| This study sought to investigate whether the variables teacher turnover and class size predicted the British Virgin Islands' public primary schools' academic performance levels based on their class five students' performance on the Primary Five Examination in English, Mathematics and Social Studies during school years 1994--1999, (except 1996) in school level analysis 1.;It also sought to summarize the collective perception of that system's principals and class five teachers relative to the effectiveness level of their schools in the areas: leadership, teacher behavior, student behavior, school climate and parental involvement, in school year 1998--1999 in school level analysis 2.;Finally, the study sought to investigate whether the variables: attendance, mobility, gender, amount of daily television viewing, hours of video games played weekly, number of adults in the home, age of parents, educational status of parents, parental involvement in their child(ren)'s school, and the family's socioeconomic status, predicted the academic performance of the class five students on the examinations in school year 1998--1999 (student level analysis).;The design called for a census of the entire target populations, given the relatively small size of the education system. Linear multiple regression was used in school analysis 1, and the student analysis to predict academic performance. Descriptive statistics were applied to school analysis 2 to describe the joint perception of the principals and class five teachers relative to the five variables.;With school analysis 1, class size was the only variable that significantly correlated with the schools' performance in all three subjects. The principals' and class five teachers' joint opinion was that their schools are average to slightly above average in their collective level of effectiveness in school analysis 2. The student analysis found that parents' educational status, parental involvement, and socioeconomic status significantly correlated with the students' performances in all subjects.;However, the most significant finding of this study was a dramatic fluctuation of the regression results across the years 1994--1999. Since class size and teacher turnover were fairly constant in their means during the period in question, the likely cause of the fluctuation in the coefficients of determination was the examination, likely suggesting that the instrument's psychometric qualities are poor, and that all other findings of this study may be unreliable since raw data from the examination underpinned the study. |
