Objective: Despite experimental manipulations that reliably elicit affective and physiological responses, the relationship between the two frequently appears small or nonexistent. We propose that this is, at least in part, due to a mismatch between the nature of the question being asked and the analytic methods applied. For example, to test if levels of affect reliably covary with physiology over time—a within-person question—one cannot apply analytic approaches that test whether people are similarly reactive across domains—a between-person question. The purpose of this paper is to compare within-person and between-person analyses testing the association between affect and physiology. Method: Participants (N = 60) recalled an event from their lives that made them angry. Self-reported anger and objective blood pressure levels were recorded at baseline, after the recall, and 5 times during recovery. Results: Between-person correlations between anger and blood pressure were nonsignificant across all phases of the study, suggesting that those least/most reactive for anger were not least/most reactive for blood pressure. These null findings held regardless of whether linear or nonlinear assumptions were modeled. In contrast, within-person multilevel modeling indicated a clear relationship, suggesting that when a person was angrier that person’s blood pressure was higher compared with when that person was less angry. Conclusion: Results suggest the importance of appropriately matching analytic strategy to the nature of the question regarding the relationships between affect and physiology. Implications for past and future research are discussed.