This study investigated the effects of introducing a family planning counseling model at clinics of Peru's Ministry of Health. Providers trained in the model presented greater quality of care and longer counseling sessions than did controls. The main effects, however, were misleading. Nearly all of the quality improvements were contributed by 37% of the trained providers; they had already been better performers at the pretest and complied with the new counseling model's requirement of job aid use at the posttest. The recognition that a majority of trained providers did not profit from training posed a challenge that led to improvements in the model. Training x Trainee Interactions should be explored in evaluations. © 2005 Sage Publications.