Robert G. Kent de Grey and Bert N. Uchino
Department of Psychology, University of Utah, Salt Lake City, UT, USA
In his Nicomachean Ethics, Aristotle is credited as stating, “a friend is a second self.” Friendships, however, entail importance beyond the pleasant affective experience of having close others. Social ties are critical to cognitive and affective development across the lifespan and have benefits beyond this as well. In a recent meta‐analysis involving more than 308,000 participants, the quality of one's social relationships was linked with reduced risk of mortality (Holt‐Lunstad, Smith, & Layton, 2010). The magnitude of this association was comparable with smoking cessation and exceeded many other conventional risk factors (e.g., physical inactivity, obesity).
Although most of the prior epidemiological work has not separated out specific relationship types (e.g., family, friends), there are strong reasons to believe that friendships in particular might be important to health. In contrast to relationships with family members, friendships are voluntary (nonbinding) relationships and thus are an important source of support across a variety of situations. Of course, this does not mean that friends have entirely positive influences. Under some circumstances friendships may adversely relate to health. In the present entry, we thus provide a general overview on links between friendships and health, including potential mechanisms, individual differences, intervention opportunities, and important future directions.
Research suggests friendships may decrease risk for health problems and early mortality. These health benefits are evident across the lifespan and can be seen in early life friendships. In a long‐term study, researchers assessed the number of friendships reported by 5,814 Scottish schoolchildren and linked them to self‐rated health 36–39 years later (Almquist, 2012). For women, a greater number of childhood friends correlated with higher ratings of health. Among male participants, those who listed no friends at the time of the initial survey reported lower adult health quality. Finally, another large‐scale study of 11,617 adolescents reported on their experiences with close friends. Fourteen years later, analysis of HbA1c (glycohemoglobin, an index of blood glucose levels associated with, and used in the monitoring of, diabetes mellitus) showed that, for female participants, having close male friendships was associated with reduced odds of elevated metabolic risk marker levels (Ehrlich, Hoyt, Sumner, McDade, & Adam, 2015).
It is also important to note that the health correlates and consequences of friendship can sometimes approach or exceed the effect size for social support from family members. For instance, in an analysis using data from the longitudinal Health and Retirement Study, 7,514 US adults were surveyed in 2006 regarding social environmental variables. These responses were used to predict the onset of six conditions (high blood pressure, heart condition, cancer, stroke, lung disease, and arthritis) in 2010. Factors related to friends were more predictive of these health outcomes than factors related to family (Hill, Weston, & Jackson, 2014). In summary, although more work is needed, the available epidemiological evidence suggests that friendships are consequential for health.
Mechanisms underlying the association between friendships and health are made salient by two major perspectives: the stress‐buffering and direct effects models. Of these theoretical approaches, the more well known is the stress‐buffering hypothesis, which holds that social support is healthy primarily because it buffers against the damaging effects of stress. Stress arises in part from the appraisal process—the evaluation of our coping abilities and the demands placed on them. Social support may decrease the perceived and objective coping burdens relative to available resources, as support has been shown to lessen the negative effects of many types of stressors.
The stress‐buffering model has been tested with regard to friends in well‐controlled laboratory studies. For instance, Christenfeld et al. (1997) examined this by stressing participants with a speech task and assigning them to receive support from a friend or stranger (confederate), or to receive no support. Social support from a friend resulted in the largest reduction in cardiovascular reactivity during stress. Compellingly, supportive behaviors were standardized across conditions, suggesting that differences were not due to friends merely knowing participants more intimately and providing better support on the basis of this knowledge. Although support from friends in such studies is generally beneficial, some studies have revealed more complex results and led to the suggestion that future work should pay closer attention to the support context, for example preferences for support.
In addition to the stress‐buffering hypothesis, another major perspective is the direct effects model. This perspective broadly emphasizes the overall benefits of integration into a social network. Social control is one important aspect of this model and suggests that embedment in a social network is healthy because it creates social roles that provide self‐esteem and meaning in life. Friends and others may also exert direct social control by attempting to influence individuals to change their health behaviors. Consistent with the social control hypothesis, an analysis of 469 elderly members of a Thai community found friends (but not family) had a direct link to health promoting behaviors such as exercise, healthful eating, and stress management (Thanakwang & Soonthorndhada, 2011).
