It’s no secret that most teachers are not fans of being evaluated based on student test scores.  This is true even though, according to the findings of an upcoming teacher survey from EdSector, the majority of teachers believe they should receive pay increases if they consistently receive outstanding evaluations from their principals . The survey also found that support for performance pay fell to less than a third of teachers if pay increases were based on student test scores. So teachers don’t mind being paid based on their performance but they don’t want their performance determined by test scores.

So why is this the case? Why are teachers so against being evaluated, even in part, based on how their students perform on standardized tests? Of course, there is no single answer but EdSector provides insights into the minds of many teachers. They highlight one teacher who said:

While she loved her job, if her pay depended on student test scores, she “would be crazy” to continue working in her school, which has a student body that is 90 percent low-income and speaks 18 different languages.

She says this because she, like many teachers, believes that standardized tests do not accurately measure the progress low-income and other traditionally disadvantaged students make during a given school year. Essentially, many teachers, just like the one highlighted, believe that their evaluations scores would suffer if they taught a class mostly made up of students from disadvantaged backgrounds.

I certainly don’t blame teachers for having such concerns. There have been many attempts to pay teachers based on their students’ performance that did penalize teachers for working with disadvantaged students.

However, times have changed and statisticians have developed tools to take on this problem. For example, value-added models have been designed to specifically isolate the impact a teacher has on their students no matter their students’ background or previous achievement. Value-added models are not perfect but they do a pretty good job of determining if a student would be better off with one specific teacher instead of another.

Despite the use of value-added models, teachers still believe they would essentially be penalized for having a class of disadvantaged students.  Even researchers who are critical of using value-added data to evaluate teachers believe teachers of disadvantaged students are penalized when they are evaluated based on value-added data.

However, findings from a recent report from the CALDER Center run counter to such criticism. The report found that a teacher’s evaluation score does not suffer if they teach more disadvantaged students. Specifically, CALDER found that teachers in North Carolina that transferred from a low-poverty school to a high-poverty school actually received higher value-added scores in the high-poverty school than they had in the low-poverty school.

Such finding provides strong evidence that teachers of disadvantaged students are not penalized if they are evaluated using high quality value-added data. Even so, as the Center’s report Building a Better Evaluation System argues, no teacher should be evaluated based only on student test scores – or any single measure for that matter. For a teacher evaluation system to yield both accurate results and provide information to help all teachers improve, multiple measures should be used, including student test scores. – Jim Hull