Empathy is the basis of teacher caring. In contrast to social perspective-taking, relatively little is known about the relationship between visuospatial perspective-taking and teacher empathy. The present paper addresses this question based on Embodied Simulation Theory and The Self to Other Model of Empathy. Visuospatial perspective-taking involves two components named Egocentric Visuospatial Transformation and Self Representation Inhibition. The present study manipulates these two components by stimulus location (wall or table) and stimulus type (inconsistent or consistent) respectively. In present study, we combine stimulus location and type to generate 4 different kinds of experimental conditions with 8 pictures showing a number in different locations and a person sitting opposite to participants. Ninety-seven teachers (24-56 years old, M = 34.08, SD = 6.39) volunteered to complete a computer-assisted task (E-prime 2.0). For each trial, participants performed visuospatial perspective-taking before Multifaceted Empathy Test (MET). MET could be used to measure both affective empathy and cognitive empathy. To measure affective empathy, participants were asked to rate how much they were feeling for an individual in each scene on a 9 points scale. To measure cognitive empathy, participants were required to infer the mental state of the subject in each scene and indicate the correct mental state from a list of 4 responses. The two aspects of empathy were each tested with 20 stimuli with positive valence and 20 stimuli with negative valence, resulting in a total of 80 pictures. Participants completed 8 practice trials followed by 80 test trials. Total experiment task was performed 10-15 minutes. The present study performed data preprocess with Excel 2016 and data analysis with SPSS 25.0. Trials with errors on the visuospatial perspective-taking task were excluded from all analyses. Data that were 3 standard deviations from the mean were omitted from the analysis of RTs.

Results showed that, for affective empathy, the analysis yielded a significant main effect of stimulus location (p < .05) and type (p < .001). Participants exhibited higher ratings of effective states after taking perspective on desk relative to perspective on wall. Participants exhibited higher ratings of affective states after taking inconsistent perspective relative to consistent perspective. For cognitive empathy, the main effects of stimulus location (p < .01) and type (p < .05), as well as interaction between stimulus location and type were significant (p < .001). Simple-effects tests revealed that participants exhibited higher accuracy after taking visual perspective on desk relative to perspective on wall under inconsistent condition. However, the analysis revealed an inverse pattern of visuospatial perspective-taking on cognitive empathy under consistent condition.

Those results suggest that two components of visuospatial perspective-taking have different impacts on teacher empathy. Egocentric Visuospatial Transformation promotes teachers’ affective empathy but weakens cognitive empathy. Self-Representation Inhibition promotes both teachers’ affective empathy and cognitive empathy. However, teachers' cognitive empathy can be improved only by egocentric visuospatial transformation under the condition of high Self-Representation Inhibition. Our results provide evidence for The Self to Other Model of Empathy, which helps us understand the effect of perspective-taking on teacher empathy.