For over 100 years, psychologists and their predecessors have struggled to uncov er the nature of human cognitive abilities. Is there a solitary facility - general intelligenc e - that underlies and is predominately responsible for an individual's performance on the multitude of human cognitive tasks? Or, at the other extreme, is there a plethora of distinct abili ties that may be deployed individually or in association? Or, is the explanation to be found some where in the middle or in some combination of these two accounts? Historically, the limits of the discussion are best represented by Spearman's (1927) two - factor theory and by Thomson's ( 1939) sampling theory. More recently, exponents of the two boundary positions include Jensen (1986) and Gardner (1983).During this century, with the refinement of factor - analytic techniques, severa l middle - ground theories were generated and have gained favour. These hierarchical theories ackn owledge the existence of numerous, relatively narrow, specific abilities along with broader, higher - order, factors that group the narrow abilities under them. The notion of a still higher - order single factor - general intelligence - has been a part of some of the hierarchical mode ls, but not of others (Royce & Powell, 1983; Vernon, 1961). Carroll's book represents a further maturation of such theories.The factor - analytic approach to cognitive abilities has not gone unchallenged, however. It has been pointed out that there are always an infinite number of mathematical soluti ons and several logical solutions for any given data set. Thus, the validity of any one model co uld not be decided upon by subsequent factor - analytic studies. Nor could ensuing factor - analyti c studies yield evidence for determining which model was a more accurate depiction of the mind's structure. Additionally, some critics have contended that any particular derived factor str ucture may be a mere artifact of the specific analysis and the set of tasks sampled, rather than being a reflection of an actual underlying cognitive structure. Some psychologists have suggested t hat the factor - analytic approach has run its course and can do no more. Due in part to these c riticisms, a parallel approach based on experimental cognitive psychology has gained populari ty during the past two decades. Carroll's book may also be viewed as an attempt to answer thes e criticisms, to incorporate the experimental work, and to demonstrate that the factor - analy tic approach continues to be useful endeavour to understand the nature of human cognitive abi lities.Carroll divides his book into three parts. The first is a historical and concept ual review of psychometric theory. Carroll begins by clearly defining and differentiating key concepts such as ability, aptitude, achievement, latent trait, and intervening variable, altho ugh recognizing that in practice it is often difficult to make distinctions. His treatment o f psychometric theory would provide even a layperson with a clear conceptual introduction as to how abilities can be differentiated and their associations recognized. Although his historical discussion furnishes a chronology of technical progress, given the controversies associated with the testing movement, the discussion appears a trifle sanitized. Carroll treats the methods and theories of authors such as Galton, Burt, Spearman, and Binet as if methods and models could be examined independently of assumptions concerning the origins of individual abilities.The second part, and the bulk of the book, is a detailed analysis and synthesis of a myriad of findings in the area of cognitive abilities. This herculean undertaking examinin g 467 datasets - 485 pages - is the primary contribution of the book. Carroll identifies over 60 specific (first - order) factors. They range in specialty from Piagetian reasoning to absolute pitch, and from lexical knowledge to choice reaction time. …
Read full abstract
Jan 1, 2005
Brian H Bornstein 
Open Access