TQB: Teacher Quality Bulletin

To use VAM or not to use VAM?

See all posts

The use of value-added models (VAM) to evaluate teachers and schools has never been short on critics, but this year the opposition has a few new leaders to rally behind. In April, the American Statistical Association (ASA) put out a statement of its concerns with the accuracy and interpretation of VAM estimates. And recently, the National Association of Secondary School Principals (NASSP) Board of Directors released its own statement condemning the use of the statistical tool.

Among its concerns, the ASA said VAMs only: measure correlation and not causation; don't necessarily predict long-range student outcomes; can't attribute much variation in student academic growth to teachers (between only 1 and 14 percent); and are risky to use in high-stakes settings.

The NASSP took a harder line, explicitly stating that VAMs should not be used for any teacher personnel decisions and rather used only to measure school improvement and program effectiveness and to inform teachers' professional development.

Economists in the field of education Chetty, Friedman and Rockoff have responded at least to the earlier statement: in November they published a statement addressing the ASA's concerns point-by-point. In response to the argument that VAMs do not measure casual impacts of teachers on student academic growth, Chetty et al. cite a number of studies that find when VAMs control for students' previous test scores, the models do capture direct effects from teachers. The researchers also note that while VAM can't predict every important long-range outcome, there are several it can predict, including college attendance, earnings and teenage pregnancy. And while not all variation in student achievement can be attributed to teachers, ChettyFriedman and Rockoff argue that what can be attributed to teachers is absolutely worth paying attention to and measuring because of teachers' direct impact on student outcomes, both immediate and long-term (i.e., an increase in lifetime earnings).

The researchers do seem to agree with the ASA's concern about VAM in a high-stakes setting— but only to say that there is not enough evidence to make a judgment on this point either way.

Each of these statements is worth a read. No matter what your opinion of VAM is, this debate increasingly matters as states start to link student academic growth to teacher evaluations.