Administered associative fluency tasks and an intelligence test under both evaluative and play atmospheres to 36 4-yr-olds. The resulting dualtrait-dualmethod matrix of correlations provided both convergent and discriminant validity evidence for the measures used. Intelligence and associative fluency scores did not correlate significantly under any combination of sets, and scores for each trait assessed separately were significantly intercorrelated despite having been obtained under different sets. The results do not support M. A. Wallach and N. Kogan's previously untested hypothesis that a play atmosphere is necessary for a valid assessment of associative fluency, indicating instead that trait dissimilarity is responsible for zero-order correlations obtained between associative fluency and intelligence. (PsycINFO Database Record (c) 2006 APA, all rights reserved). © 1969 American Psychological Association.