The boy who knew too much: a child prodigy

This is the true story of scientific child prodigy, and former baby genius, Ainan Celeste Cawley, written by his father. It is the true story, too, of his gifted brothers and of all the Cawley family. I write also of child prodigy and genius in general: what it is, and how it is so often neglected in the modern world. As a society, we so often fail those we should most hope to see succeed: our gifted children and the gifted adults they become. Site Copyright: Valentine Cawley, 2006 +

Tuesday, March 13, 2007

The Flynn Effect: are we all getting smarter?

For decades now, the average IQ of people has been on the rise, at mean rate of one third of a point per year. That is 3 points per decade. So, are we all getting smarter, then?

The short answer is no. You see the question is, for whom is the tally rising?

Research indicates that the mean IQ is rising because the IQ of those who are in the lower segment of the population is increasing. The higher up you go on the IQ scale, the less the IQ is improving. For very high IQ types, it is isn't budging, decade on decade.

Now, this presents us with a very real problem. You see, IQ tests are regularly "re-normed" to take account of this average rise...and these rising scores are reset to 100. Thus IQ points vanish, in this adjustment. This would not matter were the change in IQ the same across all IQs - but it isn't. High IQs are not really changing - but do the norms take account of this? From what I have read, and understand, the test takers are adjusting the test results, assuming that the Flynn Effect is a universal phenomenon. Since it isn't...what effect does "re-norming" have. Well, as you might have guessed, it artificially depresses the IQs of people in the upper ranges.

So, again, we have identified a source of IQ score depression for the most gifted, introduced into modern tests.

With all of these score depressing effects, (see my previous post The Great IQ Con) preferentially exerted on the most gifted, it seems that, ultimately, modern IQ tests are being designed by people who either have an agenda, or are themselves pretty stupid. Perhaps it is both, since you would have to be pretty stupid to have an agenda against the brightest among us.

What can we conclude from all of this? Well, it seems sure that whatever uses are intended for modern IQ tests, getting a true measure of the intelligence of the most intelligent is not one of them. The tests are quite useless for that - in fact, the typical modern test will give a very distorted picture of anyone who roves into the highly gifted territory and beyond. This obviously can have significant impacts on whether or not these children receive appropriate education, or not. In all, it is quite a serious issue.

Labels: , , , , ,

AddThis Social Bookmark Button
posted by Valentine Cawley @ 12:15 AM 


Anonymous Anonymous said...

As a mathematician, I am quite interested in this topic. Do you have any links or references which describe the actual mathematics used to "re-norm" the tests? I would like to see a comparison of how the old test (Stanford-Binet L-M) was scored vs. the new tests. I would like to see a reference to the research describing the trimodal distribution of IQ scores you described in a previous post. Can you point me in the right direction?

3:20 AM  
Blogger Valentine Cawley said...

I am glad you are a mathematician because I am not.

I am not privy to the mathematics used to renorm the tests - but I am aware that it squashes the bump of higher iq results into a normal curve - thus actually deforming the results, to fit dogma and convenience.

Stanford Binet gave a ratio IQ - that is a developmental rate IQ. It was very informative for highly precocious children since you could place them among intellectual peers more easily. There is a known standard conversion of ratio iq into deviation iq (I have seen it referred to - but I would have to look it up).

I am a browser - I read widely - store the nuggets of information and then synthesize it into a conclusion later - but I don't pay much heed to WHERE the information was from - so I would have to go looking again, to answer your question.

I will try to find the reference I saw to a trimodal distribution.

Thanks for your interest.

6:54 AM  
Blogger Valentine Cawley said...

By the way, here is the information that would allow you to convert the ratio results into deviation IQs.

For average IQs and above, ratio IQs appear to follow a log-normal distribution, with a standard deviation of 0.15 for the natural logarithm of the ratio of mental age to chronological age (from Scoville). Thus, for instance, a Stanford-Binet ratio IQ of 228 corresponds to a deviation IQ of 188.

After this correction, we can see that for the child mentioned below in The Great IQ Con, there would still be a massive difference in IQ on an old high ceiling test and the new low ceiling tests.

By the way, could you credit my site/me, for my thoughts if anything comes of this?


7:23 AM  
Anonymous Anonymous said...

Of course.

8:45 AM  
Blogger Valentine Cawley said...

Thanks for your decency.

10:29 AM  
Anonymous Anonymous said...

Interestingly, Flynn himself says the Flyn Effect has plateaued out in many first-world countries.

Also on the ceiling effect, most psychologists would recommewnd taking additional tests if you hit the ceiling in two or more sub-tests.

Finally you might want to look at the following article which compares three tests including the SB5 (


PS As a resgular reader of your blog I can't find anywhere else to gently request you reconsider your choice of backgrounds - the dots are, for me at least, headache-inducing.

5:45 PM  
Blogger Valentine Cawley said...

Thanks for your comment.

Yes, the matter of the Flynn plateauing...I was going to address that sooner or later: for that has interesting implications, too.

From my reading on the situation I rather formed the opinion that the ceiling effect is more serious than is given credit. For instance, if you hit the ceiling on even ONE subtest - by which I mean approach the ceiling - you don't need to bump it to be affected by its presence, the test will be making an underestimate of unknown size.

All the modern tests have relatively low ceilings - so what, I wonder, are they going to retest on? An old test? Perhaps.

I am a bit wary of changing anything on the blog lest it alter the formatting, but I will look at it. I don't suddenly want everything to be messed up.

Best wishes to you.

7:26 PM  
Blogger EbTech said...

With regards to retesting, I think the idea is to take a more difficult IQ test, which would be centered at an IQ of say 140 instead of 100. However, such tests are very hard to find. The statistical rarity of profoundly gifted individuals makes it especially difficult to give them a valid score.

2:22 AM  
Blogger Valentine Cawley said...

For practical reasons, there is unlikely ever to be a satisfactory IQ test for the profoundly - or even the exceptionally gifted. They are too rare to gather a decent comparison sample together.

We will have to settle for "Oh he/she bumps against the ceiling...I wonder how much they would bump through it, given the chance?"

10:27 AM  

Post a Comment

<< Home

Page copy protected against web site content infringement by Copyscape