Assessment of student achievement: A comparative study of student achievement using paper and pencil assessment and computerized adaptive testing (CAT)

Testing students in the 21st century is a multimillion-dollar industry. The National Center for Fair and Open Testing estimates that America's elementary and high school students complete more than 100 million standardized tests yearly (deMarrais & LeCompte).; The purpose of this study was to determine if computerized adaptive testing (CAT) could accurately predict student attainment of state-identified competencies and to study the effects on attitudes of examinees towards computerized adaptive testing and conventional paper and pencil format. A nonexperimental, exploratory, correlational research design was used in the study. Data collection tools included paper and pencil, fourth grade Michigan Education Assessment Program (MEAP) Reading test scores, a computerized adaptive test (CAT) results, and a short survey adapted from National Board on Educational Testing and Public Policy. Additional data were collected from student records and the school district's food services. A total of 267 fourth grade students in an urban school district who completed the two reading subtests of Michigan Educational Assessment Program (MEAP), a computerized adaptive test, and attitude survey participated in the study.; Study findings showed that CAT subtests (fiction, nonfiction and long passage) were statistically significant correlated in a positive direction with MEAP story and information subtests, with the long passage subtest having the highest correlations with story and information subtests. The CAT appears to be a predictor of examinees' performance on the fourth grade reading MEAP. No association was found between student test type preference and performance on either the paper and pencil MEAP reading or on the CAT.; The study found that White/Hispanic/Asian Pacific Islander students were more likely to achieve at a higher level in reading than African American students on both the paper and pencil and computerized adaptive assessments. Students who live at or below the poverty level as defined by qualifying for free and reduced lunch performed significantly lower than their peers on all subtests of the 4th grade Reading MEAP and the computerized adaptive assessment. While no gender differences were found on the MEAP reading subtests, a statistically significant difference was found on the CAT with girls having higher scores than boys.