csc@watmath.UUCP (Computer Sci Club) (05/24/84)
Lisa Chabot clouds the issue by drawing an arbitrary distinction between
"Test score" and the "ability to take such tests". I draw no such
distinction. The only reasonable measure of how well one can do on
a test, is what score one gets on the test. Thus the statement that
test scores only measure the ability to take the test is a tautalogy.
I agree that the TQ analogy goes "tallness can only measure a persons
ability to take tallness tests" but "the ability to take tallness tests
is what tallness was defined to be. (The tallness test was not defined.
Make it the length of some bone if you like.)
Lisa Chabot makes the now rather boring point that IQ scores are
affected by a large number of factors, among them cultural factors
and training. Despite this IQ scores are still highly correlated
to important factors such as success in university. The fact that
some of the things that affect IQ almost certainly do not affect
the ability to do well in University (and there are almost certainly
factors which affect success in University that do not affect IQ)
makes IQ an imperfect test (this is news?). In particular it is
of very limited value when applied to individuals. It is not
useless.
Test scores measure the ability to take tests. The ability to
take tests is determined by a large number of factors. The test
is often said to measure these factors. The factors that affect test
score are often correlated (sometimes highly correlated) to other
things. Depending on your definition of "measure" tests should not
probably not be said to measure these other things. However if
the correlation is strong the test will be a good predictor of
these other things.
William Hughescsc@watmath.UUCP (Computer Sci Club) (05/24/84)
For "should not proabably not" read "should probably not". Sorry
William