Psychology and Aging 1998, \*I. 13, No. 4, 531-543 Copyright 1998 by the American Psychological Association, Inc. 0882-7974/98/J3.00 Selection, Optimization, and Compensation as Strategies of Life Management: Correlations With Subjective Indicators of Successful Aging Alexandra M. Freund and Paul B. Baltes Max Planck Institute for Human Development The usefulness of self-reported processes of selection, optimization, and compensation (SOC) for predicting on a correlational level the subjective indicators of successful aging was examined. The sample of Berlin residents was a subset of the participants of the Berlin Aging Study. Three domains (marked by 6 variables) served as outcome measures of successful aging: subjective well-being, positive emotions, and absence of feelings of loneliness. Results confirm the central hypothesis of the SOC model: People who reported using SOC-related life-management behaviors (which were unrelated in content to the outcome measures} had higher scores on the 3 indicators of successful aging. The relationships obtained were robust even after controlling for other measures of successful mastery such as personal life investment, neuroticism, extraversion, openness, control beliefs, intelli- gence, subjective health, or age. We conducted an empirical investigation of a psychological model of successful aging, the model of selective optimization with compensation (SOC M. M. Baltes & Carstensen, 1996 P. B. Baltes & Baltes, 1980, 1990a P. B. Baltes, Dittmann- Kohli, & Dixon, 1984 Freund & Baltes, in press). In this ap- proach, successful aging, as successful development in general (P. B. Baltes, 1997 Freund & Baltes, in press Marsiske, Lang, Baltes, & Baltes, 1995), is defined as the maximization and Alexandra M. Freund and Paul B. Baltes, Max Planck Institute for Human Development, Berlin, Germany. This research was conducted within the context of the Berlin Aging Study (BASE). The study is directed by a steering committee con- sisting of P. B. Baltes, psychology H. Helmchen, psychiatry K. U. Mayer, sociology and E. Steinhagen-Thiessen, internal medicine and geriatrics. BASE is a project of the Committee on Aging and Societal Development of the Berlin-Brandenburg Academy of Sciences in col- laboration with institutes and research centers affiliated with the Free University of Berlin, Humboldt University of Berlin, and the Max Planck Institute for Human Development, where the study is coordi- nated. The study is financially supported by Federal Ministry of Family Affairs, Senior Citizens, Women, and Youth Grants 314-1722-102/9 and 314-1722-102/9a. In addition to Margret M. Baltes, Frieder R. Lang, Ulman Linden- berger, and our colleagues in the Max Planck Institute's Center for Psychology and Human Development, we thank colleagues of the Re- search Network on Successful Midlife Development (chair: O. G. Brim) of the John D. and Catherine T. MacArthur Foundation, in whose context a version of the selection, optimization, and compensation instrument used in the present study was first developed. Joy Pixley was especially helpful in developing the response format we chose for the questionnaire. Correspondence concerning this article should be addressed to Alex- andra M. Freund or to Paul B. Baltes, Max Planck Institute for Human Development, Lentzeallee 94, 14195 Berlin, Germany. Electronic mail may be sent to freund@mpib-berlin.mpg.de or to sekbaltes@mpib- berlin.mpg.de. attainment of positive (desired) outcomes and the minimization and avoidance of negative (undesired) outcomes. The SOC model provides a general framework for the under- standing of developmental change and resilience across the life span (M. Baltes & Carstensen, 1996 P. B. Baltes, 1997). It builds on the assumption that throughout the entire life span, people encounter certain opportunity structures (e.g., educa- tion) as well as limitations in resources (e.g., illnesses) that can be mastered adaptively by an orchestration of three components: selection, optimization, and compensation. Concerning selection, the limitation of resources such as time and energy inherent to human existence necessitate the selection of domains (goals) of functioning because not all opportunities can be pursued (see also Carstensen, Hanson, & Freund, 1995). Two subprocesses of selection can be identified: Whereas the term elective selection refers to regulative processes that are involved in selecting from a pool of alternative developmental pathways, the term loss-based selection occurs in response to a decline of resources or loss of previously available goal-relevant means. Loss-based selection encompasses processes such as the reconstruction of one's goal hierarchy or the search for new goals. Optimization is defined as the allocation and refinement of internal or external resources as meansof achieving higher levels of functioning in selected domains (goals). Finally, concerning compensation, when confronting loss of resources or the decline of goal-relevant means in a selected domain of functioning, substitutive processes are needed to maintain a given level of functioning in the targeted domain (see also Ba'ckman & Dixon, 1992 Dixon & Backman, 1995). The SOC model is conceptualized as a general life span devel- opmental model (P. B. Baltes, 1987, 1997 Marsiske et al., 1995). In old and very old age, however,the dynamics associated with SOC are thought to be amplified and to take on a special profile. As people approach the end of life and are more likely 531

