As states revamp their teacher evaluation systems, they continue to search for that magic number: the percentage of a teacher evaluation rating that should be based on student academic performance. Here’s how this has played out over the past month:
- The Ohio State Legislature voted to lower the weighting for student growth from 50 percent of a teacher’s evaluation to 42.5 percent. Why the seemingly random choice of 42.5 percent? Because the state Senate wanted it revised downward to 35 percent and the House wanted to keep the weighting at 50 percent. Legislators compromised on 42.5 because it lies smack dab in the middle of 35 and 50.
- In New Jersey, Governor Chris Christie signed an executive order mandating that statewide exams account for 10 percent of a teacher’s evaluation this upcoming school year rather than the previously decided upon 30 percent. It will climb to 20 percent in 2015-16.
The issue here isn’t whether 10 or 35 or 50 percent is the right amount of student growth in teacher evaluations. No one knows for sure what that number is, and no one knew it when states set their initial student growth weightings either.
Ironically, we have better evidence now than we did when states made their initial decisions. The 2013 MET Project report found that weighting student growth between 33 and 50 percent of a teacher’s evaluation score would provide the best combination of predictive power and year-to-year stability. The MET Project is by no means definitive and we could certainly use more research in this realm. But before seeing any results or carrying out their own analyses, states are pre-emptively lowering their student growth weighting. And instead of using the evidence that does exist, states are allowing political battles to drive their decisions.
–Chad Aldeman and Carolyn Chuong
This says nothing about the politics of the actual growth model selection for teacher evaluation. Many states are using Student Growth Percentiles, which are not valid for use in teacher evaluation as are often biased against schools w higher percentages of impoverished children. See here:
http://njedpolicy.wordpress.com/2013/05/02/deconstructing-disinformation-on-student-growth-percentiles-teacher-evaluation-in-new-jersey/
Also, details such as measurement error and the appeals process are crucial!
Here is something simple for everyone to consider:
In order to evaluate a teacher based on her students’ test scores, you would have to have a test that is designed to measure each child’s school learning as opposed to his learning in general. At this time there is no whole group test that can do that. If a standardized test is given as directed (no peeking) the scores generally reflect the socioeconomic background of the student. This applies to almost all standardized tests, including the SAT.
Of course, the in-school progress of a child can be measured, but at this time such testing requires the involvement of skilled professionals who would give pretests to each child in the fall and then follow that child’s progress throughout the year. At this time, there is no two-dollar group test that can assess the in-school learning of each child while evaluating the teacher at the same time. Do people really not know this, or are they pretending not to know it?
In excellent schools throughout the country, administrators make it their business to be aware of the classroom progress of every child in the school. It should be obvious to everyone that a teacher cannot be evaluated according to the progress of her students if no one is aware of that progress!
I believe that Linda is correct. How can a standardized test accurately measure student learning and growth. On testing days, my students are not themselves. They are nervous and worried about this “special” test that they have to take. I have found that my students demonstrate over and over again their learning within the classroom. However, this is not translated over to the standardized tests. I think that student progress should not determine a teachers evaluation. Students range in ability is great within a classroom. Each year, I get new students, who were nothing like my previous class. They are so different in many ways. This provides an unfair way to evaluate teachers. If we are going to use testing as a way to evaluate a teacher, then we need to create a test that accurately measures a child’s knowledge.
I disagree that standardized tests provide an accurate measure of student growth. I do agree that teachers should be held responsible for their performances. There should be alternative methods that parallel student learning with teacher evaluations. Students have various learning styles and many other factors that can alter their performances on standardized tests
Britani, It is interesting that you posted how your students are not themselves on testing days. In order for these test to truly be standardized, each child would be taking it in the exact same environment at the exact same time and given the exact same instructions in the exact same way. Our building has always had issues with temperature. In some rooms it may be extremely hot while others may be freezing. This is not a standard environment for testing. How then can we say that these test are an accurate measurement of students learning/growth and then use this information for or against a teacher?