Using data from a sample of Medicare beneficiaries and their caregivers (n = 829 dyads), we examine caregiver burden through the lens of social exchange theory, conceptualizing burden as an evaluative component or outcome of the ongoing exchange relationship between the care receiver and the caregiver. We find burden is affected by the context in which the exchange occurs, which varies by the role relationship of dyad members (spouses, adult children, or extended family and nonfamily), as well as the level of need of the care receiver, the living arrangements, and the life course timing of care provision. Gerontological research has focused heavily on the demands of caring for elderly relatives, in particular the distress and burden that these demands place on their caregivers (Biegel & Blum, 1990; Dwyer & Coward, 1992; Young & Kahana, 1989). In much of this work, caregiver burden has been conceptualized first as a contemporaneous aspect of the caregiver-care-receiver dyad and second as a one-way process. Thus, in this literature, burden is the caregiver's time-limited reaction to caring for another. Burden is viewed both as the direct consequence of providing assistance to an older relative (Montgomery, Gonyea, & Hooyman, 1985; Stull, Koloski, & Kercher, 1994) and as a direct cause of some care-receiver outcomes such as admission to a nursing home (Aneshensel, Pearlin, & Schuler, 1993; McFall & Miller, 1992; Zarit, Reever, & Bach-Peterson, 1980). We argue that caregiver burden be conceptualized as part of a reciprocal relationship that is not time limited. Looking at caregiver burden through the lens of social exchange theory, we view burden as an indicator of the ongoing exchange relationship in a caregiving dyad. The social exchange perspective provides an avenue for framing and applying a dynamic view of burden, and it responds to the challenge that gerontological research move toward theory-based explanations (Bengtson, Burgess, & Parrott, 1997; Dwyer & Coward, 1992; Kramer, 1997). We apply social exchange theory to an examination of the relative contributions of patient need, provision of informal care, receipt of formal care, perceptions of isolation, living arrangements, and life course timing to burden that are perceived by caregivers of older people who went home from the hospital and required help. The differential impact of these factors on burden will likely be tempered by ongoing relationships between members of the exchange dyad: that is, by whether the caregiver is a spouse, a child, a member of the extended family, or not a family member. SOCIAL EXCHANGE THEORY The four core assumptions of social exchange theory, described by Molm and Cook (1995), are as follows: People depend on one another for the things they value, people behave in ways that increase the outcomes they value and decrease outcomes they do not value, people engage in ongoing, mutually contingent exchanges with specific partners over time, and all outcomes obey a principle of satiation-that is, diminishing marginal utility or, conversely, increasing marginal costs. Once an outcome is achieved or acquired, its value and desirability decreases, especially at the upper thresholds. To illustrate, when only a small amount of care is required and this care does not interfere with other responsibilities and routines, small additions of time and tasks performed will be unlikely to have much of an impact on feelings of burden. However, if a great deal of care is needed that pushes the level of care the individual feels physically and emotionally capable of providing and if the demands of providing care spill over to other areas of the caregiver's life, every additional need or request may be sorely felt and have a big impact on the caregiver's perception of burden. The concepts of power and dependence are central to social exchange theory (Molm, 1991: Molm & Cook, 1995). …