CHC Theory, Cognitive Assessment, Principles of assessment of aptitude and achievement

Fluid Intelligence, Defined

Mentioning fluid intelligence at cocktail parties as if it were a perfectly ordinary topic of conversation carries with it a certain kind of cachet that is hard to describe unless you have experienced it for yourself. Part of Gf’s mystique can be attributed to Cattell’s (1987) assertions that Gf is linked to rather grand concepts such as innate ability, genetic potential, biological intelligence, mass action, and the overall integrity of the whole brain.[1] Heady stuff indeed!

At the measurement level, Gf tests require reasoning with abstract symbols such as figures and numbers.[2] Good measures of Gf are novel problems that require mental effort and controlled attention to solve. If a child can solve the problem without much thought, the child is probably making use of prior experience. Thus, even though a test is considered a measure of fluid intelligence, it does not measure fluid intelligence to the same degree for all children. Some children have been exposed to matrix tasks and number series in school or in games. Fluid intelligence is about novel problem solving and, as Kaufman (1994, p. 31) noted, wryly pointing out the obvious, a test is only novel once. The second time a child takes the same fluid intelligence test, performance typically improves (by about 5 points or 1/3 standard deviations, Kaufman & Lichtenburger, 2006). This is why reports that fluid intelligence can be improved with training (Jaeggi, Buschkuehl, Jonides, & Perrig, 2008) cannot be taken at face value.[3] Just because performance has improved on “Gf tests” because of training does not mean that Gf is the ability that has improved.

At the core of Gf is the narrow ability of Induction. Inductive reasoning is the ability to figure out an abstract rule from a limited set of data. In a sense, inductive reasoning represents a person’s capacity to acquire new knowledge without explicit instruction. Inductive reasoning allows a person to profit from experience. That is, information and experiences are abstracted so that they can be generalized to similar situations. Deductive reasoning is the ability to apply a rule in a logically valid manner to generate a novel solution. In CHC Theory, deductive reasoning is called General Sequential Reasoning. Although logicians have exquisitely nuanced vocabularies for talking about the various sub-categories of inductive and deductive reasoning, it will suffice to say that everyday problem solving typically requires a complex mix of the two.

Inductive and deductive reasoning can be found in multiple places in CHC Theory. Whenever inductive and deductive reasoning are applied to quantitative content, they are called quantitative reasoning. For mysterious reasons, inductive and deductive reasoning with quantitative stimuli tend to cluster together in factor analyses. Inductive and deductive reasoning also make an appearance in Gc. Whenever inductive and deductive reasoning tasks rely primarily on past experience and previous knowledge, they are classified as measures of crystallized intelligence. Many researchers have supposed that the Similarities subtest on Wechsler tests contains an element of fluid reasoning because inductive reasoning is used to figure out how two things or concepts are alike. If the question is something like, “How are a dog and a cat alike?” then it is very unlikely that a child arrives at the correct answer by reasoning things out for the first time. Instead, the child makes an association immediately based on prior knowledge.

Researchers are not satisfied with accepting Gf as a given. They wish to know the origins of Gf and to understand why some people are so much more adept at abstract reasoning than other people are (Conway, Cowan, Bunting, Therriault, & Minkoff, 2002). One hypothesis that is still being explored is that fluid reasoning has a special relationship with working memory. Working memory is the ability to hold information in mind while using controlled attention to transform it in some way (e.g., rearranging the order of things or applying a computational algorithm). Many researchers have noted that tests of fluid reasoning, particularly matrix tasks (e.g., WISC-IV Matrix Reasoning), can be made more difficult by increasing the working memory load required to solve the problem. Kyllonen and Christal (1990) published the provocative finding that individual differences in Gf could be explained entirely by individual differences in working memory. Many studies have attempted to replicate these finding but have failed. Most studies find that Gf and working memory are strongly correlated (about 0.6) but are far from identical (Kane, Hambrick, Tuholski, Wilhelm, Payne, & Engle, 2004).

Just as we have distinguished between statistical g and theoretical g, it is important to note that there is a difference between the Gf that is measured by Gf tests and the Gf that is written about by theorists. Some of Cattell’s hypotheses about Gf have stood the test of time, whereas others have not held up very well. For example, the heritability of Gf is not higher than that of Gc, as Cattell’s theory predicts. I mention this because it is probably not justified to claim that because a child scores well on Gf tests, the child has high innate talent or that the child’s biological intelligence is high.

Most of the effects of Gf on academic achievement are mediated by Gc (i.e., better reasoning leads to more knowledge which leads to higher achievement). However, Gf seems to have a special relationship with complex problem solving in mathematics. Because Gf tests measure abstract reasoning, it is unsurprising that they would predict performance in an abstract domain such as mathematics (Floyd, Evans, & McGrew, 2003).


[1] Horn (1985) tended to de-emphasize the biological/genetic interpretation of fluid intelligence.

[2] Test developers have tried to create Gf measures with verbal content (e.g., WJ-R Verbal Analogies or SB5 Verbal Fluid Reasoning) but find that verbal Gf tests do not always load on the same factor as traditional Gf tests (Canivez, 2008; Woodcock, 1990). It is possible that the KAIT Logical Steps subtest may be the only commercially available verbal Gf test that does not have substantial loadings on Gc (Flanagan & McGrew, 1998; Immekus & Miller, 2010), possibly because it does not use the verbal analogy format.

[3] See Moody (2009) for a discussion of other methodological problems that may have compromised the validity of the Jaeggi et al (2008) study.

This post is an excerpt from:

Schneider, W. J. (2013). Principles of assessment of aptitude and achievement. In D. Saklofske, C. Reynolds, & V. Schwean (Eds.), Oxford handbook of psychological assessment of children and adolescents (pp. 286–330). New York: Oxford.

Advertisements
Standard

3 thoughts on “Fluid Intelligence, Defined

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s