Education and the Subjective Quality of Life*

Journal of Health and Social Behavior 1997, Vol. 38 (September):275-297 We examine whether education influences subjective quality of life. If it does, what are the mechanisms by which education affects well-being? We propose that educa- tion improves well-being because it increases access to nonalienated paid work and economic resources that increase the sense of control over life, as well as access to stable social relationships, especially marriage, that increase social support. We examine the relationship between education and a variety of indicators of subjective quality of life-depression, anxiety, anger, aches and pains, malaise, and dissatis- faction. Using two representative national samples collected in 1990 and 1995, we find that the well educated have lower levels of emotional distress (including depres- sion, anxiety, and anger) and physical distress (including aches and pains and malaise), but they do not have lower levels of dissatisfaction. Education reduces dis- tress largely by way of paid work, nonalienated work, and economic resources, which are associated with high personal control; but the extent to which it reduces distress by way of marriage and social support is much more modest. We contrast distress and dissatisfaction as indicators of the subjective quality of life. Does education matter to subjective quality of life? If it does, what are the mechanisms by which education affects well-being? We pro- pose that education is valuable to individual well-being because it provides access to the two primary determinants of well-being: non- *We are indebted to the National Institute on Aging for the grant (ROI AG12393) to John Mirowsky and Catherine Ross that supported the Aging, Status, and the Sense of Control (ASOC) data collection and analysis. We are indebted to the National Science Foundation for the grant (SES- 8916154) to Catherine Ross that supported the Work, Family, and Well-Being (WFW) data collec- tion. Sampling, pretesting, and interviewing for both surveys were conducted by the Survey Research Laboratory of the University of Illinois. We thank John Mirowsky for his statistical help, Barbara Reskin for her help with the comparison process theory of satisfaction, and JHSB reviewers for their suggestions. Address correspondence to: Catherine Ross, Department of Sociology, 300 Bricker Hall, 190 North Oval Mall, The Ohio State University, Columbus, OH 43210-1353; e-mail: ross. 131 @osu.edu. alienated paid work and supportive relation- ships. Compared to the poorly educated, we expect that well-educated persons have access to nonalienated paid work that increases the sense of personal control. Well-being comes, first, from nonalienated work in which people exert control over the labor process (Kohn 1976; Marx ([1884] 1964). Work that gives people the freedom from routinization, monot- ony, and external control on the one hand, and a chance to use their skills, develop as a per- son, and learn new things on the other, theo- retically increases subjective well-being, in part by increasing perceived control. Com- pared to the poorly educated, we also expect that well-educated individuals have access to stable social relationships, especially mar- riage, that increase social support. Well-being comes, second, from primary group ties and social bonds that increase supportive relation- ships with others, especially the personal secu- rity of marriage, and the sense of having other people to talk to and turn to in times of need (Durkheim 1951; Litwak and Messeri 1989). We propose that, through these processes, 275
276 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR education improves the subjective quality of life, measured as psychological well-being and distress. We extend core economic and sociological perspectives on the meaning of education to individual well-being. We argue that educa- tion 's value extends beyond jobs, earnings, prestige, and power to people 's psychological well-being. According to human capital and status attainment theories, employers need workers who can read, write, do basic math, communicate, negotiate, solve problems, look things up, figure things out, and develop ideas; this human capital is acquired in school, and these skills, knowledge, and abilities help a person get a good job (Blau and Duncan 1967; Becker 1964; Hyman, Wright, and Reed 1975; Sewell and Hauser 1975; Spaeth 1976; Treiman and Terrell 1975). The same skills and abilities shaped by schooling, we argue, improve individual well-being through their effects on objective life conditions and social psychological resources. THE LINKS BETWEEN EDUCATION AND DISTRESS The negative association between education and psychological distress is well documented (Glenn and Weaver 1981; Kessler 1982; Lennon and Rosenfield 1992; Link, Lennon, and Dohrenwend 1993; Mirowsky and Ross 1989, 1995; Pearlin et al. 1981; Ross