http://psp.sagepub.com/Bulletin

Personality and Social Psychology

http://psp.sagepub.com/content/early/2014/04/25/0146167214529799The online version of this article can be found at:

 DOI: 10.1177/0146167214529799

published online 25 April 2014Pers Soc Psychol BullGillian M. Sandstrom and Elizabeth W. Dunn

Social Interactions and Well-Being: The Surprising Power of Weak Ties  

Published by:

http://www.sagepublications.com

On behalf of: 

  Society for Personality and Social Psychology

can be found at:Personality and Social Psychology BulletinAdditional services and information for    

  http://psp.sagepub.com/cgi/alertsEmail Alerts:

 

http://psp.sagepub.com/subscriptionsSubscriptions:  

http://www.sagepub.com/journalsReprints.navReprints:  

http://www.sagepub.com/journalsPermissions.navPermissions:  

What is This? 

- Apr 25, 2014OnlineFirst Version of Record >>

by guest on April 30, 2014psp.sagepub.comDownloaded from by guest on April 30, 2014psp.sagepub.comDownloaded from

Personality and SocialPsychology Bulletin 1 –13© 2014 by the Society for Personalityand Social Psychology, IncReprints and permissions: sagepub.com/journalsPermissions.navDOI: 10.1177/0146167214529799pspb.sagepub.com

Article

Human felicity is produced not so much by great pieces of good fortune that seldom happen, as by little advantages that occur every day.

—Benjamin Franklin

Imagine a day that begins by greeting your regular barista at the local coffee shop. You get to work and run into a col-league who you have not seen for a while, and chat about your weekend. After work, you head to yoga class where you exchange pleasantries with the girl whose hair is always a different color. Walking home afterward, you stop to chat with the guy you always see walking the pug named Wilbur. None of these people play an important role in your life, and yet a day without these kinds of interactions seems a little emptier. Past research has shown that social interactions and positive affect are mutually reinforcing, but no research has specifically examined whether interactions with these “weak ties” might be associated with positive outcomes. Are inter-actions with the coffee barista, work colleague, yoga class-mate, and dog owner part of the “little advantages that occur every day,” which Benjamin Franklin saw as central to human felicity? More precisely, can these interactions con-tribute meaningfully to our happiness, or are they inconse-quential compared to interactions with our close friends and family?

Evidence suggests that weak ties such as these—relation-ships involving less frequent contact, low emotional inten-sity, and limited intimacy—confer some important benefits. In his seminal article, Granovetter (1973) showed that weak

ties are important for diffusion of information (such as news, innovations, job openings) across a social network; informa-tion remains within isolated clusters of strong ties (i.e., close friends and family) without weak ties to bridge between them. By providing access to a breadth of perspectives and non-redundant information, weak ties have also been linked to greater creativity (Perry-Smith, 2006). The present research examines another possible advantage of weak ties—we test whether social interactions with weak ties are related to subjective well-being and belonging.

Although little work has examined the association between weak tie interactions and well-being, numerous studies have documented the association between positive affect and socializing, more generally. People report more positive affect when they are engaged in social rather than non-social activities (Pavot, Diener, & Fujita, 1990). When people recall events that occurred earlier in the day (or on the preceding day), they report more positive affect on days when they recall participating in social events (Clark & Watson, 1988; Vittengl & Holt, 1998b; Watson, Clark, McIntyre, & Hamaker, 1992), and on days when they recall feeling more connected to others (Reis, Sheldon, Gable, Roscoe, & Ryan, 2000). Furthermore, people recall having

529799 PSPXXX10.1177/0146167214529799Personality and Social Psychology BulletinSandstrom and Dunnresearch-article2014

1University of British Columbia, Vancouver, Canada

Corresponding Author:Gillian M. Sandstrom, Department of Psychology (New Museums Site), University of Cambridge, Free School Lane, Cambridge CB2 3RQ, UK. Email: gs488@cam.ac.uk

Social Interactions and Well-Being: The Surprising Power of Weak Ties

Gillian M. Sandstrom1 and Elizabeth W. Dunn1

AbstractAlthough we interact with a wide network of people on a daily basis, the social psychology literature has primarily focused on interactions with close friends and family. The present research tested whether subjective well-being is related not only to interactions with these strong ties but also to interactions with weak social ties (i.e., acquaintances). In Study 1, students experienced greater happiness and greater feelings of belonging on days when they interacted with more classmates than usual. Broadening the scope in Studies 2A and 2B to include all daily interactions (with both strong and weak ties), we again found that weak ties are related to social and emotional well-being. The current results highlight the power of weak ties, suggesting that even social interactions with the more peripheral members of our social networks contribute to our well-being.

Keywordsweak ties, subjective well-being, social relationships, social interactions, belonging

Received July 31, 2013; revision accepted March 8, 2014

by guest on April 30, 2014psp.sagepub.comDownloaded from

2 Personality and Social Psychology Bulletin

experienced greater enjoyment during socializing than dur-ing most other activities (Kahneman, Krueger, Schkade, Schwarz, & Stone, 2004; Krueger, Kahneman, Schkade, Schwarz, & Stone, 2009).

By asking people to remember their recent interactions, these past studies may have inadvertently captured predomi-nantly strong tie interactions, and may under-represent weak tie interactions. When people are asked to list the names of all of their friends and acquaintances, they tend to remember close friends and people with whom they have more frequent contact (Fu, 2005). Also, when asked to report how much time they spent with others, people are more accurate when reporting time spent with friends compared to time spent with non-friends (Eagle, Pentland, & Lazer, 2009). Thus, ret-rospective methods may be less likely to capture interactions with weak ties.

This limitation is compounded by the fact that few past studies have asked participants to report the closeness of their interaction partners. In one study, participants did rate the closeness of their interaction partners, but none of the analyses were broken down by this factor, despite the fact that 21% of reported interactions were with weak tie partners (Berry & Hansen, 1996). Therefore, whether interactions with acquaintances contribute to subjective well-being remains an empirical question.

In sum, due to a reliance on recall and the lack of distinc-tion based on closeness, it is possible that the positive effects of social interactions found in previous studies were driven largely by strong ties. Both theoretically and empirically, there is evidence to suggest that social interactions with close others should bring the greatest rewards. Theoretically speaking, Baumeister and Leary (1995) proposed that humans have a need to belong, and suffer negative conse-quences to their health and well-being when they lack fre-quent interactions with close others. Empirically speaking, in studies of daily interactions, people report more positive affect following interactions with more familiar partners (Vittengl & Holt, 1998a), and people report being less lonely when they have more intimate interactions (Wheeler, Reis, & Nezlek, 1983). Furthermore, people who have more mean-ingful conversations report greater happiness (Mehl, Vazire, Holleran, & Clark, 2010; Reis, Sheldon, et al., 2000) and a greater sense of relatedness (Reis, Sheldon, et al., 2000).

Although the original conception of the need to belong highlighted the essential role played by strong ties, more recent research suggests that a wider range of relationships may contribute to fulfilling the need to belong. Indeed, peo-ple feel a greater sense of belonging, as well as increased positive affect, after simply having a social interaction with the barista at a coffee shop (Sandstrom & Dunn, 2013a). Also, people feel more socially connected when a passerby makes eye contact instead of looking past them (Wesselmann, Cardoso, Slater, & Williams, 2012). Thus, we hypothesized that daily interactions with weak ties would be associated with greater social and emotional well-being.

Current Research

Past research has shown that people are happier on days when they socialize more, but no research has directly assessed whether interactions with weak ties are associated with positive outcomes. We predicted that people would experience greater well-being and belonging on days when they interact with more weak ties. In Study 1, we focused on the classroom context, examining students’ daily interactions with classmates. We broadened the context in Study 2, ask-ing students (Study 2A) and community members (Study 2B) to track all of their daily interactions.

Study 1: Classroom Interactions

Method

Participants. In total, 242 undergraduate students (82 males, 160 females; M

age = 19.07, SD

age = 1.78) were recruited from

eight classes with large enrollments (M = 260) at the Univer-sity of British Columbia. Three additional students were excluded from analyses because they responded to fewer than three text messages, which is the practical minimum for within-person analyses in hierarchical linear modeling (HLM). Students participated in exchange for money or course credit and also received entries into a draw for each daily report.

Procedure. In the first few weeks of the semester, participants completed measures of personality, subjective well-being, and belonging, and reported demographic details.1

On six different occasions, students were asked questions via text message right after their class ended. The EZTexting.com service was used to schedule and send the text messages.

At the end of the semester, as they entered the classroom, students received a questionnaire, asking them to list every-one that they knew in class and to classify each of these peo-ple as a strong tie or a weak tie. At the end of the class, students returned the questionnaires and were debriefed.

MeasuresPersonality. Past research has linked well-being to all five

of the Big-Five personality traits (DeNeve & Cooper, 1998; McCrae & Costa, 1991; Weiss, Bates, & Luciano, 2008). To control for these effects, we measured openness, consci-entiousness, extraversion, agreeableness, and neuroticism using the 44-item Big-Five Inventory (John & Srivastava, 1999). The subscales all demonstrated adequate reliability (αs > .76).

Number of interactions, subjective well-being, and belonging. Students reported how many different classmates they had interacted with in person, no matter how minimal the inter-action. They counted interactions that occurred right before, during, or after class, excluding conversations that were part

by guest on April 30, 2014psp.sagepub.comDownloaded from

Sandstrom and Dunn 3

of a class requirement. Happiness was measured with a sin-gle question: “How are you feeling right now?” answered on a scale from 0 to 100 (Killingsworth & Gilbert, 2010). Finally, belonging was measured with a single item from the Sense-of-Community Scale: “I feel like I belong here” answered on a scale from 1 to 52 (Davidson & Cotter, 1986).

Results

Data preparation and descriptives. Data were mostly com-plete; only 1.5% of the reports were missing and 223 partici-pants completed all six reports.

Averaging across their six reports, participants interacted with a mean of 1.97 (minimum = 0, maximum = 9.83, SD = 1.54) classmates each day. They also reported being quite happy (M = 73.52, SD = 11.10), and feeling a moderate sense of belonging in class (M = 3.87, SD = 0.72).

Analytical method. We analyzed the data using HLM to reflect the fact that there are six measurements (text messages) nested within each participant. HLM allows the slope of the relationship between the number of interactions and well-being to be different for each individual, and allows exami-nation of both within-person and between-person effects.

The following equations were used to assess the relation-ships between the number of interactions and the happiness and belonging outcome variables:

Yij j j ij ij= + +β β ε0 1 Interactions . (1)

β β β0 00 01 0j j ju= + +Interactions .

β β1 10 1j ju= + . (2)

The outcome variable—happiness or belonging—is rep-resented by Y. At the day-level, the number of interactions (person-centered) was used to predict happiness and belong-ing (Equation 1). Testing the significance of deviations from person-centered means allows us to address within-person effects (e.g., examining whether people are happier on days when they interact with more classmates than they usually do). Following recommendations by Kreft and de Leeuw (1998), we also added the number of interactions to the model at the person-level (Equation 2), this time grand-mean centered. Testing the significance of deviations from grand means allows us to assess between-person effects (e.g., whether people who have more daily classmate interactions than others report being happier). All slopes were allowed to vary randomly.

As a measure of effect size, we report pseudo-R2 (Kreft & de Leeuw, 1998), which is computed by comparing the vari-ance explained by the full model with the variance explained by a model without the variable of interest (i.e., without classmate interactions, in this particular model). Pseudo-R2 cannot be determined exactly for models with random slopes,

such as the ones in the current analyses. Thus, these effect sizes should be interpreted cautiously, as approximations.

Between-person differences in social interactions. People who, on average, interacted with more classmates than other peo-ple reported greater happiness, b = 1.30, t(240) = 3.30, p = .001, pseudo-R2 = .05, and a greater sense of belonging, b = 0.11, t(240) = 3.95, p < .001, pseudo-R2 = .04 (see Table 1).

Given that extraverted individuals tend to have more social interactions (Watson et al., 1992) and tend to be hap-pier (Pavot et al., 1990), we tested whether these effects were due to individual differences in personality. With the Big-Five personality traits (grand-mean centered) included in the model (at the person-level), the average number of classmate interactions still predicted greater happiness, b = 0.97, t(234) = 2.49, p = .01, and belonging, b = 0.06, t(234) = 2.21, p = .03.

Within-person differences in social interactions. A more strin-gent test of our hypothesis pertains to within-person effects: Do people experience greater subjective well-being and belonging on days when they interact with more classmates than they usually do? Indeed, we found within-person effects of classmate interactions on happiness, b = 0.82, t(241) = 2.50, p = .01, pseudo-R2 = .03, and on feelings of belonging, b = 0.06, t(241) = 3.32, p = .001, pseudo-R2 = .08 (see Table 1). There is no need to control for personality or other vari-ables that differ between people, but are stable within a per-son; these factors will be effectively the same for each measurement, thus they do not contribute to the within-per-son variance.

Exploratory analyses. Given that students did not report whether each interaction partner was a weak tie or a strong tie, we could not directly test the hypothesis that weak tie interactions are associated with well-being. However, we tested this hypothesis indirectly, by examining reports of classroom friendships. We reasoned that students who reported having no close friends in class must be interacting solely with weak ties. To test whether the effect of social

Table 1. Hierarchical Linear Modeling Analysis, Predicting Happiness and Belonging From the Number of Classmate Interactions in Study 1.

Happiness Belonging

Person-level (between) Intercept 73.70 (0.70)*** 3.87 (0.05)*** Interactions (mean) 1.30 (0.39)** 0.11 (0.03)***Day-level (within) Interaction slope 0.82 (0.33)* 0.06 (0.02)**

Note. Numbers represent unstandardized coefficients, with standard errors in brackets. The approximate degree of freedom is 241 for the day-level effects and 240 for the person-level effects.*p < .05. **p < .01. ***p < .001.

by guest on April 30, 2014psp.sagepub.comDownloaded from

4 Personality and Social Psychology Bulletin

interactions on happiness and belonging differed for these students (n = 133) compared with students who had some close friends in class (n = 100), we tested a model that included a dummy variable indicating whether people reported having any strong ties in class. There was no between-person effect of this dummy variable on happiness (p = .67) or belonging (p = .13), indicating that students who had only weak tie classmates did not differ from students who had some strong tie classmates in average happiness or feelings of belonging. Turning to within-person analyses, if this dummy variable moderated the effect of the number of social interactions on happiness or belonging, it would sug-gest that the effect of an additional interaction is different depending on whether a person has any strong ties in class. However, this dummy variable did not moderate the effect of classmate interactions on either happiness (p = .87) or belonging (p = .28). These findings suggest that people felt a greater sense of happiness and belonging on days when they interacted with more classmates, even if those classmates were all weak ties.

