The Journals of Gerontology Series B: Psychological Sciences and Social Sciences 59:P147-P150 (2004)
© 2004 The Gerontological Society of America
Social Control of Health Behaviors: A Comparison of Young, Middle-Aged, and Older Adults
Joan S. Tucker,
David J. Klein and
Marc N. Elliott
RAND Corporation, Santa Monica, California.
Address correspondence to Dr. Joan Tucker, RAND Corporation, 1700 Main Street, Santa Monica, CA 90407-2138. E-mail: joan_tucker{at}rand.org
 |
Abstract
|
|---|
Social control can positively influence health behaviors, but changes in social networks over time may cause older adults to experience less health-related social control. The size and composition of social control networks, and receipt of health-related social control, were examined in a probability sample of 509 household residents (aged 25–80 years) in Los Angeles County who completed a telephone survey. Compared with younger and middle-aged adults, older adults identified fewer people who attempted to influence their health behaviors and fewer health behaviors that others urged them to change. Older adults also reported less frequent social control attempts aimed at modifying their health behaviors, even after health status, health habits, and social network characteristics were controlled for. Possible explanations for these age-related differences are discussed.
Strong and consistent evidence points to an important association between social network involvement and better physical health among older adults (Blazer, 1982
; Seeman et al., 1993; Tucker, Schwartz, Clark, & Friedman, 1999). One mechanism through which the social environment can affect health is by influencing health-related behavior, a process referred to as social control (Rook, 1990; Umberson, 1987). Social control can take different forms, and in this study we refer to health-related social control as direct efforts by others to influence the health behaviors of network members. The receipt of social control has been associated with health-related behavior across several studies (e.g., Cohen & Lichtenstein, 1990; Lewis & Rook, 1999; Tucker & Anders, 2001; Umberson, 1992; Westmaas, Wild, & Ferrence, 2002), although few have focused on older adults in particular. These studies found that 70–80% of older adults experience health-related social control (Rook, Thuras, & Lewis, 1990; Tucker, 2002), and more frequent receipt of social control is associated with greater attempts to engage in healthy behavior change (Tucker, 2002).
Health-related social control may encourage older adults to make healthful lifestyle changes that can prevent or delay the onset of illness and disability, as well as assist them in modifying long-standing health habits in order to better manage chronic conditions. However, older age is associated with smaller, less frequently seen, and less proximal social networks (Ajrouch, Antonucci, & Janevic, 2001). On one hand, these changes may result in fewer individuals who are available to encourage older adults to engage in healthy behavior change and less opportunity for them to exert this influence. On the other hand, social ties that are retained over time tend to be particularly important ones (Fung, Carstensen, & Lang, 2001) and may be more likely to serve a social control function than peripheral ties (Rook & Ituarte, 1999; Tucker, 2002; Umberson, 1992). Understanding whether and how health-related social control differs by age is an important issue, given its relevance to physical health and functioning.
In this study, we used a probability sample of 509 household residents in Los Angeles County to compare younger, middle-aged, and older adults on the size and composition of their health-related social control networks, as well as the frequency of experiencing health-related social control. If we found older adults to experience less frequent health-related social control, our secondary goal was to determine whether this was due to their having smaller social control networks or other factors (e.g., physical health or health habits).
|
METHODS
|
|---|
Participants
We conducted a random digit dial (RDD) telephone survey, lasting an average of 30–45 min, with a probability sample of English-speaking household residents in Los Angeles County who were between the ages of 25–44 (n = 203), 45–64 (n = 154), and 65–80 (n = 152) years.
Procedure
We purchased lists of random telephone numbers (listed and unlisted) from a well-known firm specializing in developing RDD samples. Interviewers conducted interviews by using a computer-assisted telephone interviewing system to identify eligible households and to select respondents. Households were eligible if they contained at least one adult aged 25–80 years and an English-speaking informant. Interviewers attempted to determine whether the number was associated with a household and, if it was confirmed as a household, they attempted to conduct a 2-min screening interview with an adult household informant who could provide the information needed to select a respondent for the interview. In the screening interview, the interviewer introduced the study and asked the informant to provide information on the number of household members aged 25–44 years, 45–64 years, and 65–80 years. The interviewer then used a sampling strategy to select households and respondents in such a way as to generate a probability sample containing sufficient numbers of respondents in each of these age groups. We obtained an 83% cooperation rate and an estimated 51% response rate among cases given full field effort.
