Cattel-Horn-Carroll (CHC) Theory and the KABC-II
The Cattell-Horn-Carroll (CHC) model is a psychometric theory that rests on a large body of research accumulated over decades in literally thousands of empirical investigations, especially those that used the statistical technique of factor analysis. In this regard, CHC theory represents a data-driven theory, whereas Luria's theory was driven by the developer's - or Luria's - own clinical and neuropsychological research. That is not to imply that Luria's theory has not stood the test of empirical validation, because a wide variety of psychometric studies, including factor analysis, support its validity. But the origins of CHC theory are decidedly psychometric, in contrast to the distinctly clinical origins of Luria's model.
The origins of CHC theory trace back to two separate psychometric lineages that were merged into a single model in the late 1990s: Raymond Cattell's (1941) original two-pronged Gf-Gc theory, which was expanded and refined by Horn (1965, 1989) to include an array of abilities beyond Gf and Gc; and John Carroll's (1941, 1943, 1993) half-century of rigorous effort to satisfy "the fields need for a thoroughing survey and critique of the voluminous results in the factor-analytic literature on cognitive abilities" (Carroll, 1993, p. vii)
From the very beginning, Cattell was developing a model of intelligence and Horn was continually working to expand and refine the theory into an evolving model of cognitive abilities. Carroll, in contrast, was conducting an extensive factor-analytic survey, which he correctly perceived to be of great value to the field of cognitive abilities. But he was seemingly hesitant about viewing himself as a theorist, not presenting his theory until Chapter 16 of A survey of Factor -Analytic Studies (Carroll, 1993) and being careful to include the word survey and not theory in the book's subtitle. Nonetheless, both the Cattell-Horn and Carroll models essentially started from the same point-Spearman's (1904) g-factor theory-and ended up with remarkably consistent conclusions about the spectrum of broad cognitive abilities. Ultimately, Horn and Carroll agreed to merge their separate, but overlapping, models into a unified theory called Cattell-Horn-Carroll (CHC) theory. This merger by the two theorists was done without fanfare in a personal communication to Richard Woodcock in July 1999, but the scope, impact, applications, and key details of CHC theory have been articulated by Dawn Flanagan and her colleagues (Flanagan et al., 2000; Flanagan & Ortiz, 2001; McGraw, Woodcock, & Ford, 2002).
Spearman's (1904) ground-breaking theory focused on g-a large general factor of intelligence that accounted for test performance on nearly all mental tasks-but also included smaller, specific factors, labeled s (Spearman, 1904). Within a few years of Spearman's influential work, Cyril Burt, L.L. Thurstone, and other psychologists favored theories that hypothesized group factors in addition to g and s (Jensen, 1998). In contrast to "group ability" theorists, Cattell, who was Spearman's doctoral student, built upon his mentor's approach to intelligence. Cattell's (1941, 1957, 1963) system revolved around the concept of g, as Cattell posited two types of g abilities, not just one:
1. Fluid intelligence (Gf), the ability to solve novel problems by using reasoning, believed by Cattell to be largely a function of biological and neurological factors as well as vulnerable to the effects of aging
2. Crystallized intelligence (Gc), a knowledge-based ability that was highly dependent on education and acculturation and was resistant to the impact of aging
As Cattell's doctoral student, Horn became intrigued with Gf-Gc theory and collaborated with Cattell on a series of studies to enrich and validate the two aspects of g (Cattell & Horn, 1978; Horn & Cattell, 1966, 1967). Horn's contributions, along with Matarazzo's (1972) application of the theoretical model to Wechsler's Verbal-Performance IQ dichotomy (believed by Matarazzo to measure Gc and Gf, respectively), helped bring the dichotomous model to prominence. But almost from the beginning, Horn believed that the psychometric data, were suggesting more than just these two general abilities. He quickly identified four additional abilities (Horn, 1965, 1968): Short-Term Acquisition and Retrieval (Gsm), Long-Term Storage and Retrieval (Glr), Visual Processing (Gv), and Speed of Processing (Gs). He subsequently refined the definition and measurement of these factors and added additional factors, such that by the late 1980s to mid-1990s his model included 9 to 10 broad abilities (Horn, 1985, 1986, 1991; Horn & Hofer, 1992; Horn & Noll, 1997). The initial dichotomy had grown, but not in a hierarchy. Although the theory continued to be called Gf-Gc theory, the 9 or 10 broad abilities were treated as equals, not as part of any hierarchical structure.