Although we had no a priori hypotheses about factors that might moderate the observed effects, we ran exploratory analyses to test several possibilities. First, we tested whether the slope of the relationship between classmate interactions and well-being depended on personality (e.g., whether extra-verted people benefitted more from additional classmate interactions than did introverted people). The effect of class-mate interactions on happiness was stronger for people high in openness, b = 1.12, t(235) = 2.61, p = .01, but none of the other Big-Five personality traits qualified the effect (ps > .30). Similarly, the effect of classmate interactions on belong-ing was stronger for people high in openness, b = 0.07, t(235) = 3.28, p = .002, and marginally stronger for people low in extraversion, b = −0.03, t(235) = 1.92, p = .06, but was not qualified by any other personality factor (ps > .32).

Next, we tested whether the effects of classmate interac-tions on happiness and belonging leveled off. When a qua-dratic classmate interaction term was included in the model, it was not significant for either happiness (p = .41) or belong-ing (p = .14). This suggests that the effects are linear (i.e., that each additional classmate interaction is associated with additional happiness and feelings of belonging).3

Discussion

The present results show that interacting with more class-mates is linked to greater social and emotional well-being after class. Students who typically interact with more class-mates are happier and experience greater feelings of belong-ing. Furthermore, students are happier and experience greater feelings of belonging during classes when they interact with more classmates than usual. These results emerged even though participants described most of their classmates (64%) as weak ties, and our findings held up even for people who had no strong ties at all in class.

Of course, because participants did not classify whether each interaction partner was a strong or weak tie, we cannot directly compare the effects of interacting with weak versus strong ties. Thus, in Study 2, participants separately counted the number of interactions with weak ties and strong ties, allowing us to examine weak ties specifically, and to deter-mine the independent effects of each type of interaction. We also broadened the context of the interactions; instead of limiting the focus to interactions within a single class, par-ticipants kept track of all of their daily interactions. Study 1 was constrained to use short measures of happiness and belonging due to restrictions on the length of text messages, but we gathered a broader range of well-being measures in Study 2.

Study 2A: Daily Weak Tie Interactions (Students)

Method

Participants. A total of 58 first-year university students (15 males and 43 females, M

age = 19.22, SD

age = 3.24) received

either course credit or Cdn$30 for completing the study.

Procedure. At the beginning of the semester, participants came to the lab to complete demographic and personality questions, and receive instructions. Participants received a pair of mechanical tally counters (“clickers”) and were given detailed instructions on how to distinguish between interac-tions with strong ties and weak ties, which they were to count using red and black clickers, respectively. We provided sev-eral possible criteria for distinguishing a strong tie: “Some-one you are very close to,” “someone who you know really well (and knows you really well),” or “someone who you confide in or talk to about yourself or your problems.” For weak ties, criteria were as follows: “Someone you are not very close to,” “someone who you don’t know really well (and who doesn’t know you really well),” or “someone who you consider a friend, but would be unlikely to confide in.”

For 3 days in a row in September and 3 days in a row in November (a Tuesday, Wednesday, and Thursday each time), participants were asked to carry the clickers with them at all times and keep track of all of their social interactions. Participants counted each time that they greeted someone in person, regardless of the length of the interaction. At the end of each day, in addition to reporting the number of strong tie and weak tie interactions, participants reported their subjec-tive well-being and belonging via an online questionnaire.4

MeasuresPersonality. Participants rated their openness, conscien-

tiousness, extraversion, agreeableness, and neuroticism using the Ten-Item Personality Inventory (Gosling, Rentfrow, & Swann, 2003; see Table 2 for reliability, sample items, and response options for all measures).

by guest on April 30, 2014psp.sagepub.comDownloaded from

Sandstrom and Dunn 5

Number of interactions. Participants reported the tallies from the clickers (i.e., the number of interactions they had with strong and weak ties).

Subjective well-being. We measured subjective well-being broadly, assessing positive and negative affect, as well as subjective happiness.

Affect. To measure positive affect and negative affect, we used the Scale of Positive and Negative Experience (Diener et al., 2010). This scale was developed to measure trait-level affect, asking people to report the frequency of various feel-ings over the past 4 weeks. We adapted it to measure state-level affect, asking participants to “think about what you have been doing and experiencing today.”

Subjective happiness. Participants provided a global assessment of their own happiness, using the Subjective Happiness Scale (Lyubomirsky & Lepper, 1999). Although this measure was designed to assess trait-level happiness, past research in our lab has found that this measure can also detect short-term changes in well-being (Aknin et al., 2013).

Subjective well-being composite. In the absence of past research on weak ties to drive more specific hypotheses, we chose to use several measures designed to capture different aspects of subjective well-being, and to focus on the overall relationship between weak ties and a composite measure of subjective well-being. We standardized the positive affect, negative affect (reverse-scored), and subjective happiness measures and averaged them to create a composite measure (α = .91). For transparency, we also report the results for the individual measures in Table 3.

Belonging. To assess participants’ sense of belonging, we used several different measures that tap into related con-structs.

Social connectedness. We assessed how connected par-ticipants felt to others in general (i.e., not limited to the uni-versity community), using the Social Connectedness Scale (Lee, Draper, & Lee, 2001). We selected items with the highest reported factor loadings on the common construct, but excluded items that referred to “friends” or “people I

Table 2. Reliability, Sample Items, and Response Options for the Measures in Studies 2A and 2B.

MeasureReliability (α)

Study 2AReliability (α)

Study 2BNo. of items

No. of items reverse-scored Sample item(s) Response options

Personality O: .35C: .58E: .70A: .47N: .39

O: .23C: .78E: .81A: .62N: .44

O: 2C: 2E: 2A: 2N: 2

O: 1C: 1E: 1A: 1N: 1

“Extraverted, enthusiastic”;

“Reserved, quiet”

1 = disagree strongly7 = agree strongly

Subjective well-being .91 .95 Positive affect .90 .93 6 0 “Pleasant” 1 = very rarely or never

5 = very often or always Negative affect .87 .90 6 0 “Unpleasant” 1 = very rarely or never

5 = very often or always Subjective happiness .86 .93 4 1 “In general I

consider myself”:1 = not a very happy person7 = a very happy person

Belonging .92 .94 Social

connectedness.87 .84 11a 7 “I am able to relate

to my peers”1 = strongly disagree6 = strongly agree

Social support .73 .89 5a 0 “When I need suggestions on how to deal with a personal problem, I know someone I can turn to”

1 = definitely false4 = definitely true

Loneliness .88 .92 10a 5 “How often do you feel alone?”

1 = never4 = always

Sense of community .84 .57 8a 2 “I feel like I belong here”

1 = strongly disagree4 = strongly agree

Note. For the personality measure, reliability was assessed across the two items constituting each of the Big-Five factors: openness (O), conscientiousness (C), extraversion (E), agreeableness (A), neuroticism (N). For the remaining measures, the alpha was averaged across the different measurements. The subjective happiness measure has different response options for each question; response options for one question are provided here. Due to experimenter error, connectedness was not collected on the first 2 days in Study 2A. In Study 2B, sense of community was not included in the belonging composite due to low correlations with the other measures of belonging.aFor these scales, we selected a subset of the items (see online appendix for details).

by guest on April 30, 2014psp.sagepub.comDownloaded from

6 Personality and Social Psychology Bulletin

Table 4. Hierarchical Linear Modeling Analysis, Predicting the Belonging Composite Measure, and All Its Component Measures, From the Number of Strong and Weak Tie Interactions in Study 2A.

