Whereas supportive interactions are usually studied from the perspective of recipients alone, the authors used a dyadic design to incorporate the perspectives of both provider and recipient. In 2 daily diary studies, the authors modeled provider reports of support provision in intimate dyads over several weeks. The 1st involved couples experiencing daily stressors (n ϭ 79); the 2nd involved couples experiencing a major professional stressor (n ϭ 196). The authors hypothesized that factors relating to (a) recipients (their requests for support, moods, and stressful events), (b) providers (their moods and stressful events), (c) the relationship (relationship emotions and history of support exchanges), and (d) the stressor (daily vs. major stressors) would each predict daily support provision. Across both studies, characteristics of providers, recipients, and their relationship emerged as key predictors. Implications for theoretical models of dyadic support processes are discussed.Keywords: support provision, dyadic relationships, emotions, multilevel models, daily diary studies Social affiliation is considered to be a basic psychological need (Ryan & Deci, 2000), and a primary way adults fulfill this need is to form an intimate relationship with another person. Once formed, intimate relationships can be potent-and often primary-sources of emotional and practical support in times of stress. Indeed, it is common in social epidemiology to use marital status as a proxy measure of social support availability (e.g., Berkman & Breslow, 1983). The equation of marital status with social support has been justified by evidence that married people live longer, recover more quickly from mental and physical illnesses, and are less likely to engage in risk-taking behavior than those who are not married (for a review, see Coombs, 1991).Despite this evidence, there are indications that merely having an intimate partner does not imply that a person will receive effective support (Coyne & Bolger, 1990). Even when the support is well-intended, it can be inappropriate, untimely, or excessive (Coyne, Wortman, & Lehman, 1988). Further, when couples confront severe or chronic stressors, the partner may become overwhelmed and essentially incapable of being supportive. For example, in a study of breast cancer patients and their significant others Bolger, Foster, Vinokur, and Ng (1996) reported that the greater the patients' distress in the month immediately following diagnosis, the less likely their husbands provided support over the subsequent 6 months. This finding underscores the fact that support provision itself can be a variable process.To date, published studies on enacted support have usually focused on the perspective of the distressed, injured, or ill persons who are the recipients of the support (e.g., Neff & Karney, 2005;Sandler & Barrera, 1984;Sarason, Sarason, & Pierce, 1990). A second characteristic of the literature is that supportive acts are treated as causes rather than consequences, whether in naturalistic studies of so...