Initial Validation of Collegiate Learning Assessment Performance Task Diagnostic Instrument for Historically Black Colleges and Universities

Abstract

Aim: The purpose of this study was to develop and validate the psychometric properties of the Collegiate Learning Assessment (CLA) performance task diagnostic instrument. The CLA is currently used to assess the effectiveness of student learning at colleges and universities in the United States to measure institutional contributions to learning gained by students. Despite its wide acceptance and use in the last decade, the validity of its psychometric properties has not been established. The performance task for the CLA instrument consist of eight items designed to measure four dimensions of student learning, which are: analytic reasoning, problem solving, persuasive writing, and writing mechanics. Study Design: Quasi-experimental One-shot Cross-sectional Case study design. Place and Duration of Study: Fayetteville State University; January 2012 to May 2012. Methodology: In this study, the CLA was administered to a random sample of 254 students who were eligible to complete the CLA at a university in southeastern United States. Results: Exploratory factor analysis revealed two items per learning dimension with adequate factor loadings and internal consistency. Confirmatory factor analysis showed that a 3-factor measurement model exhibited sound and better psychometric properties (χ2(6) = 9.355, p = .155, CFI = .995, TLI = .988, RMSEA= .024) than the hypothesized 4-factor model (for a model of this complexity (χ2 = 37.075, p = .001, CFI = .97, TLI = .95, RMSEA = .08) currently used by universities to assess student learning. Also, convergent validity and construct validity was confirmed with significant unstandardized regression coefficients and approximately equivalent standardized factor loading per factor, respectively. Conclusion: The initial validity of the psychometric properties of a 3-factor CLA Performance Task Diagnostic Instrument is confirmed.