The purpose of the proposed study is to consider why and under which conditions do supervisors engage in abusive behaviours towards their subordinates. To answer my first research question, why do supervisors engage in abusive supervision, I draw on victim precipitation (e.g., Sparks, Glenn, & Dodd, 1977) and conservation of resources (COR; Hobfoll, 1989) theories to argue that certain subordinate performance-related behaviours and characteristics threaten supervisor resources leading to abuse as a stress reaction. To answer my second research question, under which conditions do supervisors engage in abusive supervision, I draw on attribution theory (Heider, 1958; Weiner, 1986). I argue that supervisors abuse subordinates when they attribute responsibility, or blame subordinates for negative performance-related behaviours and characteristics, as a means of protecting or guarding against future resource loss. To answer my research questions, I developed measures for self- and other-perceived general mental ability (GMA) and blame attributions. I obtained data from 211 supervisor-subordinate dyads in Canada and the United States. Respondents were surveyed for information about their work behaviours, characteristics, and relationships. Using Hayes (2013) PROCESS macros, I found partial support for the proposed model and offer refinements to COR and victim precipitation theories. I found relationships between both self- and supervisor-reported subordinate behaviours and characteristics and abusive supervision, largely in the direction hypothesized. I also found supervisor-reported subordinate performance behaviours and perceived GMA to share a stronger relationship with subordinate reports of abusive supervision than subordinate reported behaviours and characteristics in many instances. / October 2016