Positive and negative social exchanges and mental health across the transition to college

Abstract

There is evidence that positive and negative social exchanges have different implications for various psychological outcomes, but research has been limited by a lack of multidimensional and parallel measures and underlying processes explaining the implications are not well understood. In the present longitudinal study, we address these gaps in the literature by using parallel measures of positive and negative social exchanges to predict both positive and negative aspects of emotional health in a sample of 113 first-year college students and by testing for loneliness as a mediator. Consistent with a domain-specific effects model, increases in negative exchanges were associated with increases in negative but not positive emotional well-being and vice versa. Changes in loneliness partially mediated these effects.

Keywords

College students depression life satisfaction loneliness mediation negativity effect social support

A large body of work demonstrates that greater social support predicts better psychological health (Antonucci & Jackson, 1987; Cohen & Wills, 1985; Krause, 1996). Yet not all social interactions are positive, and recent studies concerned with the negative aspects of social relationships suggest that interpersonal tensions are a potent stressor, predicting poorer psychological health (e.g. Brown & Vinokur, 2003; Finch, Okun, Pool, & Ruehlman, 1999; Newsom, Nishishiba, Morgan, & Rook, 2003; Rook, 1984; Schuster, Kessler, & Aseltine, 1990). Intriguingly, emerging evidence implies that negative and positive social exchanges may differentially predict the positive versus negative aspects of emotional health outcomes (Ingersoll-Dayton, Morgan, & Antonucci, 1997; Rook, 2001).

To this point, however, research testing “domain specificity” is limited by measures of positive and negative exchanges that are not parallel in nature and/or through the use of outcome measures that conflate positive and negative aspects of mental health outcomes (Rook, 1990). Equally, few studies have directly tested the mechanisms by which positive and negative social exchanges might influence emotional health. In the present study, we address these gaps by using parallel measures of positive and negative social exchanges to predict both positive and negative aspects of mental health in a sample of first-year college students and by evaluating loneliness as a possible mediator of these associations.

Positive and negative social exchanges and mental health

Positive and negative social exchanges appear to be relatively distinct constructs; that is, low positivity in relationships is not equivalent to high negativity (Finch et al., 1999; Lakey, Tardiff, & Drew, 1994; Okun & Lockwood, 2003; Rook, 1984; Schuster et al., 1990; Walen & Lachman, 2000). As such, in the literature they have frequently been assessed together, and their relative import in terms of mental health weighed. There are several views on this topic. According to the “domain-specific model” of social exchanges, positive exchanges should predict the positive aspects of well-being (e.g. positive affect) and negative exchanges the negative aspects (e.g. negative affect) (Ingersoll-Dayton et al., 1997). Several studies among college students have shown that negative social interactions load on negative affectivity, whereas enacted and perceived support (i.e. positive social interactions) load on positive affectivity (Finch, 1998; Lakey et al., 1994). The domain-specific model is also consistent with theories of well-being (e.g. Diener & Emmons, 1984; Lawton, 1983), which distinguish between positive and negative aspects of well-being and postulate that distinct aspects of the environment (e.g. social interactions; Lawton, 1983) may be differentially relevant for each.

Another view is that negative exchanges are more important for health outcomes and well-being (i.e. the “negativity effect” model; Rook, 1990). Compared to positive events (like positive social exchanges), negative events (such as negative social exchanges) are believed to be more impactful because they trigger physiological arousal, convey information that is weighed more heavily, and produce more complex representations (Baumeister, Braslavsky, Finkenauer, & Vohs, 2001; Taylor, 1991). Also consistent with this view are several studies showing that negative social exchanges are more closely linked to outcomes than are supportive (i.e. positive) social exchanges (e.g. Bertera, 2005; Pagel, Erdly, & Becker, 1987; Rook, 1984; 2001). In one report on over 4500 adults, for example, negative social exchanges were more strongly associated with anxiety and mood disorder episodes than positive support (Bertera, 2005). However, because of a research focus on the negative aspects of well-being only (e.g. psychological distress), support for the negativity effect could be falsely inflated. Rook (2001) postulated a version of the negativity effects model that does account for positive well-being; in the “crossover effects model,” the negative aspects of relationships can affect both positive and negative dimensions of emotional health, whereas the positive aspects affect only the positive dimension. However, other studies have not found differential effects of positive or negative social exchanges on outcomes (Finch et al., 1999; Lepore, 1992) or have found positive exchanges to exert stronger effects than negative exchanges (i.e. a “positivity effect” model; Walen & Lachman, 2000).

From a theoretical perspective (Diener & Emmons, 1984; Lawton, 1983), we believe that the domain-specific model of social exchanges (Ingersoll-Dayton et al., 1997; Rook, 2001) may be the most appropriate for understanding the effects of social exchanges on well-being. Although “bad is stronger than good” (Baumeister et al., 2001) in general, the quasi-independence of negative and positive affects (Diener & Emmons, 1984) together with evidence that these distinct constructs may only link to valence-similar affective personality characteristics (Warr, Barter, & Brownbridge, 1983) suggests that an “affect-matching” view may be more relevant to social exchanges. That is, although negative exchanges may increase negative affect, they will not necessarily decrease positive affect. Similarly, although positive exchanges may increase positive affect, they should not necessarily decrease feelings of negative affect.

Furthermore, from a methodological perspective and as alluded to above, much of the research testing competing models of social exchanges and well-being is limited in two ways. First, although positive exchanges have been studied as consisting of multiple dimensions (e.g. informational, instrumental, and emotional support) (Antonucci & Jackson, 1987), negative exchanges have only recently been studied in this manner (Finch et al., 1999). Theoretically, contrasting these two types of exchange requires symmetry across measures (e.g. dimensions and intensity) and in the time frame assessed (e.g. over the past month) (Bertera, 2005; Rook, 1997). In addressing this limitation, the current report employed the Positive and Negative Social Exchanges Scale (PANSE) (Newsom, Rook, Nishishiba, Sorkin, & Mahan, 2005) that assesses four parallel domains of both positive and negative exchanges that are measured over an equal time interval (in the past month). We operationalize “positive social exchanges” as interactions with the network in four domains: informational support, instrumental support, emotional support, and companionship. Equally, “negative social exchanges” are assessed with four parallel domains, namely unwanted advice or intrusion, failure to provide help, unsympathetic or insensitive behavior, and rejection or neglect. Assessing both positive and negative aspects of social exchanges on well-differentiated and comparable metrics may help clarify their associations with mental health outcomes.

Second, most research has focused on negative aspects of well-being (e.g. depressed affect, psychological distress, and anxiety) (Rook, 1990) or has operationalized outcomes in a way that conflates the positive and negative aspects of well-being. Most commonly, depressive symptoms (as measured by the Center for Epidemiologic Studies Depression Scale (CES-D); Radloff, 1977) are tested but global interpretations are employed, obscuring differences across the four subscales: somatic symptoms, depressed affect, lack of well-being/lack of positive affect (i.e. “anhedonia”), and interpersonal difficulties (Hays et al., 1998; Hertzog et al., 1990). To properly test the competing models of the relevance of the positive versus negative aspects of social exchanges, it may be important to be similarly differentiated in terms of outcomes. In the current report, we focus on both negative emotional well-being (depressed affect and somatic symptoms from the CES-D) and positive emotional well-being (positive affect from the CES-D and life satisfaction).

