Abstract
I appreciate the opportunity to respond to the article by Warne, Godwin, and Smith (2013) on the question of whether there are more gifted people than would be expected in a Gaussian normal distribution. I believe the answer to that question is yes, and that answer is based on (a) data that we have collected ourselves, (b) data that are already available and quoted by the authors themselves, and (c) the nature of the measurements designed to test intelligence.
I first became aware of this phenomenon of many children who scored high on IQ tests in a study done in Champaign-Urbana, IL, the home of the University of Illinois (Gallagher, Greenman, Karnes, & King, 1960). The study was designed to discover whether the elementary schools could adapt to students with extraordinarily high IQ. We set the arbitrary figure of an IQ of 150 or above to establish the population and, on the basis of the normal curve and the school population, we expected to find six or seven students upon whom we could do a case study, to try and answer the question that we posed. We used the Stanford–Binet (L-M) as the test, and to our astonishment and frustration, we discovered 54 students in the elementary schools who scored above the cutoff figure, four of whom scored more than 200 IQ!
Because analyses of the test results were not the focus of the study, we attributed the extraordinary number of high IQs to the presence of the university with its distinguished faculty families contributing to the imbalance. Only later, upon reflection, did we consider how truly unlikely such a result would be even granting a skewed distribution because of the local population. It was easily 7 or 8 times the expectation.
Obtaining an IQ more than 150 was possible because the Binet (L-M) test at that time compared the students against other students of all ages. The current IQ tests such as the Wechsler Intelligence Scale for Children (WISC) were constructed to produce a deviation IQ, with the assumption of a normal curve built into the construction. Most of the currently used IQ tests are organized on a deviation IQ model; that is, how well does the student do when compared with students of a similar age? This is useful information for educators but not designed to explore the upper bounds of intelligence. There will be no more students scoring 180 on an IQ test because the structure of the test itself limits the upper score (Hollingsworth, 1942).
The authors themselves quote Terman (1926) and Burt (1957) as finding more students than expected at the upper levels of ability, a finding made possible by comparing the student’s performance against the whole age range of student performance. Even more relevant, the authors also quote Raven, Raven, and Court (1998) as to how suspect it is to create an intelligence test based on the normal curve, and then treat intelligence as though it formed a curve just as the tests prescribed.
Furthermore, items on which only a miniscule number of students could perform were thrown out of the test (Silverman & Kearney, 1992). For example, if you designed a mathematics abilities test for the elementary school, you would not include items requiring calculus to solve them because so few students could perform at all on such items, even though every now and then a student comes along who is able to perform on such items. So those students would clearly have their mathematical skills underestimated.
For another example, suppose that Lenny (8 years old) likes to play chess and is able to demonstrate that he can defeat all other 8-year-olds in tournament competition (deviation chess IQ). That is impressive but not nearly as impressive as his ability to compete on the level of adult chess players. This adds a different dimension to his chess IQ. It is the fact of being able to perform at an adult level that is so impressive. A true measure of high ability is how a student performs on tasks in comparison with all ages of students or adults.
The normal curve with its beautiful symmetry does not apply at either end of the spectrum of intelligence. As pointed out in an earlier article by Robinson, Zigler, and Gallagher (2000), the normal curve does not apply at the lower ends of the distribution either! Very few students (much less than the normal curve would predict) are found at the lower end of the distribution. It is hard to find anyone scoring 2 SDs or more below the mean who do not have neurological damage or genetic mishaps. So the lower bound of the scale for students of intact neurology and physiology is not 0 or 15, it is about 70 or about 2 SDs below the mean. So the normal curve is not normal at that end either despite the efforts of the test constructors to make it so.
Much of the evidence quoted from the authors of the article in support of the normal distribution of intelligence relies on data from the two extensive Scottish Mental Surveys, which in turn rely upon results from the group-administered Moray House Test (Scottish Council for Research in Education, 1933). According to the data provided by the authors of this article, the scores from more than 70,000 children who took the Moray House Test had a mean score of 36.7 and a SD of 38.4 with a total of 76 points possible on the test (p. 14 of draft). Such figures, if correct, point to a wildly skewed result (a SD larger than the mean!) and a serious ceiling effect. In no way can it be said to test the upper limits of ability of the population.
If in fact there were a normal curve of intelligence based on genetics, then what is one to make of the huge differences in the performance of various ethnic groups, noted in The Bell Curve by Herrnstein and Murray (1994) and others? Most social scientists point out that such results can be the result of differential environments and opportunities, but that means that the tests scores are influenced by the environment. How else can we explain the Flynn Effect (Flynn, 1987), which has entire generations gaining in IQ over the years?
Nor is this difference of opinion on the shape of our intelligence distribution simply a matter of a squabble between academicians and scholars. The nurturing of extraordinary ability is a matter of our flourishing as a society, or even a matter of national security (Gallagher, 2013). The good news that I hope has been demonstrated here is that we have a larger population of students who have the intellectual capabilities to change society than we had imagined. Our job is to see to it that our educational system is up to the tasks of helping such students to use their impressive gifts effectively.