<p>Someone asked the other day about what could be expected of someone with an IQ of 100. That post received a very good reply while my attempt to reply was eaten by the posting bug. :( Here I'll post my FAQ on that general issue: </p>
<p>CORRELATION OF TEST SCORES </p>
<p>There is a quite persistent set of errors in many posts here about predicting one test score from another test score. </p>
<p>1) The first error is assuming that IQ scores stay at the same level across the lifespan. There is actually abundant information to show that this is a mistaken idea. Rather than asking, "What is his IQ?" one should really ask, "What score did he obtain on what brand of IQ test on what occasion?" All longitudinal studies of IQ have shown IQ scores moving both up and down on subsequent testing of individuals who were tested at various ages. Here are some citations to studies of this issue: </p>
<p>Anastasi, Anne & Urbina, Susana (1997). Psychological Testing. Upper Saddle River, NJ: Prentice Hall. </p>
<p>Deary, Ian J. (2000) Looking Down on Human Intelligence: From Psychometrics to the Brain. Oxford: Oxford University Press. </p>
<p>Howe, Michael J. A. (1998). Can IQ Change?. The Psychologist, February 1998 pages 69-72. </p>
<p>Moriarty, Alice E. (1966). Constancy and IQ Change: A Clinical View of Relationships between Tested IQ and Personality. Springfield, IL: Charles C. Thomas. </p>
<p>Pinneau, Samuel R. (1961). Changes in Intelligence Quotient Infancy to Maturity: New Insights from the Berkeley Growth Study with Implications for the Stanford-Binet Scales and Applications to Professional Practice. Boston: Houghton Mifflin. </p>
<p>Shurkin, Joel N. (1992). Terman's Kids: The Groundbreaking Study of How the Gifted Grow Up. Boston: Little, Brown. </p>
<p>Truch, Steve (1993). The WISC-III(R) Companion: A Guide to Interpretation and Educational Intervention. Austin, TX: Pro-Ed. </p>
<p>2) The second error is assuming that any two brands of mental tests, or indeed any mental test given to the same group of test-takers more than once, will sort a group of test-takers into the same rank order. That never happens. In other words, there is no such thing as a test that has a correlation of 1.0 with any other test, and indeed there is no one brand of mental test with test-retest correlations that high. It's quite exceptional to see test-retest correlations above about .85, and that level of correlation allows for plenty of radical changes in rank order among test-takers between two instances of taking a test. For any score on any test, there is a RANGE, which can be quite broad, of expected scores on some other test. So it shows a fundamental misunderstanding of mental testing to suppose what a person's IQ score is from the person's SAT score, or the other way around, and an even greater misunderstanding to assume what a person's grade average in high school or college might be, or what occupation that person will pursue successfully, from such limited information. </p>
<p>See </p>
<p>Hopkins, Kenneth D. & Stanley, Julian C. (1981). Educational and Psychological Measurement and Evaluation. Englewood Cliffs, NJ: Prentice Hall. </p>
<p>or </p>
<p>Mackintosh, N. J. (1998). IQ and Human Intelligence. Oxford: Oxford University Press. </p>
<p>or plenty of other general titles on mental testing for more on this issue.</p>