This study examined couples' (N=94) behavior resulting from two proximal change interventions. One was a spousal "compliments intervention" to increase positivity, and the other was a "criticize intervention" to increase negativity. Participants were randomly assigned to one of the two intervention conditions or a control group. There was no main effect in affect from the pretest conflict discussion to the posttest conflict discussion between the interventions or control group. However, a manipulation check on how couples acted during either intervention produced a significant interaction effect. Pretest affect during conflict and marital satisfaction significantly predicted couples' construal of the intervention. Professionals may need to monitor how couples use specific interventions and direct the processes how the intervention is construed by the couple.