I previously wrote about what an IQ score can and can’t tell you and I followed that up with a bit more background on ability testing. Here is another step to help you understand your own and your children’s testing on standardized tests over the years.
Different ability tests have different scales or ranges of scores. The ones you can take free online generally have a higher top score than the tests you can take individually with an assessment specialist or in school. You may have noticed that the test results tell you your score but they don’t necessarily put the score in context for you, and they don’t explain how the number of people that did better or less well than you can help you see how the score relates to your real life. That’s not very helpful or satisfying.
The original Binet scales intelligence tests created in the early 1900s utilized a ratio IQ score rather than today’s standardized ‘forced bell-curve’ scores. It was called an intelligence quotient. Modern tests, most of the tests created since then, don’t use the quotient.
In those early years of intelligence testing, IQ was calculated by dividing the mental age by the chronological age and multiplying by 100 to produce a ratio IQ. Your mental age is the raw score you got on the test compared with the actual age of the others who took the same test. If your mental age is higher than your actual age, you will have an IQ above the average of 100. For example, if you scored like the average 10 year old when you were only 6, your score would be 10/6 x 100=166.7.
Scores on those tests were unstable across age ranges, and there was no ability to estimate the frequency of a particular intellectual level in the general population. A forced bell curve, otherwise known as a standard curve, used rankings instead and “smoothed the curve” so it could also use standard deviations and percentiles, something the earlier tests did not have or use. And by rankings, I mean the results from the norming sample and how many people scored higher or lower than you did.
My Background in Test and Measurement and Administering Tests
I administered the Stanford-Binet 5 for more than 12 years, and I was part of the normative sampling team for the gifted population in 2001–2002.
I later participated in norming for the Wechsler ability tests (WPPSI-IV and WISC-V) for gifted samples of children. The scores are much lower — at least initially — than each of these instruments earlier versions as well as the scores from the SB5. This is because the older versions had 12 to 15 years of a steady score rise, the Flynn Effect, and the newer versions simply arrange the final scores by who is taking the tests now. So, the children haven’t really changed overall, but for schools and organizations that have score cut-offs for inclusion, say for Mensa or a school gifted program, children who would score as gifted on an earlier version of any of these tests, may not on the newly, freshly normed tests. This is another reason I don’t like cut-offs for admission to programs that might support learners.
A Little Historical Context on IQ Tests
Over the years, such assessment techniques have proven to be quite reliable, but the effects of a gradual increase in scores, as noted by James Flynn of the Flynn Effect (about .3 points a year) makes it important that we always look at the big picture when trying to determine a child’s real “fit” compared with others of the same age or in the same classroom.
Generally, one simply cannot compare a Full Scale Score of 145 in 2013 to that of a 145 on a different same-aged child from a different peer group in 2002. Flynn and others have shown, too, that some subtests are far more sensitive to rising scores than others are, but this sensitivity affects the final score, nonetheless. I recommend Flynn’s work highly, for those who are interested, especially Are We Getting Smarter? Rising IQ in the Twenty-First Century (2012).
Current tests only go as high as about 150 for the very smartest people, and generally, only very young children have been able to get scores over 150. Both the WISC and the SB5 created those extended scores. To my knowledge, the majority of us assessment professionals has not yet scored anyone over 160 on this test nationwide, even though the test designers created an “extended score range.”
It was on the old ratio IQ tests that one could earn scores in the 170’s, 180’s and higher. Some clinicians mistakenly attributed percentile estimates to the old ratio scores, but that was technically incorrect. The authors and publishers of the old tests never actually did norm samples for the distribution of scores; they simply normed individual questions by age. As a result, there is much popular confusion over the relative meaning of different IQ scores.
What does “freshly normed” mean?
It means that because there has been a steady rise in IQ scores over time, the age cohort that helped to norm a test isn’t the same one to whom you are being compared if you take the same test many years later. If the seven-year-olds who were part of the normative sampling in 1950, for example, and no children at that time had access to television or computers, their “test savvy” would be very different from seven-year-olds taking the same test 30 or 40 or 50 years later. Their experience, their interpretation, and the culture in which they find themselves would very much affect how they approached and reacted to test items. Plus, parents and teachers tend to “prepare” children for better testing after they get their hands on the newer tests as they trickle out and into test practice workbooks.
As a society modernizes, more and more of the children experience puzzles and matrix patterns and activities that require them to figure out what’s missing, what’s wrong, what goes with what in a matching exercise, and so on. Have they actually gotten smarter?
