Despite what you might think, most teachers aren’t against testing. Most see it as an important measure of how well they are doing their job. This knowledge can help them find deficiencies in how they teach certain topics. It also allows identification of students that need help in certain areas. Sadly, the issue has been painted as teachers not wanting to own up to the results of testing. If only the results could be interpreted so simply.
If you were to look at my wife’s test scores at her previous school, you might argue that she was a weak link on her team. Compared to some of her peers, her scores languished behind. If one teacher was to be let go, based on test scores alone, she might have been the one chosen to take a hike. These facts support the idea that talent trumps experience (she had the most years teaching on her team).
Supporting the same line of thinking, my wife’s new class has among the best scores in the school, despite her being one of the least experienced teachers. The fact that she is new to the school, the 4th grade, and the curriculum has had no impact on her teaching. Her kids are doing great.
But, my wife refuses to take credit. She actually argues that she isn’t as good a teacher as she was before. Yet the data says otherwise. So what should you believe?
Well, the simple answer is to trust the data. My wife, the relatively poor teacher in her old school, is a great teacher in this new one. Perhaps this is a reflection on the rest of the teachers in this new school. Perhaps the move across country made her a better teacher (she’s a better teacher in the eastern time zone). But perhaps other factors were involved with these results.
As the most experienced teacher on a team, is it possible that she may have been given kids who were low for their grade, had behavioral issues, or some other factor that may interfere with learning? Would a reputation for excelling with difficult or struggling student lead to increased placement of these students in her classroom?
In the same instance, would her new school, not knowing who would be taking over or their ability level, give the new teacher a classroom with fewer potential issues? Would the class be loaded with a talented and gifted cluster (TAG) that left the room for reading instruction? (yes) Would children who were struggling in previous grades be shuttled toward known teachers with a reputation for success in that area? Could fewer behavior issues be expected for the new teacher’s classroom?
Asking yourself these question, is it easier to understand why my wife went from marginal to exceptional by shifting school? Perhaps the question is more complex than test scores can tell you. Yet, many people would have you believe that test scores should be the lone evaluation system.
A system based on test scores might have some issues, but the issues can be controlled for. This argument was brought forward in an LA Times report that analyzed the effectiveness of classroom teachers in LA Unified School district. The study’s author concluded that, controlling for a certain variables, teachers could be directly compared by the value they added to the students acheivement. From this data, the study’s author ranked teachers as either effective or not-effective, and the results, with teacher names, were published on the LA Times website. Thus, students, parents, and teachers had the information and perhaps could then crucify the teacher.
Obviously, this report caused quite a stir. Many argued that the results were unfair and failed to illustrate the entire picture. Others argued that the results were both a fair and effective way to delineate between good and bad teachers. Subsequent analysis of the same data by Briggs et al. called much of the analysis into question.
Among the most important assumptions in the first study was that class distribution occurred at random. But most teachers will tell you that class composition is far from random; teachers spend considerable time assembling classes often taking into account the strengths and weakness of future teachers. The reanalysis by Briggs seems to highlight this consideration. The findings suggest that in analysis of 3rd and 4th grade test scores, future teachers had an “effect” on the scores of kids they never taught in reading, and to a lesser extent in math. This data implies that the distribution is far from random (no effect would be expected if kids were randomly assigned).
In addition, the study calls into question the validity and error rate of the rankings, the inclusion and exclusion of certain variables, and the lack of consistency between the two models. Naturally, the authors of each study vigorously defend their methodology and cast doubt on the opponents (1).
Other studies have shown that the various methodologies used to catagorize and rank teachers have wide fluctautions or produce numerous errors (2,3). Do we really believe that a teacher who was good one year, would be rated very poorly the next year? Is this a fair way to evaluate teachers? Do the critics think that a system based solely on test scores can really evaluate teachers effectivness?
So yes, test scores are a valuable tool to evaluate both teachers and students. But, in this country, we don’t decide student value simply by how they scored on a single test. We value the entire student with multiple variables; this system is in place in much of our society as well. So then, is it unreasonable to ask for teachers to be evaluated in the same way?