Pomplun, M., & Omar, M. H. (2000). Score comparability of a state mathematics assessment across students with and without reading accommodations . Journal of Applied Psychology , 85 (1), 21–29. https://doi.org/10.1037/0021-9010.85.1.21

Many students with learning disabilities received the read-aloud accommodation according to what was determined to be necessary on an individual basis. Other students did not receive the read-aloud accommodation.

A random sample of 1,500 general education students and 1,609 students with learning disabilities (LD; 52% of sample), all in grade 4 in Kansas (U.S.), participated. Of the students with LD, 1,369 took the assessment with no accommodations, and 173 received the read-aloud assessment accommodation. Approximately 17% were minority students. Female students were 48% of the general education sample, 32% of the non-accommodated group of students with LD, and 24% of the accommodated group.

The math assessments from the Kansas Assessment Program were used as a dependent variable. These assessments include 30 multiple choice items, 6 multiple mark items (more than one answer could be marked correct), and 3 constructed response items with a writing requirement. EQS, a method for testing various factor-analytic models for one or more groups, was used to assess the fit of the proposed models to the three groups—general education students, students with LD receiving the accommodation, and students with LD not receiving the accommodation. The proposed model had two factors: numerical and writing.

The two-factor model was considered adequate for all three groups. The results supported factorial invariance across all three groups in terms of the weights and intercepts for the observed score-factor score regressions. It appears that the identical construct was being measured for students with LD receiving the read-aloud accommodation and students with LD not receiving the reading accommodation.