Loneliness as a mediator between social exchanges and mental health

In addition to precisely examining the links between positive and negative exchanges and both positive and negative emotional health outcomes, the current report was designed to test the possible role of a likely mediator – loneliness. To date, the underlying processes that explain the link between social exchanges and psychological outcomes have been poorly understood. Loneliness, defined as a sense of dissatisfaction arising from the feeling that one’s relationships are in some way inadequate (Fees, Martin, & Poon, 1999; Jones, 1981; Weiss, 1973), seemed a likely candidate. There were several reasons to specifically assess loneliness for this role.

First, loneliness is inherently subjective and the evaluation of network functioning as “inadequate” can encompass network size (Jones, 1981), contact (Peplau, Russell, & Heim, 1979), or relationship quality (Weiss, 1973). Thus, loneliness can be understood as a negative cognitive appraisal of one’s social network (Fees et al., 1999) and results from a perceived discrepancy between desired and actual social relationships (Gierveld, 1998; Peplau & Perlman, 1982; Shiovitz-ezra & Leitsch, 2007); it is only indirectly related to objective characteristics of the social environment (Jones, 1981; Shiovitz-ezra & Leitsch, 2007). Thus, it seems likely that frequent negative social exchanges could lead an individual to appraise his or her network negatively, thereby increasing feelings of loneliness, whereas frequent positive exchanges could influence the appraisal of the network in the opposite direction, potentially decreasing feelings of loneliness.

Second, loneliness (i.e. evaluating one’s relationships as somehow inadequate) is closely linked not just to social network processes (Cacioppo et al., 2000; Hawkley, Burleson, Berntson, & Cacioppo, 2003; Jackson, Soderlind, & Weiss, 2000; Pinquart & Sorensen, 2001; Wheeler, Reis, & Nezlek, 1983) but also to well-being (Lee & Ishii-Kuntz, 1987; Windle & Woods, 2009), including a link with greater depression (Cacioppo, Hughes, Waite, Hawkley, & Thisted, 2006; Nolen-Hoeksema & Ahrens, 2002; Perlman, Gerson, & Spinner, 1978; Rich & Scovel, 1987). Although the presence of “feeling lonely” items in measures of depression means loneliness and depressive symptoms may be conflated, the constructs appear separable (Cacioppo et al., 2006), and in the current study we removed the loneliness item from our measure of depressive symptoms (the CES-D; Radloff, 1977).

Finally, there is some existing evidence showing that loneliness may mediate between social network variables and health outcomes. Some prior work shows that loneliness mediates the links between support and physical health (Fees et al., 1999; Segrin & Domschke, 2011; Segrin & Passalacqua, 2010). Network appraisals (Newsom et al., 2005) and loneliness (Lee & Ishii-Kuntz, 1987; Stroebe, Stroebe, Abakoumkin, & Schut, 1990) have also been shown to mediate the link between aspects of the social network and mental health but only among older adults. Lee and Ishii-Kuntz (1987), for example, found that loneliness mediated large proportions of the effects of social interactions (with friends and neighbors, in particular) on morale among a large sample of adults aged 55 and older. Stroebe and colleagues (1990) found that loneliness mediated between social support and depressive symptoms among a small sample of older widowers and married men and women. However, not only did this research focus solely on older adults, but it also did not distinguish between positive and negative social interactions nor between positive and negative aspects of well-being. Whereas positive social interactions may benefit positive aspects of health (e.g. positive emotional well-being) by lessening loneliness (Segrin & Passalacqua, 2010), threats to social networks (i.e. negative social interactions) may lead to negative emotional well-being (e.g. depressed affect) by increasing loneliness (Dill & Anderson, 1999).

Newsom and colleagues (2005) tackled this issue to some extent by examining the mediating role of appraisals (rather than loneliness, per se) in the association between positive and negative social exchanges (measured with the PANSE scale) with positive well-being and psychological distress in a sample of older adults. Consistent with a crossover effects model, they found that negative exchanges were associated with lower well-being and greater psychological distress, whereas positive exchanges were related only to positive well-being; furthermore, appraisals of the exchanges appeared to mediate these links. The present study builds on the findings from Newsom et al.’s study in a number of ways: (1) we assess loneliness rather than satisfaction with specific types of exchange (as in the study of Newsom et al.), allowing us to test whether overall assessments of the network might explain associations between exchanges and mental health; (2) we do not include the (lack of) positive affect and interpersonal functioning subscales from the CES-D in our measure of negative emotional well-being, giving us a measure of negative emotional well-being not clouded by potential confounds with positive well-being and/or social exchanges; and (3) perhaps most importantly, we examine a sample of younger adults undergoing a transition by following them longitudinally over the first year of college, allowing us to generalize beyond older adults and beyond cross-sectional data. As outlined below, studying young adults undergoing the transition to college allows for a particularly ecologically valid examination of the effects of network changes on changes in mental health.

The transition to college

Prior studies of loneliness’ role in mediating links between social network factors and outcomes have been predominantly cross sectional (e.g. Lee & Ishii-Kuntz, 1987; Newsom et al., 2005). In the current report, we use longitudinal data such that we are able to model changes in networks, changes in outcomes, and the possible role of changes in loneliness. By gathering repeated measures at the individual level and controlling for the time 1 links among these phenomena, we are able to more powerfully evaluate their associations than is possible in a cross-sectional design. Because of our interest in dynamic changes, we employ data from a group of first-year university students undergoing the transition to college. Transitions (e.g. moving, divorce) normatively involve major network changes. Students may leave their parents’ home for the first time, lose touch with elements of their previous network, and are expected to create new relationships. Such changes are especially likely to trigger loneliness (Dill & Anderson, 1999), and research shows that loneliness is highest in adolescence and young adulthood (Brage, Meredith, & Woodward, 1993; Perlman, 1988; Ryan & Patterson, 1987). Studying this sample across a major transition provides an ecologically valid examination of how network changes are associated with emotional health and the possible role of loneliness.

The present study: Aims and hypotheses

The first aim of our study was to test the domain-specific model of social exchanges using parallel measures of positive and negative social exchanges to predict both positive and negative aspects of emotional health across the first year of college. Based on this model, we made the following predictions:

Changes in negative social exchanges, but not positive social exchanges, would be (positively) associated with changes in negative emotional well-being; and

Changes in positive social exchanges, but not negative social exchanges, would be (positively) associated with changes in positive emotional well-being.

Our second aim was to test whether loneliness mediated the associations between both positive and negative social exchanges and positive and negative mental health outcomes, respectively. Whereas positive social exchanges may benefit positive emotional well-being by lessening loneliness (Segrin & Passalacqua, 2010), negative social exchanges may lead to negative emotional well-being by increasing loneliness (Dill & Anderson, 1999). As such, we predicted the following:

Loneliness would mediate the associations between positive and negative social exchanges and both positive and negative emotional well-being.

Although the longitudinal design employed in this report is a substantial improvement over prior cross-sectional studies, the study is still correlational so that directionality can only be inferred. It is possible, for example, that social exchanges may have direct effects on depression and that depression may in turn influence loneliness. One recent prospective work found that, whereas loneliness at year 1 predicted subsequent differences in depressive symptoms, depressive symptoms at year 1 also predicted subsequent changes in loneliness (Cacioppo et al., 2006). Therefore, we also tested for the “reverse mediation” pathways of emotional well-being mediating between social exchanges and loneliness.

Method

Participants

Survey data for the present study were collected at two time points during students’ first year of college at a small, private university in the northeast of the United States. Time 1 data were collected by paper-and-pencil survey from 305 freshmen (a 39% response rate) during first-year orientation classes in September 2010. Time 2 data were collected via an online survey from 113 freshmen (reflecting a 37% response rate from the original sample) at the end of the spring semester (April 2011). At both time points, the survey included items on personality, social support, self-esteem, attachment, loneliness, expressive behavior, physical health/risk-taking behaviors, college adjustment, depressive symptoms, and life satisfaction.

