| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
a School of Public Health, University of Michigan, Ann Arbor
b Institute of Gerontology, University of Michigan, Ann Arbor
Benjamin A. Shaw, Department of Health Behavior and Health Education, School of Public Health, University of Michigan, 1420 Washington Heights, Ann Arbor, MI 48109-2029 E-mail: bashaw{at}umich.edu.
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Methods. Data for this study come from the first wave of the Americans' Changing Lives panel study. Complete data are available for 357 Black and 2,792 White individuals. Respondents were asked questions about their feelings of control, health status, income, social support, cognitive function, religious participation, and demographic information.
Results. The findings suggest that age has an inverse and nonlinear association with feelings of control. Moreover, this relationship is similar for Blacks and Whites. The data further reveal that, across all age groups, Blacks report a lower sense of control than Whites. Less education, less income, greater cognitive impairments, and more religiosity are associated with a lower sense of control. These factors, along with health and social support, account for 69% of the association between age and control, with no differences according to race.
Discussion. These results show that lower levels of control are associated with older age in both Blacks and Whites, and that racial disparities in feelings of control persist across all age groups. This suggests that Blacks may be at a particular disadvantage in the face of the increasing challenges of aging.
Astrong sense of personal control is essential for the maintenance of health and well-being in late life (
Mendes de Leon, Seeman, Baker, Richardson, and Tinetti 1996
). Individuals with a strong sense of control believe the changes in their social world are responsive to their choices, efforts, and actions. In contrast, people with a weak sense of control believe that events in their lives are shaped by forces outside their influence, and that they have little ability to influence the things that happen to them. Unfortunately, recent research by
Mirowsky 1995 shows that feelings of control may decline in an accelerating fashion across progressively older age groups. If we want to help older adults age successfully we must, therefore, learn more about the social and psychological factors driving this process.
Research to this point suggests that two factors are largely responsible for the observed decline in feelings of control—education and health (
Mirowsky 1995;
Wolinsky and Stump 1996). Education is thought to bolster feelings of control over the life course because people with more schooling tend to have better problem-solving abilities, as well as access to more resources. The decline in feelings of control across older age groups, therefore, may be a manifestation of the current cohort of older adults having less schooling, on average, than today's younger adults (
Mirowsky and Ross 1998). Furthermore, diminishing health and increasing impairments associated with aging may contribute to the decline in feelings of personal control by progressively reducing the range of activities in which one can exercise control (
Schulz, Heckhausen, and O'Brien 1994).
Although this research has provided valuable insight into the relationship between age and personal control, a number of key issues remain unexamined. Chief among them is the influence of race. In existing studies, race has been treated as a control variable—something to be removed rather than examined in its own right (
Mirowsky 1995;
Wolinsky and Stump 1996). This is unfortunate because race differences in feelings of control appear to exist at all ages (
Mirowsky, Ross, and Van Willigan 1996). What is less evident, however, is whether the decline of control in minority group members across different age groups is similar to what
Mirowsky 1995 and others found, or whether the relationship between age and control follows a different course among people of color. Race is a master status that affects virtually all areas of life, including education, health, and feelings of personal control (
Livingston 1994). Therefore, if race is related to feelings of personal control and the factors that erode control with advancing age, then the trajectory reported by
Mirowsky 1995 may not be the same for White and minority group members.
In addition to determining whether a similar association between age and feelings of control exists for Blacks and Whites, it is also important to see if the same factors account for this relationship in both racial groups. We suspect that relationships among education, health, and control may differ for Whites and Blacks. For example, education is a form of human capital that provides people with problem-solving skills and the resources (e.g., income) necessary for influencing the environment. However, the benefits of education may not be enjoyed equally by Blacks and Whites. In particular, due to inequitable school funding practices and discrimination in the workplace, Blacks receive less return than Whites on their educational investments (
Maxwell 1994). This suggests that there may be structurally induced racial disparities in the skill levels necessary to exercise effective control. Therefore, the relationship between education and control should be consistently weaker across the life course for Blacks than for Whites.
There may also be race differences in the relationship between health and feelings of control over the life course. For instance, lower rates of health care access and utilization among Blacks are now widely documented (
Blustein and Weitzman 1995;
White-Means 2000). Because health problems that are left unattended get worse, the downward trajectory of control should be more precipitous in Black than White groups.
Still, other factors may account for race differences in the relationship between age and control. For instance, income, social support, and cognitive function may also contribute to one's sense of control (
Caplan 1981;
Lachman and Weaver 1998). However, for many people, these resources (perhaps with the exception of social support) decline in later life (
George, Landerman, Blazer, and Anthony 1991;
Holden, Burkhauser, and Myers 1986;
Krause 1991). Declines in income and cognitive function may be especially devastating to feelings of control among the Black group, which, because of its low socioeconomic position, may already be lacking alternative sources of control. In contrast, the relatively strong social support systems of Blacks may make them less vulnerable to declines in control with advancing age (
Mutran 1985).