Certain kinds of test items that include vocabulary, analogies, and general information haven’t shown the same kind of score increase, though.
Some core pieces of intelligence have not changed as significantly as others over time.
Most notably, matrix pattern reasoning continues to show a rapid increase in populations that move from a concrete, literal way of living (say nomadic or agrarian) to one that is more complex and science-based, as with urban living and dealing with symbol systems such as money, laws, and written rules and information.
Because pretty much everyone on the scale moves up – gets older – during this time, the differences in learning ability between groups of people the same age remains fairly stable. That means that a child’s relative position in the group may still be the same. Let me give you an example. As the tests age, more children “hit the ceiling” on some of the subtests. This means that more of the children score in a range that was previously known as exceptionally or profoundly gifted, but their actual abilities aren’t in that range compared with those who earned those very high scores when the test was new.
A similar thing happens with scores at the low end of the continuum of ability. Children who need services due to low ability will score too high on the newer tests and appear to be ineligible for extra support in school. I know this is confusing, but the main point here is that you can’t very easily compare scores over time, tests, or testing occasion. And it is often counter-productive to do so!
Many tests correlate with other tests. This means that if you have a high IQ, you’re likely to be good at matrix patterns or vocabulary, too, for instance. But, a test that doesn’t include any vocabulary, such as the Naglieri Nonverbal Aptitude Test, the NNAT, misses the discovery of a very important ability: language acquisition and use.
Some very smart children who take the Naglieri Nonverbal Aptitude Test, seen by many as being more culture-free and fair than the tests that include both verbal and nonverbal assessment, don’t score as high on a strictly nonverbal test. Yet, verbal ability and vocabulary are more highly correlated with the general intelligence factor, or g. Children who are lower on their verbal abilities are more likely to have academic difficulties than children who are high on verbal abilities. They need a different kind and quality of academic support.
People have different abilities profiles , too. There really are some people who are much, much better at quantitative reasoning — math thinking, knowing how to set up the problem that needs to be solved — than others. Yes, you can teach people to be better than they naturally were, but it won’t make them the same level of “naturals” as those who had the ability and intuition for math from the get-go. You can still be a genius, though, even if math isn’t your strong suit. So, the more of a balance with different abilities an IQ test is, the more likely someone with an uneven ability profile will end up with a full scale score that understates their strengths.
Incidentally, for you high verbal people, I have always recommended the Miller Analogies Test for people who know they’re smart but don’t score high across the board and need to figure out “where they really are” compared to others, what their strengths are. Unfortunately, the company that owned it ended its publication at the end of 2023. If you think you’re more math-strong, take the Raven’s Progressive Matrices.
Different parts of the world go through periods of views about what’s important in the field of education, and tests for general intelligence, g, like the Miller Analogies Test, don’t specifically predict who is going to be a good engineer or tradesman. I’m just thinking out loud here when I present Three of my major viewpoints around intelligence and testing:
• All children are gifted.
o Testing isn’t fair, etc.
o All students need a good, i.e., (same) education.
o The need for intelligence testing is unnecessary; it’s how we teach that makes the difference.
• Not everyone needs the same kind of education.
o Women
o People of color
o People whose past education and experiences already show us who’s smart and who isn’t.
• Educational institutions at all levels need to prove their ROI, meaning Return On Investment.
o If a person takes all the right courses to be an engineer, for example, the time and money used has to make the student job ready.
o There is little way we can prove a ROI with fluffy courses in the arts and humanities.
All of these are short-sighted, of course, and yet we go back and forth, round and round, without ever really developing educational systems that work. We are not all the same in our likes, our goals, our different strengths and weaknesses and yet we keep setting up systems that assume that kindergarten through 12th grade is the right way to educate people. People grow and change throughout their lives, and we need systems them allow people to be lifelong learners.
As I wrote my follow-up book The 5 Levels of Gifted Children Grown Up: What They Tell Us (2023) that was based on the 18-20 years longitudinal case study results from 5 Levels of Gifted: School Issues and Educational Options (2005, 2009), I had some early readers tell me I should stick to acceptable score ranges for moderately gifted, highly gifted, exceptionally gifted and profoundly gifted. I won’t do that because that is more than inaccurate. It neglects to address the complexity of intelligence, personalities, behaviors, and environmental effects. And although being precise appeals to many parents, individuals, and educators, such precision vastly undermines the goal of finding the best support systems for those who are unusually gifted. I do continue to write on the topic of clear-cut labels for score ranges, just to help out those who really count it.
My current published books about the gifted: