Abstract
Background and Objectives
Research suggests that friendships are associated with better emotional outcomes. Still, little is known about the implications of daily friend encounters on emotional well-being in the context of race and age. Guided by the integrative conceptual framework for friendship research, this study considers racism and cultural beliefs associated with racial groups and different social and emotional goals associated with age and investigates whether the frequency of friend encounters and the link between friend encounters and emotional well-being in everyday life would vary by race and age.
Research Design and Methods
Black (n = 80; Mage = 53.62) and White American adults (n = 89, Mage = 52.01) from the Stress and Well-being in Everyday Life study provided background and social network information, followed by ecological momentary assessment surveys in which they reported their social encounters and mood every 3 hours for 4 consecutive days.
Results
Multilevel linear models revealed no significant differences by race or age in the frequency of friend encounters. At times when individuals were with friends, their positive mood was elevated compared to when they were not (within-person association). Yet, this association was observed only among White adults and among Black individuals who were aged 41 or younger.
Discussion and Implications
This work contributes to the conceptual framework for friendship research by considering how individuals’ race and age are linked to friendship patterns. Findings highlight the importance of everyday contact with friends for enhancing momentary emotional well-being, particularly among White individuals and younger Black adults.
Keywords: Age differences, Ecological momentary assessment, Friends, Mood, Racial differences
Friendship is one of the most important interpersonal relationships people have across adulthood. Prior studies have found that frequent contact with friends, support received from friends, or simply prioritizing friendships are associated with better emotional outcomes across adulthood (Huxhold et al., 2014; Pinquart & Sörensen, 2000). Yet, the emotional benefits of friends may vary among Black and White adults due to racism and stronger familism beliefs among Black Americans. Moreover, socioemotional selectivity theory (SST) suggests that older adults tend to focus on socially and emotionally meaningful goals, and close relationships may have a greater impact on well-being later in adulthood than earlier in adulthood (Charles & Carstensen, 2010). Social contact with friends may play a key role in fulfilling this desire for social connection, and thus may have differential associations with emotional well-being depending on individuals’ age. Given the importance of friendships for well-being, understanding the impact of friends on different populations defined by race and age may be important to understand racial health disparities.
Specifically, this study examines whether the frequency of friend encounters in everyday life and links between friend encounters and emotional well-being vary by race and age. Notably, encounters represent the structural aspect of social networks (Blieszner et al., 2019). Most studies of racial differences in social networks or social experiences rely on global reports or retrospective measures rather than actual encounters (Holt-Lunstad et al., 2010; Pinquart & Sörensen, 2000). Social network size, in general, does not always map onto contact frequency. For example, prior retrospective studies showed that Black American adults might not have large social networks, but they were more likely to activate existing ties, resulting in reports of more frequent contact than White American adults (Ajrouch et al., 2001; Alwin et al., 2018). This current study contributes to the literature by using ecological momentary assessment (EMA) to investigate the everyday lives of adults of different ages by capturing their social and emotional processes closer to the time they occurred, minimizing retrospective biases and enhancing ecological validity (Charles et al., 2016). More importantly, EMA assesses the same person repeatedly, allowing us to address the within-person processes (e.g., whether individuals’ momentary mood improves when they encounter friends compared to when they do not).
Individual Differences in Friend Encounters in Everyday Life
This study is guided by Blieszner and colleagues’ (2019) integrative conceptual framework for friendship research. This framework identifies how social–structural position (e.g., individuals’ age, racial status) and psychological dispositions (e.g., individuals’ motivation, social expectation) shape friendship-related affect, cognitions, behaviors, and ongoing interaction within broader contexts of time, space, and culture. Fiori and colleagues’ (2020) conceptual model of sociohistorical context provides further insights into how sociohistorical context (e.g., racism, culture) influences social opportunity structure (e.g., diversity of social ties) directly or indirectly (via individual’s motivations and capacities). Guided by these conceptual framings, this study investigated how individuals’ social–structural position such as race and age may play a role in ongoing friendship patterns. Although we did not measure or capture the sociohistorical context (e.g., racism, culture) in the current study, it is important to recognize that individuals are embedded in these broader contexts.
Racial Differences
In general, Americans value friendships (Lu et al., 2021) and wish to spend time with their friends; however, there may be racial subgroup differences in these values within the United States. Racial minorized groups (e.g., Black Americans) are often subjected to different forms of racism (e.g., structural, cultural, and individual-level racism), which may influence the development and maintenance of friendships among Black individuals. For instance, compared to White American adults, Black American adults may have fewer friends and less frequent contact with friends because they have lower levels of social trust (PEW, 2007). Maintaining friendships requires time, effort, and resources (Fiori et al., 2020; Huxhold et al., 2022); it may be more challenging for Black adults to get together with their friends or sustain friendships due to the greater social and economic barriers they have faced compared to White adults (e.g., lower socioeconomic status, inconsistent work schedules, lack of transportation; Galupo & Gonzalez, 2013).
Additionally, compared to White Americans, Black Americans emphasize familism and are more involved with kin ties (Falzarano et al., 2022; Taylor et al., 2013), which further dilute the time and resources for friends. Although some retrospective studies have found no Black–White differences in the frequency of contact with friends or in the likelihood of having a friend-centered social network (e.g., Alwin et al., 2018; Litwin & Shiovitz-Ezra, 2010), a majority of national survey studies have found that Black adults on average had fewer friends and less frequent contact with friends than White adults (Mouzon, 2014; Talyor et al., 2013).
Age Differences
SST suggests that as individuals age, they reduce their social networks into smaller but more intimate forms by excluding peripheral social members (e.g., casual friends; Charles & Carstensen, 2010). In line with SST, a meta-analysis of social network size found that the number of friends declines as people age across adulthood and across different cultures (Wrzus et al., 2012). Likewise, through tracking phone records, a study found that contact with friends reduced rapidly after the age of 25 and declined continuously across adulthood (Bhattacharya et al., 2016).
The current study enriches the literature by examining American adults’ everyday involvement in friendship networks. Given barriers in the social world due to racism as well as a more exclusive focus on kinship, we expected that Black Americans would report fewer friend encounters than White Americans in everyday life. We also expected older adults would report fewer friend encounters than younger adults, regardless of race.
Friendships and Mood Among Black and White American Adults
Retrospective studies showed that individuals with more friends reported better subjective well-being (e.g., happiness and life satisfaction; Pinquart & Sörensen, 2000; Wrzus et al., 2012). Furthermore, a recent EMA study conducted among adults aged 65 and older (70% were non-Hispanic White) has shown that at times when older adults encountered friends, they reported better mood compared to the times when they did not (Ng et al., 2021). The association between friendship and well-being is likely different between Black and White American adults due to group differences in life experiences historically. For instance, history of slavery and racism may lead to reduced trust or enjoyment from interacting with non-kin among Black adults. Social contact with friends may have differential associations with emotional well-being depending on individuals’ race.
