IQ Test Results..

[Reach]

Originally Posted by ieatyourlvllol

You prove a good point, Reach. In retrospect, I was a bit ambiguous when saying that "IQ pertains to only the abilities covered by the assessment material..." My intended meaning was that the relationships drawn from the results produced are direct in only the sense of the exact sectors and magnitudes of mental capacities touched upon by the test. The problem is that it's nearly impossible to determine that question 'q' covers 'x' amount of process 'a', 'y' amount of process 'b', and so on. Even if such a feat were accomplished, it would be yet more difficult to establish a system by which a universal value (such as 'g') could be obtained. Like you said, the objective measurement of the various functions of intelligence is composed by a network of correlations. My qualm is that until the network of correlations can be replaced by a [theoretically plausible] conjunction of causal interdependences, the concept, not of IQ itself, but of using it as a supposedly inarguable indicator of extrinsic elements (such as success), is complete nonsense. Perhaps in a decade or so the tests will have progressed such that the correlations are strong enough to arrive at reliable conclusions, but until then, my opinion is that IQ tests should be used primarily for fun and comparison.

You bring up some good points.

The 'g loading' of tests can be measured through factor analysis. That is, the ability (or percentage) of the test that measures g directly rather than other factors. However, even in light of this, success (among other various elements of someones life) is not entirely g loaded (it is inevitably affected by many, many things). Thus, as an indicator of such things, an IQ test will *always* be arguable in the sense that it could misrepresent someones actual position. I prefer not to underestimate nor overestimate the power of g, but recognize it as a useful tool. An interesting paper on why g does matter (at least to some degree) can be read here: http://www.udel.edu/educ/gottfredson...hygmatters.pdf


Rather than administer a plethora of various mental tasks like some IQ tests in order to try to measure g, others have simply tried to administer one type of item in progressive order of difficulty that has the highest possible g loading. Tests that do this, such as Ravens APM have been fairly successful measures of g, and also culture fair (though they have been controversial at the same time, probably because of the larger IQ gaps between racial groups and various world populations than even culture biased tests, despite it being a valid measure of success.)


It would be nice to have a test that could perfectly measure intelligence in every case, but even then it would not be a perfect predictor of any extrinsic variables. In order to do that the test would have to account for variables that have nothing to do with intelligence, though the test would certainly be highly accurate, since your innate intelligence is a factor in essentially every task you perform to some degree that would vary depending on the task complexity. I think people want to deny this fact innately, as they want an explanation as to why people are better than they are that isn't innate. Both highly intelligent and highly unintelligent individuals will often reject g despite the evidence of its existence (regardless of whether or not we can measure it properly)...attributing their ability to factors that hardly matter rather than recognize the fact they just got lucky (or unlucky XD ). The heritability factor of g is about 0.85, which means just about all of the heritability of IQ test performance and other various mental tasks is attributable to g.