The effects of testing adaptations on students' standardized test scores for students with visual impairments in Arizona

To meet requirements of Child Left Behind (NCLB) and the Individuals with Disabilities Education Act (IDEA), students with disabilities must be included in standardized assessment to measure their progress in the general curriculum (Public Law 107-110, 2002; Education Development Center, 1999). When implementing standardized assessments, all aspects of the assessment are to be standardized, to include administration procedures and time (Packer, 1989). The purpose of this study was to determine the relationship of testing modifications, a type of adaptation that may be necessary for students with disabilities, and the effects of demographic information on test scores for students with visual impairments. Ethnicity, home language, reading medium, and disability classification were considered. Typical testing modifications and possible re-occurring cluster data were analyzed. The sample consisted of 71 students in grades two through nine who attended a specialized school for the visually impaired or a public school with support from teachers of the visually impaired. Students' 2001--02 stanine scores for Total Reading, Total Mathematics, and Language from the Stanford Achievement Test, 9th edition were analyzed. A factorial analysis of variance (ANOVA) and factorial analysis of covariance (ANCOVA) were used to analyze previously collected data. A principal components analysis was completed on modification data to reduce the modifications into common reoccurring clusters. ANOVA and ANCOVA results indicated that reading medium, alone or in a cluster, has an effect on Total Mathematics and Language scores. In both the ANOVA for Language and the ANCOVA for Total Mathematics the reading medium of large print had the highest mean score followed by print then Braille. The ANOVA for Total Mathematics results showed print had a slightly higher mean score than large print, followed by Braille. When analyzing testing modification an effect was found in the area of Total Mathematics when reading medium was combined or clustered with other variables. When completing the principal component analysis the 19 variables were clustered and reduced to 4 components