Cognitive Abilities and Creative Behaviors: CAB‐5 and Consequences

The validity of six indices of divergent production is examined with reference to creative output in eight content domains: visual arts, music, literature, theater, science and engineering, business ventures, apparel design, and video and photographic work. Undergraduates ( n = 144) completed Consequences, four scales from the Comprehensive Ability Battery, and a specially developed self‐report inventory that produced eight creative content scores and a total score. Correlational analysis determined that Semantic Fluency, Ideational Fluency, Originality and Remote Consequences were substantially correlated with most creative behaviors and were uncorrelated with grade point average. A factor analysis of criteria revealed two underlying patterns of creativity, suggesting that proficiency in one creative domain is predictive of proficiency in several others. The study of creativity has traditionally been propelled by aesthetic, scientific, and economic concerns. While revolutionary breakthroughs in science, technology, and design often command our attention and respect, it is the small, evolutionary innovations that typically have the greatest immediate economic impact (Haustein, 1981). From any vantage point there has always been an interest in the prediction and measurement of creative behavior and aptitude. The available measurement approaches for creativity include cognitive abilities, personality traits, self‐reports of creative behavior and peer and teacher ratings of creative works (Hocevar, 1981). The purpose of this study is to assess the validity of an underresearched set of cognitive measures: four subtests of the Comprehensive Ability Battery (CAB‐5; Hakstian & Cattell, 1976) and the Consequences test (Christensen, Merrifield, & Guilford, 1958). Our validity strategy was correlational; criteria were eight self‐report indices of creative endeavor. The following paragraphs elaborate the underlying theory from which these measures originated and some unresolved issues. Perhaps the largest single breakthrough in the cognitive approach to understanding creativity came from Guilford's (1967) 120‐factor structure of intellect model. While commonly used intelligence tests measure convergent forms of thinking, creativity involves divergent thought processes which, according to Guilford, account for 30 of the 120 factors of intelligence. Working through a factor analytic paradigm, Guilford was able to identify over 100 out of 120 factors. Guilford's structure of intellect may be criticized for having fractured intellectual functioning into too narrowly defined units relative to what human faculties are actually engaged in ordinary intellectual endeavors. Nonetheless, the theory did provide a provocative explanation for the nature of creative thought and some interesting measurements of same. Measurements of creative thought (divergent production) involve word fluency, category formation and reformation, ideational fluency original uses for common objects (opposite of functional fixedness). One test, Consequences (Christensen et al., 1958), is of particular concern to this research project and involves a certain amount of social awareness on the part of the examinee in conjunction with divergent production. In a typical test item, examinees are presented with a hypothetical situation, “What would be the consequences if people no longer needed to sleep?” along with a few common responses. Examinees are then given five minutes to write as many consequences of the hypothetical situation as they can think of. Responses are scored by, first, eliminating responses that are redundant with the samples or other responses or that are completely irrelevant Remaining responses are then categorized as obvious and remote. The numbers of obvious and remote responses are counted to produce two scores. Consequences continues to be listed as a research instrument by its publisher. Norms in the manual (Guilford & Guilford, 1980) are only available for 331 engineers who took two forms of the test, 665 ninth‐graders, 80 twelfth graders, and a college sample of 87 cases. According to Guilford & Guilford, the consequences scores are closely related to other divergent production measures, remote more so than obvious. The CAB‐5 developed by Hakstian & Cattell (1976, 1978) who also worked in the factor analytic mode, consists of 20 subscales that appear to capture the essentials of convergent and divergent thinking in a timed battery of manageable size. The four divergent thinking scales — semantic fluency, ideational fluency, word fluency, and originality — are of particular interest to this project While the usual types of validity and norm‐building research have been done with the convergent scales (Hakstian & Cattell, 1976; Hakstian & Woolsey, 1985) little external validation work is available for the divergent scales. Hakstian & Woolsey (1985) found that semantic fluency and ideational fluency were significantly correlated with grades in an introductory psychology course (r = .33 and .30 respectively). The rationale for why divergent production would be related to grades in that course was vague. Norms for the divergent scales are currently available for 216 college and 1098 high school students. While the validation question is framed around a limited set of scales in this research project, it does represent an issue of more general concern. According to Storfer (1990), several researchers have questioned whether measures of divergent production have any external validity with actual creative works. He cited a study where teachers' ratings of their students' creativity were based on logical or convergent measures but had no relationship to divergent measures. In his review of creativity measures, Hocevar (1978) noted a generally low level of interrelationship among personality, cognitive, self‐report, and rating measures of creativity; from this trend he opined that self‐report measures of creative behavior were perhaps the most defensible indices of creativity. On the other hand, a study of approximately 500 college students showed that those that had won science prizes, or who published or exhibited their work scored higher than others on ideational fluency (Wallach & Wing, 1969). While the latter finding is reassuring, we are not convinced that prizes, publication, or exhibition represent the full range of creative behavior to be found at the college level. Prizes require a propensity to compete among the participants, and a proclivity toward subjectivity, among the judges. As academic researcher know, earth‐shattering ideas are not quickly snapped up by scientific journals, and much of what is published is indicative of the evolutionary progress mentioned at the outset of this article. There is reason to assume that there is much valuable creative effort taking place that does not reach wide audiences or that is otherwise shoved into drawers. From a statistical point of view, it is preferable to conduct a criterion related validity study that encompasses a wide (if not complete) range of criterion values and to avoid an extreme groups methodology, such as a comparison of prizewinners and others. The latter has the impact of exaggerating the predictor‐criterion effect size. Altogether, Hocevar (1981) identified 15 studies showing a positive relationship between divergent production and other measures of creativity, and 15 other studies that showed no such relationship. With the foregoing theoretical and methological points in mind, the following hypotheses were examined: 1. Consequences and CAB‐5 divergent scale scores would be significantly related to several types of creative behavior. 2. Consequences and CAB‐5 divergent scale scores would be significantly related to each other. 3. There would be a substantial interrelationship among the various types of creative behavior. Significant findings would dispel the notion among critics of divergent thinking theory (cf. Stolper, 1990) that the prediction of creativity is situationally specific. 4. Divergent measures and indices of creative behavior would not be related to college students' grade point average (GPA). While we recognize that no point is proven by failing to reject the null hypothesis, a null relationship with GPA is expected from the general theory, on the basis of findings concerning teachers' rates cited in Stolper (1990) and our own assumptions that much creative work goes unnoticed in favor of traditional academic demand. An expected null relationship thus adds an element of convergent‐discriminant validity analysis to the research plan.