The present investigation sought to examine the efficacy of the instructional curriculum described in the Direct Training Module of the PEAK Relational Training System on the language repertoires, as measured by the PEAK direct assessment, of children diagnosed with autism or related developmental disabilities. Twenty-seven children diagnosed with pervasive developmental disorders were evaluated using the PEAK direct training assessment protocol prior to assignment to control and experimental groups. Participants in the experimental group received additional language instruction derived from the curriculum programs of the Direct Training Module, while participants in the control group received treatment as usual. Both groups were then re-assessed using the PEAK direct assessment after 1 month. A repeated measures ANOVA indicated that participants in the experimental group made significantly more gains in language skills than those who were assigned to the control group, F(1, 22) = 9.684, p = 0.005. Implications for evidence-based practice and future research are discussed.