Assessing Teacher Effectiveness: How Can We Predict Who Will be a High Quality Teacher?

The first part of the study examines associations between teacher characteristics and student outcomes using data on students in grades 3 through 10 and their teachers. These data are drawn from a Florida statewide database that includes longitudinal information on nearly the entire population of teachers and students in the state. Statistical modeling procedures will be applied to the data in order to derive "value-added" scores for more than 5,000 teachers; these scores represent estimates of the contributions that individual teachers make to student achievement. To address questions regarding what teacher characteristics best predict effectiveness, associations between these effectiveness scores and a set of teacher characteristics will then be examined. Teacher characteristics in these analyses include measures of general verbal and quantitative skills, college course-taking, and certification test scores. The size of the database will allow for detailed breakdowns for sub-groups of students, teachers, and school.

The second part of the project involves the use of a mixed methods approach to compare principals' opinions of the factors that predict teacher effectiveness with the measures found to be the best predictors of value added from the first set of analyses. In addition, the degree to which principals can predict and identify which of the teachers in their own schools produce the largest value added to student outcomes is being examined. For this part of the project, value-added scores are calculated for each teacher, as in the first part of the project, and are compared with coded data from principal interviews.