Validity and reliability are the two necessary pieces that make a test that actually provides useful information about learning. Validity goes to whether or not a test actually assesses what it’s supposed to. Reliability goes to whether or not the results of a test are reproducible.
I heard an interesting podcast from Scientific American the other day. Here is the link, and following is part of the transcript.
… her studies showed that the best students were the ones most likely to choke under a high-pressure test situation. That’s because normally they use their higher memory capacity to methodically work through a problem. But when the pressure’s on, the good students resort to the same ineffective shortcuts the poor students use all the time. Another researchers shed some light on this with his study showing that anxiety actually occupies working memory, wasting it instead of devoting it to the task at hand….
Leading to this conclusion:
They say their findings suggest that high-pressure tests might not be measuring what they’re meant to, and schools might want to try de-emphasizing their importance.
Statewide tests can be a real pressure-cooker for children because so much weight is put on them by No Child Left Behind. How much do tests assess “test-taking” skills and level-headedness? If a test’s results are colored by a child’s test anxiety, that erodes the validity of the instrument. Could regular smaller quizzes or something less stressful than the big statewide test “book” that student are handed out every year be better at measuring student progress? Something to consider.