Gifted Issues Discussion homepage Forums Identification, Testing & Assessment How do Kids' Scores Translate into Adults' Scores?

I was wondering about kids' IQs and the percentages. How can you there be 2 5 year olds, 2 6 year olds, 2 7 year olds, and so on, per 100, yet only 2 per 100 18-100 year olds per 100? Does that make sense? What happens? Does that mean that most tested kids wind up not being gifted adults on an IQ test?

Another thought.. I noticed that Mensa accept the WISC GAI of 130. But what if their FSIQ is 115? Does that mean as an adult they wouldn't make it into Mensa? Does that mean in 10 years Mensa will consist of the top 20%?

I believe that some tested kids wind up not testing as "gifted" as adults in the sense that a child (especially a child younger than 8) who may have tested at say IQ 130-135 can certainly test at 116-121 (less than one standard deviation) on an adult test.

However, don't confuse absolute numbers with percentages. In your example, there can still be 2%, for example 6 out of 300 adults who were formerly five to seven year olds. As for the GAI issue, it is meant to include more people who may otherwise be excluded due a relative weakness in working memory and processing speed; it is possible, even likely, that such individuals as adults may not make the cut if tested on an instrument that didn't remove working memory and processing speed factors.

What is the population of the adults being tested? I am wondering where those numbers you mention re adults because I would expect that the number of adults undergoing IQ testing is much lower than children - testing is required for a lot of school programs but once you are grown up, I don't see where adults would be tested. So I am wondering where this question is coming from and where the data is coming from.

I know my test scores from 3 periods of my life with the last being in high school (after high school, there was no need for testing - nothing depended on those results post high-school), and the factors that impacted the scores over time.

I am confused by your question.

If you have two 6 year olds per 100 6 year olds, two 7 year olds per 100 7 year olds, etc, up to two 17 year olds per 100 17 year olds, you *also* have two 6-17 year olds per 100 6-17 year olds (there are 24 6-17 year olds in 1200 6-17 year olds, which is the same as two in 100).

How is this different from having two per 100 18-100 year olds? For every 50 people in this age group, there would one, for every 100 there would be two, for every 1000 there would be 20, and so on.

But maybe I'm missing something.

I think you may also be confusing percentages and percentiles. IQ scores are based on percentiles not percentages. For instance, in my school district 27 percent of age-eligible students have scored at the 98th percentile or above.

I am hearing two questions here:

1) is the distribution of high-IQ individuals the same in adulthood as it is in childhood?

2) will children identified as high IQ continue to be identified as such in adulthood (stability of IQ over the lifespan)?

Answers:

1) Yes. Contemporary IQ scaling is deviation-based, which means that, by definition, there will be 2% of the population at any given age that can be measured as 130+ IQ (the 98th %ile, on most cognitive assessments). This is purely a reflection of the bell curve and the test-norm development process. (Granted, there is discussion as to the applicability of the bell curve to the distribution of intelligence, as to many other human attributes.)

2) Mostly yes. The research on IQ is that it is reasonably stable from middle childhood on, but with regression to the mean, scores that are further to the right tail of the bell curve are likely to fall more on retest than those that are hovering around the mean (which are likely to continue hovering around the mean). Longitudinal studies have also found that unusually low IQs (on the left hand tail of the curve) are more stable than unusually high IQs. So it appears that the regression to the mean of outlier IQs is greater for high than for low scorers.

So to the OP's question about kids qualifying for Mensa, but then scoring below the qualification cutoff as adults: yes, that could very well happen, especially if the original qualifying score was close to the boundary. (Notice the GAI exemption does not apply to the WISC-V.) It does not necessarily follow, though, that in 10 years Mensa will consist of the top 20% instead of the top 2%, partly because not everyone who is eligible applies. (Every tested member of my FOO has qualifying scores, all by wide margins, but none of us have felt any particular inclination to join. Nothing against the organization, or anyone else joining. Just no interest.)

Frankly, the question of questionably qualified members of Mensa is probably affected more by the range of qualifying tests that are allowed than by regression to the mean. I've found pretty different numbers for the same children across combinations of the SBLM, SBV, WISC-IV, WJIII, and RIAS. Let alone the CogAT and OLSAT. A motivated person with the resources and borderline eligibility could probably test-shop until a qualifying score was obtained.

The norms for the SBLM are over four decades old now, and based on a completely different structure (mental age, rather than deviation IQ), both factors which generally result in higher numbers on the LM than on contemporary IQ tests. Mensa still accepts SBLM scores; I didn't see a comment anywhere about when the scores were obtained (perhaps others know otherwise), so one could, hypothetically, bump borderline qualifying scores over the threshold pretty easily by having an SBLM administered. (If you can find someone to do it--though Linda Silverman still likes it, so obviously some legit psychs are still making a case for using it.)

Perfect, aeh. Thank you!

That makes a lot of sense aeh. Thanks for taking the time to explain things.

I just posted in the GT Research forum a link that is relevant to your question: Converting IQ at a Given Age to an Absoloute (Rasch) Measure of Intelligence

For instance, this allows telling how smart your 145IQ 7 year-old is compared to a 120IQ gifted 13 y.o. class (about the same).

Here's a more detailed description of how it works:
To compare the absolute intelligence level of two people of different ages and IQs, first convert the IQs to standard deviations, s.d. (= z-score = sigma) = (IQ - 100)/15

Then consult the graph at the link, find the age, move up a vertical grid line to the s.d., (e.g. if the IQ is equal to +2 s.d. then find where that vertical grid line intersects the second red line above the average (black line)), then use the horizonal grid lines to find the corresponding score on the scale on the left. Note that score. Do the same for the other person's IQ and age. Now you have two comparable scores.

You can compare the difference in their scores to the size of the standard deviation at the younger person's age (vertical distance between red or black lines at that age), which allows converting the older person's score back to IQ-equivalent for the younger person's age (so that they're comparable).