relations in emotional perspective taking increased, so did the number of errors and latency to respond. Furthermore, the findings of the study indicate that the valence of emotions has an effect on the levels of deictic relational responding. On a cognitive perspective-taking task, the results of Experiment 3 revealed significant differences in response latency and correct response levels on interpersonal frames with simple and reverse levels of complexity. False beliefs and false desires increased the number of errors and latency to respond to interpersonal deictic frames, according to the findings. Overall, these protocols improved the ecological validity of RFT-based protocols of deictic frames, extended previous research on perspective taking, and opened up new research avenues.