For the present study, only those 113 participants with complete data are included in the analysis. Those students who did not complete the follow-up survey (n = 192) did not differ from the 113 participants who responded at time 2 on age, gender, ethnicity, stress levels, loneliness, depressive symptomatology, or positive or negative social exchanges. However, students who responded at time 2 were more likely to live at home at time 1 (χ ²= 9.35, p < .01) and had higher levels of life satisfaction at time 1 (M = 24.67, SD = 6.04 compared to M = 22.88, SD = 6.66, t(290) = −2.31, p < .05).

The mean age of the final sample was 17.85 (SD = .45) at time 2 and 85% (n = 96) were female. Of them, 67% identified as white, 10.7% as Asian, 8.9% as black, 6.3% as Latino, and 7.2% as “Other.” In all, 67% (n = 74) lived at home with their parents and commuted to campus, while 34.5% (n = 39) lived on campus.

Measures

Positive and negative social exchanges

Social exchanges were measured using the PANSE scale (Newsom et al., 2005). This scale consists of 24 items, with 12 items representing four dimensions of positive exchange (informational support, instrumental support, emotional support, and companionship) (e.g. “In the past month, how often did the people you know do or say things that were kind or considerate toward you?” emotional support) and 12 items representing four parallel dimensions of negative exchange (unwanted advice or intrusion, failure to provide help, unsympathetic or insensitive behavior, and rejection or neglect) (e.g. “In the past month, how often did the people you know do things that were thoughtless or inconsiderate?” insensitive behavior). Participants were asked to consider the people in their lives (partner or spouse, family members, friends, neighbors, in-laws, or others) and indicate on a five-point scale from 0 (never) to 4 (very often), how often these various exchanges had occurred over the past month. Total scores were obtained for positive and negative exchanges by calculating mean responses for each subscale, with total scores ranging from 0 to 4, and higher scores representing greater levels of positive (Cronbach’s α at time 1 and time 2 = .91 and .94, respectively) or negative (Cronbach’s α at time 1 and time 2 = .84 and .94, respectively) social exchanges.

Loneliness

Loneliness was assessed with the revised University of California, Los Angeles (UCLA) Loneliness Scale (R-UCLA; Russell, Peplau, & Cutrona, 1980), which is a 20-item measure answered on a four-point Likert-type scale from 1 (never) to 4 (often) (e.g. “There is no one I can turn to.”). The R-UCLA is the most frequently used measure of loneliness (Cacioppo et al., 2006). Positively worded items were reverse coded, and scores were summed to create a total score ranging from 20 to 80. This measure has good validity and reliability (Russell, 1996), and in the present study Cronbach’s α was high at both time points (time 1 = .91 and time 2 = .93).

Negative emotional well-being

Negative emotional well-being was assessed with two subscales from the 20-item CES-D (Radloff, 1977), for which ratings about experiences over the past week are made on a four-point Likert-type scale from 0 (rarely or none of the time) to 3 (most or all of the time). The CES-D has been found to contain four subscales: somatic symptoms/psychomotor retardation (seven items; e.g. “I did not feel like eating; my appetite was poor”), depressed affect/mood (seven items; “I felt depressed”), lack of well-being/lack of positive affect (four items; “I felt hopeful about the future,” reverse coded), and interpersonal difficulties (two items; e.g. “I felt that people disliked me”) (Radloff, 1977). The interpersonal difficulties subscale was not examined in the present study because of its potential overlap with the measure of negative social exchanges. In addition, one item from the depressed affect/mood scale (“I felt lonely”) was dropped due to its obvious overlap with loneliness, so the final depressed affect subscale consisted of six items. To be consistent with the typical calculation of the overall CES-D score, sums were created for each of the subscales. Ranges were as follows: somatic symptoms (0–21), depressed affect (0–18), and positive affect (0–12). The somatic symptoms and depressed affect subscales were combined into a mean scale labeled “negative emotional well-being” (Cronbach’s αs = .85 at time 1 and .87 at time 2).

Positive emotional well-being

Positive emotional well-being was a composite score consisting of a measure of life satisfaction and the positive affect subscale from the CES-D (Radloff, 1977). Life satisfaction was measured with the Satisfaction with Life Scale (SWLS) (Diener, Emmons, Larsen, & Griffin, 1985), which is a measure of global life satisfaction distinct from both the positive affect and loneliness components of subjective well-being. The scale consists of five items (e.g. “The conditions of my life are excellent.”) responded to on a seven-point Likert-type scale from 1 (strongly disagree) to 7 (strongly agree). A total score was obtained by summing the items, such that scores ranged from 5 to 35, with a higher score indicating a higher level of life satisfaction. The positive affect subscale from the CES-D, described above (and with the scale inverted such that higher scores represent greater positive affect), was combined with life satisfaction into a mean scale labeled “positive emotional well-being” (Cronbach’s αs = .87 at time 1 and .90 at time 2).

Control variables

In addition to gender and ethnicity, we controlled for overall levels of stress at time 1 and time 2. Stress was measured with a single item (“Within the last 12 months, how would you rate the overall level of stress you have experienced?”), rated on a five-point Likert-type scale from 0 (no stress) to 4 (tremendous stress). Finally, because relationships may be influenced by whether a student lives on or off campus (Christie & Dinham, 1991), residence was also included as a control variable, 0 (off campus) and 1 (on campus).

Analyses

Using an approach taken by Rook (2001), we conducted two hierarchical regressions in which we predicted changes in positive and negative emotional well-being from changes in positive and negative social exchanges. In this approach, control variables and time 1 variables are entered in the first step, and time 2 variables are entered in the second step. The coefficients for the second step then represent the extent to which a change in the predictor variables from T1 to T2 is associated with a change in the outcome variables from T1 to T2. In other words, we predicted each time 2 outcome from time 2 positive and negative social exchanges, with the associated time 1 outcome, time 1 positive and negative social exchanges, and time 1 loneliness also included as predictors. In addition, we controlled for gender, ethnicity, residence, and stress at times 1 and 2 in all regressions.

In a third step of each regression, we added time 2 loneliness in order to determine whether changes in loneliness mediated any effects of changes in social exchanges on changes in psychological health. The tests of the mediational hypotheses were based on Baron and Kenny’s (1986) and Kenny, Kashy, and Bolger’s (1998) definition of a mediator, which states that complete mediation occurs only when the effect of the independent variable on the dependent variable is reduced to zero when controlling for the mediator. In cases of partial mediation (in which the effect is reduced but not to zero), Sobel mediation tests (Sobel, 1982) were calculated in order to test for significance. In these instances, we also calculated the percentage of the total effect that is mediated according to Mackinnon and Dwyer’s (1993) formula: $α β / (α β + τ^{'})$ , where α represents the coefficient relating the independent variable to the mediating variable, β represents the coefficient relating the mediator to the outcome adjusted for the independent variable, and τ′ represents the coefficient relating the independent variable to the outcome adjusted for the mediator. The product αβ represents the mediated or indirect effect, while τ′ represents the nonmediated or direct effect.

Results

Descriptive

Table 1 provides descriptive information and intercorrelations among all study variables (N = 113). Most notably, stress at both time points was positively correlated with negative exchanges, loneliness, and negative emotional well-being. Stress at time 2 was also negatively correlated with positive emotional well-being at both time points. Positive and negative exchanges were negatively correlated with one another within each time point. Positive exchanges at both time points were negatively correlated with loneliness and positively correlated with positive emotional well-being. Negative exchanges were positively correlated with loneliness and negative emotional well-being, and negatively correlated with positive emotional well-being.

