As part of standards-based reforms, there is increasing emphasis on ensuring that students with moderate intellectual disabilities (ID), including students with Autism Spectrum Disorders (ASD), learn to read. There is also converging evidence that explicit teaching of letter sounds, phonics, and sight words is effective for this population, but that students' responsiveness varies. A critical part of individualizing reading instruction for students with disabilities is the reliable assessment of progress and mastery of reading skills. However, assessment of many students with ID and students with ASD is challenging because of attention, behavioral, and communication issues related to testing situation; therefore, obtaining consistent results often proves to be a difficult task. We hypothesized that alternate assessment presentation formats, as a testing accommodation, would improve the reliability, validity, and consistency of assessment performance. In this study, three different presentation formats--word lists, flash cards, and PowerPoint presentation--were used when administering proximal, curriculum-based reading assessments to determine whether a particular format increased student engagement, reduced the need for prompts, and increased accuracy of identifying known items on the test. While statistical analyses did not support the hypothesis of a format by student effect, visual analysis of the data did suggest that the number of prompts required varied by student as a function of assessment format. Most noteworthy, assessment reliability, estimated with generalizability theory, indicated that reliability increased as a function of format by student. [For the corresponding grantee submission, see ED586421.]