Effect Belonging composite Social connectedness Social support Loneliness Sense of community

Person-level (between) Intercept 0.01 (0.08) 4.53 (0.10)*** 3.35 (0.05)*** 2.08 (0.05)*** 3.23 (0.05)*** Strong ties (mean) 0.03 (0.01)*** 0.03 (0.01)*** 0.01 (0.01)* −0.02 (0.01)*** 0.01 (0.005) Weak ties (mean) 0.02 (0.01)* 0.02 (0.01)* 0.004 (0.007) −0.01 (0.005)† 0.02 (0.01)**Day-level (within) Strong tie slope 0.01 (0.003)* 0.001 (0.005) 0.01 (0.004)* −0.003 (0.003) 0.002 (0.002) Weak tie slope 0.002 (0.003) 0.003 (0.005) −0.001 (0.003) 0.002 (0.002) 0.004 (0.002)†

Note. Numbers represent unstandardized coefficients, with standard errors in brackets. The approximate degree of freedom is 57 for the day-level effects and 55 for the person-level effects.†p < .10. *p < .05. **p < .01. ***p < .001.

know,” to tap into the idea of connectedness to people in general, not just those in one’s social circle (see online appendix for details about the items we used).

Social support. Students reported the extent to which they felt they had people they could count on by answering ques-tions on the Interpersonal Support Evaluation List (Cohen, Mermelstein, Kamarck, & Hoberman, 1985). Following pre-vious research (Martire, Schulz, Mittelmark, & Newsom, 1999), we selected five items measuring emotional, informa-tional, and instrumental support.

Loneliness. We measured loneliness using the UCLA Loneliness Scale, version 3 (Russell, 1996). We selected from among the items with the strongest reported item-total correlations for student samples, while not selecting questions that overlapped too much with other constructs that we were measuring. We selected questions that referred to people in general, rather than friends and family.

Sense of community. We measured the degree to which participants felt like they were a part of the university com-munity with the Sense-of-Community Scale (Davidson & Cotter, 1986). Though this scale was originally developed to

look at sense of community within a city, we adapted it to measure sense of community at a university.

Belonging composite. As with subjective well-being, we chose to use a broad range of measures and rely upon a com-posite measure to examine the overall pattern of associations with weak ties. We standardized the social connectedness, social support, loneliness (reverse-scored) and sense of com-munity measures and averaged them to create a composite measure (α = .92). For transparency, we also report the results for the individual measures in Table 4.

Results

Descriptives. Data were mostly complete; only 1.4% of the reports were missing and 54 participants completed all six reports.

Averaging across their six reports, participants interacted with a mean of 9.65 strong ties (minimum = 0, maximum = 46, SD = 9.18) and 15.96 weak ties (minimum = 2.80, maximum = 54.83, SD = 9.45) each day.

Analytical method. We used the same analytical approach as in Study 1. To assess the relationships between the number of

Table 3. Hierarchical Linear Modeling Analysis, Predicting the Subjective Well-Being Composite Measure, and All Its Component Measures, From the Number of Strong and Weak Tie Interactions in Study 2A.

Effect Subjective well-being composite Positive affect Negative affect Subjective happiness

Person-level (between) Intercept 0.0003 (0.08) 3.61 (0.07)*** 2.02 (0.07)*** 5.08 (0.12)*** Strong ties (mean) 0.02 (0.01)** 0.01 (0.007)* −0.001 (0.005) 0.04 (0.01)** Weak ties (mean) 0.02 (0.01)** 0.01 (0.01)* −0.02 (0.01)*** 0.01 (0.01)Day-level (within) Strong tie slope 0.01 (0.005)* 0.01 (0.01)* −0.004 (0.005) 0.005 (0.004) Weak tie slope 0.02 (0.005)** 0.03 (0.01)*** −0.02 (0.005)** 0.004 (0.004)

Note. Numbers represent unstandardized coefficients, with standard errors in brackets. The approximate degree of freedom is 57 for the day-level effects and 55 for the person-level effects.*p < .05. **p < .01. ***p < .001.

by guest on April 30, 2014psp.sagepub.comDownloaded from

Sandstrom and Dunn 7

strong and weak tie interactions and the subjective well-being and belonging outcome variables, we used the follow-ing equations:

Yij j j ij j i ij= + + +β β β ε0 1 2Strong Weak j . (3)

β β β β0 00 01 02 0j j j ju= + + +Strong Weak .

β β1 10 1j ju= + .

β β2 20 2j ju= + .

(4)

By including both strong and weak tie interaction counts as predictors, we can assess the individual contribution of each. As in Study 1, the number of strong and weak tie interactions were person-centered at the day-level (i.e., within-person) and grand-mean centered at the person-level (i.e., between-per-son), and all slopes were allowed to vary randomly.

Between-person differences in social interactions. People who, on average, interacted with more weak ties than other indi-viduals reported greater average subjective well-being, b = 0.02, t(55) = 3.03, p = .004, pseudo-R2 = .04 (see Table 3), and greater average belonging, b = 0.02, t(55) = 2.19, p = .03, pseudo-R2 = .06 (see Table 4). Similarly, people who, on average, interacted with more strong ties reported greater subjective well-being, b = 0.02, t(55) = 2.71, p = .01, pseudo-R2 = .08, and greater belonging, b = 0.03, t(55) = 3.99, p < .001, pseudo-R2 = .14. Given that both weak tie and strong tie interactions were entered simultaneously into the model, this suggests that each type of interaction has an independent effect on social and emotional well-being.

These effects were not solely attributable to individual differences in personality: With the Big-Five personality traits included in the model, more daily weak tie interactions still predicted greater average subjective well-being, b = 0.01, t(50) = 2.39, p = .02, and marginally greater average belonging, b = 0.01, t(50) = 1.81, p = .08. More daily strong tie interactions still predicted greater belonging, b = 0.02, t(50) = 3.14, p = .003, but the effect on subjective well-being dropped below conventional levels of significance, b = 0.01, t(50) = 1.65, p = .11.

Within-person differences in social interactions. A more strin-gent test of our hypothesis pertains to within-person effects: Do people experience greater subjective well-being and belonging on days when they have more interactions than they usually do? Indeed, we found a within-person effect of weak tie interactions on subjective well-being, b = 0.02, t(57) = 3.36, p = .001, pseudo-R2 = .10 (see Table 3), but not on belonging, p = .62 (see Table 4).

Strong ties were also related to social and emotional well-being. On days when people interacted with more strong ties than usual, they reported greater subjective well-being, b = 0.01, t(57) = 2.26, p = .03, pseudo-R2 = .02, and a

greater sense of belonging, b = 0.01, t(57) = 2.22, p = .03, pseudo-R2 = .01.

Exploratory analyses. Although we had no a priori hypotheses about factors that might moderate the observed effects, we ran exploratory analyses to test several possibilities. First, we tested whether the within-person effects of weak tie interactions on subjective well-being and belonging were qualified by the num-ber of strong tie interactions. When a Weak Tie × Strong Tie interaction term was included in the model, the coefficient was significant for subjective well-being, b = −0.0005, t(57) = −3.98, p < .001. A plot of the simple slopes (computed via online cal-culators; Preacher, Curran, & Bauer, 2006; Shacham, 2009) shows that the effect of additional weak tie interactions is stron-ger on days when people have fewer strong tie interactions than usual (see Figure 1). The Weak Tie × Strong Tie interaction term was not significant for belonging, p = .26.