Friendships may also be related to health behaviors that are not completely voluntary. A relatively large literature suggests aspects of sleep, such as quality and duration, are related to health outcomes, including mortality rates. In a community sample of 175 US middle‐aged to older adults, participants reported how many friends/family ties they had and characterized the quality of those friendships. The number of supportive friendships (but not supportive family members) was a significant predictor of better self‐reported sleep quality (Kent, Uchino, Cribbet, Bowen, & Smith, 2015).
While the overall findings for the direct effects model suggest that friendships can contribute to positive health behaviors, relationships can have negative consequences in some circumstances (e.g., risk taking and substance abuse). In addition, overt social control may arouse resentment or serve as a source of significant stress if the individual is unwilling or unable to engage in healthy behavioral change. For instance, one study found that perceptions of direct social network control (including friends) was related to greater aging at the cellular level (Uchino et al., 2015). Similarly, research suggests that receiving highly visible social support from friends or others may harm self‐esteem or contribute to depression, both of which may be detrimental to health. Thus, high levels of direct social control and highly visible or unsolicited social support appear to relate to negative health outcomes.
Friendships may also have negative influences on health via normative influences (Rook & Pietromonaco, 1987). For example, the social interconnection of 12,067 adults in the Framingham Heart Study was repeatedly assessed along with body mass index (BMI) from 1971 to 2003. Longitudinal models showed that the chances of becoming obese increased 37% if one's spouse became obese and rose by 40% if one's adult sibling became obese, but the largest increase was a 57% greater risk of obesity if a participant had a friend who became obese during the study (Christakis & Fowler, 2007). This social contagion of bad health habits via friends is consistent with findings indicating that friends are surprisingly similar on measures such as frequency and duration of participation in organized physical activity, fast‐food habits, body image, and alcohol use (see, e.g., Fletcher, Bonell, & Sorhaindo, 2011). This phenomenon of friend similarity in alcohol use is well documented in university settings, where social norms have been found to exert large influence on drinking behaviors, even when such norms do not reflect the beliefs of most individual students.
A number of factors may influence the relationship between friendship and health. Men and women differ in important ways that may relate to friendship. One possibility is that females may be more sensitive to the quality of their relationships due to more complex relational schemata. For instance, women may be more affected by adverse interpersonal processes because their self‐esteem is more contingent on interpersonal functioning. Women may thus be more sensitive to or expend greater effort in response to relationship concerns. This may translate to health‐relevant physiological processes as well; a study of 303 normotensive adults found that conflict within close relationships predicted nocturnal blood pressure dipping for women, but not for men (Holt‐Lunstad, Jones, & Birmingham, 2009). Absence of such nightly decreases in blood pressure has been shown to predict both cardiovascular and all‐cause mortality.
Exploration of cultural differences is also an area of increasing interest in relationships and health. Importantly, friendship appears to have value and health relevance across many nations, though its specific relation to aspects of well‐being may vary. Most of this work has focused on differences between individualistic and collective (especially Asian) cultures. In such work, there is a significant distinction between explicit (i.e., seeking and using support) and implicit (i.e., the comfort of simply having close relationships) support. Due to relational concerns (e.g., worries about placing a burden on friends who are obligated to be helpful), individuals from more collective cultures may experience greater benefit during stress from implicit forms of support, in which they simply feel more connected with others.
The research reviewed thus far suggests that interventions targeting friendships might be useful in improving health outcomes. Although more work is certainly needed in this area, several intervention approaches are consistent with this possibility. “Befriending” specifically focuses on the friendship formation processes by recruiting outside relationships to serve as the basis for affirming social support over time. A meta‐analysis suggests a modest effect for such interventions in improving outcomes in different patient populations (e.g., caregivers, prostate cancer patients; Mead, Lester, Chew‐Graham, Gask, & Bower, 2010).
Another important intervention related to building friendships involves support groups. Interventions examining peer support groups suggest beneficial effects on adjustment and well‐being (Hogan, Linden, & Najarian, 2002). Researchers have also found that participation in support groups seems to influence the size and composition of one's friendship network. For instance, studies of individuals with substance abuse problems suggest that self‐help groups result in decreased contact with drug‐using network members and increased contact with support group friends (Humphreys & Noke, 1997). These network members then become crucial sources of support in offering advice and guidance to help individuals remain abstinent.
