| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
a RAND, Santa Monica, California
Joan S. Tucker, RAND, 1700 Main Street, P.O. Box 2138, Santa Monica, CA 90407-2138 E-mail: joan_tucker{at}rand.org.
Decision Editor: Margie E. Lachman, PhD
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionAppendixReferences |
|---|
It is well established that the physical benefits of a healthy lifestyle are not limited to younger individuals; engagement in healthy behaviors is also associated with lower morbidity and mortality risk for older adults (Amir 1987
; Davis et al. 1994; Kaplan and Haan 1989). Some older adults attempt to engage in a healthy lifestyle with the intention of avoiding or delaying the onset of illness and disability. Most older adults, however, have at least one chronic illness such as arthritis, hypertension, or heart disease that may require significant and long-term lifestyle changes (Collins 1997). The importance of engaging in a healthy lifestyle, combined with the potential difficulty of modifying long-standing health habits, makes it particularly critical to identify and understand factors that can facilitate healthy behavior change among older adults.
Research on the determinants of a healthy lifestyle has come from several different perspectives (e.g., Taylor, Repetti, and Seeman 1997; Tucker et al. 1995). Much of this research has examined social relationships as potentially important influences on lifestyle, largely focusing on the existence of social ties or social support and often finding weak or inconsistent associations with engagement in health behaviors (S. Cohen 1988; House, Umberson, and Landis 1988; Potts, Hurwicz, Goldstein, and Berkanovic 1992; Uchino, Cacioppo, and Kiecolt-Glaser 1996). However, relationships serve functions other than support and companionship that may be relevant to health behavior. Social control is one such function that has received little empirical attention but promises to play a key role in understanding how social relationships affect engagement in health behaviors.
Social control theory proposes that relationships serve a regulatory function such that socially integrated individuals are less likely than those who are socially isolated to engage in risky or deviant behavior (Anson 1989; Durkheim 1897/1951; Ewart 1991). Applied to health behaviors, individuals who are socially integrated should be more likely to engage in healthy behaviors and avoid unhealthy behaviors, ultimately resulting in better health and greater longevity. The social control of health behavior may operate in two basic ways (Rook, Thuras, and Lewis 1990; Umberson 1987, Umberson 1992). Direct social control involves requests, reminders, threats, or rewards from significant others that prompt individuals to engage in healthy behavior. Indirect social control involves feelings of obligation or responsibility to others that encourage engagement in a healthier lifestyle. Of course, social network members may also encourage engagement in unhealthy behavior, as the literature on peer influences on substance use clearly indicates (Hawkins, Catalano, and Miller 1992). However, the dominant focus of social control theory and research has been on the health-promoting effects of social regulation, which is also the focus of this study.
The few studies examining social control in the context of health behavior have focused primarily on direct social control, providing encouraging (although not entirely consistent) evidence for its relevance to health-related behavior. Umberson 1992 conducted the only published study using a nationally representative sample, asking respondents, "How often does anyone tell or remind you to do anything to protect your health?" Results indicated that direct social control was prospectively associated with engagement in certain health-related behaviors (e.g., cigarette smoking), but not others (e.g., alcohol consumption). Studies of married couples have indicated that intentional social control attempts (such as reminding) by one spouse are associated with greater medication adherence (Doherty, Schrott, Metcalf, and Iasiello-Vailas 1983) and abstinence from smoking (S. Cohen and Lichtenstein 1990). The use of positive social control strategies (e.g., positive reinforcement) is also associated with a partner's greater general tendency to engage in healthy behavior, although the use of negative strategies (e.g., criticizing) may backfire (Tucker and Anders 2001). A study of older adults did not find the expected associations between social control and engagement in health behaviors (Rook et al. 1990), perhaps due to the relatively low prevalence of poor health and unhealthy practices in the sample. A subsequent study by these researchers (Lewis and Rook 1999) asked participants to report on a specific situation in which someone tried to influence them to change a health-related behavior, finding that participants who received more social control reported a greater tendency to make positive behavioral changes in response to these regulatory attempts.
Although experiencing social control may have positive behavioral effects by encouraging individuals to engage in a healthier lifestyle, there has been some suggestion in the literature that it may come at a price in terms of eliciting psychological distress (Hughes and Gove 1981; Rook and Pietromonaco 1987). This dual-effects hypothesis suggests, for example, that targets of social control attempts may experience feelings such as irritation and resentment, even as they attempt to comply with others' wishes for them to modify their health habits. The potential for social control to elicit psychological distress may be particularly high for older adults who are already experiencing declines in their sense of control (Mirowsky 1995); attempts by others to regulate their health behaviors may be perceived as overly intrusive and a threat to their autonomy.