and Huber 1985; Ross and Mirowsky 1989), but the mechanisms by which education affects well-being are not. Education is rarely the focus of investigation in the sociological study of stress (Pearlin 1989). Instead, it is usually a control variable in research whose focus is on something else. This means that research has not identified the mechanisms by which edu- cation affects psychological well-being. Furthermore, when subjective well-being is measured as satisfaction, researchers find little positive effect of education. This raises the question of whether education 's effect is uni- formly positive. Well-educated persons are not more satisfied with their jobs than the poorly educated (Andrisani 1978; Gordon and Arvey 1975; Glenn and Weaver 1982; Quinn, Staines, and McCullough 1974; Ross and Reskin 1992), and they are not more satisfied with life in general (Pascarella and Terenzini 1991). Some argue that if education does not increase job satisfaction, or satisfaction over- all, maybe it has little real value to the subjec- tive quality of life, since a principal motivation for attaining a high level of formal education in the United States is access to satisfying work (Berg 1971; Quinn and Mandilovitch 1977). We examine the relationship between edu- cation and a variety of indicators of subjective quality of life. We distinguish distress- depression, anxiety, anger, aches and pains, and malaise-from dissatisfaction. Theo- retically, distress results from deprivation, whereas dissatisfaction results from depriva- tion relative to one 's expectations (Mirowsky and Ross 1989). We propose that education improves the subjective quality of life, mea- sured as psychological well-being and distress -measures not confounded by high expecta- tions among the advantaged. If education cor- relates positively with subjective well-being, what explains the association? We focus on two pathways by which education might affect individual well-being: (1) work and economic conditions, which increase personal control and (2) marriage and family conditions, which increase social support. Often educational attainment is used simply as an indicator of socioeconomic status. However, education, income, and work indi- cate different underlying concepts, so we keep the three aspects separate. Schooling indicates the accumulated knowledge, skills, values, and behaviors learned at school, in addition to being a credential that structures employment opportunities. Income and economic hardship indicate economic well-being. Work is pro- ductive activity (paid or not). Further, educa- tion, employment, and economic resources are not on the same causal level. Education is the key to one 's position in the stratification sys- tem; it shapes the likelihood of being em- ployed, the qualities of the job a person can get, and income. Combining variables from different causal levels obscures processes. If education affects psychological well-being, is its effect direct, or is it indirect by way of work or economic resources? Work and Economic Conditions Paid Work. Well-educated people are more likely to be employed and are more likely to be employed full-time (vs. part-time) than are those
EDUCATION AND THE SUBJECTIVE QUALITY OF LIFE 277 with little education (U.S. Department of Education 1992). Employment, especially full- time employment, in turn, is associated with higher levels of psychological and physical well-being (Gore and Mangione 1983; Lennon and Rosenfield 1992; Pearlin et al. 1981; Ross and Bird 1994; Verbrugge 1983). Although physical and mental health affect an individual 's likelihood of being employed, the positive asso- ciation between well-being and employment is not simply due to the selection of healthy people into the work force (Kessler, House, and Turner 1987; Ross and Mirowsky 1995). Nonalienated Work. We expect that educa- tion gives people access to nonalienated work that involves a variety of tasks, nonroutine work, and the chance for continued learning and development, which decreases distress. In nonalienated work, workers control the labor process; they have the chance to use their skills in the design and implementation of the work. Nonroutine work gives people the free- dom to use thought and independent judgment in doing different things in different ways rather than doing the same thing in the same way in a process designed and controlled by others. Skilled work gives people the chance to learn new things and develop as a person through work. Together, creative, nonroutine, independent work that gives a person control over the labor process, and work that is intrin- sic to a person 's development not external to it, are the essence of nonalienated labor. Kohn and colleagues find that control over the work process, rather than ownership of the means of production or control over the labor of others, is most important to psychological functioning (Kohn 1976; Kohn and