How divergent thinking tests mislead us: Are the Torrance Tests still relevant in the 21st century? The Division 10 debate.

Abstract

An adult observed a preschool child working assiduously on a drawing and asked, “What are you drawing?” Without looking up, the child responded, “I’m drawing the face of God.” The adult smiled and said, “But no one knows what God looks like.” The child answered, “They will in a couple minutes.” No one is really sure what creativity looks like either, but half a century ago Joy Guilford described one important piece of it—divergent production—and Guilford’s vision has shaped what many of us think creativity looks like, especially in the area of creativity assessment. The Torrance Tests are essentially divergent thinking tests based on Guilford’s model, and they are very widely used as creativity tests. Although the Torrance Tests have lost some of the edge they had quarter of a century ago, when Torrance and Presbury (1984) reported that the Torrance Tests had been used in three quarters of all recently published studies of creativity, they nonetheless remain an important force, although perhaps more in schools than in research. If they are valid measures of creativity, that’s wonderful. If they are not, it’s a huge problem. It would mean that we may think we know a lot of things—things we’ve learned from research by using a possibly invalid tool—that we don’t really know. Guilford grouped his divergent-production factors into four categories, and those four categories—fluency, flexibility, originality, and elaboration—were the basis of the Torrance Tests for many years. Torrance cautioned against the use of a composite score, recommending instead that the subscales be interpreted “in relation to one another” to get a picture of an individual’s skills: