Educational researchers have "rarely" addressed the problem of how to provide feedback on constructed responses. All participants (N = 76) read a story and completed short-answer questions based on the text, with some receiving feedback consisting of the exact material on which the questions were based. During feedback, two groups receiving feedback also rated the certainty of their response correctness--an activity we have termed "assessment certitude." Additionally, participants in one of these groups viewed their initial responses along with the feedback. All three feedback conditions showed significant gains in recall performance compared to a fourth group that received no feedback. Low ratings of assessment certitude significantly correlated with improved recall for two groups receiving feedback that did not include their original responses. Among these participants, mental reiteration of the feedback received was the most frequently used mental strategy while participants in a third group, who saw their original responses during feedback, reported using other approaches such as visualization. (Contains 4 figures and 2 tables.)