Schooler 1982; Kohn et al. 1990). We expect that the work done by well- educated people is less alienated than that done by the poorly educated, and that this work decreases distress. However, the evi- dence as to whether work characteristics explain some of the effect of education on psy- chological well-being is mixed. Lennon (1994) finds that the effect of education on depression is explained when work condi- tions-autonomy, time pressure, responsibili- ty, interruptions, physical effort, and routine- are added. Link and colleagues (1993) find that education 's impact on psychological well- being works largely through giving individu- als access to jobs involving direction, control, and planning. However, others find that the effect of education on distress remains unchanged with adjustment for job control and other occupational characteristics (Kessler 1982; Lennon and Rosenfield 1992). Economic Resources. Low levels of educa- tion increase economic hardship. Individuals with low levels of education have lower incomes than those with high levels of educa- tion (Sewell and Hauser 1975), in part because they are less likely to be employed, and if employed, more likely to hold low-level jobs. Low levels of education further deprive people of the problem-solving resources needed to cope with the stresses of economic hardship. Ross and Huber (1985) find a synergistic effect on economic hardship of low education and low income, each making the effect of the other worse. Hardship increases psychological distress; the chronic strain of struggling to pay the bills and to feed and clothe the children takes its toll, often in feelings of depression and malaise (Pearlin et al. 1981; Ross and Huber 1985). Marriage and Family Composition Marital Status. Evidence that education is positively associated with marriage is some- what mixed and indirect. The well educated are less likely to divorce, probably due to the fact that they marry later and do so under more favorable economic conditions (Glick 1984; Houseknecht and Spanier 1980). Education is negatively associated with widowhood, too, since men and women choose partners with similar levels of education (Kalmijn 1991; Qian and Preston 1993), and well-educated people live longer than those with lower levels of education (Rogot, Sorlie, and Johnson 1992). Although education is negatively asso- ciated with marital dissolution, either through divorce or widowhood, it is also negatively associated with marrying in the first place, especially for some groups. Education decreases the probability of ever marrying among Whites, while highly educated Blacks are more likely to marry than Blacks with lower levels of education (Bennett, Bloom, and Craig 1989). Women with high levels of education are more likely to postpone mar- riage, not to remarry after divorce, and never marry in the first place than are women with lower levels of education; and women who fail to marry in young adulthood attain a higher
278 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR level of education than those who do marry (McLanahan and Casper 1994). All told, how- ever, well-educated women and men are more likely to be currently married than those with- out high school degrees (Qian and Preston 1993). Overall, married people and those who live with partners are healthier and happier than those who are single, divorced, or widowed (Waite 1995). Married people have higher lev- els of physical and psychological well-being than do unmarried people (Gove, Hughes, and Style 1983; Ross, Mirowsky, and Goldsteen 1990; Ross 1995). If marriage has any nega- tive effects on well-being, it may be due to child care responsibilities, which fall dispro- portionately on women. People with children at home do not have higher levels of psycho- logical well-being than nonparents (Gore and Mangione 1983; Kessler and McRae 1982; McLanahan and Adams 1987). In many instances, parents-especially mothers-are more psychologically distressed than non- parents (Gove and Geerken 1977; Pearlin 1975). Most of the stress of children in the home is due to economic strains, excessive child care responsibilities, and difficulties arranging child care while parents are at work (Ross and Huber 1985; Ross and Mirowsky 1988). Since children often accompany mar- riage, the effect of marriage on well-being may not be uniformly positive. However, the well educated may get the benefits of mar- riage, without the stress associated with chil- dren, since well-educated women have fewer children than poorly educated women. Women with high levels of education are more likely to remain childless, to postpone having chil- dren, and to have fewer children than are those with lower levels of education (Bloom and Trussel 1984; Rindfuss, Morgan, and Swice- good 1984; Veevers 1979). Since men and women tend to choose partners with similar levels of education (Kalmijn 1991; Qian and Preston 1993), well-educated