In this thesis a prediction model using graduation rate as the performance indicator is obtained for community colleges for three cohort years, 2003, 2004, and 2005 in the states of California, Florida, and Michigan. Multiple Regression analysis, using an aggregate of seven predictor variables, was employed in determining this prediction model.…

All teacher evaluation forms completed in the arts and sciences division of a large community college were analyzed. Multiple regression was used to predict teacher rating. Items relating to teacher characteristics as perceived by the students (e.g., "tries to help students understand," etc.) had a consistant impact on rating. (Author/MLW)