An analysis of student performance of the Virginia Algebra I Standards of Learning Examination

This study was designed to predict achievement on the Virginia Algebra 1 Standards of Learning Examination and to investigate additional variables that may influence achievement on the examination. Using data from a Northern Virginia school system administration of the May 2000 exam, a multiple regression model was developed to predict student performance on the state test using four dependent variables: Algebra 1 final grade, Math 8 final grade, Stanford 9 total mathematics score, and race. Gender, free and reduced lunch, English as a second language, and Stanford 9 reading comprehension were not statistically significant variables. In the second part of the study, outliers from the multiple regression were identified and the students were contacted for interviews. These students formed two categories. Category A includes the students that scored significantly higher than the regression model predicted. Category B includes students who scored significantly lower than the model predicted. The semi-stuctured interview investigated the students' experiences in Algebra 1 and on the Algebra 1 SOL examination. Seventeen students from category A and ten students from category B participated in the interviews. The interviews were analyzed to determine the similarities and differences between these two groups. The analysis of the interviews indicated that the students whose performance exceeded the predicted score were self-motivated, independent learners who had some test anxiety. The students who performed at a level lower than their predicted score were much more dependent on their teachers as motivators and disseminators of knowledge. Both groups of students valued the social aspects of school and preferred learning in active environments that used teacher-student discourse, group work, and other investigative methods of learning.