Dissonance theory implies that relationships should exist between dissonance-reducing behaviors and measures of tension. It is suggested that dissonance-reducing behavior should be positively correlated across subjects with initial tension but negatively correlated with tension after dissonance-reducing behaviors have occurred. Thirty-six male and 36 female subjects were told that they would administer intense shocks, mild shocks, or tones to an undeserving victim. Heart rate and skin conductance were measured when the subject first received these instructions and while he delivered the shocks or tones. Skin conductance showed increasing arousal with increasing injury, particularly among females. Dissonance-reducing behaviors were negatively correlated with post-experimental ratings of conflict, but no relationship was found between dissonance-reducing behaviors and physiological measures. References and tables are included. (Author)