The following excerpt was given at the annual George Maddox Lecture held March 18 by the Center for the Study of Aging and Human Development. For a full text of Dr. George's remarks, send an e-mail to dialogue@duke.edu

As I prepared this talk, I was haunted by the image of my colleagues in the Department of Sociology murmuring, "OK, this is kind of interesting, but it doesn't sound like sociology." Second, I want to briefly describe a primary focus of my current research '" an area where I am still far from answering the question that intrigues me most.

The answer to the "Where's the sociology?" question, I think, lies in the conditions under which we develop and sustain subjective perceptions. The basic sociological premise '" to which I subscribe '" is that our perceptions of self and world and our emotional reactions to those perceptions are embedded in and vary across social environments.

One way of looking at the cumulative findings of my research on perceptions of well- being is that social structure and social location matter very little. The fact that objective conditions such as economic status, social network characteristics, and the amount of assistance offered by network members are only weakly correlated with perceived well-being could in interpreted as evidence that social location is relatively unimportant. Given the way that social structural factors are measured in the survey research that has generated these findings, I cannot disagree with that conclusion. That does not mean, however, that the social environment is unimportant.

The fundamental processes by which we process and evaluate self, others, and world develop and occur in social interaction. Consequently, if we are to understand how those processes develop and are sustained, we must study people in their social environments. For a start, we must demonstrate that different kinds of social environments are associated with distinctive perceptions and perception-generating processes. There is some sociological evidence of this kind. The work of Melvin Kohn and his colleagues stands out as a key illustration. In his earlier work, Kohn demonstrated that the occupations of fathers significantly affect styles of child- rearing and ultimately produce distinctive worldviews among their children. More recently, his research has elegantly demonstrated that key parameters of the organization of work affect a broad range of objective and subjective outcomes.

Much more research is needed that links worldviews or systems of beliefs and assumptions to the social environments within which lives unfold. This is not the sociology of conventional survey research in which individuals' characteristics, attitudes, and behaviors are extracted from the richness and histories of their environments. But it is sociology.

The limits of conventional survey-based research have been powerfully demonstrated to me through the research that my colleagues and I have conducted on the links between religion and health. As all of you know, there is now strong evidence that religious involvement is associated with a variety of positive health outcomes including delay in the onset of physical and mental illness, better course and outcome of physical and mental illness, and longer survival. Among the multiple dimensions of religious experience, attendance at religious services is the dimension of religious involvement most strongly related to health and survival.

The key issue facing this research field is identification of the mechanisms by which religion exerts its salubrious effects on health. Three primary factors have been hypothesized as explanatory mechanisms. First, religious participation may lead to healthier lifestyles, which in turn, promote health and longevity. Second, religious involvement may foster important psychosocial resources '"such as self-esteem and self-efficacy -- that have been demonstrated to benefit health. Third, religious participation may promote higher quality social support networks, which in turn benefit health and longevity. As you can imagine, I entered this field convinced that the health benefits of religious involvement would be explained by social support. I was wrong. Our research and that of many others largely failed to support the hypothesized explanatory mechanisms. Overall, it appears that health behaviors make modest contributions to explaining some health outcomes. Both social support and psychosocial resources, however, explain virtually none of the relationships between religion and health. All three sets of variables are significant predictors of morbidity and mortality; but they don't mediate the effects of religion on health.

I have puzzled over these findings long and hard. And I keep coming back to the conviction that the processes that account for the links between religion and health must be fundamentally social '" because attending religious services is the strongest predictor of health and longevity and it is the most social dimension of religious experience.

I don't think that the mechanisms that link religious attendance to health and longevity will be found until we learn more about religious congregations as social environments. What is distinctive about them? How do they affect the way we process the events and experiences of our lives? What worldviews do they foster? Is it a sense of community that is difficult to achieve in other environments '" a kind of communal support that supplements informal support networks? These are among the issues that I hope to address in future research. And, indeed, with support from the Duke Endowment, Keith Meador, Elizabeth Clipp, and I are launching a study of congregations and congregational climate that I hope will begin to answer these questions.

In closing, I want to express my gratitude for the thousands of older adults who have participated in the studies that have permitted me to pursue my research on well-being in late life. They have taught me much and I hope that I've been able to disseminate that knowledge in ways that honor and testify to the magnificence of the human spirit. They have taught me that quality of life can be high even in the presence of declining health, declining financial assets, and shrinking social networks. Most importantly they have taught me that quality of life is at least as much a function of what we believe as what we have.

Linda George is professor of sociology at Duke. The George L. Maddox Lectureship was established in 1996 in honor of Maddox, longtime director of the Duke Center for the Study of Aging and Human Development from and Chairman of the University Council on Aging and Human Development from 1982-1992.