men have fewer children, too. Social-Psychological Resources Sense of Control. Belief in personal control is a learned, generalized expectation that out- comes are contingent on one 's own choices and actions (Mirowsky and Ross 1989; Rotter 1966). The sense of powerlessness, the oppo- site, is the belief that one 's actions do not affect outcomes. It is the main form of subjec- tive alienation (Seeman 1959, 1983). Education correlates positively with the sense of control, and the sense of personal control mediates a large part of the negative associa- tion between education and distress (Mirow- sky and Ross 1989; Ross and Mirowsky 1989). The well educated have higher person- al control than the poorly educated, even adjusting for employment, job autonomy, earnings, minority status, age, marital status, sex, and household work (Bird and Ross 1993; Ross and Mirowsky 1992). People with high levels of personal control have low levels of psychological distress (Aneshensel 1992; Gecas 1989; Mirowsky and Ross 1986; Pearlin et al. 1981; Wheaton 1980, 1983), and perceived control over both good and bad outcomes correlates negatively with depression (Krause and Stryker 1984; Mirowsky and Ross 1990). High personal con- trol helps people cope actively and flexibly, to avoid problems and to prepare for those that cannot be avoided (Mirowsky and Ross 1989; Turner and Noh 1983; Wheaton 1983). In con- trast, the sense of powerlessness is demoraliz- ing in itself and interferes with active problem- solving. Social Support. Social support is the com- mitment, caring, advice, and aid provided in personal relationships, the sense of being cared for and loved, esteemed and valued as a person, and part of a network of communica- tion and obligation (Kaplan, Robbins, and Martin 1983). Little research has examined the social determinants of support. Education may provide social support by giving people access to multiple roles with independent social net- works which boost the potential for supportive relationships (Walker, Wasserman, and Wellman 1993). Education may also help peo- ple maintain supportive relationships with oth- ers directly by way of increased flexibility in dealing with problems, the ability to negotiate and compromise, and to see more than one side of an issue; and indirectly by reducing stressors of unemployment, poverty, and eco- nomic hardship which strain interpersonal relationships (Atkinson, Liem, and Liem 1986; Gore 1978). Ross and Mirowsky (1989) find that the well educated have higher levels of support than the poorly educated, but few other studies have examined the association between education and social support.
EDUCATION AND THE SUBJECTIVE QUALITY OF LIFE 279 Low levels of social support, especially emotional support, are associated with psycho- logical distress, including depression and anx- iety, and poor health (Gerstel, Riessman, and Rosenfield 1985; House, Landis, and Um- berson 1988; Kessler and McLeod 1985). Social support reduces distress directly and it interacts with stressful life events and situa- tions, buffering their negative effect (Wheaton 1983). Summary of Proposed Mechanisms by Which Education Affects Distress We propose that education gives people access to nonalienated paid work and eco- nomic resources that increase the sense of personal control, and that education gives peo- ple access to stable social relationships, espe- cially marriage, that increase social support. Through these processes, we hypothesize, education affects subjective quality of life. SAMPLES We use two representative national samples. With these we will replicate regression analy- ses in order to strengthen confidence in our results, distinguish robust findings from sample-specific ones, and comprehensively measure distress and dissatisfaction in two years. The first is the Aging, Status, and the Sense of Control (ASOC) survey. It is a 1995 telephone survey of a national probability sample of U.S. households. Respondents were selected using a prescreened random-digit dialing method that increases the rate of con- tacting eligible numbers (or decreases the rate of contacting business and nonworking num- bers) and decreases standard errors compared to the standard Mitofsky-Waksberg method while producing a sample with the same demographic profile (Lund and Wright 1994; Waksberg 1978). The ASOC survey has two subsamples, designed to produce an 80 percent oversample of persons age 60 and older. The survey was limited to English-speaking adults. The main sample draws from all households; the oversample draws only from households with one or more seniors. In the main sample, the adult (18 or older) with the most recent birthday was selected as respondent. In the oversample the senior (60 or older) with the most recent birthday was