Carroll (1993, 1997), on the other hand, conducted an in-depth, comprehensive survey of factor-analytic studies and developed a hierarchical theory composed of three levels, or strata, of abilities:
• Stratum III (General), a Spearman-like g, which Carroll (1993, 1997) considered to be a valid construct based on overwhelming evidence from factor analysis
• Stratum II (Broad), composed of 8 broad abilities that correspond reasonably closely to Horn's (1989) broad abilities and "show rough correspondences to Gardner's (1993a) seven "intelligences" Carroll, 1997, p. 127)
• Stratum I (Narrow), composed of about 70 fairly specific abilities, organized by the broad ability with which each is most closely associated, many of which indicate the person's "level of mastery, along a difficulty scale," "speed with which the individual performs tasks," or "rate of learning in learning and memory tasks" (Carroll, 1997, p. 124).
Carroll borrowed the term stratum from Cattell (1971), and used it, as Cattell had done to define levels of factor analysis. Horn's Gf-Gc theory always focused, but Horn discussed the more specific or narrow abilities as well. To Horn, the g construct (Stratum III of Carroll's model) had no place in Gf-Gc theory. Otherwise, the Carroll and Cattell-Horn theories were similar enough to warrant their merger into the new CHC theory. The precise differences between the theories have been spelled out elsewhere (Flanagan et al., 2000; Flanagan & Ortiz, 2001; McGraw et al., 2002) but are not important for understanding the development of the KABC-II or its CHC theoretical foundation. What is important is that, "Never before has a psychometric-ability model been so firmly grounded in data" (Daniel, 1997, p. 1043). That statement was directed at Carroll's theory, but it is equally applicable to CHC theory.
Figure 2.2 depicts the CHC model as it applies to the KABC-II. Carroll's Stratum III is shown in this figure because the KABC-II provides a measure of g based on the CHC model- the FCI. Not all Stratum II abilities of the CHC model are measured b the KABC-II. Separate measures of Gq or Grw are not included on the KABC-II because the authors believe that reading, writing, and mathematics belong on tests of academic achievement rather than cognitive ability tests; these CHC abilities are measured by the KTEA-II, which was conformed with the KABC-II. (The Gq narrow ability of Mathematical Achievement is measured as a secondary ability by two KABC-II subtests, Rover and Block Counting, both of which require the child to count.) The KABC-II also does not include separate scales of Auditory Processing (Ga), Processing Speed (Gs), or Decision Speed/Reaction Time (Gt) because they lacked the requisite complexity for inclusion. Both of the "speed" abilities. Gs and Gt, emerge consistently as weak measures of g in Carroll's (1993) factor-analytic survey. The KTEA-II Comprehensive Form includes a measure of Ga (Phonological Awareness), and Ga is required for success on several KABC-II subtests (e.g., Riddles, Word Order). Gs is measured to some extent at ages 7 to 18 by the KABC-II subtests that include time points (Story Completion, Triangles, and Pattern Reasoning).
As shown in Figure 2.2, Stratum 1 includes a wide array of narrow abilities, and the ones shown are merely illustrative of the approximately 70 narrow abilities that constitute the CHC theory. Excluded from this figure are all of the narrow abilities for Grw, Ga, Gs, and Gt, and some of the ones for Gc, Gv, and Glr. Narrow abilities measured by one or more of the KABC-II subtests are indicated in Figure 2.2 by blue shading. In all, 15 narrow abilities are represented in the KABC-II, including 5 associated with Gv.
For the KABC-II, the broad abilities are of primary importance for interpreting the child's cognitive profile. In developing the KABC-II, the authors did not strive to develop "pure" tasks for measuring the five CHC broad abilities. In theory, Gv tasks should exclude Gf or Gs, for example, and tests of other broad abilities, like Gc or Glr, should measure only that ability and non other. In practice, however, the goal of comprehensive tests of cognitive ability like the KABC-II is to measure problem solving in different contexts and under different conditions, with complexity being necessary to assess high-level functioning. Toward that clinical goal, the authors strove to construct measures that featured a particular ability while incorporating aspects of other abilities. For example, Rover is primarily a measure of Gv because of the visualization that is required, but it also involved Gf; Story Completion emphasizes fluid reasoning to solve the problems, but because they concern social situations, Gc is required to some extent; and so forth.
Kaufman, A., & Kaufman, N. (2004). Manual of the Kaufman Assessment Battery
for Children, 2nd Edition. Pearson Assessment.