The constructive replication of a prespecified, cognitively mediated model of rape's impact on psychosocial health is reported using longitudinal data (see Koss, Figueredo, & Prince, 2002, for a summary of model development). Rape survivors (n= 59) were assessed four times, 3 to 24 months postrape. Structural equations modeling of baseline data (intercepts) and rate of change over time (slopes) revealed that all large effects replicated, smaller effects did not. The model's central features were confirmed and showed that Psychological Problem History exacerbated Characterological Self-Blame, leading to more Maladaptive Beliefs, which determined initial Psychosocial Distress and its rate of decline. The major contributions of the study include: (a) placement in a research program designed to balance the strengths/limitations of cross-sectional and longitudinal data; (b) analysis of prerape characteristics, cognitive mediators, and multiple psychosocial distress variables in a system; and (c) a strategy for structural equations modeling in small samples.