Presentation on theme: "Participants were 1,286 general and special education students enrolled in grades third through eighth. Participants represented five (two elementary and."— Presentation transcript:
Participants were 1,286 general and special education students enrolled in grades third through eighth. Participants represented five (two elementary and three middle) schools in a rural southeastern school district. The ethnic breakdown of the sample was 94.5% Caucasian; 3.2% Hispanic; 1.4% African-American; 0.5% Asian; and 0.2% Native American. Participants came from predominantly middle to lower class SES backgrounds. Students were administered grade level M-CBM probes from the AIMSweb system during the districts regularly scheduled fall, winter, and spring benchmarks in the 2006-2007 school year. All M-CBM probes were group-administered and scored by the districts school psychologists, classroom teachers, and teaching assistants who had been thoroughly trained to administer and score these measures. Each students response was scored for digits correct. The TCAP Achievement Test was administered in the spring of 2007 by classroom teachers. The mathematics composite scores from the TCAP were used in the analyses. Researchers used a regression analysis to determine if significant correlations existed between fall, winter, and spring M-CBM benchmarks and the mathematics scaled scores on the TCAP Achievement Test. Sara B. Reynolds and Jamie Y. Fearrington, Ph.D., NCSP The Predictive Validity of Mathematics Curriculum-Based Measurement Sara B. Reynolds and Jamie Y. Fearrington, Ph.D., NCSP Appalachian State University, Boone, NC Sara McCane-Bowling, Ph.D., NCSP and Christy A. Sorrell, Ph.D., NCSP Little Tennessee Valley Educational Cooperative, Loudon, TN ABSTRACT This study analyzed correlations between Mathematics Curriculum-Based Measurement (M- CBM) benchmark scores and performance on the Tennessee Comprehensive Assessment Program (TCAP), a state mandated high stakes test. Participants were 1,286 students enrolled in grades 3-8 in a rural southeastern school system. A linear regression model was used to investigate our research questions. Specifically, this study sought to determine to what degree are the fall, winter, and spring M-CBM scores correlated with TCAP results, if there are differences in M-CBM and TCAP correlations between the various grade levels, and if there are temporal differences in M- CBM and TCAP correlations based upon the time of the benchmarking (i.e. fall, winter, and spring). METHOD BACKGROUND RESULTS and CONCLUSIONS TABLES For more information, email firstname.lastname@example.org This study suggests M-CBM benchmarking probes are moderate predictors of student performance in mathematics on the TCAP. Pearson r coefficients between M-CBM and TCAP scores for grades 3-8 ranged from.244 to.497 Spring M-CBM scores were found to be more highly correlated to TCAP scores for grades 5 and 7. This study did not find a significant difference in correlations among the range of grades assessed. The major goal of this study was to contribute to the growing body of evidence indicating that M-CBM is a valuable tool for screening students who are at risk for academic failure. Results provide empirical support for the use of M-CBM as a screener for academic needs. Additionally, M-CBM appears to be moderately capable of predicting performance on high stakes tests, such as the TCAP. Researchers should continue to investigate the tools that may serve as early indicators of intervention needs in the mathematics area. Early identification of at-risk students is needed in order to increase the mathematical proficiency of all students. Continued research into the psychometric properties of M-CBM is relevant and needed, especially given the amount of evidence in this area (as compared to other academic areas such as reading.) CohortNMeanSDPearson r Grade 3 Fall M-CBM20414.686.22.416** Winter M-CBM20923.307.71.481** Spring M-CBM20927.279.26.413** TCAP222482.5429.73 CohortNMeanSDPearson r Grade 4 Fall M-CBM22930.7911.60.465** Winter M-CBM23138.3913.75.321** Spring M-CBM23747.3217.42.244** TCAP246501.1637.39 CohortNMeanSDPearson r Grade 5 Fall M-CBM29328.7011.59.436** Winter M-CBM30836.7215.72.434** Spring M-CBM29241.3916.48.456** TCAP321507.7135.38 CohortNMeanSDPearson r Grade 6 Fall M-CBM268220.127.116.115** Winter M-CBM26332.6213.16.485** Spring M-CBM26429.9113.31.385** TCAP284522.1244.39 CohortNMeanSDPearson r Grade 7 Fall M-CBM30632.7612.59.398** Winter M-CBM30140.8615.48.418** Spring M-CBM22739.1116.15.453** TCAP323530.4344.97 CohortNMeanSDPearson r Grade 8 Fall M-CBM30835.5913.974.497** Winter M-CBM30746.5216.556.344** Spring M-CBM15442.2417.660.264** TCAP32554206147.54 Note: ** p <.001 TCAP = Tennessee Comprehensive Assessment Program. M-CBM = Math Curriculum Based Measurement. The TCAP means and standard deviations are scaled scores. M-CBM scores are raw scores of digits correct per minute. Tables 1-6 Descriptive Statistics and Correlations between M-CBM and TCAP scores by Grade Given the increased accountability required by NCLB and IDEIA, students that are not on track to meet identified goals must be identified early and data should be utilized to show evidence regarding their progress in all basic academic areas. Along with other evidenced-based tools, public schools are increasingly utilizing curriculum- based measurement as part of a three-tier problem solving model, designed to assess the general student population and provide early intervention to children determined to have educational needs beyond the scope of what the general curriculum can provide. To assist in achieving the ultimate goal of increasing the number of students who score proficient on high stakes tests, many schools are incorporating CBM within RtI models to proactively screen for skill deficits and provide early intervention to students in need. Given the particular emphasis of current educational reform, researchers should continue to investigate the effectiveness of CBM as an academic screener. A focus on M-CBM is highly important since there has been significantly less research into its utility when compared to other academic areas such as reading.