This article discusses the effects that the training of raters has on estimates of role fidelity and performance consistency of coached clients. The data support the contention that failure to measure and report indexes of role fidelity and performance consistency seriously jeopardizes the high levels of internal validity that the coached‐client analogue research model permits. It also precludes research replication. The findings also suggest that the training of raters in role‐component observation greatly enhances the reliability of the verification procedures proposed by Whiteley and Jakubowski (19