Recent research has suggested that the dual-effects hypothesis may be too simplistic a description of the range of responses to social control. A recent study of married couples, for example, found that direct social control from a spouse was associated with both positive and negative emotional and behavioral reactions by the partner, depending on the type of social control strategy that was used and the frequency of social control attempts (Tucker and Anders 2001). Other work has suggested that reactions to social control attempts may depend on the extent to which these attempts are perceived by the target as being motivated by interest, caring, or concern for them (Holmila 1991; Tucker and Mueller 2000). Indeed, research on compliance has indicated that people are more likely to comply with requests from individuals for whom they hold positive feelings (Cialdini 1994)—and the most commonly used strategy for attempting to influence others' health behaviors involves making direct requests for behavior change (Tucker and Mueller 2000). Together, this work highlights the importance of better understanding the conditions under which social control has behavioral and psychological benefits, as well as the conditions under which it is likely to backfire. The present study focuses on older adults' satisfaction with their relationships as a potentially important moderator of whether their experiences of social control tend to elicit positive or negative responses.
Certain relationships may be more relevant than others in terms of the social control of health behaviors. For example, Rook and Ituarte 1999 found that social control was more likely to be provided to older adults by family members than by friends. Two additional studies collected more detailed information on social control networks but did not specifically focus on older adults (Lewis and Rook 1999; Umberson 1992). Although both of these studies found that spouses were identified most often as social control agents, they differed in terms of the extent to which other types of relatives versus nonrelatives were identified as providing social control. A more detailed understanding is needed of the size and composition of older adults' social networks, including how these networks might differ for those who are married versus unmarried, as well as those who have children versus no children. If there is compensation within older adults' social networks for the loss or absence of these close familial ties, this may result in important subgroup differences in social control network characteristics.
The first goal of this study was to describe the size and composition of older adults' social control networks as a function of their marital and parental statuses. Previous research has indicated that unmarried individuals tend to have smaller social networks, less frequent contact with network members, and a lower proportion of kin in their networks compared with married individuals (Ajrouch, Antonucci, and Janevic 2001), decreasing the pool of potential social control agents. Unmarried individuals also lack the one person (a spouse) most likely to serve as a social control agent. For these reasons, I expected that unmarried participants in this study would report smaller social control networks compared with married participants. On the basis of previous research suggesting that social control is more likely to be provided to older adults by family members than by friends (Rook and Ituarte 1999), I also expected that among older adults who were married and/or had children, immediate family members would be identified as social control agents more often than other relatives, friends, and doctors. Further, I expected that older adults who were both unmarried and childless would be more likely than those who had immediate family ties to identify other relatives, friends, and doctors as social control agents.
The second goal was to investigate older adults' behavioral and psychological responses to others' attempts to influence their health behaviors. Four behavioral responses to social control attempts were examined: trying to engage in the desired behavior, ignoring the social control attempt, doing the opposite of what the social control agent wants, and hiding the unhealthy behavior from the social control agent. The two psychological responses to social control were positive and negative affect. Because of their stronger feelings of obligation and responsibility to others to be healthy, older adults who experienced more indirect social control were expected to have more positive (and less negative) behavioral and affective responses to others' attempts to influence their health behaviors. Any interactions of indirect social control with relationship satisfaction were expected to show even stronger effects for those with high relationship satisfaction. In contrast, older adults who experienced more direct social control from others were expected to have more positive (and less negative) behavioral and psychological responses to others' attempts to influence their health behaviors only if they were satisfied with their relationships. For those who were dissatisfied with their relationships, experiencing greater direct social control was expected to backfire, resulting in more negative (and less positive) behavioral and psychological responses to others' influence attempts.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionAppendixReferences |
|---|
). However, it should also be noted that a questionnaire was sent only to those individuals who took the initiative in returning the postcard indicating that they were interested in the study. The response rate would likely have been higher had all prospective participants received the questionnaire and follow-up reminders. It is also possible that response rates for mail surveys such as this one may be lower for older adults to the extent that they are more likely to have cognitive and/or physical impairments that would prohibit completion of a lengthy mail survey. It is a limitation of this study that little information is available on the reasons for nonresponse. Two individuals were eliminated from the analyses due to excessive missing data, resulting in a final sample size of 181. The sample was 54% male and 90% White (see Table 1 for other demographic information). Self-reports of current engagement in health-related behavior indicated that 10% of participants smoked, 7% drank excessively (three or more drinks at least three times per week, or five or more drinks at least once per week), 39% never exercised, 41% were overweight (body mass index greater than 27), and 23% did not get a yearly physical exam.
