60 undergraduate students solved four conjunctive concept attainment problems. The experimental design varied sex, type of array, and type of verbalization. The stimulus arrays were either in ordered or random form, and Ss were instructed to either verbalize the information they expected to gain prior to a card choice, the information they obtained from their card choice, or were given no verbalization instructions. Under the ordered array, males who verbalized after instance selection showed better performance than those who verbalized prior to instance selection. When the stimuli were randomly arranged, males who verbalized before instance selection were superior to those who verbalized after instance selection. Both verbalization procedures facilitated the performance of females under the ordered array; however, when the stimuli were randomly arranged, verbalization either interfered with or had no effect on performance.