A Meta-Analysis of the Relations Between Curriculum-Based Measures in Mathematics and Criterion Measures

This is the authors' abstract. We don't add key points for this paper.

This meta-analysis examined the validity of curriculum-based measures in mathematics (CBM-M) in relation to mathematics outcomes. Peer reviewed journal articles or dissertations were included if participants were in Grades 2 to 8 and the study purpose was examining the criterion validity of MCOMP and MCAP. Studies were identified via electronic databases, table of contents review, and reference and forward citation searches. Twenty-nine studies with 27,907 participants met inclusion criteria. A random-effects three-level meta-analysis of 328 correlations was conducted to estimate the average correlations between scores on the CBM-M tasks and criterion measures and to analyze the effects of potential moderating variables. The average correlation between CBM-M and criterion measures was r =.584 [95% CI: 533, .635]. Both MCAP (r =.654 [95% CI: .599, .703]) and MCOMP (r =.528 [95% CI: .462, .590]) scores were strongly correlated with criterion measures when all grades were included. Moderator analyses were conducted for each CBM-M task, yielding similar findings. Average correlations were stronger for middle compared to elementary grades and for concurrent administration of CBM-M and criterion measures. No moderating effects were found for criterion, CBM-M developer, or time of year the CBM-M task was administered. No evidence of publication bias was found.Impact StatementThis meta-analysis indicates that MCOMP and MCAP are strong predictors of mathematics achievement and thus useful for universal screening. Stronger correlations were found for middle school students as well as when the CBM-M task and the criterion measure were administered within one month or less of one another.