|
|
Social control
To obtain information on characteristics of the participants' social control networks, they identified up to 10 individuals (by initials and relationship) who they perceived to influence their health behaviors by encouraging them to engage in healthy behavior or discouraging them from engaging in unhealthy behavior. This information was used to determine the size of participants' social control networks (total number of individuals mentioned), as well as to determine whether participants identified any of the following as a social control agent: spouse, child, relative other than spouse or child, friend, and doctor.
Participants also rated their overall experience of direct and indirect social control on five- and four-item scales, respectively, that were developed for this study and are shown in the Appendix (scale anchors ranged from: 1 = never to 4 = often; 
> .80 for each scale). Evidence for the convergent and discriminant validity of the Direct Social Control scale comes from an examination of its association with other measures collected as part of the larger study. The Direct Social Control Scale correlated more strongly with the single-item measure of direct social control previously used by Rook and colleagues 1990(r = .60, p < .001) than with any of the five subscales of the Social Support Behaviors Scale (Vaux, Riedel, and Stewart 1987; rs ranged from .27 to .30, p < .001).
Behavioral responses to social control
Participants were asked to rate how often they responded in the following ways to others doing or saying things to try to get them to engage in healthy behavior or to avoid unhealthy behavior: (a) attempt to engage in the desired behavior, (b) ignore the person or do nothing, (c) do the opposite of what the person wants them to do, or (d) hide the unhealthy behavior from the person (ranging from 1 = never to 4 = often). These items have been used in previous research on married couples, finding that spouses' ratings on these items correlated moderately to strongly with their partners' perceptions of their behavioral responses to social control attempts (Tucker and Anders 2001).
Affective responses to social control and psychological well-being
Affective responses to social control were assessed by means of a 12-item mood scale (Brunstein, Dangelmayer, and Schultheiss 1996). Participants rated how often they felt positive affect (e.g., happy, loved, valued) and negative affect (e.g., resentful, irritated, embarrassed) when others attempted to influence their health behaviors. Items were rated on a 4-point scale (ranging from 1 = never to 4 = often) and averaged to form separate measures of positive and negative affect ( > .80).
Relationship satisfaction
Participants were asked to rate their level of satisfaction with each of the following individuals or groups, as relevant: (a) their spouse, (b) their child(ren), (c) other relatives, and (d) friends and peers (ranging from 1 = extremely dissatisfied to 7 = extremely satisfied). These items were averaged to form an overall measure of relationship satisfaction ( = .79).
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionAppendixReferences |
|---|
|
Associations of Social Control With Behavioral Responses
Table 4 shows results from separate hierarchical regression analyses predicting each of the four behavioral responses to social control from overall relationship satisfaction, direct social control, and indirect social control. These variables were simultaneously entered into each regression model at Step 1, controlling for sex, level of education, and marital status (note that age was not included as a control variable because of its lack of association with the other study variables). The two Social Control x Relationship Satisfaction interaction terms were then added to the models at Step 2. Results indicated that older adults who experienced more direct and indirect social control tended to report more frequent attempts to engage in healthier behavior, regardless of their level of relationship satisfaction. In addition, older adults who experienced more indirect social control reported that they ignored social control attempts or did nothing less frequently; again, this was the case regardless of their level of relationship satisfaction. However, relationship satisfaction moderated associations between social control and hiding unhealthy behavior, with this moderating effect operating differently for the two types of social control.
|
, additional analyses were conducted to determine how the relationships of direct and indirect social control with hiding unhealthy behaviors changed at high, medium, and low levels of relationship satisfaction. The regression lines for the measure of hiding unhealthy behaviors on direct and indirect social control for the three levels of relationship satisfaction were calculated and are plotted in Fig. 1. The three levels of relationship satisfaction were the mean value for the sample, a high value 1 standard deviation above the mean, and a low value 1 standard deviation below the mean. Older adults who experienced more direct social control reported more frequent hiding of their unhealthy behaviors if they had low relationship satisfaction and less frequent hiding if they had high relationship satisfaction. In contrast, participants who experienced more indirect social control reported more frequent hiding of their unhealthy behaviors if they had high relationship satisfaction, but less frequent hiding if they had low relationship satisfaction. There were no significant associations between experiences of either type of social control and doing the opposite in response to social control attempts by others.
