I'm not sure why the test publishers are so insistent on keeping subtests that are poor measures of 'g'. It messes up the predictive validity and construct validity of the IQ score. It's important to remember that IQ subtests are *not* designed to measure specific cognitive abilities. They are designed to measure components of the statistical concept of 'g'. Too many different types of cognitive skills are involved in each subtest. There are many other more 'pure' tests of things like working memory and processing speed. If you really want to break down a child's strengths and weaknesses outside of general verbal and non-verbal measures of 'g', other tests give you a much more useful measure.

Last edited by Jool; 12/20/08 10:48 AM.