Cross-cultural friendship research provides some insights into group differences in experiences with friends. For example, Lykes and Kemmelmeier (2014) found that lack of contact with friends was more closely linked to loneliness in individualistic societies than in collectivistic societies. Conversely, the lack of contact with family was more closely linked to loneliness in collectivistic societies than in individualistic societies. A study using data from 99 countries found that the associations between friendship and well-being (i.e., happiness and subjective well-being) were stronger in individualistic than collectivistic cultures (Lu et al., 2021). These findings suggest that White adults (who are more likely to endorse individualism) may be more reactive to their encounters with friends than Black adults (who are more likely to endorse familism which is a subtype of collectivism; Oyserman et al., 2002). Therefore, we hypothesized that friend encounters would substantially impact mood throughout the day, particularly among White American adults.
Friendships and Mood Among Adults of Different Ages
The emotional health benefits of friends not only vary between Black and White adults, but may also vary depending on individual’s age. In line with SST, studies have shown that friends (e.g., activities with friends, perceiving friendship as important, support from friends) are more strongly associated with better subjective well-being among older adults than younger adults (e.g., Huxhold et al., 2014; Lu et al., 2021), indicating friends may fulfill older adults’ need to connect with others, which is a crucial factor in achieving better emotional well-being in older adulthood. On the contrary, given that younger adults focus on achieving information- and status-related goals (Charles & Carstensen, 2010), contacts with friends (which may involve more conflicts than older adult’s friendships) may generally be less closely associated with their emotional well-being.
Older adults often report better emotional well-being owing to adopting attention, coping, and emotion regulation skills to avoid negative events (Charles & Carstensen, 2010); however, the strength and vulnerability integration model suggests that age-related improvements in emotional well-being may not occur among individuals who are chronically stressed (e.g., racial minorized groups; Charles, 2010). Specifically, when individuals are experiencing chronic stress, they are not able to use those skills to avoid stress, and thus, may not experience age-related improvements in emotional well-being. Putting this model into context, Black individuals may show fewer age-related benefits from contacting friends compared to White adults given their disadvantaged status.
As such, this study examined whether age moderated the associations between friend encounters and mood among White and Black individuals. We predicted that the link between friend encounters and better mood would be stronger among older adults compared to young adults. However, such age-related effects would be observed among White adults, but not Black adults.
The Current Study
The current study explored racial and age differences in daily experiences with friends among a sample of Black and White American adults aged 33–91 years old. In this study, we were interested in the within-person associations (e.g., whether individuals’ momentary mood improves when they encounter friends compared to when they do not). We tested the following hypotheses:
H1: Black Americans would report fewer friend encounters than White Americans, and older adults would report fewer friend encounters than younger adults.
H2: From the within-person analysis, friend encounters would be associated with better mood, particularly among White adults.
H3: Within-person associations between friend encounters and mood would be stronger for those who were older, particularly among White adults.
Method
Sample and Procedures
This study drew on data from the Stress and Well-being in Everyday Life (SWEL) study collected from March 2018 through March 2020. These participants were from the Detroit area who participated in at least one wave of the longitudinal Social Relations and Health Study (SRS; three waves of data in 1992 [aged from 8 to 93], 2005, 2015; original sample description in Ajrouch et al., 2001; Birditt et al., 2020). The Black sample was matched with a subset of the larger sample of White participants on key characteristics including age, gender, education, social network size, and hypertension status using propensity scores (Guo & Fraser, 2015). Notably, the data collection was halted due to COVID-19 resulting in slightly incomplete matching in education between Black and White participants. A total of 238 adults (109 Black and 129 White adults) participated in the baseline interviews in their homes or over the phone. Among these 238 participants, 169 (80 Black and 89 White adults; Mage = 52.86, SD = 12.79) participated in a 4-day experience sampling. This study used this subsample (n = 169) for data analyses. Compared to those who did not participate in EMAs (n = 69), participants who completed EMAs were younger; t = 6.14, p < .001, more educated; t = −3.04, p = .003, and healthier; t = −2.06, p = .04.
Interviewers showed the participant how to complete the experience-sampling surveys on a smartphone provided to participants. The smartphone administrated six surveys each day across 4 days at set times (at waking, 9 a.m., 12 p.m., 3 p.m., 6 p.m., and 9 p.m.). Participants showed high response rates—they completed an average of 17.19 EMA reports during the 4-day study period (SD = 6.52, ranged from 1 to 26; more information about EMA in Supplementary Table 1). Fifteen participants completed fewer than five EMAs. We considered all participants in analyses because multilevel models are well suited to unbalanced data. Since the survey at waking did not ask for participants’ social encounters information, we excluded these surveys (n = 414) in the analyses.
Participants were paid $30 for the baseline assessment and $50 for completing the 4-day EMAs. In addition, some participants received add extra $10 incentive per day toward the end of the study to encourage participation. Prior to the study, ethical approval was obtained from the local ethics review board, and informed consents were signed by participants.
Baseline Interview Measures
Racial status
Adult participants from Wave 1 of the Social Relations Study (SRS) were asked: “Are you White, Black, Native American, Asian, Hispanic, or another race?”. If participants indicated multiple race categories, they were asked to indicate the one category that best described their race. In the same wave, child participants’ race was reported by interviewers through observation. In Waves 2 and 3 of the SRS, participants were asked the open-ended question, “What is your ancestry or ethnic origin?”. Responses were coded by the study team (see original sample racial measures in Ajrouch et al., 2001; Birditt et al., 2020). The current sample was drawn from individuals who endorsed the “Black” or “White” categories on these measures from prior waves. For purposes of the analyses, racial status was coded as 1 (Black) and 0 (White) in this current study.
Age
Participants provided their date of birth in month, day, and year. Then, the interviewer checked whether the date of birth provided by the participant matched with the date of birth provided in the previous wave. Based on this information, we calculated the participant’s age by subtracting the date of birth from the baseline interview date, and rounding down to the nearest whole number. Age was a continuous variable measured in number of years.
Social convoy
In the baseline interview, participants indicated their social ties by closeness using three concentric circles (i.e., innermost, middle, and outermost circle; Antonucci, 1986). In total, participants listed a total of 1,819 social convoy partners (mean = 10.76, SD = 6.40, range 0–39; this was referred to as overall social convoy size). Only a few participants (n = 16) reported having more than 20 convoy partners. To avoid study fatigue, participants only specified details of up to 20 convoy partners including information about the relationship type. This subsample elaborated on details of a total of 1,704 social convoy partners (M = 10.14, SD = 4.78, ranged from 0 to 20). We generated a categorical variable to indicate these social partners’ relationship types: 1 (friends), 2 (romantic partners; e.g., spouses, romantic partners), 3 (children; e.g., child, step-child, children-in-law), 4 (other family members/relatives; e.g., siblings, parents, grandchild, cousin, uncle), and 5 (other social partners; e.g., godparent, coworkers, neighbors). On average, each participant reported 2.10 friends (SD = 2.51, range 0–10; more information about social convoy is in Supplementary Table 2).
