Gifted Issues Discussion homepage Forums Identification, Testing & Assessment Flynn effect

The same is true, by the way, for kids who need support services at the low end of the ability spectrum, e.g., the scores drift away from average at that end, too. This means that [when the identification test is new] a child with learning difficulties, a slower learner, will score HIGHER than the cut-off score needed for him to qualify for special education help in school. Flynn himself uses the example of death row cases where prisoners whose IQ's are too low are considered incapable of understanding that what they did was wrong; they're spared the death penalty. But when tests are new, more of them miss that low-end cut-off and are put to death.

The more I learn about how unreliable IQ test scores can be, the more I realize that descriptions of young gifted people are still probably the most reliable way to figure out how to meet their needs.

Schools throughout the United States, including those in Colorado, frequently utilize subjective measures as part of their criteria for identifying children as gifted. How accurate are measures such as behavioral rating scales in ascertaining whether a child is gifted? The research does not strongly support rating scales as an accurate means of identifying gifted children.

While it may be true that most gifted children exhibit characteristics such as sensitivity and perfectionism, the reverse is not as true. All children who are sensitive and perfectionistic are not necessarily gifted. It is much like the mathematical analogy that all squares are rectangles, but all rectangles are not squares.

Locally, in [local] Districts, a child can be identified as gifted with a combination of any two of the following items: a behavioral rating scale such as the SIGS, high grades or other performance measures, high achievement on tests such as the [NCLB tests], or an ability test such as a group ability test or an IQ test. IQ tests usually are not administered and are only considered when private testing has been done.

So, what does the research show? The well known and local Gifted Development Center's (GDC) work on behavioral characteristics of giftedness is often cited as proof that parents are good at identifying whether their children are gifted. What the GDC studies found was that, using the Characteristics of Giftedness Scale that they developed for use in their own work,

"parents of all the children who scored in the gifted range indicated that their children manifested at least 13 of the 16 characteristics in the original scale"

and

"In this study, 84% of the children whose parents indicated that they fit three-fourths of the characteristics tested above 120 IQ."

Neither of these two statements is the same as saying that parents or teachers who identify students with behavioral characteristics such as curiosity and a good memory, two of the characteristics on their scale, have identified which children are gifted.

The first statement indicates, again, that all squares (gifted children) are rectangles (possessing certain characteristics), not that all children who possess those characteristics are gifted. The second statement indicates that 84% of parents who believe that their children are gifted strongly enough to spend over $1000 having an assessment of their aptitude done at the GDC were close enough to right that their children had IQs at or above the 91st percentile.

In terms of the behavioral scale most commonly used locally, the SIGS (Scales for Identifying Gifted Students), it shows only weak to moderate correlation with intelligence scores (.38 to .67 as correlated with the Weschler Intelligence Scale for Children III / WISC III) and it has not been correlated with the most recent version of the WISC, which itself is so outdated that it is about to be renormed. A correlation coefficient of less than .5 is considered a weak correlation and the correlation coefficient should be greater than .8 to show a strong relationship between variables.

What needs to be kept in mind when using subjective measures such as behavioral rating scales is that they should not be given as much weight as more quantitative measures of giftedness such as ability testing. Additionally, when dealing with what is a often considered a desirable label such as "gifted," teachers with no formal education in giftedness and even parents can easily misidentify non-equilateral rectangles as squares.