Table 1.

Percentages, or means and standard deviations, paired samples t-tests, and intercorrelations among all study variables (N = 113).

	M (%)	SD	Paired samples t test	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
1. On campus	34.5	n/a		–
2. Female	85	n/a		−.01	–
3. Nonwhite	33.6	n/a		−.08	−.07	–
4. Stress, T1	2.44	.91	−10.98***	.04	.12	−.12	–
5. Stress, T2	3.40	.79	−10.98***	.06	.15	−.10	.41***	–
6. Positive exchanges, T1	3.07	.70	.46	.01	.11	−.12	−.03	−.01	–
7. Positive exchanges, T2	3.04	.74	.46	−.07	.13	.19*	.12	.04	.36***	–
8. Negative exchanges, T1	1.17	.64	−2.43*	−.15	−.07	−.07	.23*	.16†	−.29**	−.06	–
9. Negative exchanges, T2	1.39	.92	−2.43*	.00	.08	−.21*	.20*	.25**	−.11	−.23*	.33***	–
10. Loneliness, T1	36.34	9.97	−1.29	−.01	−.18†	−.11	.29**	.24*	−.34**	−.26**	.47***	.37***	–
11. Loneliness, T2	37.64	11.72	−1.29	.08	−.09	.01	.17†	.31**	−.40***	−.54***	.27**	.51***	.65***	–
12. Negative emotional well-being, T1	5.71	3.37	−1.90†	−.01	.12	−.18†	.25**	.30**	−.22*	−.07	.45***	.41***	.49***	.31**	–
13. Negative emotional well-being, T2	6.01	3.28	−1.90†	.08	.17†	−.16	.28**	.44***	−.23*	−.11	.43***	.50***	.51***	.48***	.86***	–
14. Positive emotional well-being, T1	16.84	4.16	1.98†	.13	−.02	−.02	−.08	−.20*	.38***	.32***	−.39***	−.39***	−.69***	−.51***	−.44***	−.48***	–
15. Positive emotional well-being, T2	16.11	4.60	1.98†	−.06	−.04	−.03	−.09	−.27**	.44***	.51***	−.24*	−.35***	−.55***	−.71***	−.41**	−.47***	.69***	–

Note. Ranges: stress (0–4); positive and negative social exchanges (0–4); loneliness (20–80); negative emotional well-being (0–19); positive emotional well-being (2–24). Although ethnicity was controlled for in all analyses, “nonwhite” is used here for purposes of interpreting correlations.

†p < .10.

*p < .05.

**p < .01.

***p < .001.

Time 1 and 2 variables were highly correlated with one another (e.g. stress at time 1 and time 2), indicating stability across time. However, we also ran paired samples t tests to determine whether there were any significant changes in stress, loneliness, positive and negative social exchanges, or positive and negative emotional well-being over time. At time 2, students reported significantly higher levels of stress and higher levels of negative exchanges compared to time 1. Furthermore, there were trends such that students reported higher levels of negative emotional well-being and lower levels of positive emotional well-being at time 2 compared to time 1.

Primary mediation analyses

Negative emotional well-being

Table 2 shows the final two steps of a three-step hierarchical linear regression predicting negative emotional well-being at time 2, with background variables and time 1 variables (stress, negative emotional well-being, positive and negative exchanges, and loneliness) entered in step 1, time 2 stress and time 2 positive and negative exchanges added in step 2, and time 2 loneliness entered in step 3. As seen in step 2 and in line with our predictions, controlling for the effect of increases in stress on increases in negative emotional well-being (β = .16, p < .01), increases in negative exchanges were associated with increases in negative emotional well-being across the two time points (β = .12, p < .05). In contrast, changes in positive exchanges were not associated with changes in negative emotional well-being (β = .001, p = .987).

Table 2.

Hierarchical linear regression predicting negative emotional well-being at time 2, from positive and negative social exchanges at times 1 and 2, loneliness at time 1, negative emotional well-being at time 1 (step 2) and loneliness at time 2 (step 3).

Variables	B	SE B	β	Squared part correlation
Step 2
On campus	.56	.35	.08	.001
Ethnicity	.05	.12	.02	.000
Female	.88	.47	.09†	.007
Stress, T1	−.11	.22	−.03	.001
Negative emotional well-being, T1	.69	.06	.70***	.281
Positive exchanges, T1	−.03	.26	−.01	.000
Negative exchanges, T1	.07	.31	.01	.000
Loneliness, T1	.04	.02	.11†	.007
Stress, T2	.73	.26	.16**	.017
Positive exchanges, T2	.00	.25	.00	.000
Negative exchanges, T2	.45	.22	.12*	.009
Step 3
On campus	.56	.34	.08†	.006
Ethnicity	.06	.12	.03	.001
Female	.89	.46	.09†	.008
Stress, T1	−.10	.21	−.03	.000
Negative emotional well-being, T1	.71	.06	.73***	.294
Positive exchanges, T1	.09	.26	.02	.000
Negative exchanges, T1	.15	.30	.03	.000
Loneliness, T1	.00	.02	.00	.000
Stress, T2	.50	.26	.11†	.007
Positive exchanges, T2	.37	.27	.08	.004
Negative exchanges, T2	.24	.22	.06	.002
Loneliness, T2	.06	.02	.22**	.015

Note. Step 1 (with just control variables and time 1 predictors) is not shown here for purposes of presentation; ΔR ²from step 2 to step 3 = .015, p < .01; total adjusted R ²at step 3 = .81.

Consistent with our hypotheses and as shown in step 3 of the regression, we found that changes in loneliness partially mediated the association between changes in negative exchanges and changes in negative emotional well-being. For illustrative purposes, we explain this first mediation effect in terms of Baron and Kenny’s (1986) definition of mediation: we first established that there was a significant relationship between changes in negative exchanges and changes in negative emotional well-being (in step 2), as mentioned above (β = .12, p < .05); then, we showed that variations in changes in negative exchanges significantly accounted for variations in changes in loneliness (B = 3.29, SE = .97, β = .25, p < .01; not shown in table), and variations in changes in loneliness (the mediator) significantly accounted for variations in changes in negative emotional well-being while controlling for changes in negative exchanges (β = .22, p < .01); finally, when the path between loneliness and negative emotional well-being was controlled, the relationship between negative exchanges and negative emotional well-being diminished to nonsignificance (from β = .12, p < .05 to β = .06, p = .282). This decrease was significant according to the Sobel test statistic (z = 2.10, p < .05), confirming partial mediation. Furthermore, according to Mackinnon and Dwyer’s (1993) definition, 46.5% of the effect of changes in negative exchanges on changes in negative emotional well-being were mediated by changes in loneliness across the first year of college.

Positive emotional well-being

Table 3 shows the final two steps of a three-step hierarchical linear regression predicting positive affect at time 2, with background variables and time 1 variables (stress, positive emotional well-being, positive and negative exchanges, and loneliness) entered in step 1, time 2 stress and time 2 positive and negative exchanges added in step 2, and time 2 loneliness entered in step 3. Consistent with expectation and as seen in step 2, increases in positive exchanges were associated with increases in positive emotional well-being over time (β = .32, p < .001). In contrast, changes in negative exchanges were not associated with changes in positive emotional well-being (β = −.05, p = .502). Furthermore, as can be seen in step 3, the addition of time 2 loneliness to the model (β = −.45, p < .001) reduced the effect of positive exchanges to nonsignificance (β = .15, p = .061). According to the Sobel test, the decrease was significant (z = 3.31, p < .001). Approximately 53% of the total effect of changes in positive exchanges on changes in positive emotional well-being were mediated by loneliness.