Participant-level covariates
Prior studies have shown that women (Kalmijn, 2003) and those who have higher SES (Bianchi & Vohs, 2016) tend to have more frequent contact with friends than men and those who have lower SES. Unmarried individuals have more frequent contact and may benefit more emotionally from contacting friends (Gillespie et al., 2015). Individuals who have poorer health typically have less frequent contact with friends (Ha et al., 2017). Therefore, we adjusted for these participant-level factors: gender 1 (male) and 0 (female), year of education (0 to 17+), marital status recoded as 1 (married/living with a partner) and 0 (not married), self-rated health recorded as 1 (poor) to 5 (excellent), and overall social convoy size.
EMA Measures
Social encounters
Every 3 hours, participants indicated the number of people they encountered in the past 3 hours. They then reported whether those encounters included convoy members. Around 12% social convoy encounters involved friends. Furthermore, participants could indicate up to five individuals they had encountered but were not listed in the social convoy (i.e., non-convoy social partners) during the prior 3 hours (more information about EMA social encounters in Supplementary Table 3). Participants answered additional questions on these non-convoy social encounters, including their relationship type (i.e., romantic partner, child, sibling, another family member, friend, acquaintance, service provider, stranger, and other). Participants indicated the mode of contact (i.e., text/email/social media, phone, or in-person) for each social encounter (both convoy and non-convoy encounters). In this study, we considered all the social encounters regardless of mode of contact, but we focused on in-person encounters in sensitivity analyses. We categorized these non-convoy types of social relationships using the same categories as convoy members.
We generated sum scores to represent the number of encounters for each relationship type for each 3-hour assessment. Similar to previous studies, we then considered a series of dichotomized variables of encounters representing whether the participant encountered a particular type of social partner or not in a given assessment (e.g., any encounters with friends; 1 [yes] and 0 [no]; Alwin et al., 2018; Ng et al., 2021, 2022; Taylor et al., 2013). Additionally, we estimated sensitivity analyses with the continuous variables of encounters.
Mood
Participants rated their positive and negative moods during the past 3 hours. They rated the extent to which eight positive mood items (i.e., energetic, loved, happy, calm, content, excited, proud, and optimistic) and eight negative mood items (i.e., tired, worried/anxious, tense/stress, irritated, lonely, angry, bored, and sad) described them on a scale from 1 (not at all) to 5 (a great deal). We calculated averages to generate a positive mood score (between-person reliability: α = 0.90) and a negative mood score (between-person reliability α = 0.87) for each 3-hour assessment.
Assessment-level covariates
This study adjusted for assessment-level variables measured in the EMAs because this was critical for uncovering within-person associations between friend encounters and well-being. These variables included encounters with other types of social partners (including romantic partners, children, other family members, and other social partners; 1 [yes] and 0 [no] for each relationship type), whether participants engaged in working/school tasks in the past 3 hours, 1 (yes) and 0 (no), and time of the day (i.e., morning [9 a.m. and 12 noon assessments], midday [3 p.m. and 6 p.m. assessments], and evening to bedtime [end-of-day assessments]).
Analytic strategy
First, we calculated descriptive statistics for background characteristics and other participant-level variables to detect differences between Black and White participants using t tests (for continuous variables) or chi-squared tests (for categorical variables).
To test our hypotheses, we estimated multilevel models. Given the nested data structure, we first considered three-level models (assessments nested within days and nested within participants). We present more parsimonious two-level models (assessments [Level 1] nested within participants [Level 2]) because patterns of findings were the same using two-level or three-level models. To examine H1 regarding the racial and age differences in the frequency of encounters with friends throughout the day, we treated race 1 (Black) and 0 (White) and age as separate predictor variables, any encounters with friends within every 3-hour (Level 1; dichotomized variable) as the outcome, and controlled for covariates (continuous covariates were grand-mean centered) in two separate multilevel logistic models. We also considered both race and age simultaneously in the same model. In a supplementary analysis, we tested the sum score of friend encounters every 3 hours (Level 1; continuous variable) using a multilevel linear model.
With respect to the second hypothesis regarding the within-person associations between encounters with friends and mood, we first estimated a multilevel linear model, treating any friend encounters (Level 1; dichotomized variable) as a predictor and positive and negative moods (Level 1) as two separate outcomes. To distinguish within-person changes from between-person differences, we utilized a series of within-between random effects models (Hamaker & Muthén, 2020). Specifically, the time-varying predictor (i.e., dichotomized variable of friend encounters) was recoded to its person-mean across the study days (Level 2) and the deviation from its person-mean at each assessment (Level 1). Given that participants could have encountered multiple types of social partners in the same 3-hour period and these encounters also contributed to mood, we adjusted for the time-varying variables of encounters with nonfriend social partners including romantic partners, children, other family members, and other social partners in the analyses. Again, the model adjusted for the same covariates. To test whether the within-person associations between friend encounters and mood differed among Black and White American adults, we considered the stratified models between Black and White adults. To test the slope differences (i.e., whether the size of the effect was statistically different between Black and White adults), we created interaction terms by multiplying race with every single variable (variable of interests, covariates) in the full sample model.
Lastly, to examine whether age (Level 2) may moderate the association between friend encounters and mood, we treated age as the moderator by including the interaction term of friend encounters × age (grand-mean centered) in the models. The moderating effect of age was tested in the full sample as well as race-stratified samples. For significant interaction terms involving a continuous moderator of age, we used the Johnson–Neyman technique to probe the simple slope for different ages (Carden et al., 2017). All models were performed in STATA 16.
Results
On average, participants listed two friends (SD = 2.51; range 0–10) as social convoy members. Around two-thirds of participants reported at least one friend encounter over the 4-day observation period. Furthermore, participants reported having encounters with friends in 18% of the assessments (Black adults 19%, White adults 17%). Table 1 presents descriptives for variables of interest by racial status. t Tests and chi-square tests revealed that White adults were better educated than Black adults and had more encounters with romantic partners in everyday life but did not differ in other background characteristics, daily encounters, or mood. For the bivariate correlations, see Supplementary Materials.
Table 1.