Next, we tested whether the slope of the relationship between interactions and well-being differed for people who had different personality traits (e.g., whether extraverted people benefitted more from additional interactions than did introverted people). The effects of weak tie and strong tie interactions on subjective well-being were not qualified by any of the Big-Five personality traits (ps > .39 and ps > .43, respectively). However, the within-person effect of weak tie interactions on belonging was stronger for people low in extraversion, b = −0.01, t(52) = 3.26, p = .002 (other ps > .45). In other words, each additional weak tie interaction resulted in a greater increase in belonging for people low in

Figure 1. The Weak Tie × Strong Tie interaction in Study 2A.Note. Both variables are person-centered and thus represent deviations from a person’s usual number of daily interactions. The simple slopes are plotted at −1 SD and +1 SD from the mean number of daily strong tie interactions.

by guest on April 30, 2014psp.sagepub.comDownloaded from

8 Personality and Social Psychology Bulletin

extraversion than for people high in extraversion. Also, the within-person effect of strong tie interactions on belonging was stronger for people low in openness, b = −0.01, t(52) = −2.48, p = .02, and people high in agreeableness, b = 0.01, t(52) = 2.70, p = .01 (other ps > .14).

Finally, we tested whether the effects of social interac-tions leveled off, by adding quadratic strong and weak tie interaction terms to the model. The more interactions some-one has, the smaller the effect of each additional interaction; the quadratic weak tie interaction term was significant for subjective well-being, b = −0.0003, t(57) = −2.50, p = .02, but not belonging, b = 0.0001, t(57) = 0.64, p = .53, whereas the quadratic strong tie interaction term was significant for both subjective well-being, b = −0.0003, t(57) = −2.67, p = .01, and belonging, b = −0.0004, t(57) = −3.65, p = .001.

Discussion

The present results suggest that the benefits of weak tie inter-actions are not restricted to the classroom. Students who usu-ally have more daily interactions with weak ties than others are happier and experience greater feelings of belonging. Furthermore, on days when students interact with more weak ties than usual, they are happier.

Studies 1 and 2A are limited in that they both use student populations, thus restricting the generalizability of the find-ings. The daily context of a university student differs from the daily context of a community member in several respects. For one thing, students spend many hours on campus in close proximity to other students, thus potentially exposing them to more frequent opportunities for interaction. Furthermore, university students have more weak ties in their social net-works. Past research has shown that social networks grow smaller as people age (Carstensen, 1992; Fredrickson & Carstensen, 1990; Wrzus, Hänel, Wagner, & Neyer, 2013), and that the reductions are mostly in the periphery of the social network (i.e., weak ties; Lang, 2000; Lang & Carstensen, 1994). Although often studied in older adults, some evidence suggests that this reduction of weak ties may begin in late adolescence (Carstensen, 1992; Wrzus et al., 2013). Given these differences in the number of weak tie relationships and in opportunities for interactions, we ran a small follow-up study to explore the critical issue of general-izability: Do the effects of weak ties only appear for univer-sity students, or can they be detected in a broader community sample?

Study 2B: Daily Weak Tie Interactions (Community Members)

Method

Participants. Clickers were mailed to 53 community members (16 males and 37 females), all older than 25. Participants were paid Cdn$30 for completing the study.

Procedure. Participants were recruited via print and online classified advertisements and were phone screened by a research assistant. If the participant was eligible to complete the study,5 they learned about the purpose of the study, received an overview of the procedure, and were told how much they would be paid. Upon confirming their interest in completing the study, the procedures were explained in more detail, and the criteria for distinguishing between strong ties and weak ties were described.

After the phone screening, but before being sent the click-ers, participants completed an online questionnaire, rating their personality and reporting the same demographic infor-mation as in Study 2A. This questionnaire was used to solicit consent, and to assess commitment toward the study before incurring the cost of shipping the clickers. After participants received the clickers, they used them for 6 days over the course of approximately 2 weeks, counting all of their social interactions with both strong and weak ties and filling out the same online questionnaire as in Study 2A.

MeasuresPersonality. Participants again rated their personality

using the TIPI (Gosling et al., 2003; see Table 2 for sample items, response options, and reliability for all measures).

Number of interactions, subjective well-being, and belonging. Participants reported the interaction tallies from the clickers.6 We measured subjective well-being and belonging broadly, using the same scales as in Study 2A, and creating the same composite measures (subjective well-being: α = .95; belong-ing: α = .94). We omitted the sense of community measure from the belonging composite due to its low reliability and low correlations with the other measures of belonging (|r|s < .26, ps > .10).

Results

Attrition and descriptives. Although we aimed to recruit as many participants as in Study 2A, there was a high rate of attri-tion. Three people claimed that they did not receive the click-ers, and three people decided not to participate after receiving the clickers. In addition, three people completed less than three surveys, which is the practical minimum for use in within-person analyses with HLM. Finally, three people sub-mitted the majority of their reports in the morning, rather than at the end of the day, thus preventing us from seeing the effect of a day’s interactions on that day’s well-being. After remov-ing these participants from all analyses, the final sample was 41 community members (11 males and 30 females), for whom data were mostly complete; only 3.25% of the reports were missing and 35 participants completed all six reports.

Averaging across their reports, participants interacted with a mean of 6.70 strong ties (minimum = 0.33, maximum = 22.50, SD = 5.81) and 11.40 weak ties (minimum = 0, max-imum = 74.67, SD = 12.96) each day.

by guest on April 30, 2014psp.sagepub.comDownloaded from

Sandstrom and Dunn 9

Between-person differences in social interactions. People who, on average, interacted with more weak ties than other peo-ple reported significantly greater average feelings of belong-ing, b = .02, t(38) = 2.89, p = .01, pseudo-R2 = .05 (see Table 6), and slightly (but non-significantly) greater subjec-tive well-being, b = .01, t(38) = 1.45, p = .16 (see Table 5). The average number of strong tie interactions was not related to average feelings of belonging, b = .001, t(38) = 0.03, p = .98, or average subjective well-being, b = −.02, t(38) = −1.15, p = .26.

Within-person differences in social interactions. A more strin-gent test of our hypothesis pertains to within-person effects: Do people experience greater subjective well-being and belonging on days when they have more interactions than they usually do? Indeed, people reported greater belonging, b = .002, t(40) = 2.84, p = .01, pseudo-R2 = .001 (see Table 6), though not greater subjective well-being, b = −.0004, t(40) = −0.29, p = .77 (see Table 5), on days when they inter-acted with more weak ties than usual.

Strong tie interactions were also related to well-being. On days when people interacted with more strong ties than usual, they reported greater subjective well-being, b = 0.01,

t(40) = 2.77, p = .01, pseudo-R2 = .02, and a greater sense of belonging, b = 0.01, t(40) = 2.14, p = .04, pseudo-R2 = .004.

Discussion

The results of this small follow-up study suggest that the associations between weak tie interactions and social and emotional well-being are not limited to university students. In a community sample, people who interacted with more weak ties than others or interacted with more weak ties than usual felt a greater sense of belonging.

General Discussion

The current studies provide initial, convergent evidence that interactions with people on the periphery of our social net-works may contribute to our social and emotional well-being. In Study 1, students who had, on average, more interactions with their classmates than others reported greater feelings of belonging and greater happiness. After classes in which they interacted with more classmates than usual, students reported greater feelings of belonging and greater happiness. We found similar results when we expanded the scope to include

Table 6. Hierarchical Linear Modeling Analysis, Predicting the Belonging Composite Measure, and All Its Component Measures, From the Number of Strong and Weak Tie Interactions in Study 2B.

