(from ``East Coast Noesis'' #139)
Factual Error #5: Kevin reiterates the standard factor-analytic line on g (general intelligence) that has been around in roughly the same form since the time of Moses.
Correction: Unfortunately, factor analysis of IQ test statistics is not the last word on intelligence. The relative homogeneity of a limited class of psychometric instruments (IQ tests) could too easily preclude fine-scale statistical distinctions among possible factor structures, and some of these factor structures could imply that different aspects of intelligence are measured by other kinds of test . . . e.g., Mega-style tests, which differ from ordinary IQ tests in item complexity. In effect, Kevin is saying that we can inject what amounts to a whole new dimension of complexity into test items, and yet rest assured that these items are measuring the exact same kind of intelligence that low-complexity items measure! That's quite a leap.
While certain preliminary analyses of Mega-style test data may indicate the presence of a dominant general factor, this obviously need not be the same general factor evoked by ordinary IQ tests. For example, ordinary IQ tests emphasize mental quickness or facility; Mega-style tests de-emphasize quickness and emphasize an ability to handle logical complexity. Computationally speaking, these are distinct intellectual ``metafactors'' involving computational time and space respectively.
Regarding the former, Kevin seems to think that time restrictions could be removed from all IQ tests without significantly changing the statistics. Unfortunately, it has yet to be proven that there is any such thing as an untimed IQ test . . . i.e., that the time parameter is irrelevant to IQ (actually, the time parameter is relevant by fiat, a fact known to anyone who has taken a standard IQ test and left blocks of items untouched because he's been ``too thorough'' regarding other items). This implies that Kevin, in bold defiance of psychometric protocol, recognizes only the space meta-factor of IQ.
Kevin is simply going to have to get into the underlying factor structure of intelligence to get at this issue, and that's all there is to it. Until he does, he has no business equating z-scores on his tests to ``IQ scores.'' That's why we've reformulated our admissions requirements to include both IQ and Mega-style IEQ (Intellectual Efficiency Quantitation), and why the Paul Maxims and Rick Rosners of our Society--and yes, even the Kevin Langdons and Chris Coles--must learn to live in peace and harmony with each other.
If Kevin wants to argue with this careful piece of scientific and
administrative reasoning, then he'd better start doing his own homework for a change.
Because otherwise, it will remain the sad duty of someone--who will of course be labeled a
schizoid, self-deifying megalomaniac--to kick his sorry, self-important rear all over
these pages yet again.