Links: BETWEEN EDUCATION AND DISTRESS The negative association between education and psychological distress is well documented (Glenn and Weaver 1981; Kessler 1982; Lennon and Rosenfield 1992; Link, Lennon, and Dohrenwend 1993; Mirowsky and Ross 1989, 1995; Pearlin et al. 1981; Ross and Huber 1985; Ross and Mirowsky 1989), but the mechanisms by which education affects well-being are not. Education is rarely the focus of investigation in the sociological study of stress (Pearlin 1989). Instead, it is usually a control variable in research whose focus is on something else. This means that research has not identified the mechanisms by which edu- cation affects psychological well-being. Furthermore, when subjective well-being is measured as satisfaction, researchers find little positive effect of education. This raises the question of whether education 's effect is uni- formly positive. Well-educated persons are not more satisfied with their jobs than the poorly educated (Andrisani 1978; Gordon and Arvey 1975; Glenn and Weaver 1982; Quinn, Staines, and McCullough 1974; Ross and Reskin 1992), and they are not more satisfied with life in general (Pascarella and Terenzini 1991). Some argue that if education does not increase job satisfaction, or satisfaction over- all, maybe it has little real value to the subjec- tive quality of life, since a principal motivation for attaining a high level of formal education in the United States is access to satisfying work (Berg 1971; Quinn and Mandilovitch 1977). We examine the relationship between edu- cation and a variety of indicators of subjective quality of life. We distinguish distress- depression, anxiety, anger, aches and pains, and malaise-from dissatisfaction. Theo- retically, distress results from deprivation, whereas dissatisfaction results from depriva- tion relative to one 's expectations (Mirowsky and Ross 1989). We propose that education improves the subjective quality of life, mea- sured as psychological well-being and distress -measures not confounded by high expecta- tions among the advantaged. If education cor- relates positively with subjective well-being, what explains the association? We focus on two pathways by which education might affect individual well-being: (1) work and economic conditions, which increase personal control and (2) marriage and family conditions, which increase social support. Often educational attainment is used simply as an indicator of socioeconomic status. However, education, income, and work indi- cate different underlying concepts, so we keep the three aspects separate. Schooling indicates the accumulated knowledge, skills, values, and behaviors learned at school, in addition to being a credential that structures employment opportunities. Income and economic hardship indicate economic well-being. Work is pro- ductive activity (paid or not). Further, educa- tion, employment, and economic resources are not on the same causal level. Education is the key to one 's position in the stratification sys- tem; it shapes the likelihood of being em- ployed, the qualities of the job a person can get, and income. Combining variables from different causal levels obscures processes. If education affects psychological well-being, is its effect direct, or is it indirect by way of work or economic resources? Work and Economic Conditions Paid Work. Well-educated people are more likely to be employed and are more likely to be employed full-time (vs. part-time) than are those EDUCATION AND THE SUBJECTIVE QUALITY OF LIFE 277 with little education (U.S. Department of Education 1992). Employment, especially full- time employment, in turn, is associated with higher levels of psychological and physical well-being (Gore and Mangione 1983; Lennon and Rosenfield 1992; Pearlin et al. 1981; Ross and Bird 1994; Verbrugge 1983). Although physical and mental health affect an individual 's likelihood of being employed, the positive asso- ciation between well-being and employment is not simply due to the selection of healthy people into the work force (Kessler, House, and Turner 1987; Ross and Mirowsky 1995). Nonalienated Work. We expect that educa- tion gives people access to nonalienated work that involves a variety of tasks, nonroutine work, and the chance for continued learning and development, which decreases distress. In nonalienated work, workers control the labor process; they have the chance to use their skills in the design and implementation of the work. Nonroutine work gives people the free- dom to use thought and independent judgment in doing different things in different ways rather than doing the same thing in the same way in a process designed and controlled by others. Skilled work gives people the chance to learn new things and develop as a person through work. Together, creative, nonroutine, independent work that gives a person control over the labor process, and work that is intrin- sic to a person 's development not external to it, are the essence of nonalienated labor. Kohn and colleagues find that control over the work process, rather than