Descriptive Statistics for Sample
| Variable | Full sample (N = 169) | White adults (n = 89) | Black adults (n = 80) | t or χ2 | |||||
|---|---|---|---|---|---|---|---|---|---|
| Mean/proportion | SD | Min. | Max. | Mean/proportion | SD | Mean/proportion | SD | ||
| Participant characteristics | |||||||||
| Age | 52.86 | 12.79 | 33.00 | 91.00 | 53.62 | 13.90 | 52.01 | 11.46 | 0.81 |
| Female | 0.66 | 0.64 | 0.69 | 0.42 | |||||
| Education in years | 14.40 | 1.89 | 10.00 | 17.00 | 14.74 | 1.91 | 14.01 | 1.80 | 2.55** |
| Married | 0.55 | 0.61 | 0.49 | 2.42 | |||||
| Self-rated health | 3.73 | 0.94 | 1.00 | 5.00 | 3.83 | 0.97 | 3.61 | 0.89 | 1.52 |
| Convoy size | 10.76 | 6.40 | 0.00 | 39.00 | 10.54 | 5.74 | 11.01 | 7.08 | −0.48 |
| Number of friends | 2.10 | 2.51 | 0.00 | 10.00 | 2.38 | 2.53 | 1.79 | 2.47 | 1.54 |
| Number of romantic partners | 0.65 | 0.48 | 0.00 | 1.00 | 0.72 | 0.45 | 0.58 | 0.50 | 1.97* |
| Number of children | 2.15 | 1.80 | 0.00 | 8.00 | 2.06 | 1.76 | 2.25 | 1.85 | −0.70 |
| Number of other family members | 5.01 | 3.65 | 0.00 | 17.00 | 4.73 | 3.42 | 5.31 | 3.88 | −1.04 |
| Number of other social partners | 0.18 | 0.55 | 0.00 | 4.00 | 0.10 | 0.30 | 0.26 | 0.72 | −1.92 |
| Daily experiencesa | |||||||||
| Friend encounters | 0.18 | 0.21 | 0.00 | 0.86 | 0.17 | 0.19 | 0.19 | 0.23 | −0.51 |
| Romantic partner encounters | 0.40 | 0.36 | 0.00 | 1.00 | 0.46 | 0.37 | 0.34 | 0.35 | 2.15** |
| Children encounters | 0.32 | 0.30 | 0.00 | 1.00 | 0.33 | 0.32 | 0.31 | 0.29 | 0.30 |
| Other family member encounters | 0.30 | 0.27 | 0.00 | 1.00 | 0.29 | 0.28 | 0.25 | 0.27 | −0.50 |
| Other social partner encounters | 0.29 | 0.26 | 0.00 | 1.00 | 0.30 | 0.27 | 0.28 | 0.24 | 0.47 |
| Positive mood | 3.10 | 0.66 | 1.50 | 4.70 | 3.07 | 0.63 | 3.14 | 0.70 | −0.68 |
| Negative mood | 1.47 | 0.36 | 1.01 | 3.33 | 1.49 | 0.38 | 1.44 | 0.35 | 1.03 |
aDaily experiences variables reported here were aggregated into participant level.
*p = .05. **p < .05.
Racial and Age Differences in Friend Encounters in Everyday Life
The model treating race and age simultaneously as predictors yielded the same pattern of findings as models treating race and age as separate predictors. For parsimony, we present results from the model considering race and age as predictors simultaneously. As shown in Table 2, two-level models revealed that race did not predict the likelihood of friend encounters (dichotomous variable; odd ratio [OR] = 0.88, p = .65) nor the log-transformed sum score of friend encounters (B = −0.05, p = .59; in Supplementary Table 5). In other words, Black adults had a similar frequency of friend encounters compared to White adults.
Table 2.
Multilevel Logistic Models Predicting Any Friend Encounters From Racial Status And Age
| Variable | Any friend encounters | |||
|---|---|---|---|---|
| B | SE | OR | ||
| Fixed effects | ||||
| Intercept | −1.85 | *** | 0.38 | |
| Black | −0.13 | 0.29 | 0.88 | |
| Age | 0.01 | 0.01 | 1.01 | |
| Covariates | ||||
| Female | −0.06 | 0.31 | 0.95 | |
| Education | 0.00 | 0.08 | 1.00 | |
| Married | −1.04 | *** | 0.30 | 0.36 |
| Health | 0.27 | 0.16 | 1.32 | |
| Network size | 0.05 | 0.02 | 1.05 | |
| Working | 0.25 | 0.27 | 1.29 | |
| Morning | ||||
| Afternoon | 0.41 | ** | 0.13 | 1.51 |
| Evening | 0.49 | * | 0.20 | 1.63 |
| Random effects | ||||
| Intercept variance (Level 2: Participant) | 2.25 | *** | 0.44 | |
Notes: Participant n = 169. Number of assessments n = 2,491. Continuous predictor and covariates were grand-mean centered.
*p < .05. **p < .01. ***p < .001.
Likewise, age did not predict the likelihood of friend encounters (dichotomous variable; OR = 1.01, p = .52) nor the number of friend encounters (B = 0.00, p = .50), indicating older adults had a similar frequency of encounters with friends compared to young adults in everyday settings. We also explored the potential race × age interaction effect on the frequency of friend encounters. Such an interaction effect was nonsignificant.
Friendships and Mood Among Black and White American Adults
As shown in Table 3, for the full sample, within-person associations of encounters with friends (B = 0.06, p = .04), romantic partners (B = 0.07, p = .02), other family members (B = 0.07, p = .01), and other social partners (B = 0.09, p < .001) were associated with elevated positive mood, except encounters with children (B = 0.02, p = .42). That was, at times when individuals were with friends, romantic partners, family members (other than children, e.g., sibling, extended family members), or others social partners (e.g., church members, neighbors), their mood was elevated compared to when they were not.
Table 3.
Multilevel Linear Models Predicting Positive Mood From Friend Encounters
| Variable | Full sample (N = 169) | White adults (n = 89) | Black adults (n = 80) | Race difference p values |
||||||
|---|---|---|---|---|---|---|---|---|---|---|
| B | SE | B | SE | B | SE | |||||
| Intercept | 2.93 | *** | 0.14 | 2.73 | *** | 0.17 | 3.20 | *** | 0.20 | <.001 |
| Within-person effects | ||||||||||
| Friend encounters | 0.06 | * | 0.03 | 0.08 | * | 0.04 | 0.03 | 0.05 | .40 | |
| Romantic partner encounters | 0.07 | * | 0.03 | 0.05 | 0.04 | 0.09 | * | 0.05 | .51 | |
| Children encounters | 0.02 | 0.03 | 0.00 | 0.04 | 0.04 | 0.04 | .53 | |||
| Other family member encounters | 0.07 | ** | 0.03 | 0.04 | 0.03 | 0.10 | * | 0.04 | .30 | |
| Other social partner encounters | 0.09 | *** | 0.03 | 0.05 | 0.03 | 0.13 | *** | 0.04 | .11 | |
| Between-person effects | ||||||||||
| Friend encounters | 0.58 | * | 0.23 | 0.52 | 0.35 | 0.70 | * | 0.31 | .70 | |
| Romantic partner encounters | −0.10 | 0.17 | 0.33 | 0.21 | −0.80 | ** | 0.27 | .001 | ||
| Children encounters | 0.03 | 0.17 | −0.05 | 0.21 | 0.32 | 0.27 | .28 | |||
| Other family member encounters | 0.15 | 0.18 | 0.39 | 0.23 | −0.20 | 0.28 | .10 | |||
| Other social partner encounters | −0.36 | 0.19 | −0.20 | 0.23 | −0.54 | 0.30 | .37 | |||
| Covariates | ||||||||||
| Black | 0.09 | 0.10 | — | — | — | — | — | |||
| Age | 0.01 | 0.00 | 0.00 | 0.00 | 0.01 | * | 0.01 | .17 | ||
| Female | 0.08 | 0.10 | 0.03 | 0.13 | 0.11 | 0.15 | .68 | |||
| Education | −0.01 | 0.03 | −0.01 | 0.03 | −0.01 | 0.04 | .95 | |||
| Married | 0.15 | 0.12 | 0.07 | 0.16 | 0.34 | 0.18 | .26 | |||
| Health | 0.15 | ** | 0.05 | 0.14 | * | 0.06 | 0.16 | 0.08 | .85 | |
| Network size | 0.02 | * | 0.01 | 0.02 | 0.01 | 0.02 | * | 0.01 | .62 | |
| Working | −0.11 | *** | 0.03 | −0.12 | ** | 0.04 | −0.11 | * | 0.04 | .85 |
| Morning | (Ref.) | (Ref.) | (Ref.) | |||||||
| Afternoon | 0.05 | * | 0.02 | 0.06 | * | 0.03 | 0.04 | 0.03 | .67 | |
| Evening | 0.01 | 0.03 | 0.03 | 0.04 | −0.02 | 0.04 | .31 | |||
| Random effects | ||||||||||
| Intercept VAR (Level 2: Participant) | 0.33 | *** | 0.04 | 0.29 | *** | 0.05 | 0.35 | *** | 0.06 | — |
| Residual VAR | 0.22 | *** | 0.01 | 0.21 | *** | 0.01 | 0.23 | *** | 0.01 | — |
Notes: For overall sample, total number of assessments = 2,491. Among these 2,491 assessments, n = 543 were without any social encounters. For White adults, number of assessments = 1,341. For Black adults, number of assessments = 1,150. Continuous covariates were grand-mean centered. Race differences indicate p value of interaction terms with race using a whole sample.
