We were heartened to find not one–but two–education studies this month using data gleaned from twins to answer research questions about teacher quality and classroom performance. This is a novel approach for education research, although it's quite common in other kinds of social science research. The working theory is plausible (though not foolproof): by comparing how twins perform under different teachers, researchers can pull out real differences attributable to schools and not students' backgrounds.

A Netherlands study takes advantage of a policy in Dutch schools where twins are always assigned to different classrooms to examine the impact of teacher experience on student achievement. Researchers were able to compare the academic outcomes for 495 twin sets in grades 2, 4, 6 and 8.

Early in their schooling, twins' scores in grade 2 show meaningful (and statistically significant) differences based on the experience of their teachers. For each additional year of experience one twin's teacher had over the other's, there was about 1.5 percent of a standard deviation's growth; that may  sound small, but over time that difference adds up. 

The researchers think that an examination of twins' performance is an especially robust idea for early grades because there's little chance that students would have been sorted into classes based on their academic abilities at so young an age.

That might explain why in higher grades, the impact of teacher experience decreases: the effect on students in grades 4 and 6 is only found in reading and all gains due to teachers' years of experience disappear in grade 8.

The most interesting finding here may be that there was no evidence of the 'plateau effect' from experience that American studies frequently find at about year five in a teacher's career. Rather, the effect of experience found in this study is linear, meaning that a 20-year veteran is consistently more effective than a 15-year veteran who is consistently more effective than a 10-year veteran, and so on.

Back in the U.S., another twin study also looks at the relationship between teacher characteristics and student outcomes. The research considers master's degrees (where no relationship with student outcomes has ever been soundly documented), national board certification (where a relationship has been found that correlates with increased student outcomes) and teacher experience. Of the three factors, teacher experience seemed to be the teacher characteristic most strongly related to student outcomes. Unlike the Netherlands study, though, the effect of a teacher working longer tapers off, as almost every American study has shown, after about a five-year career.