The literature on multiple roles and identities tends either to ignore strategies of role-system organization altogether or to assume that all people organize a salience hierarchy, through which they assign more importance to some roles and selves than to others. Drawing on our reading of William James and George Herbert Mead, we argue that the way people organize their roles and identities is an empirical issue, not an established fact, and that it is a live option for people to create a nonhierarchical pattern of self-organization. We offer findings from two studies of role balance. Using planned comparisons, we confirm hypotheses that people who maintain more balance across their entire systems of roles and activities will score lower on measures of role strain and depression and higher on measures of self-esteem, role ease, and other indicators of well-being. We end with some cautions and suggestions concerning the further exploration of role balance.
Key Words: identity theory, multiple roles, role balance, role involvement, role strain, self-esteem.
In the literature on multiple roles and selves, there are two distinct shortcomings. The first is insufficient attention to the systemic organization of roles and selves. The second is the assumption by those who do theorize about such systems that roles and selves are always hierarchically organized. We discuss each problem in turn and then offer a phenomenology of role-system organization that will obviate some difficulties signaled in our critique. We then present two empirical studies of role balance-a neglected type of rolesystem organization.
ASSUMPTIONS ABOUT MULTIPLE ROLES AND SELVES
Neglecting Goode's Legacy: The Atomistic Tendency
In his seminal article, Goode (1960) claimed that role strain-any felt difficulty in carrying out a role-arises only when a person's total role system is overdemanding. Role strain is not, therefore, role-specific in its origin; it is a product of the individual's myriad interests and his or her navigation through the entire system of typical activities.
Despite this holistic, systemic animus of the theory, much of the literature since then has reverted to an atomistic framework. Each role is seen as a thing in itself, separable from its embeddedness in an organized system of roles. (See O'Neil & Greenberger, 1994, and the accompanying critique by Marks, 1994.) To be sure, research after Goode has provided an important corrective to his cynical statement that role strain is "normal"; we are cheered by findings suggesting that multiple roles may be good for one's health (Verbrugge, 1986) and for one's psychological wellbeing (Baruch & Barnett, 1986). Similarly, there has been useful research on the quality of certain roles (for example, Barnett, 1994) and on the enhancing effects of certain role combinations on well-being (for example, Barnett, Marshall, & Pleck, 1992; Menaghan, 1989). Role-specific experience-role quality-does make a difference within a given role and within each of several roles in combination. But Goode explicitly adopted a holistic rather than an atomistic level of analysis, calling our attention to the way our roles function as a single pattern or system. That is, the pattern of organization itself may generate important consequences.
A psychological counterpart of atomistic role theories is seen in some work on multiple selves. For Linville (1987), people with many selfaspects are buffered against stress from negative events because they have the option of refocusing on whichever selves have remained unaffected by any particular event. While Linville sees selfaspects to be systemically embedded in a "larger associative network" (p. 664), she lays so much emphasis on the distinctness of each self-aspect that she verges on the atomistic tendency just noted. Similarly, Markus and Nurius (1986) point to a system of self-conceptions within the person, but they offer no analysis of how the organization of this overall system may make a difference. …