"Good teachers do well with students at all levels of achievement, and there is no evidence that teacher education or performance on a certification examination contributes to quality teaching."
In The Market for Teacher Quality (NBER Working Paper No. 11154), co-authors Eric Hanushek, John Kain, Daniel O'Brien, and Steven Rivkin use a unique dataset from the Texas School Microdata Panel to measure teacher quality by the annual growth in each student's scores on the mathematics section of the Texas Assessment of Academic Skills. The dataset links detailed student, teacher, and school characteristics in grades 4 through 8 for the school years 1995/6 to 2000/1 in a major Texas urban district.
Their results confirm that good teachers increase student achievement. The average student who has a teacher at the 85th quality percentile can expect annual achievement gains that are 0.22 standard deviations greater than the average student with a median teacher. Good teachers do well with students at all levels of achievement, and there is no evidence that teacher education or performance on a certification examination contributes to quality teaching. The authors believe that the latter findings raise "serious questions both about the desirability of requiring or rewarding with higher pay those with a post-graduate degree and about the efficacy of the existing certification procedures in Texas and similar systems in other states."
First-year teachers have much lower performance on average than other teachers. After that, teacher performance improves markedly, peaking in a teacher's fourth year. In Texas, 10 percent of teachers with 0-2 years of experience, and 7 percent of all teachers, leave each year. Because first-year teachers are at a disadvantage, to the extent that higher turnover in urban school districts increases the proportion of first-year teachers, high turnover may be part of the explanation for their poorer performance.
Although people who make education policy worry that high quality teachers leave urban schools in disproportionate numbers, the authors find no evidence of this. They conclude that the teachers who do change districts are roughly equivalent in quality, and that the students benefit from having a same race teacher, quality held constant.
The authors also find substantial variation in teacher quality within schools. This has two implications for state accountability programs. Many states measure quality by aggregating student scores for an entire school. According to the authors, any program that "aggregates performance to the school level or across years misses the majority of the variation in the quality of instruction." This "weakens the incentives for good teachers to enter and remain in teaching, for ineffective teachers to leave, and for all teachers to put forth greater effort.
-- Linda Gorman