ownership of the means of production or control over the labor of others, is most important to psychological functioning (Kohn 1976; Kohn and Schooler 1982; Kohn et al. 1990). We expect that the work done by well- educated people is less alienated than that done by the poorly educated, and that this work decreases distress. However, the evi- dence as to whether work characteristics explain some of the effect of education on psy- chological well-being is mixed. Lennon (1994) finds that the effect of education on depression is explained when work condi- tions-autonomy, time pressure, responsibili- ty, interruptions, physical effort, and routine- are added. Link and colleagues (1993) find that education 's impact on psychological well- being works largely through giving individu- als access to jobs involving direction, control, and planning. However, others find that the effect of education on distress remains unchanged with adjustment for job control and other occupational characteristics (Kessler 1982; Lennon and Rosenfield 1992). Economic Resources. Low levels of educa- tion increase economic hardship. Individuals with low levels of education have lower incomes than those with high levels of educa- tion (Sewell and Hauser 1975), in part because they are less likely to be employed, and if employed, more likely to hold low-level jobs. Low levels of education further deprive people of the problem-solving resources needed to cope with the stresses of economic hardship. Ross and Huber (1985) find a synergistic effect on economic hardship of low education and low income, each making the effect of the other worse. Hardship increases psychological distress; the chronic strain of struggling to pay the bills and to feed and clothe the children takes its toll, often in feelings of depression and malaise (Pearlin et al. 1981; Ross and Huber 1985). Marriage and Family Composition Marital Status. Evidence that education is positively associated with marriage is some- what mixed and indirect. The well educated are less likely to divorce, probably due to the fact that they marry later and do so under more favorable economic conditions (Glick 1984; Houseknecht and Spanier 1980). Education is negatively associated with widowhood, too, since men and women choose partners with similar levels of education (Kalmijn 1991; Qian and Preston 1993), and well-educated people live longer than those with lower levels of education (Rogot, Sorlie, and Johnson 1992). Although education is negatively asso- ciated with marital dissolution, either through divorce or widowhood, it is also negatively associated with marrying in the first place, especially for some groups. Education decreases the probability of ever marrying among Whites, while highly educated Blacks are more likely to marry than Blacks with lower levels of education (Bennett, Bloom, and Craig 1989). Women with high levels of education are more likely to postpone mar- riage, not to remarry after divorce, and never marry in the first place than are women with lower levels of education; and women who fail to marry in young adulthood attain a higher 278 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR level of education than those who do marry (McLanahan and Casper 1994). All told, how- ever, well-educated women and men are more likely to be currently married than those with- out high school degrees (Qian and Preston 1993). Overall, married people and those who live with partners are healthier and happier than those who are single, divorced, or widowed (Waite 1995). Married people have higher lev- els of physical and psychological well-being than do unmarried people (Gove, Hughes, and Style 1983; Ross, Mirowsky, and Goldsteen 1990; Ross 1995). If marriage has any nega- tive effects on well-being, it may be due to child care responsibilities, which fall dispro- portionately on women. People with children at home do not have higher levels of psycho- logical well-being than nonparents (Gore and Mangione 1983; Kessler and McRae 1982; McLanahan and Adams 1987). In many instances, parents-especially mothers-are more psychologically distressed than non- parents (Gove and Geerken 1977; Pearlin 1975). Most of the stress of children in the home is due to economic strains, excessive child care responsibilities, and difficulties arranging child care while parents are at work (Ross and Huber 1985; Ross and Mirowsky 1988). Since children often accompany mar- riage, the effect of marriage on well-being may not be uniformly positive. However, the well educated may get the benefits of mar- riage, without the stress associated with chil- dren, since well-educated women have fewer children than poorly educated women. Women with high levels of education are more likely to remain childless, to postpone having chil- dren, and to have fewer children than are those with lower levels of education (Bloom and Trussel 1984; Rindfuss, Morgan, and Swice- good 1984; Veevers 