Reply to Glen Wooten

Kevin Langdon


Glen Wooten has brought up a number of questions concerning high-range testing and the organizational framework of the Mega Society. As these matters are of concern to all Mega members I am replying to Mr. Wooten's letter in this issue.

Dear Dr. Hoeflin,
I just read Noesis 155 and wanted to interpose some of my thoughts. In Chris Cole's article on high range testing, he stated half the answers to the Titan Test were "easily available" on the Internet. I suspect this is an exaggeration to make his point.

Unfortunately, this is not an exaggeration. The test hackers are a real problem these days.

Anyway, let's not forget, the use of computers is not allowed! Even when the tests were first designed, there was no way to prevent dishonest testees from using an assortment of software programs or their own written programs to solve problems.

My tests never prohibited the use of computers and it seems unwise to me to do so, for exactly the reason you cite; it unnecessarily opens an avenue for cheaters. Also, I strongly suspect that people who use computers to solve the problems on these tests are more likely to make conceptual or computational errors somewhere along the way than those who solve the problems without such aids.

An inspection of all people who have qualified for membership in the Mega Society via your tests reveals a superabundance of individuals with educational backgrounds in computer science, mathematics and physics--the people most likely to have knowledge of cheating methods using computers.

There's an assumption there about the direction of causality. People who are good at abstract thinking are in demand for their skills in these fields. The members of Four Sigma, Prometheus, and Mega who qualified by means of the LAIT, which does not prohibit the use of computers, also include a very high percentage of computer scientists, mathematicians, and physicists.

Furthermore, computers cannot be used to cheat on my tests or other, similar instruments. Programming a computer to understand one of these problems is a much trickier proposition than just solving it oneself.

Not unlike the present, there was also no way to enforce the no-collaboration rule. Again, people with certain educational backgrounds, especially at certain universities, are likely to know talented professors or other students in their respective fields with the ability to assist them.

There is a natural reluctance on the part of highly intelligent people to compromise intelligence tests. Individual assistance is unlikely to be a major factor, but collaboration of groups is more problematical.

We also can't forget several individuals, some current Mega Society members, who submitted an unknown number of answer sheets using pseudonyms. Any excuse for doing so is highly suspect.

The Mega Society has voted, more than once, that members' membership status is not subject to review subsequent to admission.

There were also two people who qualified for the Mega Society who admitted to spending three months on a single test. What do those scores purport to measure?

You don't suppose that these people might have had lives that interfered with sitting down and putting in a hundred hours to complete the Mega Test (not unusual)? What they measure is how many items people can get right. And on the spatial half of the test that turns out to be a respectable measure of g.

The bottom line is that the current Mega Society is trapped in a glass house. Some of the most suspicious-looking people of all are already members.

Some very suspicious-looking people are to be found in all the high-IQ societies, including those which have no problem with the limits of measurement.

Casting stones at the six thousand truly qualified potential members seems silly,

It might, indeed, be silly if anybody were doing it! We'd be delighted to have those six thousand people as members.

especially given the fact the society is considering merging with the psychometrically dubious Pi Society. (No offense intended.)

Pi has only accepted three members on a basis other than the tests Mega accepts, but some people have been admitted through scores submitted after the date that Mega has chosen as the cutoff date for acceptance of these scores. And Mega grandfathered in members of the 606 Society, which made use of highly questionable standards, at the time of its founding. It seems important to try to unify the 99.9999th-percentile societies, if it's possible to do so without compromising our admission standards.

I think a new, more respectable society should be formed which requires supervised test scores in addition to your tests. This standard would be expected of all members. It should not only cut down on the number of cheaters but would add much needed respect and legitimacy to the society.

The supervised tests don't have enough ceiling, so we'd have to settle for scores considerably below our cutoff level, which might, as you suggest, catch a cheater or two but which would also be likely to screen out highly intelligent people who don't react well to the conditions of supervised tests.

For income-challenged persons, the CTMM could be used. It isn't as respectable as the Wechsler, but ceiling scores should work just fine.

What's this preoccupation with "respectability"? What are we trying to prove?

The problem with using standard tests is that they simply don't have enough ceiling. And this is particularly a problem with respect to the CTMM.

If Mr. Wooten wants to found his own society, he is free to do so, but I think it would be counterproductive.

I also think Ferguson's formula should be implemented so that people have the option of qualifying via supervised tests only, providing ceilng scores on at least two respectable tests (using only tests administered in adulthood) are submitted. This would help eliminate complaints similar to those of Paul Maxim.

I don't give a rat's ass about Mr. Maxim's complaints. He's not a member of Mega and he has his own private crusades which frequently involve the affairs of the higher-IQ societies.

The old Mega Society Constitution specifically provided that the Ferguson formula could be used, but it has never actually been used for admitting members. And, of course, care must be taken to apply it in mathematically sound ways.

Chris Langan is the only Mega Society member I'm familiar with who has achieved a ceiling score on a well-respected supervised test in addition to a high score on one of your tests. Again, it adds legitimacy to the claim of an IQ score in the Mega range and I think it's a standard all "true" Mega Society members would want to espouse.

And, again, the inadequate ceilings of the standard tests make them unsuitable for this purpose at the Mega level. Prometheus could use the Ferguson formula with the standard tests, though. And, of course, Chris Langan is not the only Mega member who has made high scores on standard tests.

Although I personally fully support your tests' continued use and believe they are far superior to any other unsupervised tests, complaints have been raised by Grady Towers, Paul Maxim and Marilyn vos Savant, among others. These complaints involve questions about the "true" IQ of members of societies that don't require supervised test scores. Grady Towers went so far as to suggest some Mega members couldn't qualify for Mensa!

I disagree with Grady's conclusion. None of the members I've come into contact with would have the slightest difficulty qualifying for Mensa on any reasonably well-designed IQ test. As for the complaints, they vary in their validity. It is true that there have been irregularities in the admission procedures of a number of high-IQ societies in the past, including Mega. It may be that a few members have cheated without detection, but there's no evidence that anyone has.

Another complaint is that the unsupervised tests don't correlate well with standardized IQ tests.

In fact they correlate very respectably with certain tests, particularly those known to have high g loadings.

An alleged reason is that unsupervised tests don't measure intelligence so much, but rather educational background skills such as mathematics.

This criticism may fairly be brought to bear on Dr. Hoeflin's tests, but mine don't depend on particular backgound skills, beyond a very rudimentary skill set shared by most high school graduates.

The same may be said of the "old" SAT and its somewhat low correlation with intelligence tests.

The SAT correlates with IQ tests as well as they correlate with one another. It was originally developed as an IQ test and is considered to be a cognitive ability test by psychometricians.

However, the method I have outlined should satisfactorily allow testing to continue.

We need some new tests and better approaches to test security, but the approach suggested by Glen Wooten would compromise our admission standards.

If you have any thoughts on these suggestions, please let me know. To my knowledge, few realistic ideas have been suggested regarding admittance criteria and certainly no action has been forthcoming.

What needs to be addressed is how to do it right.

Unfortunately, most members seem quite content with the quietus of the position of membership officer.

The duties carried out by the membership officers of our sister societies are performed by the Administrator.