General intelligence

From EvoWiki

General Intelligence (g, g factor or general cognitive ability) is the highest order hierarchical factor that can be extracted from a diverse body of cognitive tests using factor analysis.

Using the factor analytic model he himself invented, psychologist Charles Spearman applied its methods of tests of mental abilities which had long been known to be positively correlated, i.e., mental test scores are, to some degree or another, predictive of other mental test scores, either on different test batteries or test subsets. What the factor analysis of these test scores showed was that they all had something in common, general intelligence, and that it could explain a large portion of their total variance (typically 70 or 75%).

Challenges

Critics, such as Stephen Jay Gould, have attempted to argue that the extraction of a general factor is equivalent to other methods (Gould mentions the Thurstonian group factors), and consequently g is little more than statistical voodoo. While no knowledgeable factor analyst would deny the mathematical equivalence of competing methods, other considerations strongly support g:

a) g best represents the fact that differences between individuals are typically greater than differences between abilities of individuals. Various orthogonal factors, like Thurstone's do not suggest this.

b) some tests are more g-loaded than others; that is, draw more heavily on g.

c) g is not tied to any particular type of item content or acquired ability. It emerges from all mental tests, even if they ostensibly appear different.

d) no test has been constructed that measures alternative orthogonal factors but not also a g factor; that is, tests designed to measure the "pure" forms of these primary multiple factors also measure g, to some degree or another. However, there are pure measures of g which do not measure other factors.

f) more than any other factor, g has the most "real world" predictive validity. This includes attained socio-economic status, education, and a host of other variables, both in Western societies, and abroad.

g) g has a number of biological correlates. For example, brain size is intrinsically correlated with IQ to about r = +.4 (Andreasen et al. 1993, Egan et al. 1994, Raz Torres et al. 1993, Wickett at al. 1994, Willerman et al. 1991, Wickett et al. 1996, Schoenemann 1997), with the amount of glucose metabolized during the thirty-five-minute testing period of the RAPM to between r = -.7 and -.8 (Haier 1993), and brain nerve conduction velocity to r = +.37 (Reed & Jensen 1991).

h) g is highly heritable (between .5 and .75), and much more so than any other factor.

Consequently, Gould's doctrinaire disparagements must be seen as simply that.

The Genetics of G

In recent years new studies from quantitative geneticists have helped to identify segments of DNA contributing to high versus low IQs. These include EST00083 (a mtDNA marker, Plomin et al. 1994), CTG B33 (a transnucleotide repeat, Plomin et al. 1994), the HLA A (B) allele (Plomin et al. 1994, Jacomb et al. 1999, Turic et al. 2001), IGF2R gene (Chorney et al. 1998, Plomin 1999), and exon 2 of CTSD gene (Payton et al. 2003)