Several recent intersecting conversations lead me to this post: The North “credit recovery” issue, increasing discussions about using performance funding for Colorado higher ed and/or K12, evaluations of ProComp and other teacher incentive pay programs and Alex Oom’s valuable recent post.
If we want to incentivize or reward educational performance in some form (and we do), we need to pay careful attention to how we do that. Nearly any output or outcome measure can potentially be “gamed” or cheated. We see this with No Child Left Behind, where state tests are the key to school evaluation. As a result, states have produced considerable improvement on those tests, while not showing much improvement on NAEP, the national test that was not “dumbed down” to show greater proficiency of students.
It is also true that no single measure comes near being perfect. In addition to cheating or gaming, reliance upon a single measure (and test scores are the one that most of us would lean towards), makes the assumption that this measure is capturing appropriately what we want to capture. Currently, for state tests like CSAP, this is not the case, and we clearly need to find more, better tests.
In some ways, this is an obvious point – who can oppose multiple measures of evaluation?