Friday, January 17, 2003

Dishonest Testing

Critics of the SAT and other standardized tests claim that the SAT does not test "real" ability and advocate various alternative ways to test ability.

These alternatives take various forms, but the thing that tends to tie them together is their low reliability, that is, the high variablity of test scores. Standard deviations of scores for these tests, in other words, tend to be high.

That fact alone allows their proponents to claim that these tests are "more fair" and less biased than standardized written tests.

How can this true? Because when the test is administered the examiners can norm the test and set a passing grade in the usual way. A certain percentage of the students will fail, but a student with marginal ability is more likely to make a passing score just by chance. This is especially true if he or she is able to retake the test or take a similar test one or more times after initially flunking. With repeated testing they are able to pass more often just by chance if the test score variablity is high. This is not possible, or is much more difficult, with a standardized exam.

(Of course, the examiners have failed in their primary duty, which is to provide the students and others with an accurate assesment of their ability. When it comes to important things like determining which medical students have an adequate knowledge of medicine, for example, this failure can get to be quite serious.)

In addition, just by chance the class ranking of students taking such tests tends to be different from those of standardized examinations. Ipso facto, these alternative exams are testing a "different kind of knowledge" according to their advocates. This enables some students who don't usually do well in standardized examinations to look better.

That there is no basis whatsoever in educational psychology for the idea that there are "different kinds of knowledge" does not phase them one bit.

The other fact about some of these alternative evaluation methods that endears them to egalitarian educational buffs is their subjectivity. Subjectivity provides examiners with enough wiggle room to insure that the "disadvantaged" students pass at a reasonable rate.

So, whenever one hear critics of the SAT or other standardized written exams talk about other evaluation methods that test "real" ability, or test "different kinds of knowledge" or the like, it should be apparent what the game is.