Doctoral Dissertations

Date of Award

8-1999

Degree Type

Dissertation

Degree Name

Doctor of Education

Major

Educational Administration

Major Professor

Gerald C. Ubben

Committee Members

Mary Jane Connelly, John Ray, Arnold Saxton

Abstract

Using 4th grade student achievement math scores and Tennessee Value- Added Assessment System (TVAAS) math teacher effectiveness estimates for grades 5-8 (matched to individual students) as predictor variables, this study's purpose was to ascertain the residual and cumulative effects of these teachers on student 9th grade Competency scores. Students from 2 large districts who were 9th graders in the fall of 1997 with complete information were included. From the base model ANOVA, partial sums of squares for teacher estimates for all grades were significant, indicating the presence of residual effects free of partial confounding by other variables: 5th, F=8.75, p-value = 0.0031; 6th, F=14.82, p-value=0.0001: 7th, F=25.75, p-value = 0.0001; 8th, F=73.43, p-value = 0.0001. The differential effect of 5th and 6th grade teachers on 4th grade prior achievement levels produced mean Competency 9th grade scores for students in the bottom quartile ranging from 59-65 (se=0.63 to 0.81), depending on the level of teacher effectiveness students encountered in 5th grade. Competency means for similar low-achieving students assigned to 4 consecutive very ineffective teachers were 53.6 (se=0.733) versus 71.8 (se=.939) for 4 very effective teachers (70 is required for passing). Although a significant ethnic effect was apparent, it paled in comparison to this student Competency score response to 4 similarly effective teachers. Substituting a success variable in place of the grade Competency score as the dependent variable allowed the determination of passing probabilities for various sequences of teachers with each quartile of 4th grade student prior achievement. As the assumed cut score was raised, student passing probability at 9th grade grew increasingly more dependent on 4th grade scores and teacher effectiveness estimates for grades 5-8: F=82.12 at 60 versus 129.83 at 80 (p-values=0.0001). Although all children benefited from highly effective teachers, the lower 50% benefited most. A change in level of effectiveness of teacher sequence increased the probability of passing to at least 50/50 for at least one quartile of students for each assumed cut score. Conclusions: (1) Tennessee's current high stakes student test is holding students accountable for something beyond their control, and the punitive policies associated with it should be reviewed for appropriateness. (2) the availability of effective teachers for lower achieving students is a critical determinant in their ability to pass. (3) Broad variability in teacher effectiveness suggests a need for improved teacher preparation and additional staff development opportunities for practicing teachers with improved accountability measures for teacher preparation.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS