Statistics of the top scorers

© Nov. 2010 Paul Cooijmans

Introduction

This report contains some information on the very highest test scores having occurred from 1995 to present. The cutoff for inclusion is I.Q. 170, which corresponds to protonorm 620 according to the current norms. The I.Q. range 170-179 is tentatively called "pervasive intelligence" (a few scores are actually in the 180+ range, called "exceptional intelligence", but it is not a priori so that higher scores have extra significance in this report). Considered are only scores on tests containing at least two different item types, when items are classified by their appearance as either verbal, numerical, spatial, or logical. Such heterogeneous tests have proven to be more robust, valid, and reliable than are homogeneous tests (with only item type). Tests of which the norms are still of doubtful quality, and/or that may have too low ceilings, have not been considered.

The scores

There are 16 scores in this range, following above criteria. The scores range from I.Q. 170 to 182. Their exact distribution, the norming of tests in this range, and the question whether higher scores within this range also mean greater ability, are not topics of this report. Those matters are dealt with, when possible, in the statistical reports for the tests in question, and in the report on the norming of protonorms to norms. When norms change, the number of scores that fall at or above I.Q. 170 may naturally change. Considering the rareness of these scores, it should best be assumed that the current norms within this range are not good enough to distinguish well between the corresponding performances, and that this will improve with future renormings.

The incidence of these scores among all of the obtained scores since 1995 is about 1 in 222; if homogeneous one-sided tests were included, it would rise to about 1 in 67.

Number of scores per candidate

The 16 scores have been obtained by 11 candidates:

Sex

All of the scores have been obtained by males.

Scores per country

Country# scores
United States*****
Belgium****
Finland***
Germany**
The Netherlands*
United Kingdom*

Scores by age when test was taken

Age class# scores
45-49**
40-44*
35-39*****
30-34**
25-29***
20-24***

Scores by birth year

Birth year# scores
1960-64**
1965-69******
1970-74****
1975-79***
1980-84 
1985-89*

Scores per test

Test# scores
Cooijmans Intelligence Test - Form 1****
The Nemesis Test***
Test For Genius - Revision 2004***
Long Test For Genius**
Short Test For Genius*
Cooijmans Intelligence Test - Form 2*
Reason Behind Multiple-Choice - Revision 2008*
Grail Test*

Scores per year occurred

Year# scores
2010*
2009**
2008*
2007***
2006 
2005 
2004**
2003*
2002*
2001*
2000*
1999**
1998*
1997 
1996 
1995 

Scores by educational level (insofar known)

Level# scores
Secondary school completed**
Bachelor**
Master*****
Doctor**

Disorders

Four scores are from candidates who report having a psychiatric or neurological disorder (it appears to concern disorders of the neurotic kind, and in no case of the psychotic kind), and one score is from a candidate reporting the presence of such disorder among parents and/or siblings.

Three scores are from candidates who have been observed displaying neurotic behaviour; no candidates have been observed displaying psychotic behaviour (which otherwise is not uncommon among test candidates).

The majority of scores therefore is from candidates with no disorders in themselves or relatives.

Scores by Order of Thoth grade

This grade indicates the number of tests one has taken to date (in parenthesis).

Grade# scores
Neophytus (1 to 4 tests)***
Candidatus Minor (5-9)********
Candidatus (10-14)****
Candidatus Maior (15-19)*

Personality scores and observed behaviour

Insofar candidates have taken personality tests like Personality Scales for Intelligent Adults and Gifted Adults' Inventory of Aspergerisms, their scores are on the whole remarkably normal, around average, the only thing that stands out being that the majority of them are (compared to high-range test candidates) below average on indicators of deviance. A few are above average on those though.

Regarding observed behaviour, remarkable is the absence of negativity and rudeness, and the on the whole positive, polite, constructive attitude.

Statistics of the top scorers - earlier report

© February 2006 Paul Cooijmans

The scores

After 11 years and close to 3000 submissions to my tests by close to 1500 people the time has come to take a look at those who scored highest. This report considers only the top scorers on tests with a mixture of item types, as I currently believe those give the best measure of intelligence. One-sided tests are less satisfactory for various reasons that are not the subject of this report.

The cut-off for inclusion in this study is I.Q. 170 (SD=15). That is about the point where scores get really rare, as can be seen in this distribution. The only mixed-item tests on which scores of I.Q. 170 and higher have been reached are the Short Test For Genius, Long Test For Genius, Test For Genius Revision 2004, Cooijmans Intelligence Test Form 1 and Cooijmans Intelligence Test Form 2. The TFG versions provide 5 scores, the CIT versions 6. The 11 I.Q.s are below, represented by the year in which they were achieved:

(That the year 2000 is missing may be because I stopped scoring tests for a while then, so fewer submissions came in.)

These scores belong to 9 people (males), two of whom having two scores in this range. There exist no scores higher than 178, except for on one-sided tests where 181 is the top score; It is like in athletics, where a specialist on the 100 meter runs faster than a decathlon athlete. It is NOT so that the one-sided tests are normed too generously; It is a natural phenomenon that higher scores are possible in a more narrow range of abilities, but with lower validity with regard to intelligence.

Countries

Here are the testees' countries of origin, ranked first by the number of scores from that country and then by the average of those scores:

Ages

The ages at the time of the scores:

Remarks

Most of these scores date back to before I was collecting as much personal information as I do now, so there is limited data otherwise. Regarding the important matter of the correlation between I.Q. and psychiatric disorders and disposition therefor (which has been found by me to be negative over the high range), only two of these testees have submitted the P.S.I.A. (personality test), and their profiles are very similar and largely average, except that both are a full standard deviation below the mean in Deviance, the factor that reflects disposition for psychiatric disorders. So that at least is in line with what I found before; A negative correlation between I.Q. and deviance implies indeed that the very highest scorers will be low on deviance.

These two also reported their educational level, Ph.D. in both cases. This is the highest educational level, and educational level, according to my own statistics to be published soon, correlates negatively with deviance as well.

Two others reported their score on the GAIA questionnaire which too measures disposition for psychiatric disorders, and the average of their scores is 18, which is somewhat below the mean.

In their communication with me all these people have been civilized and polite, without weirdness or negativity. All in all the top scorers seem to confirm what has slowly been dawning to me over the past years: that, within the high range of intelligence, higher I.Q.s tend to come with a greater likelihood of being psychiatrically and socially normal.

Of course psychiatric and social problems do occur in highly intelligent persons, but it appears that 1) those problems are not caused by being more intelligent than others but by being deviant and/or having disorders, and 2) those problems become rarer at higher I.Q. levels.

Also, the sense of isolation some high-I.Q. individuals experience, and the idea "If two people differ more than 15 points in I.Q. they cannot communicate", may really be rooted in deviance rather than intelligence.

A related matter is that of Asperger Syndrome; There exists a more or less popular viewpoint that a diagnosis of Asperger is "turning high I.Q. into a disorder", and that the symptoms of Asperger are really the normal features of highly intelligent children and adults. I am quite confident that this viewpoint is wrong, that Asperger is a real disorder, and that many high-I.Q. children and adults do not display significant Aspergoid symptoms at all.

In the 20th century it has been suggested that high intelligence, not just "giftedness" (98th centile) but a high subset thereof, tends to go with problems in social adjustment, isolation and so on. Researcher Leta Hollingworth has contributed to that viewpoint, and in the world of I.Q. societies it has been propagated by Grady Towers. It is a sympathetic viewpoint and I have believed in it myself in the past, but in the light of the data from my tests I have come to see it as partly wrong.

I now think the group meant by Hollingworth and Towers is really the "intermediate" group, say roughly the 140-160 I.Q. range, and even in that group I think the problems are not related to intelligence but to deviance and disorder. The cause of this partly wrong idea is probably that in Hollingworth's days there were no tests that could discriminate in the high range of adult intelligence (as my tests do), so there was no way of knowing where a group of subjects was located within the high range. It was all based on Stanford-Binet childhood I.Q.s back then, and those do not reveal one's true position within the high range. You can pave the roads with high childhood I.Q.s, but not with high adult I.Q.s.