Table 3.

Hierarchical linear regression predicting positive emotional well-being at time 2, from positive and negative social exchanges at times 1 and 2, loneliness at time 1, positive emotional well-being at time 1 (step 2) and loneliness at time 2 (step 3).

Variables	B	SE B	β	Squared part correlation
Step 2
On campus	.001	.63	.00	.000
Ethnicity	.024	.22	.08	.004
Female	−.44	.86	−.04	.001
Stress, T1	−.43	.40	−.08	.005
Positive emotional well-being, T1	.60	.11	.53***	.116
Positive exchanges, T1	.79	.48	.12	.011
Negative exchanges, T1	1.05	.55	.15†	.015
Loneliness, T1	−.03	.05	−.07	.002
Stress, T2	−.77	.47	−.13	.011
Positive exchanges, T2	1.87	.45	.32***	.071
Negative exchanges, T2	−.26	.38	−.05	.002
Step 3
On campus	.00	.57	.00	.000
Ethnicity	.20	.20	.06	.003
Female	−.58	.78	−.05	.002
Stress, T1	−.46	.36	−.09	.005
Positive emotional well-being, T1	.59	.10	.52***	.112
Positive exchanges, T1	.50	.44	.08	.004
Negative exchanges, T1	.77	.50	.11	.008
Loneliness, T1	.06	.05	.12	.005
Stress, T2	−.19	.44	−.03	.001
Positive exchanges, T2	.88	.47	.15†	.012
Negative exchanges, T2	.25	.37	.05	.002
Loneliness, T2	−.17	.04	−0.45***	.063

Note. Step 1 (with just control variables and time 1 predictors) is not shown here for purposes of presentation. ΔR ²from step 2 to step 3 = .076, p < .01; total adjusted R ²at step 3 = .44.

Reverse mediation analyses

We also tested for reverse mediation (i.e. we tested whether positive emotional well-being mediated between positive social exchanges and loneliness and whether negative emotional well-being acted as a mediator of the association between negative social exchanges and loneliness). After the introduction of time 2 positive emotional well-being to a model predicting time 2 loneliness from time 2 positive and negative social exchanges (with the control variables ethnicity, gender, stress at times 1 and 2, positive and negative social exchanges at time 1, loneliness at time 1, and positive emotional well-being at time 1), an initially significant effect of time 2 positive social exchanges on time 2 loneliness (B = −5.74, SE = 1.12, β = −.37, p < .001) remained highly significant (B = −3.71, SE = 1.12, β = −.24, p < .01). Similarly, after the introduction of time 2 negative emotional well-being to a model predicting time 2 loneliness from time 2 positive and negative social exchanges (with the corresponding control variables), an initially significant effect of time 2 negative social exchanges on time 2 loneliness (B = 3.29, SE = .97, β = .25, p < .01) remained highly significant (B = 2.71, SE = .96, β = .21, p < .01). These results imply that it is very unlikely that well-being mediates the relationship between social exchanges and loneliness.

Discussion

To our knowledge, this report is among the first to longitudinally examine how changes in the frequency of positive and negative social exchanges may predict changes in emotional or mental health outcomes. Consistent with predictions, the report extends prior research in several key ways. First, we uncovered the expected effects in which changes in positive and negative social exchanges during the first year of college were differentially associated with distinct mental health outcomes. Consistent with the domain-specific effects model, increases in negative exchanges were associated with increases in negative emotional well-being but not with changes in positive emotional well-being. Conversely, increases in positive exchanges were associated with increases in positive emotional well-being but not with changes in negative emotional well-being. Second, our analyses indicated that changes in loneliness partially mediated the significant social exchange effects, suggesting that feelings of loneliness are central to the links these processes show with emotional well-being. In the space below, we revisit these results more fully, concentrating on the implications our findings have for the contrasting predictions stemming from domain-specific and negativity effect models and for the understanding of the processes by which positive and negative social exchanges may predict outcomes.

Revisiting the domain-specific and negativity effect models: Predicting the positive and negative aspects of mental health

The first contribution of the current report lies in its ability to speak to ongoing debates regarding (a) the relative importance of negative versus positive social exchanges and (b) the types of outcomes that each should (hypothetically) be associated with. The literature to date has been inconsistent, with some studies suggesting that negative exchanges are more important (e.g. Bertera, 2005; Pagel et al., 1987; Rook, 1984; 2001), other works failing to find differences (Finch et al., 1999; Lepore, 1992), and still others reporting stronger effects for positive exchanges (Walen & Lachman, 2000). Most studies have, however, concentrated on the negative aspects of well-being (e.g. psychological distress) despite conceptualizations of this construct as having both positive and negative aspects (Diener & Emmons, 1984; Lawton, 1983), have measured positive and negative exchanges in noncomparable ways (Ingersoll-Dayton et al., 1997), or have focused solely on older adults (e.g. Ha, 2010; Ingersoll-Dayton et al., 1997; Rook, 1990) and/or cross-sectional data (e.g. Newsom et al., 2005).

Given the complexity of these phenomena, it is unsurprising that there are several competing models in the area. On the one hand, the negativity effect model (Newsom et al., 2003; Rook, 1990) suggests that negative exchanges are more detrimental for both positive and negative aspects of well-being. Relatedly, some researchers posit a crossover type or asymmetrical effect in which negative exchanges predict both aspects of emotional health, whereas positive exchange characteristics only predict the positive dimension of emotional health (Rook, 2001). Conversely, the domain-specific model suggests that positive exchanges should predict the positive aspects of well-being, whereas negative exchanges should predict the negative aspects (Finch et al., 1999; Ingersoll-Dayton et al., 1997; Lakey et al., 1994; Rook, 1990).

Broadly speaking and as expected, our data are consistent with the domain-specific model. Changes in positive exchange metrics robustly predicted changes in positive emotional well-being but were not linked to negative emotional well-being. Conversely, negative exchange metrics predicted changes in negative emotional well-being but were unrelated to positive emotional well-being. Changes in stress were covaried in all models, so it is notable that these effects remained even after controlling for stress that could presumably (at least in part) be generated by negative exchanges. A few prior studies have provided evidence of domain specificity, linking positive and negative exchanges to positive and negative affects (Ingersoll-Dayton et al., 1997), spousal support and coping with arthritis (Manne & Zautra, 1989), and support, goal strivings, well-being, and distress (Ruehlman & Wolchik, 1988). It has been noted, however, that interpretations of prior work are limited insofar as measures of positive and negative exchanges have not typically been operationalized in a manner that ensured they were directly comparable (Ingersoll-Dayton et al., 1997). By employing comparable measures of positive and negative exchanges, the current report counters this possibility and thus extends the literature.

The question, then, is why does this domain specificity occur? It has previously been acknowledged that measurement-based accounts are possible (i.e. the greater covariation between positive exchanges and positive outcomes, and negative exchanges and negative outcomes, reflects aspects of measurement, see Ingersoll-Dayton et al., 1997). However, it is also possible that positive exchanges are simply more relevant to the generation of positive affect and well-being than they are to the alleviation of somatic symptoms or depressive symptoms; the obverse may apply for negative exchanges. Our results are in contrast to the findings of Newsom et al. (2005) whose results supported a negativity effect model (i.e. negative exchanges were related to less well-being and greater psychological distress, whereas positive exchanges were related only to positive well-being). However, their study was cross sectional in design. It could be that by controlling for initial levels of positive and negative emotional well-being and, effectively, assessing change over time, our analyses reduce methodological biases and thus offer a more accurate picture of the effects of positive and negative exchanges on the positive and negative aspects of emotional well-being. Alternatively, because their participants were older adults, it may be that there is something specific to the transition to college such that domain specificity accounts for the association between social exchanges and well-being. Life span emotions research suggests that emotional experiences become more intermingled or complex with age (Magai, Consedine, Krivoshekova, McPherson, & Kudadjie-Gyamfi, 2006), creating the possibility that age-related variation in the relations between the positive and negative aspects of emotional experience may be relevant.

