Do Student Tests Say Anything About the Teacher?
Posted by SBuck | Education, Politics | March 23, 2010
Diane Ravitch said the following at a recent event:
“Every test publisher says, to my knowledge, and everyone who’s involved in the psychological testing business says, the tests should be used only for the purpose for which they’re intended. If a test is intended to measure a 5th grader’s ability to read, it’s not developed and designed to measure whether the teacher is effective. There are all kinds of reasons why students’ scores are higher or lower which may have nothing to do with the effectiveness of the teacher but rather with a million other things.”
What if we want to know whether the 5th grade teacher was effective at teaching the 5th graders to read? What exactly would be covered on the test other than items measuring the 5th graders’ ability to read? I understand the point that there are lots of reasons the 5th graders might not do very well (low SES, bad family circumstances, homelessness, illiteracy at the start of the grade, etc.), but none of those factors have anything to do with how the test itself would be written, would they?
It’s really, really simple. Distinguish between *levels* and *gains*, amounts and changes, stock and flow, etc. This is why “teacher effectiveness” is being tied often to “value added”, not to prior learning. Ravitch buries this distinction in ambiguous language in the above quotation.
She does make a good point that “why students’ scores are higher or lower” may have to do with “a million other things.” For example, a natural disaster could strike town and disrupt the lives of several of the children in a teacher’s classroom. This might interrupt progress in learning, and the teacher could do nothing about it. But does this possibility tell against the validity of using many students’ pre- and post-test scores over years to measure a teacher’s effectiveness relative to other teachers?
Yes, but what about the *test* itself would require that you use it merely to determine “levels” rather than “gains” as compared to prior tests?
You may be begging the question. I think the lady’s saying you don’t use the test for that purpose.