Ability, demographic and personality predictors of creativity

Abstract

In all 10, 415 adult Finish managers attending an assessment centre completed a battery of tests including a personality trait measure (revised Personality Research Form), three ability tests and a well established measure of divergent thinking (DT) (Consequences Test). Gender, but not age nor education was a significant predictor of DT. Two of the three ability tests were correlated with DT. Various facet scores were correlated with DT but at the domain level it was only the Extraversion factor that proved significant. A regression indicated that bright, Extraverted males did better at the DT test. Implications and limitations are considered.

Introduction

This study involves a psychometric investigation of creativity (Plucker & Renzulli, 1999), usually defined as the production of both novel and useful ideas, concepts or products. Many investigations of the creativity construct have utilised tests of divergent thinking (DT) (Aliotti et al., 1975, Barron and Harrington, 1981, Batey and Furnham, 2006, Chandler and Pengally, 2006, Feist, 1998, Gelade, 1995). It is usually agreed that DT is necessary (but not sufficient) for creativity but that it is only one component, others being originality and flexibility (Batey & Furnham, 2006). Both cognitive and non-cognitive correlates of creativity as DT have been investigated. This study sought to examine the ability, personality and demographic correlates of creativity as measured by the Consequences Tests extensively used in creativity research as a criterion measure of creativity.

Following on from the early work on fluency (Hargreaves, 1927), Guilford, 1950, Guilford, 1967 was one of the first to operationalise creativity in terms of tests of DT. Nearly all of these tests require people to produce several ideas in response to a specific prompt in a specific time period. DT tests are most commonly quantitatively scored for the number of responses provided by the participant (fluency). They may be scored for statistical infrequency of response (originality). Wallach and Kogan (1965) scored their sample for “uniqueness”; answers provided by only one participant, while others have suggested scoring scales whereby no points are allotted for common responses, with increasingly higher scores allotted for statistically infrequent responses (Torrance, 1974). There is no agreement on the best scores though fluency is the most common.

DT responses may be scored for the number of categories included in a participant’s answers or flexibility (Torrance, 1974). Although rarely used, DT responses may be scored for comprehensiveness of answer or elaboration (Torrance, 1974). An alternative to the quantitative-based approach involves the use of ratings of DT responses by judges. The most popular method employed has been the Consensual Assessment Technique (Amabile, 1982, Hennessey and Amabile, 1988). DT tests possess good concurrent validity with other creativity tests (Plucker, 1999).

This study used the Consequences Test (Christensen, Merrifield, & Guilford, 1958). It contains a number of questions like “what would be the consequences if everyone suddenly lost the ability to read and write?” or “what would be the consequences if none of us needed food any more to live?” Participants are given a specific time limit either per problem or for all problems. Responses as for other DT tests may be assessed quantitatively or qualitatively usually done by consensual rating techniques where a pool of expert and/or trained judges make a range of specific judgments with respect to issues like overall quality, originality and realism as well as complexity, use of principles or the number of positive vs negative outcomes. Perhaps the best known scoring technique is that of Hennessey and Amabile (1988) who specified six principles while others (Mumford, Marks, Connelly, Zaccaro, & Johnson, 1998) have added others.

Numerous researchers have attempted to examine the relationship between DT and intelligence. Batey and Furnham (2006) have recently conducted an exhaustive review of the area and concluded that they are modestly related with correlations in the area of r = .20 to r = .40. This is the case for samples as diverse as architects and air force officers, to ordinary and gifted school children. The relationship between intelligence and DT, however, has been demonstrated to be non-linear (Guilford, 1967), such that at low levels of intelligence there are significant relationships between the two constructs, however, above a quotient of approximately 120 the correlation between intelligence and DT is non-significant (Guilford, 1981, Torrance, 1962, Yamamoto, 1964). However Preckel, Holling, and Wiese (2006) reported data that speak against threshold theory. Hence whilst we examine the relationship between DT and intelligence we do not posit a specific hypothesis.

Studies have specifically examined cognitive ability (i.e. intelligence) correlates of the DT Consequences Test. Mumford et al. (1998) tested over 1800 military personnel and found measures of intelligence correlated on the order of r = .21 to r = .29 with Consequences Test ratings of quality, originality, realism and complexity. In a more sophisticated study of 110 military leaders, Vincent, Decker, and Mumford (2002) found intelligence correlated with the Consequences Test idea generation on the order of r = .25. Their model showed divergent thinking was significantly related to intelligence, expertise and idea generation which through idea implementation predicted leader performance. More recently Batey and Furnham (in press) found DT fluency for the Consequences Test to be unrelated to intelligence. Furnham and Bachtiar (2008) confirmed this in a study of four different measures of creativity; intelligence (as assessed by the Wonderlic Personnel Scale) was a non-significant correlate of all four measures.

The study of the personality traits associated with DT has also been well documented (Batey & Furnham, 2006). Research utilising the Eysenck Personality Questionnaire (EPQ: Eysenck & Eysenck 1975) found positive significant relationships between DT and Psychoticism (Aguilar-Alonso, 1996, Kline and Cooper, 1986, Woody and Claridge, 1977). Significant positive relationships have been observed with regards to Extraversion (Aguilar-Alonso, 1996). Research employing the Five Factor Model paradigm has tended to find consistent positive relationships of DT to Openness to Experience (King et al., 1996, McCrae, 1987, Wuthrich and Bates, 2001) and Extraversion (King et al., 1996, Martindale and Dailey, 1996). These results received wider support in the comprehensive meta-analysis of Feist (1998). Furnham and Bachtiar (2008) found Extraversion and Openness from the Big Five accounted for 47% of the variance in predicting DT. They note however that the reason why Extraversion may be such a consistent predictor is because in their study the creativity test was administered in a group setting.

There is a growing interest in the relationship between DT scores and “dark-side” measures of personality or sub-clinical measures of psychopathology. Researchers have found relationships between hypomania and DT (Furnham et al., 2008, Schuldberg, 2000–2001) in addition to schizotypy and DT (Batey and Furnham, in press, Cox and Leon, 1999, Green and Williams, 1999).

This study used a different but well established personality measure notably the Jackson PRF (Jackson, 1984). This modified test has been shown to measure four of the five Big Five dimensions of personality. Based on the previous literature it is predicted that Extraversion will be a significant predictor of DT.

This study set out to examine demographic, personality trait and disorder as well as intelligence correlates and predictors of DT as measured by the Consequences Test in a large Finnish working population.

The study is unique in three ways. First personality was measured using the Finnish version of the Jackson Personality Research Form (Jackson, 1984), which measures 16 personality variables at the facet level (rather than domain level). Few studies on creativity have used this well established test which may yield interesting insights. Next, this study utilised three different intelligence tests that measured abstract, verbal and numerical reasoning. It was therefore possible to explore whether different, but related, features of intelligence predicted creativity. Third, it has been speculated that there are demographic correlates of creativity. This study explored gender and age correlates of creativity.