1979). Since men and women tend to choose partners with similar levels of education (Kalmijn 1991; Qian and Preston 1993), well-educated men have fewer children, too. Social-Psychological Resources Sense of Control. Belief in personal control is a learned, generalized expectation that out- comes are contingent on one 's own choices and actions (Mirowsky and Ross 1989; Rotter 1966). The sense of powerlessness, the oppo- site, is the belief that one 's actions do not affect outcomes. It is the main form of subjec- tive alienation (Seeman 1959, 1983). Education correlates positively with the sense of control, and the sense of personal control mediates a large part of the negative associa- tion between education and distress (Mirow- sky and Ross 1989; Ross and Mirowsky 1989). The well educated have higher person- al control than the poorly educated, even adjusting for employment, job autonomy, earnings, minority status, age, marital status, sex, and household work (Bird and Ross 1993; Ross and Mirowsky 1992). People with high levels of personal control have low levels of psychological distress (Aneshensel 1992; Gecas 1989; Mirowsky and Ross 1986; Pearlin et al. 1981; Wheaton 1980, 1983), and perceived control over both good and bad outcomes correlates negatively with depression (Krause and Stryker 1984; Mirowsky and Ross 1990). High personal con- trol helps people cope actively and flexibly, to avoid problems and to prepare for those that cannot be avoided (Mirowsky and Ross 1989; Turner and Noh 1983; Wheaton 1983). In con- trast, the sense of powerlessness is demoraliz- ing in itself and interferes with active problem- solving. Social Support. Social support is the com- mitment, caring, advice, and aid provided in personal relationships, the sense of being cared for and loved, esteemed and valued as a person, and part of a network of communica- tion and obligation (Kaplan, Robbins, and Martin 1983). Little research has examined the social determinants of support. Education may provide social support by giving people access to multiple roles with independent social net- works which boost the potential for supportive relationships (Walker, Wasserman, and Wellman 1993). Education may also help peo- ple maintain supportive relationships with oth- ers directly by way of increased flexibility in dealing with problems, the ability to negotiate and compromise, and to see more than one side of an issue; and indirectly by reducing stressors of unemployment, poverty, and eco- nomic hardship which strain interpersonal relationships (Atkinson, Liem, and Liem 1986; Gore 1978). Ross and Mirowsky (1989) find that the well educated have higher levels of support than the poorly educated, but few other studies have examined the association between education and social support. EDUCATION AND THE SUBJECTIVE QUALITY OF LIFE 279 Low levels of social support, especially emotional support, are associated with psycho- logical distress, including depression and anx- iety, and poor health (Gerstel, Riessman, and Rosenfield 1985; House, Landis, and Um- berson 1988; Kessler and McLeod 1985). Social support reduces distress directly and it interacts with stressful life events and situa- tions, buffering their negative effect (Wheaton 1983). Summary of Proposed Mechanisms by Which Education Affects Distress We propose that education gives people access to nonalienated paid work and eco- nomic resources that increase the sense of personal control, and that education gives peo- ple access to stable social relationships, espe- cially marriage, that increase social support. Through these processes, we hypothesize, education affects subjective quality of life. SAMPLES We use two representative national samples. With these we will replicate regression analy- ses in order to strengthen confidence in our results, distinguish robust findings from sample-specific ones, and comprehensively measure distress and dissatisfaction in two years. The first is the Aging, Status, and the Sense of Control (ASOC) survey. It is a 1995 telephone survey of a national probability sample of U.S. households. Respondents were selected using a prescreened random-digit dialing method that increases the rate of con- tacting eligible numbers (or decreases the rate of contacting business and nonworking num- bers) and decreases standard errors compared to the standard Mitofsky-Waksberg method while producing a sample with the same demographic profile (Lund and Wright 1994; Waksberg 1978). The ASOC survey has two subsamples, designed to produce an 80 percent oversample of persons age 60 and older. The survey was limited to English-speaking adults. The main sample draws from all households; the oversample draws only from households with one or more seniors. In the main sample, the adult (18 or older) with the most recent birthday was selected as respondent. In the oversample the senior (60 or older) with the most recent birthday was