Measures
Participants rated how often they were urged by others to modify each of the 14 health-related behaviors listed in Table 1 (1 = never to 4 = often). If the question was applicable, the participants were asked how many individuals in each of the 10 categories of social control agents listed in Table 2 urged them to change the behavior. On the basis of this information, we calculated the following: (a) social control network size (the number of different individuals identified as social control agents); (b) social control network composition (whether participants identified a particular type of social control agent, such as friend or spouse); and (c) number of health behaviors targeted for social control attempts.
A five-item general measure of health-related social control (adapted from Tucker, 2002) asked how often someone engages or offers to engage in healthy behavior with you; does things that make it easier for you to engage in a healthy behavior; gives you advice or information on how to stay healthy; tells or reminds you to engage in a particular health-related behavior; and provides encouragement for you to engage in healthy behavior (1 = never to 4 = often; 
=.82).
Control variables included demographic characteristics (gender, education, race, marital status, number of children, and number of persons living in household), physical health (Physical Component Summary scale from the Medical Outcomes Study SF-12; Ware, Kosinski, & Keller, 1996; =.79), and a health behavior index. We created the health behavior index by summing the number of unhealthy behaviors (out of 10) reported by participants: currently smokes, has engaged in binge drinking in the past month, is overweight (body mass index 
25), does not exercise, has not had both a medical and dental exam in the past year, sleeps <6 or >9 hr per day, eats fruits and vegetables less than once per day, never drives within the speed limit, never makes time to relax, and never wears sun protection. Descriptive information on these self-reported health behaviors is provided in the Appendix.
Statistical Methods
We used design weights inversely proportionate to the probability of inclusion in all analyses to obtain unbiased estimates for the population of all eligible household residents of Los Angeles County. In our analyses, we corrected for design effects induced by the design weights by using the linearization method (StataCorp., 2001). We did not construct nonresponse weights, given that a simple model of nonresponse (based on participation rates as a function of the requested respondent's age) found no evidence of differential nonresponse (p =.65). We used the weight-corrected equivalent of the chi-squared test of homogeneity (design-based F test; StataCorp., 2001) and one-way analysis of variance for age-group comparisons on the main study variables. In cases in which we found an overall age-group difference, we conducted pairwise comparisons by using a Bonferroni-adjusted value of p <.016.
|
RESULTS
|
|---|
Overall, 47% of participants had social control networks of fewer than 3 people; 32% had networks of 3–5 people; 13% had networks of 6–9 people; and 8% had networks of 10–26 people. There was a significant age-group difference on the size of participants' social control networks (Table 3). Older adults had smaller networks than middle-aged adults (p =.01) and marginally smaller networks than younger adults (p =.06). Given that race and gender varied by age, we conducted a multiple regression analysis to confirm that older adults had smaller networks than middle-aged (p <.01) and younger (p <.07) adults, even after we controlled for race and gender. Excluding health care providers resulted in more pronounced age-related differences, with older adults reporting smaller networks than the other groups (ps <.01).
We did not find age-related differences in the percentage of respondents identifying a spouse or partner, sibling, or other relative as a source of social control (Table 2). However, middle-aged and older adults were more likely to identify a child, and younger and middle-aged adults were more likely to identify a parent. We found similar results when we restricted the sample to those who had the particular social tie in question, except that younger and middle-aged adults no longer differed in terms of identifying a child. Younger and middle-aged adults were more likely than older adults to identify friends as a source of social control. Finally, older adults were more likely than middle-aged adults, and both of these age groups were more likely than younger adults, to identify medical doctors or nurses as a source of social control.
