The current study aggregated methods from the cardiovascular reactivity and dietary restraint paradigms in an attempt to advance our understanding of stress-induced eating behavior. Seventy-seven female subjects completed a protocol consisting of distinct baseline, stress-induction, and recovery phases during which we monitored blood pressure, heart rate, and self-reported affect. Food was inconspicuously made available to participants during the recovery phase. Our results replicated the restraint x affect level interaction observed in the restraint literature, while showing that physiological measures could further explain distressed eating behavior. Physiological arousal was found reliably to predict reduced food consumption, but only among unrestrained eaters. Analysis of the recovery data showed that food consumption was associated with impaired physiological recovery rates for restrained but not for unrestrained participants. We believe that our results help to reconcile findings in the stress, eating, and dietary restraint fields and offer support for recently developed theories of stress-induced overeating.