Test Instructions and Identification of Creative Scientific Talent

Abstract

The performance of 16 scientists told to “Be creative” in working on three open-ended tasks was compared with the performance of 15 scientists given neutral instructions. Under neutral instructions, the relation of performance (as measured by a summary score reflecting the judged quality and originality of response and response frequency) on the tasks and rated on-the-job creativity was an inverted U ( rho = –.17). In comparison to this group, “Be creative” instructions improved the test performance in the scientists rated in the upper third of the creativity distribution, depressed the performance of the lower rated men, and did not affect the performance of the middle group ( rho = +.71). The neutral instructions usually given may substantially decrease the power of tests to discriminate among more and less creative individuals.