A predictive validity study of computer adaptive placement tests for Tennessee higher education institutions
Assessment of basic skills proficiency allows institutions of higher education to recognize and appropriately recommend remedial/developmental coursework for underprepared students. Traditionally, placement assessment has been administered through paper/pencil examinations, but modern measurement theory has introduced computer adaptive testing that is tailored to the examinee. The question of the effectiveness and efficiency of basic skills assessment through computer adaptive testing gave impetus to this study.
The predictive validity of three assessment instruments designed for placement of underprepared students was determined specifically for the remedial/developmental program in Tennessee. Computerized Placement Tests (CPTs) developed by ETS and Computer-Adaptive Placement Assessment and Support System (COMPASS) developed by ACT were the computer adaptive instruments evaluated in addition to the paper/pencil instrument, the Academic and Assessment Placement Program (AAPP) Pretest, currently administered by Tennessee Board of Regents colleges. The concept of computer adaptive testing was introduced through a review of proficiency testing, an outline of advantages and disadvantages of adaptive testing, and a brief history of developmental education. Examinee results of the currently administered paper/pencil instrument were correlated with student grades in addition to the two computer adaptive tests, CPTs and COMPASS, to determine the instrument with the best predictive validity.
The Pearson product-moment correlation coefficient and independent sample chi-square tests were the statistical tools utilized to determine the relationships between the three assessment instruments and examinee success in remedial/developmental mathematics courses. The design of the study was structured by three levels of remedial/developmental mathematics since grades were collected specifically in the three courses, Basic Mathematics, Elementary Algebra, and Intermediate Algebra.
Results of the predictive validity study for Basic Mathematics indicated further investigation is necessary to determine if the currently administered AAPP pretest or the computer adaptive tests best predict student success. CPTs provided the best predictive validity for success in the Elementary Algebra course, according to chi-square analyses and correlation coefficients. A small number of placement scores collected for Intermediate Algebra limited possible conclusions from the analysis. The literature review and results of the analyses indicate computer adaptive testing as a viable method of assessment for consideration by the Tennessee remedial/developmental program.
Thesis97b.D39.pdf
5.2 MB
Adobe PDF
5e71529af0f1ebdf7306ed807ccb2c6b