Kettler, R. J., Dickenson, T. S., Bennett, H. L., Morgan, G. B., Gilmore, J. A., Beddow, P. A., Swaffield, S., Turner, L., Herrera, B., Turner, C., & Palmer, P. W. (2012). Enhancing the accessibility of high school science tests: A multistate experiment . Exceptional Children , 79 (1), 91–106. https://doi.org/10.1177/001440291207900105

This study, part of the Operationalizing Alternate Assessment for Science Inquiry Skills (OAASIS) project, examined various assessment enhancements—both accommodations and modifications, clearly defined—that were made to sets of science test items resulting in the creation of a new test for research purposes. These enhancements were implemented in a package approach, combining several accommodations and modifications together simultaneously. For instance, all participants in the enhanced test condition received read-aloud and added white space; nearly all (93%) received eliminated answer choice; a majority (at least 50%) received shortened item stem and simplified vocabulary; a large minority (at least 20%) received simplified answer choices, changed graphic, and/or bulleted text instead of passage format; and small minorities (10% or less) each received difficult vocabulary defined, added graphic, and/or eliminated graphic.

Four hundred high school students from three U.S. states (South Carolina, South Dakota, and Wyoming) participated, composed of students without disabilities (n=228), students with disabilities who were eligible to take the AA-MAS (n=88), and students with disabilities not eligible for the AA-MAS (n=84). Disability categories of participants were various, yet researchers indicated that about 69% of students had specific learning disabilities. Additional demographic details including sex and ethnicity of students were also reported.

Scores for two 20-item sets of computer-delivered science items on high school biology content, common to all three states, were compared across experimental conditions. The Accessibility Rating Matrix (ARM; Beddow, Elliott, & Kettler, 2009) served to evaluate the science tests. A Maze reading fluency measure provided participant skill data. Student participants also completed a survey about their perceptions of the test enhancements.

All students' scores on enhanced tests averaged about 2 points higher than on the unenhanced form, and differences by participant groups were similar, indicating no differential benefits in total score for students with or without disabilities. For the 40 enhanced items, 7 showed a differential boost, and 11 showed a minimal differential boost, over the original item format. Enhancements which seemed beneficial included shorter item stems, and enhancements that did not seem beneficial included bulleted text and simplified graphics, according to the specific items which had those enhancements. The test-takers mostly indicated that the enhancements seemed to decrease the difficulty of testing, but when viewing specific original items and enhanced items, they indicated that each item had a similar level of difficulty. Limitations of the study were reported, and future research directions were suggested.