|
|
|
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionAppendixReferences |
|---|
), the larger social control networks among married participants in this study could be explained by the inclusion of spouses in these networks. It is unclear the extent to which the inclusion of the spouse in the social control networks might be associated with greater health benefits for married older adults. Certainly, one of the most consistent findings in the literature on social relationships and health is that married individuals tend to live longer and healthier lives compared with those who are unmarried (e.g., Berkman and Syme 1979; House, Robbins, and Metzner 1982; see review by Burman and Margolin 1992). Although there are several explanations for this effect (Tucker, Friedman, Wingard, and Schwartz 1996), it is possible that married individuals tend to be healthier because they feel greater responsibility to engage in a healthy lifestyle owing to their marital role obligations (indirect social control) as well as to their spouse's efforts to monitor and positively influence their health-related behaviors (direct social control).
Married participants in this study were most likely to identify their spouse as a social control agent, followed by their children, other relatives, friends, and doctor. This finding is consistent with previous research indicating that older adults are more likely to receive social control from family than friends (Rook and Ituarte 1999). In contrast, Umberson 1992 reported that married individuals most often identified their spouse as a social control agent, followed by unrelated individuals, parents, and children. Given that Umberson's sample was aged 24 years or older (mean age = 54 years), it is possible that these differences reflect developmental changes in social control networks. Compared with younger adults, older individuals are less likely to have a parent available to monitor and regulate their health behavior, but their children may be more likely to assume this role. The narrowing of social networks at older ages, including the discontinuation of relationships that are less close or important (Lang 2000, Lang 2001), may also play a role in the greater likelihood of social control agents being relatives than nonrelatives.
Two interesting findings emerged regarding the composition of older adults' social control networks. Although spouses and children were most often identified as social control agents by older adults who had these social ties, between 40% and 45% of older adults with these ties did not mention them as having a positive influence on their health behaviors. It is interesting to note that of the participants who had these ties but did not identify them as social control agents, 19% mentioned someone else (e.g., a friend or other relative) as a social control agent. The failure to identify spouses and children as having a positive influence on health behaviors may be partially due to a lack of awareness on the part of some older adults as to the impact that these family members have on their lifestyle. However, it may also reflect untapped resources within the social networks of many older adults; namely, immediate family members who are not currently playing an active role in the health-related aspects of their older relatives' lives but are potentially available to encourage and assist them in initiating and maintaining healthy lifestyle changes. Of course, it may be the case that some of these older adults led such exemplary lifestyles that spouses and children had no reason to engage in social control attempts. Sorting through these explanations is beyond the scope of this study, although it would be useful for future research to explore the reasons why a substantial minority of older adults fail to identify immediate family members as exerting a positive influence on their lifestyle.
A second interesting finding regarding these networks is that their composition differed somewhat as a function of marital and parental statuses. Friends and doctors were identified more often as social control agents by older adults who lacked both a spouse and children, although they were not more likely to identify other relatives as social control agents. Previous research on older adults has demonstrated that social ties other than marriage become stronger predictors of mortality risk with increasing age, at least for women (Tucker, Schwartz, Clark, and Friedman 1999; Yasuda et al. 1997), suggesting that the absence of certain social ties may lead to compensation within social networks in ways that have relevance to health status. Results from the present study also suggest that compensation may be occurring in the social control networks of older adults in the absence of immediate family members. Friends and doctors may choose to play a more active role in monitoring and influencing the health behaviors of an older adult if they know that immediate family is not available to take on this responsibility or because their assistance has been solicited. It is also possible that these network members are not more likely to engage in social control attempts with older adults lacking immediate family; rather, the influence of friends and doctors may be more apparent to older adults lacking immediate family or they may have a stronger impact on the behavior of this group. Gaining a better understanding of social control networks is important; for example, additional analyses (not shown) indicate that although these three groups did not differ in terms of social control network size or feelings of indirect social control, those lacking both a spouse and children reported experiencing significantly less direct social control than the other two groups. This effect may be largely due to differences in the composition of social control networks across groups.
