Gifted Issues Discussion homepage Forums Identification, Testing & Assessment Most accurate- FSIQ, GAI, GAI with extended norms?

I'd read that, also. I wonder if working memory can improve with age.

Then I still don't understand why, for instance, my ds's tester considered the GAI to be the best measure in ds's report, when the tester also stated perfectionism most likely being behind the PSI score.

Admittedly, there is still much I don't understand about IQ measurements and the tests to obtain them.

Zen Scanner you missed my point entirely. My point is that the notion that gifted but slow processing subjects have inflated PRI scores because their very slowness gives them more time doesn't hold water. The block design section is timed so 'slowness' is penalized.

Frank22 welcome to the board, btw.

It's ultimately the discretion of the tester, but there are a number of guidelines between Pearson's http://pearsonassessmentsupport.com/support/index.php?View=download&EntryID=353 and this article+research from the Gifted Development Center http://www.gifteddevelopment.com/About_GDC/whoaregiftd.htm .

I did not. Perhaps you missed my second entry on the subject.
"Sherman, if you would... "

My kid is on the other end of the perfectionism continuum, and wrecklessness hits hard on visual transfer tasks. Glancing at a 3 by 3 grid meaning to check position 2,2 and instead looking at position 1,2 is a problem.

The best measure of scholastic achievement to my knowledge is g, and not working memory. Some scholars have argued that working memory and g are one and the same, but the methods used to support this hypothesis are shoddy. Moreover, digit span forward is a poor measure of g, and I believe it was Terman who seriously considered removing it from the Stanford-Binet (although it was ultimately retained because of the clinically rich data it provided on the subject).

Thank you Zen Scanner and Madeinuk for the salutations.


Working memory increases with age up to a certain point, after which it begins to decline. However, up-to-date IQ tests (such as the WAIS) are age-normed and take into account such decline. The average digit span for a twenty year old, for example, is 7 digits, while the average for a person in his or her sixties would be about 5.5.

Thanks for the information, and yes, welcome to the boards!

Frank22, which IQ test do you think is the most reliable for identifying giftedness in children and adults?

You're welcome, and thank you for the greetings.


All IQ tests that purport to provide a so-called "quotient IQ," derived by dividing one's "mental age" by his or her "chronological age" are defunct measures of intelligence. Extant IQ tests rely on deviation IQ, which is based on one's z score (the amount of deviation he or she has from the mean). The present day understanding of g is akin to the understanding of the gene between the time of Gregor Mendel and the discovery of the DNA double helix by Watson and Crick; we know that g exists empirically due to the correlation between any and all mental abilities, no matter how diverse. However, we do not know what it is that physiologically causes for g, although it must be a property or properties of the brain that are involved in all thoughts involving a conscious choice. According to Dr. Arthur Jensen, who passed away in late October of 2012, with a concerted effort from pyschometricians and stable funding, this problem may be solved in the next one or two decades.

That being said, such classifications as "gifted" and "superior" are not meant to pigeonhole the subject, but instead provide an argot that is accessible to both the clinician and layperson alike, and it is known that all IQ tests will measure the g-factor in addition to a sort of "fudge-factor" of non-g factors; the more tests one administers with higher g-loadings and ever more differing mental abilities, the more one minimizes this fudge-factor and measures g more accurately.

Although I am not familiar with the vast welter of IQ tests available, the Wechsler intelligence tests are the most widely used IQ tests at the moment. They are excellently normed, and in certain age groups measure full-scale IQ up to 210 for the WISC-IV. I would also prefer the WAIS-IV over the WAIS-III for other reasons; the addition of "digit span sequencing" has made the digit span task much more highly g-loaded (the g-loading of digit span on the WAIS-IV is 0.72, as compared to 0.59 for the WAIS-III). Arithmetic on the WAIS-IV has also been made to include simpler but longer calculations, therefore increasing its working memory demands and g-loading. Also, the WAIS-IV measures full-scale IQ up to 160, as compared to 155 for the WAIS-III.

In general however, many of the IQ tests in present use (such as the Stanford-Binet, Woodcock-Johnson, Differential Abilities Scale, Cognitive Abilities Test, Otis-Gamma Test, and so on and so forth) are easily capable of assessing IQ's several standard deviations from the mean, and a child who scored a full-scale IQ of 130 or more did not attain such a score through luck or "having a good day," etc.

Thanks, Frank22!