A multivariate analysis of the effects of school *attendance on mathematics achievement test scores of West Tennessee fifth-grade students accounting for socioeconomic status, age, gender, ethnicity, language, and educational status

School attendance has become a topic of increasing national importance as factors impacting mathematics achievement receive greater scrutiny. This study investigated the effects of school attendance of 111 West Tennessee fifth grade students on mathematics achievement test scores accounting for socioeconomic status, age, gender, ethnicity, language, and educational status. Attendance data were gathered from 2004-2005 school attendance records using a pretest-posttest design with absences categorized as either excessive or non-excessive (Shutt, 2000). Mathematics achievement was assessed using ThinkLink Learning (2004) a scientific, research based, predictive assessment system for students (PASS). Does poor school attendance significantly affect academic gains of fifth grade elementary students in mathematics? A One-Way Analysis of Covariance was conducted to address that first question using two levels of the independent variable (low attendance and high attendance), math posttest scores as the dependent variable, and gender, age, socioeconomic status, educational status, English language learners, and ethnicity as the covariates. Results initially indicated no significant differences between the two levels of attendance. Data were further disaggregated to students with high absenteeism; this allowed examining only data of students with ten or more absences. Results indicated a relationship between students with high absenteeism and achievement. Do the covariates of ethnicity, English language learners, race, gender, socioeconomic status (SES) and educational status affect the relationship of attendance and achievement? A path analysis was conducted to address the second question. On average, the higher a student scored on the pretest, the more likely they were to score higher on the posttest once the other variables in the equation were controlled. White students were significantly more likely to score higher on the posttest than Black or Other students once the other variables in the equation were controlled. Implications for current research are discussed.