Contradictory findings about whether agentic women are penalized or rewarded persist in gender and leadership research. To account for these divergent effects, we distinguish between agentic traits that people believe female leaders ought to possess (i.e., agency prescriptions) and ought not possess (i.e., agency proscriptions). We draw on expectancy violation theory to suggest that an agentic advantage is elicited when women are perceived to violate agency prescriptions (e.g., competence), whereas an agentic disadvantage is elicited when they are perceived to violate agency proscriptions (e.g., dominance). We first developed and validated a new, six-factor measure of agency in Studies 1 and 2, CADDIS (i.e., C ompetent agency, Ambitious agency, D ominant agency, D iligent agency, I ndependent agency, and Self-assured agency). We theorized that these agency factors represented distinct agency prescriptions and proscriptions for men and women. In Studies 3-5, we found that this six-factor conceptualization of agency not only reconciles existing tensions within the gender and leadership literature, but also leads to a different understanding of past conclusions-an agentic advantage occurs when women are perceived to possess competent agency, diligent agency, and independent agency, and an agentic disadvantage occurs when women are perceived to possess dominant agency. (PsycInfo Database Record (c) 2022 APA, all rights reserved).