My job status, plus my take on teacher evaluations

So despite my fears about losing my job to budget cuts, I will still be employed next year. My classes will be bigger, and I'll probably teach multiple grades, but I'll still be teaching.

However, many great teachers at my school did lose their jobs. All of these teachers had years of experience over me, and I feel conflicted about keeping my job at their expense. The whole situation got me thinking more about how (and why) a teacher's job performance is evaluated, and what factors principals should consider when making decisions about whom to lay off.

Before I continue, let me restate my belief that this round of layoffs was avoidable. The Texas legislature could have taken steps to adequately fund public schools, but its members chose not to. One might even argue that teacher evaluation systems are unnecessary in this case because we really shouldn't have to lay off anyone--and instead of talking about the best way to go about laying people off, we should be questioning the necessity of the layoffs. While I certainly sympathize with this argument, it would be unrealistic for schools not to have a system in place to ensure that teachers will be treated fairly and humanely in the event that layoffs cannot be prevented.

So I was both curious and concerned when I saw that the Texas state Senate recently approved a bill that would create a teacher evaluation system to measure the effectiveness and quality of public school teachers in the state. I was curious because few details have been released about what factors the state would consider in its evaluations; I was concerned because, most likely, the evaluation system being considered will rely primarily on student scores on standardized assessments.

Why does this concern me? First, student performance on standardized tests is too unpredictable to be a reliable indicator of good teaching. Too many factors outside the teacher's control (student characteristics, non-random student assignments, effects of other teachers, out-of-school factors) can influence student performance. In particular, teachers tend to receive lower "effectiveness" ratings when they are assigned English language learners, students with learning disabilities, and low-income students than when they teach more affluent and more advantaged students.

Second, statistical models that are most often used to measure teacher effectiveness ("value-added" models) have unreasonable high error rates-- a study by the U.S. Department of Education found that one commonly used value-added model had a 25% error rate with 3 years of data and a 35% error rate with one year of data. Up to 70% of teachers considered "highly effective" one year may be found "ineffective" the next, through no fault of their own.

Third, there are other negative consequences to tying teacher employment status to test scores. These consequences include a narrowing of the curriculum to only the subjects being tested, further warping teaching into test prep, discouraging teachers from working in the neediest schools, encouraging cheating and other ways of gaming the system, and pitting teachers against one another rather than fostering collaboration.

For the sake of comparison, I did some research about teacher evaluation systems that have been implemented or overhauled recently in other states. In Virginia, 40% of a teacher's evaluation is now based on student value-added; in Colorado, Tennessee, Delaware, Washington D.C.,  Rhode Island, Georgia, Florida, and Illinois, it's 50%. All of these changes have occurred in the last two years. It's reasonable to expect that Texas would pursue something similar.

I'll save my thoughts about alternative systems for evaluating teachers for another blog post, but let me just say that absolutely zero research supports the tying of test scores to teacher evaluations. There is no empirical evidence that doing so will help teachers improve or help students learn. But then again, since when has Texas let facts get in the way of what it wants to believe?