Success has also been shown for relationship interventions that teach individuals the skills necessary to acquire responsive support (Hogan et al., 2002). In one intervention, adolescents were assigned to enroll in a semester‐long course that focused on friendship development as well as positive peer and teacher relationships (Eggert, Thompson, Herting, Nicholas, & Dicker, 1994). Students also learned how to elicit support regarding personal problems. Results of this study showed that individuals in the intervention had an increase in their friendships and in self‐esteem. Moreover, compared with adolescents not provided with the intervention, individuals with social skills training showed a trend toward less drug use, a decrease in drug‐related problems, and an increase in grade point average.
More generally, friends may be an important source of support for individuals attempting healthy behavioral change. In one study, researchers combined a support intervention with standard behavioral treatment (SBT) for weight loss (Wing & Jeffery, 1999). Participants either came alone or were asked to bring three friends who were interested in weight loss. Importantly, individuals who participated in the support intervention with friends were more likely to complete the program and maintain their weight loss over the 10‐month period. One interesting aspect of this study was the finding that individuals who recruited three friends to join them for the protocol were less likely to have participated in prior organized weight‐loss programs. These data suggest an additional benefit of recruitment with friends as it may increase participation for individuals who may not wish to go alone. In summary, although more work is needed, the existing literature highlights the promise of friendship‐based interventions to improving health outcomes.
As reviewed in this entry, friendships can have both positive and negative influences on health. Of course, this is true of family relationships as well, and future work that directly contrasts different types of relationships and the functions they serve will be important to advance theory and highlight unique intervention opportunities. Future work should also examine the utility of considering positive and negative aspects of relationships in combination. Corresponding to the colloquial concept of the “frenemy,” an ambivalent friend is one characterized by both high levels of positivity and negativity. Although one might ask why we would continue to be friends with such ambivalent ties, research indicates that individuals have a sense of commitment to them based on a long history of interactions. Importantly, existing research indicates that friendship ambivalence is quite common and related to negative health‐relevant biological outcomes (Holt‐Lunstad, Uchino, Smith, & Hicks, 2007).
It is also impossible to ignore the fact that an increasing proportion of social interaction takes place online. Scientific examination of differences between online and offline friendships is still in its infancy. People may be more willing to self‐disclose when getting to know someone online vs. in person, which could facilitate friendship formation. The Internet may also preserve friendships that would otherwise deteriorate over time. In one study, the number of Facebook friends was a stronger predictor (vs. offline interpersonal network size) of perceived social support, less physical illness, and greater well‐being among participants with a high number of objective life stressors (Nabi, Prestin, & So, 2013). While causality in such research presently remains largely unknown, research indicates that Facebook members who use the site multiple times daily score eight percentage points higher than average Americans on a measure of social support (Hampton, Goulet, Rainie, & Purcell, 2011). The same researchers interpret their findings to mean that “someone who uses Facebook multiple times per day gets about half the boost in total support that someone receives from being married or living with a partner” (Hampton et al., 2011, p. 35). While this claim requires a great deal of additional research before substantiation, such a bold statement hints at the potential importance of this emerging research area.
As mentioned above, much existing epidemiological research does not separate specific relationship types (e.g. friends, family, coworkers). Future work should distinguish between relationship categories to show how friendships may differ from other social ties in terms of health relevance. When relationship types are aggregated, researchers may not be able to detect effects that pertain to specific kinds of social ties (see, for example, Kent et al., 2015).
Friendship may serve as an important source of social support during stress, promote healthy behaviors, and contribute to healthful social norms, all of which may link to positive health outcomes. However, unsolicited, intrusive, or highly visible support attempts, as well as direct social control, may be mechanisms by which friendships can adversely influence health. Additional research is needed to more fully understand the health consequences and mechanisms associated with friendship, which can rival family relationships in terms of health relevance.
Robert G. Kent de Grey, MS, is a PhD candidate in social and health psychology at the University of Utah studying relationships and health. He also researches in the Cancer Control and Population Sciences Program at the Huntsman Cancer Institute. Specific subareas of interest include online social networks and virtual interactions, as well as proenvironmental attitudes and behaviors. He serves as a consulting environmental and social psychologist to the University of Utah Sustainability Office.
Bert N. Uchino, PhD, is a professor of social and health psychology at the University of Utah. His graduate studies at The Ohio State University were supervised by Dr. John T. Cacioppo. Uchino also researches the influence of social factors on aging processes as part of the Interdepartmental Program in Neuroscience at the University of Utah. In 2004, Yale University Press published his book Social Support and Physical Health: Understanding the Health Consequences of Relationships.