Bonnie Cramond begins with the question, “Trust creativity tests to do what?” as if I hadn't been specific about the things we shouldn't trust creativity tests to accomplish. But as I stated in my opening paragraph, I was writing about the widespread use of divergent-thinking tests in the selection of students for gifted/talented (G/T) programs.
I recommended having experts evaluate the creativity of actual products, like poems and artworks, as a possible alternative to divergent-thinking tests. Cramond is concerned, correctly, that judges of creative products might sometimes misjudge their creativity. We know that shifts in the opinions of experts have occurred in the past, although such changes are far more likely when dealing with genius-level, paradigm-shifting creative products (as in Cramond's examples of the works of Van Gogh, Einstein, Edison, and Maya Angelou) than in judging the creativity of student work.
Expert Opinion Is the Best Criterion
While we must acknowledge that the collective opinion of experts in a field is imperfect, it is, in the end, the best criterion we have. Expert opinion determines which theories, paintings, compositions, novels, scientific papers, and other creative products are, at any given point in time, recognized as the most creative and therefore influential (Kuhn 1970). Expert opinion establishes the standards by which creative work is judged.
For this reason, a good paper-and-pencil test of creativity would need to predict creative behavior, as judged by experts, or it would not be a valid measure of creativity. Consensual assessment of creative products, in contrast, goes right to the creative behavior we care about and measures the creativity of the actual products (poems, theories, artworks) that students have created, not some variables that a given theory predicts are related to creativity. This technique has been extensively researched by Amabile (1983) and others (see Baer 1993 for a summary), and it has been shown to be reliable, although certainly not perfect.
If one must have some measure of a particular kind of creativity, consensual assessment of creative products introduces the least amount of measurement error. I am not sure, however, that for most placement decisions, such as entrance to gifted/talented programs, any measure of creativity is necessary. All students, those in G/T programs and those not in G/T programs, should be taught the skills and attitudes that lead to creativity (as both Cramond and I propose). We know that teaching a variety of creativity-relevant skills can help increase actual creative behavior (Baer 1994, in press). So do we need a test that can predict possible future creative behavior?
Focus on Present Achievements
This brings us to the aptitude-achievement distinction, a slippery slope if ever there was one. Conceptually the distinction is clear, but when it comes to assessment, we are stuck with using some measure of achievement—that is, something students can do now—to estimate their aptitude for doing something in the future.
There certainly may be students who do not currently write creative poetry but who one day will do so, as Cramond suggests. The best predictor of future behavior, however, is past behavior; although this, too, is an imperfect gauge. If a test could add significantly to our ability to make such predictions, it would certainly be interesting—although probably not very useful in making placement or other educational decisions.
To teach students effectively we need to meet them where they are now (or a slight stretch beyond, in what Vygotsky termed their “zone of proximal development”). Knowing what great things students might one day achieve doesn't tell us what they need now and doesn't do much to inform instruction. Even if we have reason to believe that a student has great talent in, say, mathematics, and even if we could predict with some confidence that he or she will one day be a knowledgeable mathematician, we still need to help that student learn the math he or she doesn't yet know (Of course, how fast he or she learns may cause us to adjust our pace).
Are the Torrance Tests of Creative Thinking good predictors of future creative performance? Studies yield many different conclusions, but if one looks at all of the available evidence, it is hard to be very optimistic (see Baer 1993 for summary). Part of the problem lies in defining the criteria that will be accepted as evidence of creativity. Creative achievements that have been used as evidence for the Torrance Tests' validity include such things as subscribing to a professional journal and changing one's religious affiliation, either one of which may or may not be related to creativity. It depends on one's theory of creativity.
The Overlap Problem
There is also the problem of overlap between what the Torrance Tests measure and what intelligence tests measure. Even if the Torrance Tests were modestly correlated with creative achievement, they might not provide information beyond that already provided by standard intelligence tests.
In fairness the jury is still out as far as determining the validity of the divergent-thinking tests as measures of creativity. Studies suggest that “the validity of the Torrance instruments must be considered equivocal at best” (Kogan 1983). Recent research suggesting that the skills underlying creative performance on different tasks are not related does not bode well for any general test of creativity. Nonetheless, divergent-thinking tests may eventually be shown to have some ability to predict actual creative performance. Right now, however, we simply don't have a test we can trust.
Teach Creativity to All Students
If we are going to make creativity an important part of the education of all students—both those in G/T programs and those not in G/T programs—then why use our already limited resources testing students' divergent-thinking skills? Gifted/talented programs must employ diverse assessment procedures, but not ones that have questionable validity and little connection to the skills one actually needs to succeed in such programs.
There are special cases, of course, in which some kind of creativity assessment might be appropriate. For example, if a school had a G/T program devoted exclusively to nurturing poets, then it might be worthwhile to assess students' creativity by having experts evaluate the students' poems. But most G/T programs are not so narrowly focused, and success in most of them does not typically require the refined divergent-thinking skills that most creativity tests try to measure.
Let's spend our creativity-training resources teaching all students, both in and out of G/T programs, to be more creative thinkers, not administering tests of doubtful validity to determine students' creativity quotients.
