Teacher-Student Matching Increases Teacher Effectiveness
If gifted students end up being taught by more qualified teachers, then estimates of the effect of teacher qualifications on student achievement are likely to be higher than if teachers and students were randomly assigned.
If gifted students end up being taught by more qualified teachers, then estimates of the effect of teacher qualifications on student achievement are likely to be higher than if teachers and students were randomly assigned. To test that theory, authors Charles Clotfelter, Helen Ladd, and Jacob Vigdor use statewide classroom data from North Carolina to estimate the effect of teacher qualifications on student achievement for fifth graders in the 2000/1 school year.
In Teacher-Student Matching and the Assessment of Teacher Effectiveness, (NBER Working Paper No. 11936), they review data on the qualifications of 3,842 fifth-grade teachers from the state's 117 school districts, including licensure test scores, undergraduate institution attended, advanced degrees, and the number of years of teaching experience. Data on the fifth grade students include gender, ethnicity, fourth-grade achievement test scores, parental education, free lunch status, and self-reported time spent on homework, television watching, and personal computer use.
The results suggest that positive matching occurs both across schools and within schools. Teachers with better qualifications - that is, more experience, degrees from more highly ranked schools, higher licensure test scores, National Board Certification, and advanced degrees -- are more likely to work in schools with students who are more likely to be white, ineligible for subsidized lunches, to have college educated parents, and to have scored better on the prior year's state assessment test.
To explore the effects of teacher qualifications without the bias created by such positive matching, the authors turn to a subset of schools that distribute students across classrooms in ways that balance observable student characteristics. They find that "the only teacher qualifications that consistently predict improved student performance are experience and licensure test scores." A single standard deviation increase in licensure score increases predicted math achievement by 1 to 2 percent. Students assigned to highly experienced teachers improve their math scores by roughly one tenth of a standard deviation and their reading scores by slightly less than one tenth of a standard deviation.
Because teacher experience produces larger gains in math achievement for students with highly educated parents, reallocating the strongly qualified teachers to less advantaged students would probably reduce students' mean achievement scores. The pervasiveness of positive matching (between teachers and students) likely results from four forces: teachers seeking amenable working conditions; parents desiring to maximize the quality of their children's education; administrators seeking to maximize achievement; and administrators seeking to please vocal parents
-- Linda Gorman