*p < .05. **p < .01. ***p < .001.
With regard to the race-stratified models, findings revealed a within-person association of friend encounters and elevated positive mood occurred among White adults (B = 0.08, p = .03) but not among Black adults (B = 0.03, p = .52). However, for Black adults, the within-person associations of encounters with romantic partners (B = 0.09, p = .046), other family members (e.g., extended relatives, niece, nephew, uncle; B = 0.10, p = .01), and other social partners (e.g., church members, neighbor; B = 0.13, p = .001) were associated with elevated positive mood; these associations were not significant among White adults.
Additionally, we tested whether the slopes were statistically different between Black and White adults. Interaction terms involving within-person effects were nonsignificant (race difference column in Table 3). In other words, the within-person association of encounters with friends on positive mood was not statistically different between Black and White adults, albeit the effect was significantly different from zero in White adult sample. This could be due to the small sample sizes in detecting significant interaction effects.
Although not our focus, we observed a between-person association of friend encounters on increased positive mood (B = 0.70, p = .03) and a between-person association of romantic partner encounters on lower positive mood (B = −0.80, p = .003) for Black adults.
With regard to negative mood, the within-person associations of friend encounters as well as other social encounters were not associated with negative mood in the full sample or the race-stratified models (Supplementary Table 6).
Friendships and Mood Among Adults of Different Ages
We tested whether age conditioned the association between friend encounters and mood by including the interaction term of friend encounters × age in the models among full and race-stratified models. Interaction term of friend encounters × age on positive mood was nonsignificant in the full sample (B = −0.00, p = .28; finding not shown). The race-stratified model revealed the interaction effect was nonsignificant among White adults (B = −0.00, p = .95; finding not shown) but it was significant among Black adults. (B = −0.01, p = .05; Table 4). Using the Johnson–Neyman technique (Carden et al., 2017), we found that the within-person association of friend encounters on positive mood was significantly different from zero for Black adults who were aged 41 or younger (region of significance on Figure 1). Seemingly, the within-person association of friend encounters and elevated positive mood occurred among younger Black adults but not among older Black adults.
Table 4.
Multilevel Linear Models Predicting Positive Mood From Contacts with Friends and Age
| Variable | Full sample (N = 169) | White adults (n = 89) | Black adults (n = 80) | ||||||
|---|---|---|---|---|---|---|---|---|---|
| B | SE | B | SE | B | SE | ||||
| Intercept | 2.91 | *** | 0.15 | 2.60 | *** | 0.18 | 3.18 | *** | 0.20 |
| Age | 0.01 | 0.01 | 0.01 | 0.01 | 0.02 | 0.01 | |||
| Within-person effects | |||||||||
| Friend encounters | 0.07 | * | 0.03 | 0.08 | * | 0.04 | 0.04 | 0.05 | |
| Friend encounters × age | −0.00 | 0.00 | −0.00 | 0.00 | −0.01 | * | 0.00 | ||
| Romantic partner encounters | 0.07 | * | 0.03 | 0.06 | 0.04 | 0.07 | 0.05 | ||
| Romantic partner encounters × age | −0.00 | 0.00 | −0.00 | 0.00 | −0.01 | 0.00 | |||
| Children encounters | 0.02 | 0.03 | −0.00 | 0.04 | 0.03 | 0.04 | |||
| Children encounters × age | 0.00 | 0.00 | 0.00 | 0.00 | 0.01 | 0.00 | |||
| Other family | 0.07 | * | 0.03 | 0.04 | 0.03 | 0.10 | ** | 0.04 | |
| Other family encounters × age | −0.00 | 0.00 | −0.00 | 0.00 | −0.00 | 0.00 | |||
| Other encounters | 0.09 | *** | 0.03 | 0.05 | 0.03 | 0.14 | *** | 0.04 | |
| Other encounters × age | −0.00 | 0.00 | −0.00 | 0.00 | −0.00 | 0.00 | |||
| Between-person effects | |||||||||
| Friend encounters | 0.54 | * | 0.24 | 0.55 | 0.36 | 0.71 | * | 0.32 | |
| Friend encounters × age | −0.01 | 0.02 | 0.00 | 0.03 | −0.01 | 0.03 | |||
| Romantic partner encounters | −0.14 | 0.18 | 0.32 | 0.21 | −0.84 | ** | 0.28 | ||
| Romantic partner encounters × age | −0.02 | 0.01 | −0.02 | 0.01 | −0.01 | 0.02 | |||
| Children encounters | −0.01 | 0.18 | 0.01 | 0.23 | 0.22 | 0.28 | |||
| Children encounters × age | 0.01 | 0.01 | 0.03 | 0.02 | −0.03 | 0.03 | |||
| Other family | 0.15 | 0.18 | 0.39 | 0.23 | −0.13 | 0.28 | |||
| Other family encounters × age | −0.00 | 0.01 | −0.02 | 0.02 | 0.04 | 0.02 | |||
| Other encounters | −0.35 | 0.19 | −0.19 | 0.23 | −0.53 | 0.30 | |||
| Other encounters × age | 0.00 | 0.01 | 0.00 | 0.02 | 0.00 | 0.03 | |||
| Black | 0.09 | 0.10 | — | — | — | — | |||
| Covariates (not shown in the table) | — | — | — | ||||||
| Random effects | |||||||||
| Intercept variance (Level 2: Participant) | 0.32 | *** | 0.04 | 0.26 | *** | 0.04 | 0.29 | *** | 0.05 |
| Residual variance | 0.22 | *** | 0.01 | 0.21 | *** | 0.01 | 0.22 | *** | 0.01 |
Notes: Continuous covariates were grand-mean centered. Covariates including gender, education, marital status, self-rated health, social convoy size, work status, and time of the day were not shown in the table.*p <.05. **p <.01. ***p <.001.
Figure 1.