Effect Belonging composite Social connectedness Social support Loneliness Sense of community

Person-level (between) Intercept −.01 (.13) 4.47 (.17)*** 3.27 (.10)*** 2.03 (.09)*** 2.30 (.06)*** Strong ties (mean) .001 (.02) −.01 (.03) .01 (.02) .003 (.01) .004 (.009) Weak ties (mean) .02 (.01)** .02 (.01)* .01 (.005)** −.01 (.004)* .005 (.002)*Day-level (within) Strong tie slope .01 (.003)* .01 (.004)** .003 (.002) −.004 (.005) .00003 (.003) Weak tie slope .002 (.001)** .003 (.002)† .001 (.001) −.002 (.001)* .002 (.001)

Note. Numbers represent unstandardized coefficients, with standard errors in brackets. The approximate degree of freedom is 40 for the day-level effects and 38 for the person-level effects. Sense of community is not included in the belonging composite due to low reliability and low correlations with the other measures.†p < .10. *p < .05. **p < .01. ***p < .001.

Table 5. Hierarchical Linear Modeling Analysis, Predicting the Subjective Well-Being Composite Measure, and All Its Component Measures, From the Number of Strong and Weak Tie Interactions in Study 2B.

Effect Subjective well-being composite Positive affect Negative affect Subjective happiness

Person-level (between) Intercept .002 (.12) 3.69 (.12)*** 1.67 (.09)*** 5.23 (.19)*** Strong ties (mean) −.02 (.02) −.01 (.02) .01 (.02) −.05 (.03)†

Weak ties (mean) .01 (.01) .01 (.01) −.01 (.005) .02 (.01)†

Day-level (within) Strong tie slope .01 (.005)** .03 (.01)*** −.0003 (.01) .01 (.01) Weak tie slope −.0004 (.001) .003 (.002) .003 (.001)* −.0001 (.002)

Note. Numbers represent unstandardized coefficients, with standard errors in brackets. The approximate degree of freedom is 40 for the day-level effects and 38 for the person-level effects.†p < .10. *p < .05. **p < .01. ***p < .001.

by guest on April 30, 2014psp.sagepub.comDownloaded from

10 Personality and Social Psychology Bulletin

all daily interactions. In Study 2A, students who had, on average, more daily weak tie interactions than others reported greater feelings of belonging and greater happiness. Furthermore, they reported greater happiness on days when they interacted with more weak ties than usual. Finally, in Study 2B we found evidence that the effects are not limited to students. In particular, community members who had, on average, more weak tie interactions than others reported greater feelings of belonging. Furthermore, people reported greater feelings of belonging on days when they interacted with more weak ties than usual. As the first studies in social psychology to focus explicitly on the positive outcomes associated with weak tie interactions, the present research points to the important role played by the peripheral mem-bers of individuals’ social networks in daily life.

The results of our exploratory analyses suggest fruitful areas for further research. We found that the effect of each additional weak tie interaction was greater on days when peo-ple had fewer daily weak tie interactions or fewer daily strong tie interactions. Although we did not predict these effects, it makes sense intuitively that on days when people have fewer interactions, they might benefit the most from having more weak tie interactions. In addition, we explored whether the effects depended on personality, but found little evidence for moderation. Consistent with past research (Fleeson, Malanos, & Achille, 2002), we found evidence that the effect of weak ties was not limited to extraverts; indeed, having interactions with a broad range of network members might be especially beneficial for those who are low in extraversion.

Limitations

Although the overall pattern of results suggests that weak tie interactions are broadly associated with subjective well-being and belonging, the strength and significance of the effects var-ied across studies and measures. This variability may be due to differences in the social contexts and types of interactions experienced by the different samples. For example, in a recent study in our lab, students classified 57% of their interactions as being “just for fun” (Sandstrom & Dunn, 2013b). Although we do not have comparable data for a community sample, we suspect that community members have fewer social, elective interactions, and more instrumental, required interactions (e.g., with service providers and colleagues at work). Unfortunately, in the current studies we only had information about the number of interactions, and not about the length of the interactions, or the content or emotional quality of the interactions. This information would allow us to determine the relative strength of positive and negative interactions on one’s well-being, and to weigh the quantity of interactions against the quality. Future research is needed to investigate how these interaction attributes differ between contexts, and how these differences shape the effects on well-being.

One important limitation of the current studies is that the methods are correlational and thus causal direction cannot be

determined. We suspect that interacting with more weak ties results in people feeling better, but it is also possible that feel-ing better causes people to interact with more weak ties; indeed, these possibilities are not mutually exclusive, and future work should test both causal directions. For example, participants could be randomly assigned to participate in an intervention designed to enhance daily mood or sense of belonging, and changes in daily interactions could be assessed to determine whether people interact more with weak ties as a result of feeling better. Conversely, one could test whether interactions with weak ties cause greater well-being by encouraging people to interact with more (or fewer) weak ties and measuring resultant changes in well-being. Initial research in our lab, however, suggests that simply instructing people to alter their social interactions might be ineffective (see Note 1); instead, it may be necessary to creatively struc-ture situations that foster (or limit) weak tie interactions.

Implications

The present studies have implications for future research link-ing social interactions to psychological outcomes. The vast majority of past studies examining the association between social interactions and subjective well-being relied upon ret-rospective reports. Our findings reveal the shortcomings of this methodology. Students in Study 2A reported interacting with approximately 9.5 strong ties and 16.5 weak ties each day. Previous studies have not always reported the average number of daily interactions, but two studies with student populations found means of 6 to 10 interactions per day (Asendorpf & Wilpers, 1998; Berry & Hansen, 1996). Thus, by explicitly asking participants to count weak tie interactions as they occurred, issues related to recall were reduced in the current studies: We captured more than double the number of social interactions that were reported in previous studies. Consideration of the effects of these interactions has been missing in previous research, and future research might ben-efit by examining the full spectrum of social relationships.

The current studies also suggest a possible modification to the theoretical account of the need to belong (Baumeister & Leary, 1995). Originally, this need was thought to be satis-fied only by frequent interactions with close others. However, the current studies are consistent with more recent research suggesting that even minimal relationships can play a role in fulfilling the need to belong (Sandstrom & Dunn, 2013a, Wesselmann et al., 2012). Thus, the present results join a chorus of other new research in pointing to the value of re-examining the role played by our broader social networks in fulfilling the fundamental human need to belong.

The results of Study 2B have implications for socio-emo-tional selectivity theory (SST; Carstensen, 1992), which pro-poses that the reason why people retain fewer weak ties as they age is because they find these relationships less fulfilling. In contrast, we found evidence that these relationships are ful-filling: People of various ages (though younger than those

by guest on April 30, 2014psp.sagepub.comDownloaded from

Sandstrom and Dunn 11

usually studied with respect to SST) experienced a greater sense of belonging on days when they interacted with more weak ties. Thus, although SST suggests that the pruning of weak ties is a functional decision in service of emotion regula-tion, it is possible that the decision is actually a result of people underestimating the benefits of weak ties. Indeed, past research suggests that young adults underestimate the emotional bene-fits of interacting with people they do not know well (Dunn, Biesanz, Human, & Finn, 2007). Future research might test whether older adults make similar—and perhaps exacer-bated—affective forecasting errors, leading them to prune their social networks at a cost to their overall well-being.