Loneliness: The active ingredient?

As noted, the processes underlying the links between patterns of social exchange and emotional health remain poorly understood; we hypothesized that loneliness – a subjective evaluation of one’s social network as inadequate in some sense – would help explain how changes in positive and negative social exchanges are linked to emotional health. This prediction was substantially upheld. Specifically, our Sobel testing showed that changes in loneliness mediated between 46% and 53% of the effects, linking changes in positive and negative exchanges to changes in both positive and negative emotional well-being. To our knowledge, this demonstration is unique and the presence of significant partial mediation across both positive and negative aspects of network exchanges and emotional health stands as a robust testament to the centrality of the experience of loneliness during the transition to college.

This pattern of findings is consistent with recent suggestions that social support may benefit health because it lessens loneliness (Segrin & Passalacqua, 2010) and that threats to social networks may increase loneliness and thus lead to depressive symptoms (Dill & Anderson, 1999). Prior studies examining the links between social exchange parameters and health (Fees et al., 1999; Segrin & Domschke, 2011; Segrin & Passalacqua, 2010) have previously provided evidence indicative of loneliness mediation. To this point, however, only a few studies have previously suggested such mediation with respect to mental health, and both of these studies used samples of older adults. A study among bereaved older adults has demonstrated that emotional loneliness mediates the links between social support and two adaptational outcomes – somatic complaints and depression (Stroebe et al., 1990), and a study among adults aged 55 and older showed that loneliness mediated the effects of social interactions (with friends and neighbors, in particular) on morale (Lee & Ishii-Kuntz, 1987).

In the context of the current work, it is worth recalling that the feeling of loneliness is the result of a fundamentally subjective evaluation of the network as inadequate or discrepant with desires in some way (Fees et al., 1999; Gierveld, 1998; Jones, 1981; Peplau & Perlman, 1982; Peplau et al., 1979; Shiovitz-ezra & Leitsch, 2007; Weiss, 1973). Hence, in large part, it is not the presence or absence of positive and negative exchanges that appears relevant. Instead, it is the participants’ evaluations of their networks that are crucial, since loneliness mediated the effects of positive and negative exchanges. In other words, it is not the nature of social exchanges per se that impact outcomes but rather the extent to which they contribute to feelings of isolation. Such an interpretation is consistent with commentary from the few prior mediational studies examining loneliness in the context of mental health outcomes among older adults (Lee & Ishii-Kuntz, 1987; Newsom et al., 2005; Stroebe et al., 1990). For example, Stroebe et al. (1990) noted that although bereaved older adults experienced their friends as very helpful, they could not replace the lost loved one. Furthermore, our results imply that merely interacting with others is not enough to stave off loneliness and maintain good mental health; rather, some interactions (i.e. negative exchanges) with network members may actually increase feelings of loneliness, perhaps by making the potential loss of those partners more salient (e.g. conflicts with new friends) or by highlighting other interpersonal losses (e.g. the reduced contact with family and old friends) that have occurred as a result of the transition to college.

Limitations, directions for future research, and implications

Although these data represent a useful addition to the developing understanding of the links between positive and negative social exchanges and the positive and negative aspects of emotional health, several considerations should be borne in mind. Most obviously, it should be recalled that the current sample was moderately sized, convenience-based, and predominantly comprised of female and commuting college students. Although we controlled for both gender and commuting status, some of the effects seen here may be specific to the sample in question and/or to the specific context of the transition to college.

As noted, most of the research assessing social exchanges and well-being (and in particular research using the PANSE; Newsom et al., 2005) has been conducted with older adults. However, the measure was originally designed based on qualitative studies with both older and younger adults (Newsom et al., 2003), and we believe that the content of the items in the PANSE is equally applicable to older and younger adults (e.g. “In the past month, how often did the people you know … include you in things they were doing? … let you down when you needed help?” etc.). Furthermore, we believe that the concentration of research among older adults is driven by a desire to pinpoint interventions for the elderly who are experiencing physical health declines, not by a belief that social exchanges are not equally important for well-being in younger adults. In fact, most of the key theories regarding social support and health (e.g. Carstensen, 1992; Kahn & Antonucci, 1980) are life span theories thought to apply equally at different ages.

The specificity with which we assessed both network exchanges and the positive and negative aspects of emotional health represents a useful extension over prior work. However, the parallel nature and identical time frame of the network exchanges do not necessarily mean that they are “optimally balanced” (Newsom et al., 2005: 310). It could be, for example, that “emotional support” is not as positive as “unsympathetic or insensitive behavior” is negative (or vice versa). This issue may be exacerbated by the fact that negative exchanges tend to be reported much less frequently than positive exchanges (see the means in Table 1). To some extent, this may reflect denial, self-protection, or face-saving on the part of the participants (i.e. why would we consider someone a friend if our interactions are consistently negative?), but it may be that positive interactions are in fact more frequent. If that is the case, then any nonsignificant effects of negative interactions could potentially be due to their lower frequency, not to the fact that they are less powerful predictors. However, this seems unlikely since the negative exchanges (unlike positive exchanges) were strongly predictive of negative emotional well-being. If the frequency of exchanges were affecting their impact on well-being outcomes, one would think it would affect positive and negative well-being equally. Given that this is not the case, such an interpretation seems unlikely.

Furthermore, future work might benefit from examining the possible role of specific types of negative and positive interactions (Newsom et al., 2005) (e.g. criticism from a romantic partner) rather than aggregating across social partners and contexts. Similarly, there may be utility in differentiating between social and emotional loneliness and perhaps also between interactions with romantic partners versus interactions with casual friends. Interactions with romantic partners would presumably influence emotional loneliness, since emotional loneliness is theorized to stem from the lack of a close, intimate attachment to another person (Weiss, 1973). In contrast, interactions with a network of casual friends or acquaintances would presumably influence social loneliness, since social loneliness is thought to arise from the lack of a network of friends who share common interests and activities (Green, Richardson, Lago, & Schatten-Jones, 2001; Weiss, 1973).

As mentioned in the introduction, although the longitudinal design employed in this report is a substantial improvement over prior cross-sectional studies, directionality among the variables remains uncertain. It is possible, for example, that depression may influence loneliness and/or social support. However, we tested for reverse mediation and did not find any evidence to support the idea that emotional well-being mediated the association between social exchanges and loneliness. More advanced analyses (e.g. cross-lagged panel analysis) would be a better test of this idea. Although our data were longitudinal, more than two assessments over the course of several years (rather than two assessments over the course of 8 months) would have allowed us to detect a more stable pattern of loneliness, well-being, and social exchanges. Furthermore, although the naturalistic method used in the present study has certain advantages, creative experimental work (e.g. manipulating exposure to negative interactions) could potentially disentangle cause and effect. More generally, research is needed to sort out which model (i.e. domain-specific, negativity effect, etc.) operates most commonly, or which model operates under which conditions. As stated earlier, for instance, most of the research on this topic has focused on older adults; more research focusing on younger adults could potentially shed light on whether domain specificity is age specific. Meta-analyses as well as research testing for moderating effects (of age, type of measure, etc.) are necessary to move this body of research forward.