Johnson–Neyman plot of the simple slope coefficients of the within-person association of friend encounters on positive mood across the range of ages. At age 41 or younger, the confidence bands no longer contain zero (i.e., region of significance), indicating the simple slope coefficients of the association of friend encounters on positive wood were significant. CI = confidence interval.
The interaction term of friend encounters × age predicting negative mood was nonsignificant in the overall and race-stratified models (not shown). Lastly, we tested the three-way interaction by including the interaction term friend encounters × age × race on mood in the full sample; however, it was nonsignificant (not shown).
Sensitivity Tests
A series of sensitivity tests were conducted to ascertain the findings, including re-estimating models by (a) using continuous versions of the social predictor variables, (b) excluding covariates that were matched in recruitment, (c) adjusting for the total number of social encounters, or just a single category of “nonfriends” encounters instead of four dichotomized variables of different encounters, (d) differentiating close versus less close friends, (e) selecting assessments that involved at least one in-person encounter, and (f) excluding participants with low response rates. Patterns of results from these re-estimations remained unchanged. Overall, patterns of results from these re-estimations remained unchanged (Supplementary Materials for more details).
Discussion
This study contributes to the integrative conceptual framework for friendship research proposed by Blieszner et al. (2019) by considering how individuals’ race and age may play a role in ongoing friendship patterns. Moving beyond the global or retrospective report of friendships, this study contributes to the literature by investigating how often Black and White American adults encounter friends throughout the day using EMA. Furthermore, we examined whether friend encounters were associated with emotional well-being differently between Black and White adults and among individuals of different ages. Findings suggest that White adults and a selective age group of Black individuals (younger Black adults) may benefit more emotionally from encounters with friends.
Individual Differences in Friend Encounters in Everyday Life
Racial differences
Contrary to what we hypothesized, there were no differences in the likelihood or number of friend encounters throughout the day between Black and White adults. Notably, Black and White samples in this study were matched on key characteristics (i.e., age, gender, health, network size) during the recruitment. Therefore, it is possible that the Black–White differences that may exist in the population for any number of characteristics have been substantially eliminated by the study design.
Alternatively, according to the hierarchical compensatory model, individuals first reach out to core family members (e.g., spouse and children) and then turn to friends and other social relationships for contact and support if core family is unavailable (Cantor, 1979). Despite the emphasis on familism, Black Americans are less likely to endorse their spouses as important ties (Alwin et al., 2018) and have lower marital quality (Raley et al., 2015). As a result, it is possible that Black adults reach out to friends for contact and support just as much as their White counterparts, leading to the observed finding in the current study. Notably, other retrospective studies have found that Black adults reached out more often to their friends for contact and support than White adults in daily life (e.g., Mouzon, 2014). In line with the findings in this study, a network typology study has found that Black and White adults are equally likely to have friend-centered network and some studies have shown no Black–White differences in frequency of contact with friends (e.g., Alwin et al., 2018; Litwin & Shiovitz-Ezra, 2010).
Age differences
In this study, age was not associated with friend encounters when other covariates (e.g., social network size, health condition, work status) were taken into consideration. It is possible that older individuals contact friends just as much as younger individuals to reduce loneliness and to fulfill their need to belong when their social networks begin to shrink (e.g., loss of a spouse in older age; Chen & Feeley, 2014). Moreover, compared to younger adults, older adults may have more time and fewer distractions (e.g., from work, child-rearing, and caregiving to aging parents) to connect with close friends during retirement years (Kail & Carr, 2020).
Although not our focus, we observed that marital status was a consistent and more predictive factor (compared to age and racial status) for friendly encounters, such that unmarried individuals were more likely to encounter friends than married individuals. This finding was consistent with the literature that showed unmarried adults (e.g., widowed, divorced, never married) are more engaged with friends (Sarkisian & Gerstel, 2016).
Friendships and Mood Among Black and White American Adults
Consistent with prior EMA studies with older adult samples (Ng et al., 2021), this study showed that adult individuals reported better mood when they encountered friends compared to when they did not. When subgroups were examined, this within-person association was present for White adults only. Additional research with a larger sample size is needed to explore this possibility as the racial differences did not reach statistical significance in direct group comparisons.
Although White adults reported a greater positive mood when they encountered friends, Black adults reported a greater positive mood when they encountered multiple kin relationships (i.e., romantic partners, other family members) compared to when they did not. These findings reflect the important role friends served for White adults and further illustrate the familism value among Black Americans (Falzarano et al., 2022; Taylor et al., 2013). Present findings also suggest potential impacts of racism which may prevent Black individuals from benefiting emotionally from contacting with friends. Despite having a similar likelihood or frequency of friend encounters, it is possible that maintaining friendships or developing trust with friends demands time and effort that could be more stressful for Black adults than White adults. Interestingly, encounters with other social partners boosted momentary mood for Black adults, perhaps because contact with church members was included in this category. Literature has shown that contacts with church-related ties have a positive impact on Black adults’ sense of well-being (Nguyen, 2020). However, additional research is needed to explore this possibility as this “other social partners” category also included coworkers, acquaintances, among others.
Similar to another EMA study (Ng et al., 2021), this study also did not reveal within-person associations of friend encounters with negative mood. After all, friends are rarely negative because people choose to be around them. It could also be due to a generally low rating of negative mood among the sample leading to a lack of variability in negative mood for analyses.
The current study also provided insights into encounters with romantic partners, outside the scope of our examination of friendships. Black adults who had more encounters with romantic partners had lower positive mood than Black adults who had fewer encounters with romantic partners during the study period. Despite greater familism values, literature has consistently shown that Black couples report poor marital quality and have greater marital instability than White couples due to socioeconomic disadvantages (Raley et al., 2015). It is possible that, during the study period, when Black adults encountered romantic partners, those encounters involved conflicts, stress, or tension that may link to reduced positive mood.
Friendships and Mood Among Adults of Different Ages
Although SST suggests contacts with close partners are salient to older adults’ emotional well-being due to their prioritization of social and emotional goals (Charles & Carstensen, 2010), our study did not reveal age differences in the within-person association between friends and positive mood among the full sample or White stratified sample. Indeed, friends could serve similar emotional functions (e.g., fulfilling individuals’ general needs to belong) across different stages of adulthood (Blieszner et al., 2019).
Interestingly, however, younger Black adults (aged 41 and younger) reaped more immediate emotional benefits when they were in contact with friends than older Black adults, suggesting the emotional importance of friends among younger Black adults. Again, tying back to the racism literature, as evidenced in the formation of peer groups in younger adulthood (e.g., Goins, 2011), younger Black adults appear to rely more on friends for contact and support to combat racism and empower each other. Conversely, older Black adults may have more substantial support from family members or churches in later adulthood.
