3.
Part II
4. REGRESSION MODELS ONLY FOR THE EPGY STUDENTS 4.1 CFA (WCFA) and CST06 as predictors of CST07 4.2 Results from a Title I school in district 2 that was outside the Effectiveness Study but which had EPGY students 4.3 Regressions adding the number of correct second-attempts, CSA, as a predictor of CST07 4.4 Correlation between CST06 and CFA 4.5 Correlation between re-scaled correct first-attempts and Math CST scores 4.6 Analysis of the 2006 to 2007 gain on Math CST by CST06 scores and CFA values PREDICTING THE PROFICIENTY LEVELS OF THE EPGY STUDENTS USING MULTIVARIATE CLASSIFIERS 5.1 Minimum-distance classifier 5.2 Bayesian classifier 5.3 Learning curves PREDICTION MODEL FOR EPGY STUDENTS ON THE MATH CST 2008 OLS model HLM model 6.1 OLS and HLM prediction models using correct first-attempts Coefficients from California Standard Math Test 2006-2007 Prediction result 3
5.
6.
6.2 OLS and HLM prediction models that include WCFA as a predictor 7. CONCLUSIONS Acknowledgements References Appendix METHOD OF STATISTICAL ANALYSIS A1 Effect sizes Cohens'd Our modification of d A2 Three-level hierarchical linear model A3 Binomial Analysis of Changes in proficiency level
4
1. INTRODUCTION Stanford University's Education Program for Gifted Youth (EPGY) conducted a randomized treatment experiment (RTE) during the 2006-2007 school year to test the efficacy, for Title I students, of the EGPY Kindergarten through Grade 5 Mathematics Course Sequence (Math K-5). All eight participating schools had a full K-5 sequence of instruction. While the EPGY curriculum was originally developed for gifted students in any grade, the students in this study were not selected as gifted but were simply the students in the Title I schools. This report is divided into two parts. Part I is concerned with measuring the effectiveness of the EPGY curriculum relative to the controls and consists of sections 2 and 3. In Section 2 we describe the research design and the data collection procedures for the RTE. Section 3 gives the main results of the RTE, focused on analyzing the Mathematics scores from the 2007 California Standards Test (Math CST07) for the EPGY (E) and control (C) students in several ways. Part II reports several analyses we made that are not directly focused on measuring the effectiveness of the EPGY curriculum and consists of sections 4 to 6. Section 4 is devoted to developing linear regression models that relate the variables measured in the EPGY program to the Math CST07 scores for the E group. Section 5 presents the results of applying multivariatenormal covariate models to the E group's test results. In Section 6 we use several different regression models to predict the Math CST08 scores for EPGY using CST07 scores, CFA (WCFA) values and average latencies as predictors. The details of some of the statistical methods we used are briefly reviewed in the Appendix.

上一页下一页