The Construct Validity of Curriculum-Based Measurement of Reading: An Empirical Test of a Plausible Rival Hypothesis

• Published in: Journal of school psychology 1998, Vol.36 (4), p.399-415
• Main Authors: Kranzler, John H, Brownell, Mary T, Miller, M.David
• Format: Article
• Language: English
• Subjects: Children; Cognitive Processes; Construct Validity; Curriculum Based Assessment; Curriculum-based measurement; Elementary Secondary Education; Information processing; Intelligence; Psychometrics; Reading Research; Reading Skills; Reliability
• ISSN: 0022-4405; 1873-3506
• DOI: 10.1016/S0022-4405(98)00018-1

Research has confirmed that curriculum-based measurement (CBM) of oral reading fluency and measures of reading comprehension are highly correlated, as predicted by developmental theories of reading. Research on CBM, however, has only begun to rule out plausible alternative explanations of this relationship—an important aspect of a strong program of construct validation (e.g., Messick, 1989). This study investigated one such rival hypothesis by examining the relative roles of general cognitive ability, speed and efficiency of elemental cognitive processing, and oral reading fluency in the prediction of reading comprehension. Results of simultaneous multiple regression analyses substantiate the construct validity of CBM oral reading fluency. These findings indicate that the significant relationship between oral reading fluency and reading comprehension cannot be explained by general cognitive ability or by processing speed and efficiency. CBM oral reading fluency also did not correlate significantly with any of the processing speed and efficiency tasks. Interestingly, however, CBM oral reading fluency accounted for less variance in reading comprehension ( r 2 = .17) than expected based on the results of previous research and less than that explained by general cognitive ability ( r 2 = .24). When controlling for psychometric g and processing speed in the regression analyses, CBM oral reading explained 11% of the variance in reading comprehension. Implications of these results for further research on the construct validity of CBM are discussed.