Finally, religion may play a somewhat different role in accounting for the association between age and control. Because religion bolsters feelings of control (
Krause and Tran 1989), it is expected that religiosity may suppress the negative association between age and control. Given that Blacks are more involved in religion than Whites (
Taylor 1986), the "protective" effects of religion may be greater for members of this minority group.
Taken as a whole, the discussion above leads to the following hypotheses:
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
). This study used a stratified, multistage, area probability sample of noninstitutionalized persons 25 years of age and older living in the coterminous United States. Blacks and persons 60 years of age and older were oversampled. Face-to-face interviews for the first wave of data collection took place between May and October 1986. A total of 3,617 interviews were completed successfully, representing a 68% response rate. The current study uses only those respondents who self-identified as Black or White. After using listwise deletion of missing values to deal with item nonresponse, complete data were available for 3,149 study participants. Approximately 11.4% (n = 357) of the sample identified themselves as Black. The average age of the sample as a whole was 46.7 years (SD = 16.40), with a range of 25 to 96 years. Fifty-two percent were classified as female, and 68.8% were married. Study participants reported having completed an average of 12.47 years of schooling (SD = 3.0 years). These descriptive statistics, as well as the findings presented in the Results section, are based on weighted data.
Measures
The measure of control appears in Table 1 . A high score represents a strong sense of personal control. The internal consistency reliability estimate for this short scale is .718. Race is represented as a binary variable (1 = Black; 0 = non-Hispanic White). People who self-identified as American Indian, Hispanic, Asian, or "other" were excluded. Age was based upon self-reports from respondents and is scored in a continuous format. Education is scored in a continuous format reflecting the total number of years of completed schooling. Physical health was measured as a simple count of the number of major chronic conditions experienced in the last year out of a list of 10. This list included arthritis/rheumatism, lung disease, hypertension, heart attack and heart trouble, diabetes, cancer/malignant tumor, foot problems, stroke, fractures or broken bones, and loss of urine beyond one's control. Income was measured as total pre-tax family income from all sources in the 12 months prior to the survey. Respondents chose from among 10 ordinal categories of income ranging from "less than $5,000" to "$80,000+." The social support measure also appears in Table 1 . A high score indicates more support. Cognitive impairment was measured as the number of incorrect responses to eight questions from the Short Portable Mental Status Questionnaire (
Pfeiffer 1975) covering the date, mother's maiden name, day of the week, current president, former president, serial three test, and self-reported age and birth date. Table 1 also includes the measure of religiosity. This measure is coded so that higher scores indicate greater religious involvement. The internal consistency reliability estimate for this scale is .753. And finally, the models in this study were estimated after the effects of sex and marital status were controlled statistically. Both are represented by binary variables (1 = female, 0 = male; 1 = married, 0 = otherwise).
|
), the key independent variable—age—is represented as a squared term throughout. Whether the lower order term of age should also be included in the equations is a source of some debate. The prevailing rule of thumb is that higher order terms should be tested only when the lower order terms are also included in the equation. However,
Allison 1977 proposes that if the predictor in question is measured at the ratio level, inclusion of the higher order term only is appropriate. In the case of this study, the variable of interest—age—is a ratio-level measure and, therefore, only the squared term is included in all models.
The following equations were estimated:
 | (1) |
 |
 |
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Age and Control
As shown in Model A of Table 2 , the findings from estimating 1 replicate the main results of
Mirowsky 1995. More specifically, feelings of control decrease in an accelerated fashion across progressively older age groups (b = -1.38 x 10-3; p < .001). Also in Model A, it is important to note that Black race is negatively associated with feelings of control (b = -1.07; p < .001). This suggests that, regardless of age, Black respondents reported a lower sense of control than Whites.
|
Effects of Intervening Variables
Results from the analyses testing the effects of various intervening variables on the relationship between age-squared and control are also presented in Table 2 . The findings from the additive model appear in Model C. These results show that a greater sense of control is associated with more education (b = .17; p < .001), more income (b = .16; p < .001), and fewer cognitive impairments (b = -.16; p < .001). However, in contrast to what was anticipated, a higher degree of religiosity is associated with a lower sense of control (b = -.14; p < .001). In addition, after accounting for the effects of all of the other predictors, health and social support are not associated with feelings of control. Taken together, the data in Table 2 reveal that the intervening variables account for 69% of the relationship between the age-squared term and control.
Next, racial differences in the effects of these predictors on control were tested with interaction terms. As shown in Model D of Table 2 , race differences appear to exist only with regard to the impact of education on feelings of control (b = -.10; p < .05). Separate analyses within each racial subsample (not shown here) show that education is significantly associated with feelings of control for Whites (b = .18; p < .001), but not Blacks (b = .07; not significant). The impact of the other predictors on feelings of control, however, appear not to differ by race. Furthermore, because the effect of the age-squared term on control does not change considerably with the inclusion of any of the interaction terms (Model C vs Model D), no racial differences appear to exist in the effects of these predictors on the relationship between age-squared and control.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Of particular interest are the effects of education and religion. The data indicate that, regardless of age, less education is associated with less perceived control among Whites only. This would lead us to believe that education should be more useful for explaining why Whites, but not Blacks, experience an inverse association between age and control. However, this was not the case. Instead, the findings indicate that there were no race differences in this relationship. The complex linkages among age, race, education, and control represent an important area for further inquiry.
