The purpose of this study was to explore whether school grade configuration had a relationship with eighth-grade reading and mathematics proficiency and ninth-grade enrollment in a selective high school. I analyzed data for two cohorts of students enrolled in K-8 and 6-8 schools located in the Baltimore City Public School System. Since the early 20^th century educators and researchers have identified the middle-grade years as potentially crucial for youth development. Findings from prior research on the impact of grade configuration on student achievement have been inconsistent. Some researchers identify clear, consistent academic advantages for students enrolled in K-8 schools. Other researchers identify no such performance difference.

Analytic methods consisted of both descriptive and inferential statistics. The primary analytic method was hierarchical generalized linear modeling (HGLM). Student-level variables included were gender, fifth grade reading proficiency, fifth grade mathematics proficiency, over-age status, eligibility for free or reduced-price meals and the number of years enrolled in the same school between fifth and eighth grades. School-level variables included grade configuration, school size, student and teacher climate measures, the percent of over-age students, the percent of students eligible for free or-reduced price meals, cohort, and the percent of fully certified teachers. Student-level variables were consistent for all data models. I added school-level variables in stages.

Inferential analyses of student data revealed significant differences in the fifth-grade performance and demographic characteristics of students enrolled in K-8 and 6-8 schools. Compared with students enrolled in K-8 schools, 6-8 students in both cohorts were less likely to be proficient in mathematics and/or reading in fifth grade. Middle school students were more likely to be over-age for grade. At the school-level, there were few significant differences between the schools attended by cohort 1 and cohort 2 students. HGLM analyses revealed no unique, direct relationship between school grade configuration and the study outcomes once all variables were included in the model. Fifth grade proficiency levels were highly predictive of all outcomes. The study findings indicate that changing school grade configuration for middle-grade students may not be sufficient to improve student achievement in the absence of other conditions.