Growth in Mathematics Achievement: Analysis With Classification and Regression Trees

Using a recently developed statistical technique often referred to as classification and regression trees (CART), the author classified students into groups with differential rates of growth in mathematics achievement during middle and high school according to individual and family variables, and characterized students who grew fast and those who grew slow in mathematics achievement. Data analysis of the Longitudinal Study of American Youth (LSAY) showed that age (being younger in the same grade cohort) was critically important for fast growth in mathematics achievement. Whereas younger White and Asian students grew at the best rate in mathematics achievement, older White and Asian students with low family socioeconomic status (SES) grew at the worst rate. Hispanic, Black, and other students were sandwiched in-between. One in 3 socially disadvantaged students in the sample overcame the negative impacts of low family SES and large family size and progressed to the 2nd best rates in mathematics achievement; 1 in 7 socially advantaged students did not take advantage of high family SES and small family size. Boys and girls shared the fast end of growth in mathematics achievement; boys predominantly occupied the slow end of growth.