Limitations and Future Directions
There are several limitations in the current study that should be acknowledged. First, similar to other friendship studies, there are methodological challenges in defining friendship. The term “friend” may have different meanings for different racial groups (Adams et al., 2000). Future studies could provide a clear definition of friends or adopt more subtle friend categories (e.g., close friends, casual friends, church friends). Second, this study has not considered the quality of encounters or capture what was going on during the encounters. Prior research has shown that satisfaction with friends was a better predictor of subjective well-being than the quantity of friends (Gillespie et al., 2015). Future research could examine specific contexts involving friends (e.g., conflicts, activities, or support exchanges with friends). As more social interactions occur online, especially during the COVID-19 pandemic, future studies can utilize social media data and computational social science approaches to explore friendship patterns across different countries (Ng, 2020, 2022, 2023). Moreover, a 4-day study may not accurately capture individuals’ typical life, a longer or follow-up study could be conducted to ascertain these findings. Finally, the sample was based on matching techniques and focused on the U.S. adults who resided in the greater Detroit area; thus, the findings may have limited generalizability toward the population level. Future research that uses larger national probability samples may be a better representative of the population and could provide more comprehensive results on Black–White differences.
Instead of relying on retrospective reports by asking participants to recall social and emotional experiences over long periods spanning weeks, months, or a year, this study studied adults’ social encounters and mood closer to the time these events occurred by using ecological reports. This technique reduced retrospective bias and enhanced the ecological validity of the findings. Rather than recruiting participants from a single stage of the life course, we recruited a spectrum of participants of different ages which offers a more comprehensive view of adult friendships. Although findings have to be replicated in future studies using a larger national sample, our study highlights the importance of everyday contact with friends for enhancing momentary well-being, particularly among White individuals and younger Black adults.
Supplementary Material
Acknowledgments
We would like to thank Sae Hwang Han for his feedback on this manuscript and Angela Turkelson for her efforts in managing the SWEL data sets. The content is solely the responsibility of the authors and does not necessarily represent the views of the NIDILRR.
Contributor Information
Yee To Ng, Department of Physical Medicine and Rehabilitation, University of Michigan, Ann Arbor, Michigan, USA; The Institute for Social Research, University of Michigan, Ann Arbor, Michigan, USA.
Karen L Fingerman, Department of Human Development and Family Sciences, The University of Texas at Austin, Austin, Texas, USA.
Kira S Birditt, The Institute for Social Research, University of Michigan, Ann Arbor, Michigan, USA.
Funding
This work was supported by grants from the National Institute on Aging: Racial Disparities in Health: The Roles of Stress, Social Relations and the Cardiovascular System (multiple PIs: K. S. Birditt; K. Najarian; R01AG054371) and Social Relations, Aging and Health: Competing Theories and Emerging Complexities (PI: T. C. Antonucci; R01AG045423). Y. T. Ng was supported by a postdoctoral fellowship award funded by the University of Michigan’s Advanced Rehabilitation Research Training Program in Community Living and Participation from the National Institute of Disability, Independent Living, and Rehabilitation Research, Administration for Community Living (NIDILRR; grant no. 90ARCP0003).
Conflict of Interest
None declared.
Data Availability
This study was preregistered on the Open Science Framework (https://osf.io/d4jgy/). Data and additional documentation of sample and measures are available upon request to the corresponding author.
References
- Adams, R. G., Blieszner, R., & De Vries, B. (2000). Definitions of friendship in the third age: Age, gender, and study location effects. Journal of Aging Studies, 14(1), 117–133. doi: 10.1016/s0890-4065(00)80019-5 [DOI] [Google Scholar]
- Ajrouch, K. J., Antonucci, T. C., & Janevic, M. R. (2001). Social networks among blacks and whites: The interaction between race and age. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 56(2), S112–S118. doi: 10.1093/geronb/56.2.s112 [DOI] [PubMed] [Google Scholar]
- Alwin, D. F., Thomas, J. R., & Sherman-Wilkins, K. J. (2018). Race, social relations and the life course. In Alwin D. F., Felmlee D. H., & Kreager D. A. (Eds.), Social networks and the life course (pp. 285–313). Springer. doi: 10.1007/978-3-319-71544-5_14 [DOI] [Google Scholar]
- Antonucci, T. C. (1986). Measuring social support networks: Hierarchical mapping technique. Generations: Journal of the American Society on Aging, 10(4), 10–12. [Google Scholar]
- Bhattacharya, K., Ghosh, A., Monsivais, D., Dunbar, R. I., & Kaski, K. (2016). Sex differences in social focus across the life cycle in humans. Royal Society Open Science, 3(4), 160097. doi: 10.1098/rsos.160097 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bianchi, E. C., & Vohs, K. D. (2016). Social class and social worlds: Income predicts the frequency and nature of social contact. Social Psychological and Personality Science, 7(5), 479–486. doi: 10.1177/1948550616641472 [DOI] [Google Scholar]
- Birditt, K. S., Sherman, C. W., Polenick, C. A., Becker, L., Webster, N. J., Ajrouch, K. J., & Antonucci, T. C. (2020). So close and yet so irritating: Negative relations and implications for well-being by age and closeness. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 75(2), 327–337. doi: 10.1093/geronb/gby038 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Blieszner, R., Ogletree, A. M., & Adams, R. G. (2019). Friendship in later life: A research agenda. Innovation in Aging, 3(1), igz005. doi: 10.1093/geroni/igz005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cantor, M. H. (1979). Neighbors and friends: An overlooked resource in the informal support system. Research on Aging, 1(4), 434–463. doi: 10.1177/016402757914002 [DOI] [Google Scholar]
- Carden, S. W., Holtzman, N. S., & Strube, M. J. (2017). CAHOST: An excel workbook for facilitating the Johnson-Neyman technique for two-way interactions in multiple regression. Frontiers in Psychology, 8(1293), 1–7. doi: 10.3389/fpsyg.2017.01293 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Charles, S. T. (2010). Strength and vulnerability integration: A model of emotional well-being across adulthood. Psychological Bulletin, 136(6), 1068–1091. doi: 10.1037/a0021232 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Charles, S. T., & Carstensen, L. L. (2010). Social and emotional aging. Annual Review of Psychology, 61, 383–409. doi: 10.1146/annurev.psych.093008.100448 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Charles, S. T., Piazza, J. R., Mogle, J. A., Urban, E. J., Sliwinski, M. J., & Almeida, D. M. (2016). Age differences in emotional well-being vary by temporal recall. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 71(5), 798–807. doi: 10.1093/geronb/gbv011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chen, Y., & Feeley, T. H. (2014). Social support, social strain, loneliness, and well-being among older adults: An analysis of the Health and Retirement Study. Journal of Social and Personal Relationships, 31(2), 141–161. doi: 10.1177/026540751348872 [DOI] [Google Scholar]
