The present study looked at the internal reliability of the PEAK Relational Training Assessment, using a split-half method of measurement. The reliability of the assessment questions within each of the four factors, within the PEAK Relational Training Assessment was estimated through this process. Eighteen participants, between the ages of 26 months and ten years old were included in the study. All participants had been diagnosed with either a language based or developmental disability, including autism, seizure disorder, Down syndrome and related language disorders. The PEAK Relational Training Assessment (PEAK-D) was administered by a direct-care provider for each of the 18 participants and during standard instructional periods. Results indicate that for each of the 18 participants, there was a strong correlation between scores when one half of the items in each factor were compared to the remaining half. Results did show internal reliability for the PEAK-D when using split-half methodology.