Why might weak ties contribute to well-being over and above the well-studied benefits of strong ties? One intriguing possibility is that weak ties promote well-being by contribut-ing diversity to the social network. Just as a diverse financial portfolio makes people less vulnerable to market fluctua-tions, a diverse social portfolio might make people less vul-nerable to fluctuations in their social network. Indeed, the diversity of one’s social network has been identified as a pro-tective factor against disease development and mortality across a broad range of illnesses (Berkman, 1995; Cohen & Janicki-Deverts, 2009). There may be added value in having a wide circle of weak ties who offer companionship in differ-ent contexts (e.g., some acquaintances at the dog park, differ-ent acquaintances at yoga class).

Conclusion

The current research suggests that we should not underesti-mate the value of our acquaintances—interactions with weak ties are related to our subjective well-being and feelings of belonging. Although further research is needed to examine causality, the current results are consistent with the idea that the more peripheral members of our social network shape our day-to-day happiness. So, chat with the coffee barista, work colleague, yoga classmate, and dog owner—these interactions may contribute meaningfully to our happiness, above and beyond the contribution of interactions with our close friends and family.

Acknowledgments

The authors wish to thank the following people for assisting with data collection: Richard Abasta, Kate Block, Lysandra Chan, Carrie Cheung, Amanda Chui, Jen Fatkin, Ashton Hung, Audrey Kim, Kostadin Kushlev, Stephanie Kwong, Victoria Lau, Genevieve Lorenzo, Jenn McDermid, Ben Pierce, Kate Rogers, Jael Van Bentum, Tess Walker, Jen Won, Fontayne Wong. We also wish to thank Lauren Human for contributing to the study design, Carl Falk and Christiane Hoppmann for statistical advice, and Jill Allen for reading a previous version of the manuscript.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

This work was supported by the Social Sciences and Humanities Research Council of Canada [grant numbers 752-2011-2192, 410-2011-0582].

Notes

1. Data were collected as part of a broader study designed to experimentally investigate the effects of in-class interactions on happiness. Participants randomly assigned to the experimental condition had the opportunity to meet several classmates, and on several occasions during the semester were encouraged to keep in touch with one another. This manipulation was inef-fective; there was no difference in the average number of daily in-class interactions for the experimental group (M = 1.97, SD = 1.73) compared with the control group (M = 1.96, SD = 1.30), t(240) = .06, p = .96. As a result, the current study ignores the condition assignment. Importantly, the groups also did not differ in their average daily reports of happiness (p = .72) or belonging (p = .56).

2. By asking the belonging question on a 1 to 5 scale, and the hap-piness question on a 0 to 100 scale, we may have confused some of the participants. Nine responses to the belonging question that were greater than five, and 16 responses to the happiness ques-tion that were more than three standard deviations below the mean were removed from analyses. We also asked students how much they enjoyed class, which is not relevant to the current research question.

3. We also tested a model that included the dummy code indicating whether participants had strong ties in class. There was no inter-action between this dummy variable and the quadratic terms for either happiness (p = .78) or belonging (p = .26), suggesting that there is no difference in the amount of leveling off for people with versus without strong ties in class.

4. In addition to the measures listed, we also measured Flourishing (Diener et al., 2010) as an index of subjective well-being and Positive Relations With Others (Ryff, 1989) as an index of belonging. These measures were included on an exploratory basis, and were only assessed twice, rendering them unsuitable for within-person analysis with HLM, which effectively requires at least three measurements.

5. People were considered eligible if they were above the age of 25 (i.e., older than the average university student), had daily social interactions, and had daily access to a computer to complete the online questionnaires.

6. Unfortunately the fields used to gather the number of daily interactions in the online survey could be left blank. On occa-sions when participants reported a non-zero number of strong tie interactions, we assumed that they had interacted with zero weak ties when that field was left blank (26 participant reports). Similarly, when participants reported a non-zero number of weak tie interactions, we treated missing strong tie interactions as zero (21 reports). In contrast, if both the number of weak tie and strong tie interactions were blank, they were assumed to be missing rather than zero. When blank interaction reports are excluded rather than assumed to be zero, the results remain similar. There are between-person effects of weak tie interac-tions on belonging, b = .01, t(38) = 2.53, p = .02, but not sub-jective well-being, p = .65, and no between-person effects of strong tie interactions on either belonging, p = .64, or subjective

by guest on April 30, 2014psp.sagepub.comDownloaded from

12 Personality and Social Psychology Bulletin

well-being, p = .32. There was a within-person effect of weak tie interactions on belonging, b = .004, t(40) = 2.14, p = .04, but not subjective well-being, p = .28, and a marginal within-person effect of strong tie interactions on subjective well-being, b = .01, t(40) = 1.88, p = .07, but not belonging, p = .98.

References

Aknin, L. B., Barrington-Leigh, C. P., Dunn, E. W., Helliwell, J. F., Burns, J., Biswas-Diener, R., . . .Norton, M. I. (2013). Prosocial spending and well-being: Cross-cultural evidence for a psychological universal. Journal of Personality and Social Psychology, 104, 635-652. doi: 10.1037/a0031578

Asendorpf, J. B., & Wilpers, S. (1998). Personality effects on social relationships. Journal of Personality and Social Psychology, 74, 1531-1544. doi:10.1037/0022-3514.74.6.1531

Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motiva-tion. Psychological Bulletin, 117, 497-529. doi:10.1037/0033-2909.117.3.497

Berkman, L. F. (1995). The role of social relations in health promo-tion. Psychosomatic Medicine, 57, 245-254.

Berry, D. S., & Hansen, J. S. (1996). Positive affect, negative affect, and social interaction. Journal of Personality and Social Psychology, 71, 796-809. doi:10.1037/0022-3514.71.4.796

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

Clark, L. A., & Watson, D. (1988). Mood and the mundane: Relations between daily life events and self-reported mood. Journal of Personality and Social Psychology, 54, 296-308. doi:10.1037/0022-3514.54.2.296

Cohen, S., & Janicki-Deverts, D. (2009). Can we improve our physical health by altering our social networks? Perspectives on Psychological Science, 4, 375-378. doi:10.1111/j.1745-6924.2009.01141.x

Cohen, S., Mermelstein, R., Kamarck, T., & Hoberman, H. (1985). Measuring the functional components of social support. In I. G. Sarason & B. R. Sarason (Eds.), Social support: Theory, research and applications (pp. 73-94). The Hague, The Netherlands: Martinus Nijhoff.

Davidson, W. B., & Cotter, P. R. (1986). Measurement of sense of community within the sphere of city. Journal of Applied Social Psychology, 16, 608-619. doi:10.1111/j.1559-1816.1986.tb01162.x

DeNeve, K. M., & Cooper, H. (1998). The happy personality: A meta-analysis of 137 personality traits and subjective well-being. Psychological Bulletin, 124, 197-229.