These limitations noted, the current data may provide some guidance when seeking to intervene among at risk groups. Prior work shows that both social network characteristics (Hays & Oxley, 1986) and loneliness (Pascarella, 2006) are associated with freshmen’s successful adaptation to college and that loneliness is inversely linked to academic persistence, particularly among first-year students (Nicpon et al., 2006; Polewchak, 2002 ). As such, understanding the links among social exchanges, loneliness, and emotional well-being may help pinpoint areas of intervention to improve student adaptation and retention. Given that loneliness mediated most effects in our models, interventions to reduce feelings of loneliness are clearly warranted. Interventions focused on increasing positive exchanges and/or changing individuals’ “appraisals” of their networks may increase the positive aspects of emotional health both directly and indirectly, by reducing the experience of loneliness.

Footnotes

Authors’ note

An earlier version of this research was presented at the 2011 Annual Convention for the American Psychological Association in Washington, DC.

Funding

This research was supported in part by a Faculty Development Grant from the Office of Research and Sponsored Programs at Adelphi University.

References

Antonucci

T. C.

Jackson

J. S.

(1987). Social support, interpersonal efficacy, and health: A life course perspective. In Carstensen

L. L.

Edelstein

B. A.

(Eds.), Handbook of clinical gerontology (pp. 291–311). Elmsford, NY, USA: Pergamon Press.

Baron

R. M.

Kenny

D. A.

(1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182.

Baumeister

Bratslavsky

Finkenauer

Vohs

(2001). Bad is stronger than good. Review of General Psychology, 5, 323–370.

Bertera

E. M.

(2005). Mental health in U.S. adults: The role of positive social support and social negativity in personal relationships. Journal of Social and Personal Relationships, 22, 33–48. doi:10.1177/0265407505049320

Brage

Meredith

Woodward

(1993). Correlates of loneliness among midwestern adolescents. Adolescence, 28, 685–693.

Brown

S. L.

Vinokur

A. D.

(2003). The interplay among risk factors for suicidal ideation and suicide: The role of depression, poor health, and loved ones’ messages of support and criticism. American Journal of Community Psychology, 32, 131–141.

Cacioppo

J. T.

Ernst

J. M.

Burleson

M. H.

McClintock

M. K.

Malarkey

W. B.

Hawkley

L. C.

(2000). Lonely traits and concomitant physiological processes: The MacArthur social neuroscience studies. International Journal of Psychophysiology, 35, 143–154.

Cacioppo

J. T.

Hughes

M. E.

Waite

L. J.

Hawkley

L. C.

Thisted

R. A.

(2006). Loneliness as a specific risk factor for depressive symptoms: Cross-sectional and longitudinal analyses. Psychology and Aging, 21, 140–151.

Carstensen

L. L.

(1992). Social and emotional patterns in adulthood: Support for socioemotional selectivity theory. Psychology and Aging, 7, 331–338. doi:10.1037/0882-7974.7.3.331

10.

Christie

N. G.

Dinham

S. M.

(1991). Institutional and external influences on social integration in the freshman year. The Journal of Higher Education, 62, 412–436. doi:10.2307/1982003

11.

Cohen

Wills

T. A.

(1985). Stress, social support, and the buffering hypothesis. Psychological Bulletin, 98, 310–357.

12.

Diener

Emmons

R. A.

(1984). The independence of positive and negative affect. Journal of Personality and Social Psychology, 47, 1105–1117.

13.

Diener

Emmons

R. A.

Larsen

R. J.

Griffin

(1985). The satisfaction with life scale. Journal of Personality Assessment, 49, 71–75.

14.

Dill

J. C.

Anderson

C. A.

(1999). Loneliness, shyness, and depression: The etiology and interrelationships of everyday problems in living. In Joiner

Coyne

J. C.

(Eds.), The interactional nature of depression Advances in interpersonal approaches (pp. 93–125). Washington, DC, USA: American Psychological Association.

15.

Fees

B. S.

Martin

Poon

L. W.

(1999). A model of loneliness in older adults. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 54, P231–P239.

16.

Finch

J. F.

(1998). Social undermining, support satisfaction, and affect: A domain-specific lagged effects model. Journal of Personality, 663, 315–334.

17.

Finch

J. F.

Okun

M. A.

Pool

G. J.

Ruehlman

L. S.

(1999). A comparison of the influence of conflictual and supportive social interactions on psychological distress. Journal of Personality, 67, 581–621.

18.

Gierveld

J. D. J.

(1998). A review of loneliness: concept and definitions, determinants and consequences. Reviews in Clinical Gerontology, 8, 73–80. doi:10.1017/S0959259898008090

19.

Green

L. R.

Richardson

D. S.

Lago

Schatten-Jones

E. C.

(2001). Network correlates of social and emotional loneliness in young and older adults. Personality and Social Psychology Bulletin, 27, 281–288. doi:10.1177/0146167201273002

20.

J. H.

(2010). The effects of positive and negative support from children on widowed older adults’ psychological adjustment: A longitudinal analysis. The Gerontologist, 50, 471–481.

21.

Hawkley

L. C.

Burleson

M. H.

Berntson

G. G.

Cacioppo

J. T.

(2003). Loneliness in everyday life: cardiovascular activity, psychosocial context, and health behaviors. Journal of Personality and Social Psychology, 85, 105–120.

22.

Hays

J. C.

Landerman

L. R.

George

L. K.

Flint

E. P.

Koenig

H. G.

Land

K. C.

Blazer

D. G.

(1998). Social correlates of the dimensions of depression in the elderly. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 53, P31–P39.

23.

Hays

R. B.

Oxley

(1986). Social network development and functioning during a life transition. Journal of Personality and Social Psychology, 50, 305–313.

24.

Hertzog

Van Alstine

Usala

P. D.

Hultsch

D. F.

Dixon

(1990). Measurement properties of the center for epidemiological studies depression scale (CES-D) in older populations. Psychological Assessment, 2, 64–72.

25.

Ingersoll-Dayton

Morgan

Antonucci

(1997). The effects of positive and negative social exchanges on aging adults. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 52, S190–S199.

26.

Jackson

Soderlind

Weiss

K. E.

(2000). Personality traits and quality of relationships as predictors of future loneliness among American college students. Society, 28, 463–470. doi:10.2444/sbp.2000.28.5.463

27.

Jones

W. H.

(1981). Loneliness and social contact. Journal of Social Psychology, 113, 295–296.

28.

Kahn

R. L.

Antonucci

T. C.

(1980). Convoys over the life course: Attachment, roles, and social support. In Baltes

P. B.

Brim

O. B.

(Eds.), Life-span development and behavior (Vol. 3, pp. 253–268). New York, NY: Academic Press.

29.

Kenny

D. A.

Kashy

D. A.

Bolger

(1998). Data analysis in social psychology. In Gilbert

Fiske

S. T.

Lindzey

(Eds.), Handbook of social psychology (Vol. 1, pp. 233–265). Boston, USA: McGraw-Hill, USA. doi:10.1002/pits.10035

30.

Krause

(1996). Neighborhood deterioration and self-rated health in later life. Psychology and Aging, 11, 342–352.

31.

Lakey

Tardiff

T. A.

Drew

J. B.

(1994). Negative social interactions: Assessment and relations to social support, cognition, and psychological distress. Journal of Social Clinical Psychology, 13, 42–62.

32.

Lawton

M. P.

(1983). Environment and other determinants of well-being in older people. The Gerontologist, 23, 349–357.

33.

