Johnson, E. S., Kimball, K., Brown, S. O., & Anderson, D. (2001). A statewide review of the use of accommodations in large-scale, high-stakes assessments . Exceptional Children , 67 (2), 251–264. https://doi.org/10.1177/001440290106700208
Johnson, E. S., Kimball, K., Brown, S. O., & Anderson, D. (2001). A statewide review of the use of accommodations in large-scale, high-stakes assessments. Exceptional Children, 67(2), 251–264. https://doi.org/10.1177/001440290106700208
Accommodations that are permitted for the Washington Assessment for Student Learning were included in the analysis. These include aids (dictionaries, physical supports or assists, isolate portions of the test, clarify directions), scribe (answer orally, point, use voice recognition technology, sign answers, or use word processor), large print/Braille, and oral delivery (read-aloud).
Students in grades 4 and 7 from across the State of Washington (U.S.) were included in the analysis; these included students in general education, special education (n=14,738), Section 504 students (n=811), bilingual students (n=5,540), and migrant students (n=1,077). Altogether, data from 147,080 students were analyzed.
Performance on the listening, reading, writing, and math sections of the Washington Assessment for Student Learning was analyzed.
Only a small percentage of students in the 'met the standard' category received accommodations. However, many of those in special populations who met the standard received an accommodation. When scores for accommodated students in special populations were compared to the performance of a random subset of students in general education, results indicated that the students in general education performed significantly higher than students who had read-aloud accommodations for math, and students who had a scribe for the writing assessment. Students in special education who received accommodations outperformed, on average, special education students receiving no accommodations. In interpreting these results, however, it is important to note that these analyses were post-hoc (non-experimental).