- Falzarano, F., Moxley, J., Pillemer, K., & Czaja, S. J. (2022). Family matters: Cross-cultural differences in familism and caregiving outcomes. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 77(7), 1269–1279. doi: 10.1093/geronb/gbab160 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fiori, K. L., Windsor, T. D., & Huxhold, O. (2020). The increasing importance of friendship in late life: Understanding the role of sociohistorical context in social development. Gerontology, 66(3), 286–294. doi: 10.1159/000505547 [DOI] [PubMed] [Google Scholar]
- Galupo, M. P., & Gonzalez, K. A. (2013). Friendship values and cross-category friendships: Understanding adult friendship patterns across gender, sexual orientation and race. Sex Roles, 68(11), 779–790. doi: 10.1007/s11199-012-0211-x [DOI] [Google Scholar]
- Gillespie, B. J., Lever, J., Frederick, D., & Royce, T. (2015). Close adult friendships, gender, and the life cycle. Journal of Social and Personal Relationships, 32(6), 709–736. doi: 10.1177/0265407514546977 [DOI] [Google Scholar]
- Goins, M. N. (2011). Playing with dialectics: Black female friendship groups as a homeplace. Communication Studies, 62(5), 531–546. doi: 10.1080/10510974.2011.584934 [DOI] [Google Scholar]
- Guo, S. & Fraser, M. W. (2015). Propensity score analysis: Statistical methods and applications (2nd ed.). Sage Publications. [Google Scholar]
- Ha, J. H., Kahng, S. K., & Choi, N. (2017). Reciprocal effects between health and social support in older adults’ relationships with their children and friends. Research on Aging, 39(2), 300–321. doi: 10.1177/0164027515611182 [DOI] [PubMed] [Google Scholar]
- Hamaker, E. L., & Muthén, B. (2020). The fixed versus random effects debate and how it relates to centering in multilevel modeling. Psychological Methods, 25(3), 365–379. doi: 10.1037/met0000239 [DOI] [PubMed] [Google Scholar]
- Holt-Lunstad, J., Smith, T. B., & Layton, J. B. (2010). Social relationships and mortality risk: A meta-analytic review. PLoS Medicine, 7(7), e1000316. doi: 10.1371/journal.pmed.1000316 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Huxhold, O., Fiori, K. L., & Windsor, T. (2022). Rethinking social relationships in adulthood: The differential investment of resources model. Personality and Social Psychology Review, 26(1), 57–82. doi: 10.1177/10888683211067035 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Huxhold, O., Miche, M., & Schüz, B. (2014). Benefits of having friends in older ages: Differential effects of informal social activities on well-being in middle-aged and older adults. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 69(3), 366–375. doi: 10.1093/geronb/gbt029 [DOI] [PubMed] [Google Scholar]
- Kalmijn, M. (2003). Shared friendship networks and the life course: An analysis of survey data on married and cohabiting couples. Social Networks, 25(3), 231–249. doi: 10.1016/s0378-8733(03)00010-8 [DOI] [Google Scholar]
- Kail, B. L., & Carr, D. C. (2020). Structural social support and changes in depression during the retirement transition:“I get by with a little help from my friends”. The Journals of Gerontology: Series B, 75(9), 2040–2049. doi: 10.1093/geronb/gbz126 [DOI] [PubMed] [Google Scholar]
- Litwin, H., & Shiovitz-Ezra, S. (2010). Social network type and subjective well-being in a national sample of older Americans. Gerontologist, 51, 379–388. doi: 10.1093/geront/gnq094 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lu, P., Oh, J., Leahy, K. E., & Chopik, W. J. (2021). Friendship importance around the world: Links to cultural factors, health, and well-being. Frontiers in Psychology, 11, 3568. doi: 10.3389/fpsyg.2020.570839/ [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lykes, V. A., & Kemmelmeier, M. (2014). What predicts loneliness? Cultural difference between individualistic and collectivistic societies in Europe. Journal of Cross-Cultural Psychology, 45(3), 468–490. doi: 10.1177/0022022113509881 [DOI] [Google Scholar]
- Mouzon, D. M. (2014). Relationships of choice: Can friendships or fictive kinships explain the race paradox in mental health? Social Science Research, 44, 32–43. doi: 10.1016/j.ssresearch.2013.10.007 [DOI] [PubMed] [Google Scholar]
- Ng, Y. M. M. (2020). Re-examining the innovation post-adoption process: The case of Twitter discontinuance. Computers in Human Behavior, 103, 48–56. doi: 10.1016/j.chb.2019.09.019 [DOI] [Google Scholar]
- Ng, Y. M. M. (2022). Using data visualizations to study digital public spaces. First Monday, 27(4). doi: 10.5210/fm.v27i4.12586 [DOI] [Google Scholar]
- Ng, Y. M. M. (2023). Twitter intermittent and permanent discontinuance: A multi-method approach to study innovation diffusion. Computers in Human Behavior, 138, 107482. doi: 10.1016/j.chb.2022.107482 [DOI] [Google Scholar]
- Ng, Y. T., Huo, M., Gleason, M. E., Neff, L. A., Charles, S. T., & Fingerman, K. L. (2021). Friendships in old age: Daily encounters and emotional well-being. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 76(3), 551–562. doi: 10.1093/geronb/gbaa007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ng, Y. T., Huo, M., Han, S. H., Birditt, K. S., & Fingerman, K. L. (2022). Older adult’s marital status, conversation and well-being in everyday life. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 77(3), 499–512. doi: 10.1093/geronb/gbab112 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Nguyen, A. W. (2020). Religion and mental health in racial and ethnic minority populations: A review of the literature. Innovation in Aging, 4(5), 1–13. doi: 10.1093/geroni/igaa035 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Oyserman, D., Coon, H. M., & Kemmelmeier, M. (2002). Rethinking individualism and collectivism: evaluation of theoretical assumptions and meta-analyses. Psychological Bulletin, 128(1), 3–72. doi: 10.1037/0033-2909.128.1.3 [DOI] [PubMed] [Google Scholar]
- PEW Research Center. (2007). Americans and social trust: Who, where and why. https://www.pewresearch.org/social-trends/2007/02/22/americans-and-social-trust-who-where-and-why/#fn-414-4
- Pinquart, M., & Sörensen, S. (2000). Influences of socioeconomic status, social network, and competence on subjective well-being in later life: A meta-analysis. Psychology and Aging, 15(2), 187–224. doi: 10.1037//0882-7974.15.2.187 [DOI] [PubMed] [Google Scholar]
- Raley, R. K., Sweeney, M. M., & Wondra, D. (2015). The growing racial and ethnic divide in US marriage patterns. The Future of Children/Center for the Future of Children, the David and Lucile Packard Foundation, 25(2), 89–109. doi:10.1353/foc.2015.0014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sarkisian, N., & Gerstel, N. (2016). Does singlehood isolate or integrate? Examining the link between marital status and ties to kin, friends, and neighbors. Journal of Social and Personal Relationships, 33(3), 361–384. doi: 10.1177/0265407515597564 [DOI] [Google Scholar]
- Taylor, R. J., Chatters, L. M., Woodward, A. T., & Brown, E. (2013). Racial and ethnic differences in extended family, friendship, fictive kin, and congregational informal support networks. Family Relations, 62(4), 609–624. doi: 10.1111/fare.12030 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wrzus, C., Wagner, J., & Neyer, F. J. (2012). The interdependence of horizontal family relationships and friendships relates to higher well-being. Personal Relationships, 19(3), 465–482. doi: 10.1111/j.1475-6811.2011.01373.x [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
This study was preregistered on the Open Science Framework (https://osf.io/d4jgy/). Data and additional documentation of sample and measures are available upon request to the corresponding author.