Definitive Evidence on the Value-added Approach

This post is also available in: Spanish

A study just published by professors from Harvard and Columbia provides new evidence on the validity of using an approach known as “value-added analysis” to evaluate teacher effectiveness.

Using this approach with 2.5 million students in grades 3-8 over 20 years, the authors find that good teachers raise the test scores of their students significantly more than do poor teachers.

Further, they show that once this approach controls for all the logical variables that might affect student achievement (e.g., income, gender, race, etc.) there is very little margin for additional biases due to other variables that are harder to observe. Finally, they find that these differences in student achievement between the students of the worst and best teachers have lasting impacts on students’ success later in their life. When the authors followed the students of the best teachers into their adult life, they found that they were more likely to enroll in college and have higher wages.

This study is widely seen among experts as offering definitive evidence of the validity of value-added models. It remains to be seen, however, whether these findings will continue to hold as school systems increasingly attach high-stakes to test scores, thereby increasing the incentives for cheating or other types of behavior that do not encourage learning. At the end of an Education Next piece on the subject, you can find a healthy debate on this and other key questions raised by the study.