Consistent with social control theory, this study found that experiencing more direct social control was associated with more frequent attempts to engage in healthy behavior change. However, for older adults with low relationship satisfaction, experiencing more direct social control had the potential to backfire: These adults had a greater tendency to experience negative affect, as well as to hide their unhealthy behaviors from others. This latter finding raises two interesting possibilities. One possibility is that older adults who are dissatisfied with their relationships experience more negative types of direct social control, such as nagging or threats. Another possibility is that these dissatisfied individuals put a more negative spin on others' social control attempts, such as perceiving them as being selfishly motivated or a threat to their autonomy. In either case, the experience of direct social control may exact a toll for these older adults, even as it is effective in prompting them to engage in healthy behavior change. It is unclear why experiencing more direct social control was unrelated to positive affect, particularly in light of previous work finding that young spouses who experienced more frequent direct social control attempts from their partner reported greater positive affect (Tucker and Anders 2001). It may be the case that the social control attempts typically experienced by older adults tend to be relatively mundane (e.g., reminders to take medications) or come to be expected as a natural part of their everyday interactions with family and friends. As such, experiencing them more frequently might not necessarily make the individual feel more loved, valued, inspired, and so forth.
Consistent with expectations, older adults who had stronger feelings of obligation and responsibility to others to stay healthy reported more frequent positive responses to others' attempts to influence their health behaviors. In the case of positive affect, this association was particularly strong for those with high relationship satisfaction. However, an unexpected finding is that they were also more likely to have certain negative responses to social control attempts if they had high relationship satisfaction and less likely to have these negative responses if they had low relationship satisfaction. It is not entirely clear why social control might elicit negative responses among older adults who have a strong sense of obligation to others and are satisfied with their relationships. It is possible that attempts by family and friends to regulate their health behaviors indicate to them that they are not fully living up to their responsibilities and obligations or have disappointed their loved ones. If this is the case, then these social control attempts would likely make them feel bad and encourage them to hide their unhealthy behaviors, although not necessarily encourage them to ignore the person attempting to influence their behavior or do the opposite of what this person wants them to do. Much has been written of the health risks associated with having low social support or high conflict within personal relationships. That satisfying relationships may also have adverse effects on the well-being of older adults under certain conditions is a provocative finding and worthy of further investigation.
Several limitations of this study should be noted. The sample was mostly White and may not have been representative. The response rate for completing the questionnaire was less than optimal, and the reasons for nonresponse are not known. Another limitation is the study's cross-sectional design; further research is needed to tease apart causal associations between the receipt of social control and engagement in health-related behavior. This study also relied exclusively on retrospective self-report measures of social control and health-related behaviors. Although this has been the standard practice for research on social control in the health domain, future research would benefit from the use of alternative assessment techniques, such as behavioral observation or diary methodology. The measures of social control were also developed specifically for this study, and additional work is needed to further establish their reliability and validity. Finally, it is a limitation that relationship satisfaction was not assessed with specific reference to social control agents. Not all of the individuals encompassed in the relationship satisfaction measure served as social control agents for the respondents, and certain social control agents such as doctors were not included in the relationship satisfaction measure. Nonetheless, this measure is probably a reasonable proxy for older adults' relationship satisfaction with social control agents; this assumption is bolstered by the high internal reliability of the measure, indicating that individuals who were satisfied (or dissatisfied) with one relationship type tended to be satisfied (or dissatisfied) with other types of relationship, presumably including social control agents.
In summary, results from this study suggest that attempts by others to regulate the health behaviors of older adults may have positive behavioral consequences, even for those with low relationship satisfaction. Relationship satisfaction is more important to consider in terms of understanding the potentially negative consequences of social control attempts, such as eliciting negative affect or encouraging older adults to hide their unhealthy behaviors from others. The impact of relationship satisfaction on responses to social control appears more complex than originally anticipated, operating differently depending on whether the social control involves direct attempts by others to regulate the older adults' health behaviors or their feelings of obligation and responsibility to engage in a healthier lifestyle.
|
Acknowledgments |
|---|
I thank Robert Leeman for assistance with data collection and Sherry Anders for her assistance in developing the social control scales.
Received for publication June 25, 2001. Accepted for publication December 27, 2001.
|
Appendix |
|---|
|
TOPAbstractMethodsResultsDiscussionAppendixReferences |
|---|
Direct Social Control
Indirect Social Control
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionAppendixReferences |
|---|
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]
|
|
|
|
 |
|
 J. S. Tucker, M. N. Elliott, and D. J. Klein Social Control of Health Behavior: Associations With Conscientiousness and Neuroticism Pers Soc Psychol Bull, September 1, 2006; 32(9): 1143 - 1152. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