Lee

G. R.

Ishii-Kuntz

(1987). Social interaction, loneliness, and emotional well-being among the elderly. Research on Aging, 9, 459–482. doi:10.1177/0164027587094001

34.

Lepore

S. J.

(1992). Social conflict, social support, and psychological distress: evidence of cross-domain buffering effects. Journal of Personality and Social Psychology, 63, 857–867.

35.

Mackinnon

D. P.

Dwyer

J. H.

(1993). Estimating mediated effects in prevention studies. Evaluation Review, 17, 144–158. doi:10.1177/0193841X9301700202

36.

Magai

Consedine

N. S.

Krivoshekova

Y. S.

McPherson

Kudadjie-Gyamfi

(2006). Emotion experience and expression across the adult lifespan: Insights from a multi-modal assessment study. Psychology & Aging, 21, 303–317.

37.

Manne

S. L.

Zautra

A. J.

(1989). Spouse criticism and support: Their association with coping and psychological adjustment among women with rheumatoid arthritis. Journal of Personality and Social Psychology, 56, 608–617.

38.

Newsom

J. T.

Nishishiba

Morgan

D. L.

Rook

K. S.

(2003). The relative importance of three domains of positive and negative social exchanges: A longitudinal model with comparable measures. Psychology and Aging, 18, 746–754.

39.

Newsom

J. T.

Rook

K. S.

Nishishiba

Sorkin

D. H.

Mahan

T. L.

(2005). Understanding the relative importance of positive and negative social exchanges: Examining specific domains and appraisals. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 60, P304–P312.

40.

Nicpon

M. F.

Huser

Blanks

E. H.

Sollenberger

Befort

Kurpius

S. E. R.

(2006). The relationship of loneliness and social support with college freshmen’s academic performance and persistence. Journal of College Student Retention Research Theory Practice, 8, 345–358.

41.

Nolen-Hoeksema

Ahrens

(2002). Age differences and similarities in the correlates of depressive symptoms. Psychology and Aging, 17, 116–124.

42.

Okun

M. A.

Lockwood

C. M.

(2003). Does level of assessment moderate the relation between social support and social negativity? A meta-analysis. Basic and Applied Social Psychology, 25, 15–35.

43.

Pagel

M. D.

Erdly

W. W.

Becker

(1987). Social networks: We get by with (and in spite of) a little help from our friends. Journal of Personality and Social Psychology, 53, 793–804.

44.

Pascarella

(2006). How college affects students: Ten directions for future research. Journal of College Student Development, 47, 508–520.

45.

Peplau

Perlman

(1982). Perspectives on loneliness. In Peplau

Perlman

(Eds.), Loneliness: A Sourcebook of Current Theory Research and Therapy (pp. 1–20). New York, USA: Wiley.

46.

Peplau

Russell

D. W.

Heim

(1979). The experience of loneliness. In Frieze

I. H.

Bar-Tal

Carroll

J. S.

(Eds.), New Approaches to Social Problems Applications of attribution theory (pp. 53–78). San Francisco, USA: Jossey-Bass.

47.

Perlman

(1988). Loneliness: A life-span, family perspective. In Milardo

R. M.

(Ed.), Families and Social Networks (pp. 190–220). Thousand Oaks, CA, USA: Sage.

48.

Perlman

Gerson

A. C.

Spinner

(1978). Loneliness among senior citizens: An empirical report. Essence: Issues in the Study of Ageing, Dying, and Death, 2, 239–248.

49.

Pinquart

Sorensen

(2001). Influences on loneliness in older adults: A meta-analysis. Basic and Applied Social Psychology, 23, 245–266. doi:10.1207/S15324834BASP2304_2

50.

Polewchak

J. L.

(2002). The effects of social support and interpersonal dependency upon emotional adjustment to college and physical health. Unpublished doctoral dissertation, Virginia Consortium for Professional Psychology.

51.

Radloff

L. S.

(1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401. doi:10.1177/014662167700100306

52.

Rich

A. R.

Scovel

(1987). Causes of depression in college students: A cross-lagged panel correlational analysis. Psychological Reports, 60, 27–30.

53.

Rook

K. S.

(1984). The negative side of social interaction: impact on psychological well-being. Journal of Personality and Social Psychology, 46, 1097–1108.

54.

Rook

K. S.

(1990). Parallels in the study of social support and social strain. Journal of Social and Clinical Psychology, 9, 118–132. doi:10.1521/jscp.1990.9.1.118

55.

Rook

K. S.

(1997). Positive and negative social exchanges: Weighing their effects in later life. The Journals of Gerontology, Series B: Psychological Sciences and Social Science, 52B, S167–S169.

56.

Rook

K. S.

(2001). Emotional health and positive versus negative social exchanges: A daily diary analysis. Applied Developmental Science, 5, 86–97. doi:10.1207/S1532480XADS0502_4

57.

Ruehlman

L. S.

Wolchik

S. A.

(1988). Personal goals and interpersonal support and hindrance as factors in psychological distress and well-being. Journal of Personality and Social Psychology, 55, 293–301.

58.

Russell

Peplau

L. A.

Cutrona

C. E.

(1980). The revised UCLA loneliness scale: Concurrent and discriminant validity evidence. Journal of Personality and Social Psychology, 39, 472–480.

59.

Russell

D. W.

(1996). UCLA loneliness scale (Version 3): Reliability, validity, and factor structure. Journal of Personality Assessment, 66, 20–40.

60.

Ryan

M. C.

Patterson

(1987). On loneliness and the elderly. Journal of Gerontological Nursing, 13, 22–27.

61.

Schuster

T. L.

Kessler

R. C.

Aseltine

R. H.

(1990). Supportive interactions, negative interactions, and depressed mood. American Journal of Community Psychology, 18, 423–438.

62.

Segrin

Domschke

(2011). Social support, loneliness, recuperative processes, and their direct and indirect effects on health. Health Communication, 26, 221–232.

63.

Segrin

Passalacqua

S. A.

(2010). Functions of loneliness, social support, health behaviors, and stress in association with poor health. Health Communication, 25, 312–322.

64.

Shiovitz-ezra

Leitsch

S. A.

(2007). The role of social relationship in predicting loneliness: The national social life, health, and aging project. Social Work Research, 34, 157–168.

65.

Sobel

M. E.

(1982). Asymptotic confidence intervals for indirect effects in structural equation models. Sociological Methodology, 13, 290–312. doi:10.2307/270723

66.

Stroebe

Abakoumkin

Schut

(1990). The role of loneliness and social support in adjustment to loss: a test of attachment versus stress theory. Journal of Personality and Social Psychology, 59, 1128–1139.

67.

Taylor

S. E.

(1991). Asymmetrical effects of positive and negative events: the mobilization-minimization hypothesis. Psychological Bulletin, 110, 67–85.

68.

Walen

H. R.

Lachman

M. E.

(2000). Social support and strain from partner, family, and friends: Costs and benefits for men and women in adulthood. Journal of Social and Personal Relationships, 17, 5–30. doi:10.1177/0265407500171001

69.

Warr

P. B.

Barter

Brownbridge

(1983). On the independence of positive and negative affect. Journal of Personality and Social Psychology, 44, 644–651.

70.

Weiss

(1973). Loneliness: The experience of emotional and social isolation. Cambridge, MA: MIT Press.

71.

Wheeler

Reis

Nezlek

(1983). Loneliness, social interaction, and sex roles. Journal of Personality and Social Psychology, 45, 943–953.

72.

Windle

Woods

(2009). Variations in subjective well-being: The mediating role of a psychological resource. Ageing and Society, 24, 583–602.