It is no secret that good teachers can make a world of difference in their students' lives. Most people presumably would like to make sure that disadvantaged students have access to these good teachers. As a result, figuring out how to sort the good teachers from the bad has become an important public policy question.
One common measure of a schoolteacher's effectiveness is known as "value-added." This metric essentially measures the average test-score gain for a teacher's students over the course of one year, accounting for differences such as students' starting scores. However, many people (as well as teachers' unions) have argued against using test score-based metrics to measure a schoolteacher's performance. After all, teachers who raise students' test scores could be merely teaching to the test and not providing any life-long benefits for their students. Thus, the present system whereby decisions pertaining to a school's faculty are made primarily based upon educational attainment and seniority should be left as is.
Recent academic research throws cold water on this argument. Last month, three economists Raj Chetty and John Friedman of Harvard University and Jonah Rockoff of Columbia University published a paper entitled "The Long Term Impacts of Teachers." This paper made use of a data set of 2.5 million students in grades three to eight from a "large urban school district" over 20 years.
Contrary to the teachers' unions' claims, higher value-added teachers have a significant impact on their students' adult lives. Students who have higher value-added teachers are less likely to have children as teenagers, are more likely to attend college, earn higher incomes, live in higher-quality neighborhoods as adults and save more for their own retirement.
As an example, consider the effects of a teacher's value-added on their students' future earnings. A one standard-deviation improvement in a teacher's value-added for a single grade will increase each of their students' lifetime earnings by an average of $25,000. Furthermore, replacing the bottom 5 percent of teachers (as measured by three years of value-added data) with average teachers would result in a gain of $190,000 per teacher per year, based on the present value of their students' increased future earnings.
Obviously, students from lower-income backgrounds are at a disadvantage relative to other students. Nonetheless, the study shows that higher value-added teachers can make a difference even for disadvantaged students.
Furthermore, a 2006 paper "What Does Certification Tell Us About Teacher Effectiveness?" by Dartmouth economics professor Douglas Staiger, Rockoff and Harvard's Thomas Kane shows that a teacher's educational attainment is, at best, an extremely poor predictor of that teacher's effectiveness. The authors find that "classroom performance during the first two years, rather than certification status, is a more reliable indicator of a teacher's future effectiveness."
Obviously, test scores should not be the only measure schools use to evaluate teachers other research has shown student feedback and outside evaluations of a teacher's classroom performance can also be useful measures of teacher quality. Furthermore, there are many other unsettled questions regarding education policies that have little to do with testing: how schools should be funded, the proper role of the federal government in education and so on.
Regardless of how one answers these questions, two things are abundantly clear: First, the persons responsible for hiring and firing a school's faculty should be given the authority to relieve a teacher on the basis of poor value-added results. Second, information about a teacher's previous value-added results should be made available to other schools that are considering hiring that teacher. The status quo, whereby decisions regarding faculty employment are made primarily on the basis of educational certification and seniority, is unacceptable and needs to change.