Diener, E., Wirtz, D., Tov, W., Kim-Prieto, C., Choi, D., Oishi, S., & Biswas-Diener, R. (2010). New well-being measures: Short scales to assess flourishing and positive and negative feelings. Social Indicators Research, 97, 143-156. doi:10.1007/s11205-009-9493-y

Dunn, E. W., Biesanz, J. C., Human, L. J., & Finn, S. (2007). Misunderstanding the affective consequences of everyday social interactions: The hidden benefits of putting one’s best face forward. Journal of Personality and Social Psychology, 92, 990-1005. doi:10.1037/0022-3514.92.6.990

Eagle, N., Pentland, A. S., & Lazer, D. (2009). Inferring friendship network structure by using mobile phone data. Proceedings

of the National Academy of Sciences, 106, 15274-15278. doi:10.1073/pnas.0900282106

Fleeson, W., Malanos, A. B., & Achille, N. M. (2002). An intrain-dividual process approach to the relationship between extraver-sion and positive affect: Is acting extraverted as “good” as being extraverted? Journal of Personality and Social Psychology, 83, 1409-1422. doi:10.1037/0022-3514.83.6.1409

Fredrickson, B. L., & Carstensen, L. L. (1990). Choosing social partners: How old age and anticipated endings make people more selective. Psychology and Aging, 5, 335-347.

Fu, Y. (2005). Measuring personal networks with daily contacts: A single-item survey question and the contact diary. Social Networks, 27, 169-186. doi:10.1016/j.socnet.2005.01.008

Gosling, S. D., Rentfrow, P. J., & Swann, W. B. (2003). A very brief measure of the Big-Five personality domains. Journal of Research in Personality, 37, 504-528. doi:10.1016/S0092-6566(03)00046-1

Granovetter, M. S. (1973). The strength of weak ties. American Journal of Sociology, 78, 1360-1380.

John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: History, measurement, and theoretical perspectives. In L. A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research (Vol. 2, pp. 102-138). New York: Guilford.

Kahneman, D., Krueger, A. B., Schkade, D. A., Schwarz, N., & Stone, A. A. (2004). A survey method for characterizing daily life experience: The day reconstruction method. Science, 306, 1776-1780. doi:10.1126/science.1103572

Killingsworth, M. A., & Gilbert, D. T. (2010). A wandering mind is an unhappy mind. Science, 330, 932. doi:10.1126/science.1192439

Kreft, I. G. G., & de Leeuw, J. (1998). Introducing multilevel mod-eling. Thousand Oaks, CA: Sage.

Krueger, A. B., Kahneman, D., Schkade, D., Schwarz, N., & Stone, A. A. (2009). National time accounting: The currency of life. In A. B. Krueger (Ed.), Measuring the subjective well-being of nations: National accounts of time use and well-being (pp. 9-86). Chicago, IL: University of Chicago Press.

Lang, F. R. (2000). Endings and continuity of social relation-ships: Maximizing intrinsic benefits within personal networks when feeling near to death. Journal of Social and Personal Relationships, 17, 155-182. doi:10.1177/0265407500172001

Lang, F. R., & Carstensen, L. L. (1994). Close emotional relation-ships in late life: Further support for proactive aging in the social domain. Psychology and Aging, 9, 315-324.

Lee, R. M., Draper, M., & Lee, S. (2001). Social connectedness, dysfunctional interpersonal behaviors, and psychological distress: Testing a mediator model. Journal of Counseling Psychology, 48, 310-318. doi:10.1037/0022-0167.48.3.310

Lyubomirsky, S., & Lepper, H. S. (1999). A measure of sub-jective happiness: Preliminary reliability and construct validation. Social Indicators Research, 46, 137-155. doi:10.1023/A:1006824100041

Martire, L. M., Schulz, R., Mittelmark, M. B., & Newsom, J. T. (1999). Stability and change in older adults’ social contact and social support: The Cardiovascular Health Study. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 54, S302-S311. doi:10.1093/geronb/54B.5.S302

McCrae, R. R., & Costa, P. T. (1991). Adding Liebe und Arbeit: The full five-factor model and well-being. Personality and Social Psychology Bulletin, 17, 227-232. doi:10.1177/014616729101700217

by guest on April 30, 2014psp.sagepub.comDownloaded from

Sandstrom and Dunn 13

Mehl, M. R., Vazire, S., Holleran, S. E., & Clark, C. S. (2010). Eavesdropping on happiness. Psychological Science, 21, 539-541. doi:10.1177/0956797610362675

Pavot, W., Diener, E., & Fujita, F. (1990). Extraversion and happi-ness. Personality and Individual Differences, 11, 1299-1306. doi:10.1016/0191-8869(90)90157-M

Perry-Smith, J. E. (2006). Social yet creative: The role of social relationships in facilitating individual creativity. The Academy of Management Journal, 49, 85-101. doi:10.5465/AMJ.2006.20785503

Preacher, K. J., Curran, P. J., & Bauer, D. J. (2006). Computational tools for probing interaction effects in multiple linear regression, multilevel modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics, 31, 437-448.

Reis, H. T., Sheldon, K. M., Gable, S. L., Roscoe, J., & Ryan, R. M. (2000). Daily well-being: The role of autonomy, competence, and relatedness. Personality and Social Psychology Bulletin, 26, 419-435. doi:10.1177/0146167200266002

Russell, D. W. (1996). UCLA Loneliness Scale (Version 3): Reliability, validity, and factor structure. Journal of Personality Assessment, 66, 20-40. doi:10.1207/s15327752jpa6601_2

Ryff, C. D. (1989). Happiness is everything, or is it? Explorations on the meaning of psychological well-being. Journal of Personality and Social Psychology, 57, 1069-1081. doi:10.1037/0022-3514.57.6.1069

Sandstrom, G. M., & Dunn, E. W. (2013a). Is efficiency overrated? Minimal social interactions lead to belonging and positive affect. Social Psychological & Personality Science, 5, 436-441. doi:10.1177/1948550613502990

Sandstrom, G. M., & Dunn, E. W. (2013b). [Using cell phones to study social interactions]. Unpublished raw data.

Shacham, R. (2009). A utility for exploring HLM 2 and 3 way inter-actions [Computer software]. Retrieved from http://www.sha-cham.biz/HLM_int/index.htm

Vittengl, J. R., & Holt, C. S. (1998a). Positive and negative affect in social interactions as a function of partner familiar-ity, quality of communication, and social anxiety. Journal of Social & Clinical Psychology, 17, 196-208. doi:10.1521/jscp.1998.17.2.196

Vittengl, J. R., & Holt, C. S. (1998b). A time-series diary study of mood and social interaction. Motivation and Emotion, 22, 255-275. doi:10.1023/A:1022388123550

Watson, D., Clark, L. A., McIntyre, C. W., & Hamaker, S. (1992). Affect, personality, and social activity. Journal of Personality and Social Psychology, 63, 1011-1025. doi:10.1037/0022-3514.63.6.1011

Weiss, A., Bates, T. C., & Luciano, M. (2008). Happiness is a personal(ity) thing: The genetics of personality and well-being in a representative sample. Psychological Science, 19, 205-210. doi:10.1111/j.1467-9280.2008.02068.x

Wesselmann, E. D., Cardoso, F. D., Slater, S., & Williams, K. D. (2012). To be looked at as though air: Civil atten-tion matters. Psychological Science, 23, 166-168. doi:10.1177/0956797611427921

Wheeler, L., Reis, H., & Nezlek, J. B. (1983). Loneliness, social interaction, and sex roles. Journal of Personality and Social Psychology, 45, 943-953. doi:10.1037/0022-3514.45.4.943

Wrzus, C., Hänel, M., Wagner, J., & Neyer, F. J. (2013). Social net-work changes and life events across the life span: A meta-analysis. Psychological Bulletin, 139, 53-80. doi:10.1037/a0028601

by guest on April 30, 2014psp.sagepub.comDownloaded from