This dissertation explores when and how the social comparisons that employees make with respect to their LMX (leader-member exchange) relationships affect their work performance and behaviors. The study introduces the concept of LMX social comparison across dyads (LMXAD) in which a follower compares the quality of his/her supervisory relationship to other leader-member dyads outside of the workgroup (e.g., my leader-myself vs. other leaders-other colleagues). Thus, the study sheds light on LMX social comparison processes at a dyadic level (e.g., our relationship vs. their relationships) as opposed to the individual level (e.g., my relationship vs their relationships, when followers share a same leader) to highlight the importance and saliency of leader-member dyadic comparisons. Drawing upon Thibaut & Kelley (1959)’s social exchange theory, the study, which collected data from 318 employees in Korean companies, empirically supported the positive effects of LMXAD on work performance, organizational citizenship behavior (OCB), and the negative effects of LMXAD on counterproductive work behavior (CWB), beyond LMX and LMX social comparison within group (e.g., my leader-myself vs. my leader-coworkers). Furthermore, results suggest upward counterfactual thoughts with regards to the current LMX relationship, mediates the relationship between LMXAD and work performance and CWB. Individual LMX and causal attributions also have a moderating effect by weakening the negative effects of LMXAD on upward counterfactual thoughts.