How to grade teachers fairly using student achievement

Victor Joecks

When we want to know how good someone is at something, we make a judgment based on the facts.

A realtor is judged (and paid) based on how many properties she sells. A baseball player is judged based on his batting average and home runs. A defense lawyer is judged based on how many clients he gets acquitted. But teachers, who have the very important job of helping to educate our children, aren’t judged on their students’ performance. Why?

Some people, like Betty Buehler, the author of this letter to the Review-Journal, argue that it’s not fair to judge teachers by their students’ achievements, because there are too many factors outside teachers’ control.

There are factors affecting achievement over which teachers have absolutely no control. That’s why individual student achievement is a terrible way to evaluate teachers. Some of these factors are the skills level with which the student enters the class (some are years below grade level; many are at least a couple of years below); student aptitude (which a colleague calls “the elephant in the room,” because it’s politically incorrect to suggest some kids have higher IQs than others); student transiency; student attendance; whether the kid does homework; whether the kid pays attention in class; whether the kid ever picks up a book and reads; whether the parent values education; whether the parent asks about school and helps the student to the best of his or her ability.

Buehler is right that simply testing students at the end of the year and rewarding or punishing a teacher based on those results isn’t a fair or accurate test of a teacher’s ability. But that doesn’t mean we should abandon testing altogether. It means we should find a better test.

And a better measuring tool, called Value Added Testing, is available.

The value-added approach to testing is very simple. It focuses on changes in test scores over time, rather than on a single test score at a given moment. The whole point of school is for children to learn and progress over time. Yet, most current testing regimes do not measure progress. Virtually every city newspaper publishes the average test scores of local schools at some point during the year, often accompanied by an article that highlights the best schools (those with the highest average scores), and the worst schools (those with the lowest). But are the children in the schools with the highest scores actually learning more than those in the schools with the lowest? Not if the high-scoring schools are just taking in children with the highest IQs from the most enriched environments and simply spitting them out at the same advanced levels years down the road.

A value-added system can use the exact same tests as a standard system. The only difference is that the value-added system keeps track of every student’s previous scores (whether or not they were at the same school), so that the change in their scores from year to year can be calculated. In this kind of system, good schools are not necessarily the ones with the highest average test scores. They are the ones that show the greatest average increases in their students’ scores. They are the ones that are “adding the most value.”

Under their current systems, many states estimate and publish changes in schools’ average test scores at particular grade levels. But that is not the same as value-added analysis. They are making comparisons across cohorts. For example, they might show that this year’s seventh grade had higher reading scores than last year’s seventh grade at a particular school. That might mean the school is improving. But it might also mean that this year’s seventh grade was simply ahead of last year’s at the start. The essential element of the valueadded approach is that it measures the progress of each and every child as he or she moves through school.

Because of the need to keep track of students across time and place, even if they change schools, value-added regimes require a more sophisticated data collection system than most current regimes. But this problem is by no means insurmountable. Tennessee has had a value-added system since 1991 andhas been testing every student in grades 2-8 each year in math, reading, language arts, science, and social studies. (Testing high school students began in 1995.) Every year, each student’s results are calculated and compared to his or her scores from the previous year. Since 1993, school and district-wideaverages have been reported to the public. The average score increases of all the students taught by each teacher are reported to administrators for use in evaluations and personnel decisions.

Buehler also goes on to suggest that better-performing teachers should get paid more, which echoes the recommendations of Matthew Ladner of the Goldwater Institute in his recent report, New Millennium Schools: Delivering Six-Figure Teacher Salaries in Return for Outstanding Student Learning Gains.

Unfortunately, right now in Nevada teachers are rewarded for obtaining degrees that have little to no correlation with student achievement, not for being superior teachers.