| NCLB has forced schools to raise student achievement. The purpose of the study was to examine the relationship between mathematics curriculum and student achievement on the ACTAAP 8th grade Mathematics Benchmark exam. The population identified was all schools in Arkansas containing 8 th grade. A simple random sample was drawn which encompassed 3 different mathematics curricula: traditional, standards-based, and balanced ( N = 150). Student mathematics achievement data from a three year period (2005--2007) were collected on all schools that responded through the Arkansas Department of Education website in conjunction with the NORMES website, and included the percentage of students in the school that scored in each of four performance levels (below basic, basic, proficient, and advanced) for mathematics on the ACTAAP 8th grade Benchmark exam. The design was an ex post facto hierarchical ANOVA with one repeated measure dependent variable, the mathematics achievement on the ACTAAP 8 th grade Benchmark exam for 3 years. The two independent variables were math curriculum and proficiency level. Because the original interval scores (exam scores) were categorized into ordinal data (performance levels), the date were analyzed both as interval and as ordinal. The results of the two-between, one-within hierarchical ANOVA revealed statistically significant differences for the percents of the nested effect of performance level within math curriculum and the interaction of performance level within math curriculum and time, but not for the main effect of math curriculum. The cumulative logit analysis revealed that standards-based math curriculum was significantly different in 2005 from the other two types of curriculum; however, there were no differences between the types of math curriculum in 2006 and 2007. These results suggest that mathematics curriculum does not account for 8th grade students' achievement as measured by the ACTAAP Benchmark exam; however, the research contributed new information with the strength of utilizing a large sample size representative of an entire state. |
