This study examined the effects of experimentally provided social support on stress reactions and task performance in alone and co-acting problem solving situations. The reconstructed social support questionnaire and trait anxiety scale were administered to 264 female junior college students. Fifty-four of them, with low assessed social support scores and high trait anxiety scores, were selected as subjects. A2 (support: provided vs. not-provided) ×2 (task situation: alone vs. co-acting) factorial design was used. Dependent variables were physical state anxiety (pulse rate), cognitive state anxiety and cognitive interference as stress reactions, and correct word puzzle solutions. Though neither main effects of the two factors nor their interaction were significant for any dependent variables, re-analysis of perceived social support based on subjects' impression ratings to the experimenter who offered support suggested that perceived support might reduce cognitive state anxiety and facilitate task performance, while provided support might increase cognitive interference. These findings and suggestions were discussed from the viewpoint of support exchanges between provider and recipient.