Additionally, for both groups, greater involvement in religion is associated with a lower sense of control. Moreover, religion accounts for some of the observed association between age and control for both groups. The direction of these findings, however, was not anticipated. Based on previous research, we expected to see the opposite effect (i.e., more religion bolsters control;
Krause and Tran 1989). Perhaps the explanation for these unexpected findings may be found by turning to the literature on primary and secondary control (
Rothbaum, Weisz, and Snyder 1982). Primary control refers to efforts aimed at changing the external world. Following
Skinner 1996, secondary control involves efforts to change the external world by shifting responsibility to a benevolent other. Simply put, the negative association between religion and control may reflect the relinquishing of control to a powerful but benevolent other (i.e., God).
In the process of exploring these as well as other theoretical explanations for the differences in control over the life course, researchers would be well advised to take into account the limitations in the work we have presented. Three are identified briefly below.
As discussed above, our finding of a negative association between religion and control may reflect a relinquishing of primary control in favor of secondary control. However, the items used to measure control in the current study all refer to primary control. Much more could be learned by developing measures of secondary control and testing them within this framework.
Another limitation involves drawing conclusions about the effects of age on control using cross-sectional data. Two issues raise concern in this respect. First, the direction of causality between the predictors and control is unknown. For example, with only one wave of data, the degree to which increasing religion affects feelings of control cannot be distinguished from the degree to which declines in control may cause changes in religiosity. In addition, with only one wave of data, we cannot be sure if the observed patterns in control are associated with age, period, or cohort effects.
Finally, there are limitations in some of the study measures. For instance, greater insight into the relationship between age and control may be achieved by using better measures of religion and health. In addition, the measure of social support used in this study requires some explanation. As described in Table 1 , levels of social support from various sources (e.g., spouse, child, friends, and relatives) were assessed separately. Therefore, because not all respondents reported having relationships with people from each source, we constructed a measure that reflects support from the person most likely to be closest to the respondent (see
Krause, Goldenhar, Liang, Jay, and Maeda 1993, for a similar approach). Thus, for respondents who were married or living with a partner, the measure reflects support from this partner; for those without a significant other but who had children, the measure reflects support from their children; and for those who were without both a significant other and children, the measure reflects support from friends and relatives. This measure is based on an implicit hierarchy reflecting the differential importance of various sources of support. This means that if a significant other is present, then it is assumed that the assistance he or she provides matters most. If a respondent does not have a significant other, but has children, it is assumed that support from a child is more important than support from the remaining source assessed in this study (i.e., friends and relatives). This measurement strategy, which differentially weights support scores in favor of close family members, is based on socioemotional selectivity theory (
Carstensen 1992). According to this perspective, as people age, emotional support from family members becomes increasingly important.
The measure of control used in our study could also be improved. All three indicators available to us were keyed in the direction of externality (i.e., less control). In addition, the three items ask about feelings of personal control connected to only negative outcomes (e.g., bad things and death). However, more sophisticated instruments for measuring control include items keyed towards both external and internal locus of control, as well as both positive and negative outcomes (
Mirowsky and Ross 1991). Nevertheless, the fact that the pattern of decline in feelings of control found in our data set is similar to those found in previous research by other investigators suggests that our measure may capture at least part of the same conceptual domain assessed by more complete measures of control.
Although there are limitations in the work we have done, perhaps the greatest contribution of our research lies in our effort to cast the study of race and control in a life course perspective. We hope that this research sparks further interest in examining the significance of race, as well as other social structural factors such as gender and socioeconomic status, in the relationship between age and control.
|
Acknowledgments |
|---|
Received for publication April 3, 2000. Accepted for publication November 7, 2000.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. Cairney and N. Krause Negative Life Events and Age-Related Decline in Mastery: Are Older Adults More Vulnerable to the Control-Eroding Effect of Stress? J. Gerontol. B. Psychol. Sci. Soc. Sci., May 1, 2008; 63(3): S162 - S170. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Krause Age and Decline in Role-Specific Feelings of Control J. Gerontol. B. Psychol. Sci. Soc. Sci., January 1, 2007; 62(1): S28 - S35. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. S. London and N. A. Myers Race, Incarceration, and Health: A Life-Course Approach Research on Aging, May 1, 2006; 28(3): 409 - 422. [Abstract] [PDF]
|
|
|
|
|
|
Y. Jang, A. Borenstein-Graves, W. E. Haley, B. J. Small, and J. A. Mortimer Determinants of a Sense of Mastery in African American and White Older Adults J. Gerontol. B. Psychol. Sci. Soc. Sci., July 1, 2003; 58(4): S221 - 224. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Romualdi, D. Balding, I. S. Nasidze, G. Risch, M. Robichaux, S. T. Sherry, M. Stoneking, M. A. Batzer, and G. Barbujani Patterns of Human Diversity, within and among Continents, Inferred from Biallelic DNA Polymorphisms Genome Res., April 1, 2002; 12(4): 602 - 612. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
