The main purpose of this study is to examine whether a self-handicapping strategy is used in private conditions where the subject's behavior is not publicly known. A 2 (contingency of success) ×2 (selfinvolvement of task) ×2 (publicity of self-handicapping behavior) factorial design was used. 80 male undergraduates experienced (a) contingent or non-contingent successes in (b) high or low self-involvement tasks. While anticipating the second task, they were allowed to choose between tapes that would either enhance or inhibit their performance in (c) public or private conditions. Except for one condition, the result suggested that three situational factors, that is, non-contingent success, existence of audience, and high self-involvement task, promoted self-handicapping. Contrary to the prediction, subjects in private condition who experienced contingent successes in low self-involvement tasks were most likely to choose the debilitating tapes. This choice was interpreted not as self-handicapping because of the reasons they reported and the correlation between contingency perception and choice.