We found an age-group difference in the number of health behaviors that others urged participants to change (Table 3), with older adults reporting fewer behaviors than the other age groups (ps <.001). When including only behaviors that others urged them to change often, older adults still reported significantly fewer health behaviors targeted for social control attempts than younger adults (p <.001) and marginally fewer than middle-aged adults (p <.02). Compared with older adults, younger adults were urged more frequently to decrease their alcohol and caffeine consumption, middle-aged adults were urged more frequently to use sun protection, and both groups were urged more frequently to eat healthier, change their sleep habits, drive slowly, visit a doctor or dentist, relax, and take care of themselves (Table 1).
Finally, we found an overall age-group difference in the frequency of experiencing health-related social control based on the five-item measure (Table 3), with older adults experiencing less frequent social control than the other age groups (ps <.001). A hierarchical linear regression analysis determined whether these differences could be accounted for by demographic characteristics, physical health, health-related behaviors, or social network characteristics. Step 1 included two age dummy variables (reference group = older adults), gender, race, education, physical health, and the health behavior index. Older adults experienced less frequent social control compared with younger (b =.40, p <.001) and middle-aged (b =.35, p <.001) adults, after we controlled for these other factors; F(7, 498) = 5.65, p <.001, and R2 =.06. When we added social control network size and two indicators of social ties (marital status and number of children) to Step 2 of the model, older adults remained less likely than younger (b =.33, p <.001) and middle-aged (b =.23, p =.01) adults to experience social control; F(10, 488) = 9.80, p <.001, and R2 =.15.
|
DISCUSSION
|
|---|
In this study we found that older adults received less frequent health-related social control, were urged to change fewer health behaviors, and had smaller social control networks compared with younger and middle-aged adults. One possible explanation is that older adults are less aware of, or less willing to acknowledge, the influence or assistance that they receive from others. However, this does not appear to be the case in that additional analyses (not reported) did not find age-related differences on perceived social support, as measured by a nine-item version of the Medical Outcomes Study Social Support Survey (Sherbourne & Stewart, 1991). Although the present study is limited by its cross-sectional design, these findings raise the possibility that the scope and frequency of health-related social control decline as individuals get older.
Older adults in this study reported fewer individuals who attempted to influence their health-related behaviors compared with younger and middle-aged adults. This is consistent with prior work showing that older age is associated with smaller and less frequently seen social networks (e.g., Ajrouch et al., 2001). However, the less frequent health-related social control reported by older adults could not be explained by their smaller social control networks (nor by their poorer health status or lesser tendency to engage in health-compromising behaviors).
We raise several possibilities that may be worth exploring in future research aimed at understanding age-related differences in health-related social control. First, this study provides a glimpse at how social control network composition differs for younger, middle-aged, and older adults. For example, older adults are less likely to identify friends as social control agents. Perhaps health-related social control occurs less frequently in long-term friendships, or perhaps older adults simply have fewer friends in their networks (Levitt, Weber, & Guacci, 1993). Such age-related differences in network composition may have important implications for the frequency of experiencing influence attempts. Second, the less frequent social control experienced by older adults may have to do with ageist stereotypes suggesting that older adults are set in their ways, curmudgeonly, and severely impaired (Hummert, 1990). Such stereotypes may discourage network members from attempting to influence the health habits of older family members and friends. Third, individuals may see a more pressing need or obligation to regulate the health habits of younger network members. Within families, for example, the provision of health-related social control may be strongly motivated by the need to help children develop sound health practices and avoid hazardous situations, and to ensure that young and middle-aged adults can adequately fulfill their financial and familial obligations. In the absence of extenuating circumstances, regulating the health behaviors of older adults may be a lesser priority for family members. Efforts to increase health-related social control will have to consider and potentially address these and other factors that may impact the ability and motivation of network members to monitor and exert a positive influence on the health habits of older adults.
|
Acknowledgments
|
|---|
The research reported in this article was funded by Grant R01AG17621 from the National Institute on Aging.
