The Mega Test is an untimed, unsupervised, "take-home" IQ "power test". As Chris Langan, the Chairman of the Mega Foundation, puts it, test-takers "lay protracted siege to 48 difficult problems". It isn't accepted by psychometrists as a valid IQ test, for two reasons. First, among psychometrists, problem-solving speed is considered to play a pivotal role in intelligence and in IQ measurements, and is the reason that conventional IQ tests have tight time limits. Since no time constraints are applied to solving problems on the Mega Test, psychometrists consider its "IQ" measurements to be denatured, since they depend upon persistence as well as speed. And second, the fact that it is unsupervised opens the garden gate to collusion and cheating. Proponents of the Mega Test argue that cheating on the test's tougher problems isn't practical because individuals who are smart enough to help are too few and far between. They also argue that the problems on the test are more akin to the kinds of problems that one may encounter in a real-world setting than those on an IQ test. They have argued that its poor correlations with conventional IQ tests are the result of its much-higher ceiling than most conventional IQ tests.
    I would look for signs of cheating in the form of "outlier" scores below the thresholds of the problem-solving curves given in this document as an indication that a few low-scoring test-takers had had help. I would expect that the easier verbal analogies would lend themselves to cheating more readily than difficult non-verbal problems. However, the data doesn't mirror this.
    Chris Langan has suggested that the term Intellectual IEQ (Intellectual Efficacy Quantitation) be used to describe power test "IQ's", as well as conventional IQ scores. Second, the Mega Test results don't correlate well with other aptitude tests such as the SAT, the GRE, and the WAIS-R Wechsler Adult Intellgence Test – Revised. (See references 1 and 2 for correlation tables.)