532 FREUND AND BALTES to encounter health-related constraints and other losses in plas- ticity and reserve capacity (M. M. Bakes & Carstensen, 1996 P. B. Baltes et al., 1984), a limitation of resources and associated compensatory needs become more evident than in younger ages. This decline is also apparent in subjective reports on the nature of life span development. When adults are asked about their expectations of changes during the adult life span, the ratio of expected gains to losses becomes increasingly less favorable and less controllable with age (Heckhausen & Baltes, 1991 Heckhausen, Dixon, & Baltes, 1989). One strength of the SOC model is that it can be applied at several levels of analysisand with different emphases of content. In the present context, the scope of its application is "aging well,'' that is, we asked, on a correlational level, whether older people who report life-management strategies consistent with the SOC model also report higher levels of functioning in sub- jective indicators of successful aging (i.e., satisfaction with one's own aging, positive emotions, and the absence of loneli- ness). Note that there was no apparent content overlap in the items used to measure the predictor (SOC) and the outcomes investigated. Preventing such confounding of item content is important in the study of successful aging, especially when self- reports are the sole medium of assessment (P. B. Baltes & Baltes, 1990b). The operationalization of the SOC model can proceed along various methods of data collection, experimentation, and levels of analysis. For instance, the focus can be on a specific behav- ioral domain (e.g., job performance Abraham & Hansson, 1995), on personal functioning in a more general and systemic sense such as everyday functioning or longevity (M. M. Baltes & Lang, 1997), or on individual versus social and collective SOC activities. In the present study, we wanted to (a) provide a test of the usefulness of a domain general self-report measure of SOC and (b) test the hypothesis that SOC is related to various subjective indicators of successful aging. Table 1 is a summary of the approach chosen to formulate SOC-related items. As successfully done in past work in the field of child development and the study of action-control beliefs (e.g., Skinner, Chapman, & Baltes, 1988), we proceeded within an action-theoretical framework to identify items for assessing selection, optimization, and compensation. In our approach (P. B. Baltes, Baltes, Freund, & Lang, 1995 Freund & Baltes, in press), elective selection is defined as developing, choosing and committing oneself to goals or preferred behavioral outcomes. Loss-based selection encompasses processes such as the recon- struction of one's goal hierarchy or the search for new goals. Optimization and compensation, on the other hand, refer to goal- related means. Whereas optimization denotes the means of achieving a selected goal or higher levels of functioning in a selected domain, compensation refers to means that are aimed at counteracting (actual or impending) loss in goal-relevant means to maintain goal attainment. Table 1 shows one possible kind of operationalization of the three components (selection, optimization,and compensation). The table provides a list of behaviors we consider prototypical instances of selection, optimization, and compensation from an action-theoretical perspective. These examples form the basic framework for the items included in our study. Table 2 (see next page) shows the specific items used to operationalize SOC. The purpose of this article is not to specify in more detail how we conceptualize the relations among selection, optimiza- tion, and compensation as interactive components of an adaptive system (see M. M. Baltes & Carstensen, 1996 P. B. Baltes & Baltes, 1990a Marsiske et al., 1995). We should note, however, that there are several ways to view the combinatorial dynamics. First, selective optimization with compensation can be con- ceived of as one single "integrative" process of adaptive mas- tery. This approach has been the center of past discussions (P. B. Baltes & Baltes, 1990a). Second, on a lower or more micro- level of aggregation, the facets of SOC can be viewed as sepa- rate processes, each contributing to successful development. In addition, we consider the possibility that the processes might unfold their impact on successful development primarily in in- teraction or coordinated modulation with each other. In the pres- ent study, we attempted to operationalize SOC in these ways outlined, namely as one combined process of selective optimiza- tion with compensation as the four individual processes of elec- tive selection, loss-based selection, optimization, and compensa- tion and in interaction with each other. The nature of interaction was restricted, however, to statistical interaction. An illustration Table 1 Selection, Optimization, and Compensation Embedded in an Action-Theoretical Framework (Baltes, Baltes, Freund, & Lang, 1995) Compensation (means and resources for counteracting loss and decline in goal-relevant means) Selection (goals and preferences) Optimization (goal-relevant means) Elective selection Specification of goals Goal system (hierarchy) Contextualization of goals Goal commitment Loss-based selection Focusing on most important goal or goals Reconstruction of goal hierarchy Adaptation of standards Search for new goals Attentional focus Seizing the right moment Persistence Acquiring new skills and resources Practice of skills Effort and energy Time allocation Modeling successful others Substitution of means Use of external aids and help of others Use of therapeutic intervention Acquiring new skills and resources Activation of unused skills and resources Increased effort and energy Increased time allocation Modeling successful others who compensate Neglect of optimizing other means