We thank Rosa Elena Garcia and the RAND Survey Research Group for their assistance in data collection.
|
Footnotes
|
|---|
Decision Editor: Margie E. Lachman, PhD
Received for publication September 4, 2003.
Accepted for publication March 1, 2004.
|
References
|
|---|
- Ajrouch, K. J., Antonucci, T. C., Janevic, M. R. (2001). Social networks among Blacks and Whites: The interaction between race and age. Journal of Gerontology: Social Sciences, 56B,S112-S118.
- Blazer, D. (1982). Social support and mortality in an elderly community population. American Journal of Epidemiology, 115,684-694.[Abstract/Free Full Text]
- Cohen, S., Lichtenstein, E. (1990). Partner behaviors that support quitting smoking. Journal of Consulting and Clinical Psychology, 58,304-309.[Medline]
- Fung, H. H., Carstensen, L. L., Lang, F. R. (2001). Age-related patterns in social networks among European Americans and African Americans: Implications for socioemotional selectivity across the life span. International Journal of Aging Human Development, 52,185-206.
- Hummert, M. L. (1990). Multiple stereotypes of elderly and young adults: A comparison of structure and evaluations. Psychology and Aging, 5,182-193.[Medline]
- Levitt, M. J., Weber, R. A., Guacci, N. (1993). Convoys of social support: An intergenerational analysis. Psychology and Aging, 8,323-326.[Medline]
- Lewis, M. A., Rook, K. S. (1999). Social control in personal relationships: Impact on health behaviors and psychological distress. Health Psychology, 18,63-71.[Medline]
- Rook, K. S. (1990). Social networks as a source of social control in older adults' lives. In H. Giles, N. Coupland, & J. Wiemann (Eds.), Communication, health and the elderly (pp. 45–63). Manchester, England: University of Manchester Press.
- Rook, K. S., Ituarte, P. H. G. (1999). Social control, social support, and companionship in older adults' family relationships and friendships. Personal Relationships, 6,199-211.
- Rook, K. S., Thuras, P. D., Lewis, M. A. (1990). Social control, health risk taking, and psychological distress among the elderly. Psychology and Aging, 5,327-334.[Medline]
- Seeman, T. E., Berkman, L. F., Kohout, F., Lacroix, A., Glynn, R., Blazer, D. (1993). Intercommunity variations in the association between social ties and mortality in the elderly: A comparative analysis of three communities. Annals of Epidemiology, 3,325-335.[Medline]
- Sherbourne, C. D., Stewart, A. L. (1991). The MOS Social Support Survey. Social Science and Medicine, 32,705-714.
- StataCorp. (2001). Stata Statistical Software: Release 7.0. College Station, TX: Author.
- Tucker, J. S. (2002). Health-related social control within older adults' relationships. Journal of Gerontology: Psychological Sciences, 57B,P387-P395.
- Tucker, J. S., Anders, S. L. (2001). Social control of health behaviors in marriage. Journal of Applied Social Psychology, 31,467-485.
- Tucker, J. S., Schwartz, J. E., Clark, K. M., Friedman, H. S. (1999). Age-related changes in the associations of social network ties with mortality risk. Psychology and Aging, 14,564-571.[Medline]
- Umberson, D. (1987). Family status and health behaviors: Social control as a dimension of social integration. Journal of Health and Social Behavior, 28,306-319.[Medline]
- Umberson, D. (1992). Gender, marital status and the social control of health behavior. Social Science and Medicine, 34,907-914.
- Ware, J. E., Jr., Kosinski, M., Keller, S. D. (1996). A 12-item short-form health survey: Construction of scales and preliminary tests of reliability and validity. Medical Care, 34,220-233.[Medline]
- Westmaas, J. L., Wild, T. C., Ferrence, R. (2002). Effects of gender in social control of smoking cessation. Health Psychology, 21,368-376.[Medline]
This article has been cited by other articles:
|
|
|
|
 
T. A. Schroepfer
Social Relationships and Their Role in the Consideration to Hasten Death
Gerontologist,
October 1, 2008;
48(5):
612 - 621.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
Abstract
Methods
