A wealth of evidence has revealed that deficits on a commonly used measure of executive-function, the Wisconsin Card Sorting Test (WCST), can be improved through a variety of methods of instruction in patients with schizophrenia. Relatively little is known however, which of these remediation methods produce strongest and most durable effects and whether these effects generalize to other, untrained executive-function measures. Two of the most commonly studied methods for remediation on the WCST, step-by-step didactic instruction on the task and utilization of self-monitoring strategies, have both been shown to improve WCST performance, yet have never been directly compared. Thirty-four participants with schizophrenia were randomly assigned to one of three experimental conditions: Condition A consisted of didactic training which incorporated a detailed account of changing sorting principles throughout the test; Condition B consisted of a self-monitoring strategy that required that participants verbalize their strategies out loud after each card sort; Condition C was a non-trained control group that received the same outcome assessments as the two training groups at identical time intervals without intervening training. Patients were assessed with the WCST and two other executive-function tests immediately prior to training, immediately after training and at a 1-month follow-up. Results revealed: (1) participants assigned to the didactic and self-monitoring conditions made significant gains on the WCST relative to a no-intervention control condition; (2) the effects of self-monitoring, but not didactic training, were evident at a one-month follow-up; and (3) only participants assigned to the self-monitoring condition showed generalization to a second, non-trained measure of executive-function. The significance of these results for implementation of strategies for comprehensive and sustained programs of remediation are discussed.