Does Shared Positivity Make Life More Meaningful? Perceived Positivity Resonance is Uniquely Associated with Perceived Meaning in Life

Abstract

Pleasantness and meaningfulness are sometimes seen as opposing pursuits. Yet past research has found that the pursuit of meaning often leads to pleasure. In four longitudinal studies—three observational, one experimental, ranging from 5 weeks to 18 months—we investigated an inverse process, whereby specific kinds of pleasant states can foster a sense of meaning in life. We hypothesized that perceptions of positivity resonance, a form of co-experienced positive affect characterized by mutual care and synchrony, are experienced as particularly meaningful in the moment and, over time, build social resources (e.g., supportive relationships and communities) that foster an enduring sense of meaning in life. Results indicate that perceived positivity resonance is associated with perceived meaning both between- and within-persons, links that emerge independently of overall pleasant emotion and social interaction quantity. Perceived social resources mediate between-persons links, and changes in perceived social resources mediate longitudinal links. Overall, these findings suggest that co-experienced, caring, and synchronous pleasant states may be uniquely suited to cultivating a person’s sense that life is meaningful.

What does a good life look like, and how can people live well? Within psychology, there are two traditions of research on the topic. One tradition focuses on “hedonic” states like pleasure or positive affect. Although the value of such states is apparent to all who have experienced them, some researchers argue that the fact that someone feels good does not entail that they are flourishing in the fullest sense (Ryff & Keyes, 1995; Waterman, 2013). Hence, a second tradition focuses on “eudaimonic” states and processes. Although definitions and operationalizations of “eudaimonia” vary, perhaps the most quintessential eudaimonic construct is perceived meaning in life (Steger et al., 2013).

Meaningfulness is central to the ordinary concept of a good life (King & Napa, 1998; Prinzing, De Freitas, et al., 2021), and is thought to require transcending the self—i.e., becoming part of, and finding significance in, something larger (Frankl, 1966; George & Park, 2014; Schnell, 2011). Thus, meaning is often sharply contrasted with pleasure (Frankl, 1971; Kim et al., 2014; Wolf, 1997). Many activities that are positively associated with a sense of meaning in life are negatively associated with pleasant states, and vice versa (Baumeister et al., 2013). Moreover, people often find meaning in unpleasant events (Murphy & Bastian, 2019; Tov & Lee, 2016), and even in suffering or trauma (Park, 2010). For these reasons, some have argued that pleasure and meaning are “not entirely compatible” pursuits (Baumeister, 2018), and even that “discomfort is the price of admission to a meaningful life” (David, 2016, p. 131).

Yet, in the wake of ardent debates between the hedonic and eudaimonic traditions of psychological research (Biswas-Diener et al., 2009; Kashdan et al., 2008; Waterman, 2008), there seems to be growing agreement that—although distinct and sometimes antagonistic—hedonia and eudaimonia are typically positively associated (Dwyer et al., 2017; Huta & Ryan, 2010; Kashdan et al., 2008). This invites the question of precisely how pleasant states and a sense of meaning in life might be related.

According to a traditional view, meaning leads to pleasure, but not the other way around. For instance, Victor Frankl (1971) argued that the direct pursuit of pleasure is bound to fail. To live well, he argued, one must first discover a meaningful purpose for one’s life, and then pleasant states will follow naturally. Empirical research has largely supported this traditional view (Mauss et al., 2011; Sheldon et al., 2019; Titova & Sheldon, 2021), leading to the development of the Eudaimonic Activity Model (Sheldon, 2018). According to this model, to achieve hedonia, it is often most effective to not aim for it directly, but to instead prioritize activities that one finds meaningful (e.g., helping others or developing one’s talents). The pursuit of these “eudaimonic goals” typically satisfies a person’s basic psychological needs (e.g., for competence, autonomy, and relatedness), which in turn produces hedonia.

Although numerous studies have supported the Eudaimonic Activity Model (Martela & Ryan, 2016; Niemiec et al., 2009; Sheldon et al., 2012, 2019; Steger et al., 2008), there is reason to suspect that the relationship between hedonia and eudaimonia is more complex. Cross-cultural research suggests that the traditional view described above applies primarily to individualistic Westerners (Ford et al., 2015). People in more collectivist cultures tend to be more successful in their pursuit of hedonia because they are more likely to pursue it socially. Other research has found that, in Western contexts as well, socially-engaged strategies for improving one’s well-being (e.g., spending more time with family) are more successful than non-social strategies (e.g., finding a better job; Rohrer et al., 2018)—perhaps because people typically experience more pleasant states (Boothby et al., 2014; Reis et al., 2017) and a stronger sense of meaning (Choi et al., 2017) when they are with others. These findings suggest that pleasant experiences with others might support eudaimonia in ways that solo-experienced pleasant states do not. In this paper, therefore, we investigate whether socially embedded and co-experienced positive affect characterized by mutual care and synchrony—termed “positivity resonance”—can support a sense of meaning in life. We theorized that, because of its inherently social and self-transcendent character, positivity resonance might be especially well-suited to fostering a sense of meaning, both in the moment and over the longer term.

Meaning in Life and the “Bigger (Social) Picture”

Theorists across disciplines have argued that assessing whether a life is meaningful involves trying to see how it connects with things beyond itself or fits into the “bigger picture” (Baumeister & Landau, 2018; Frankl, 1966; Prinzing, 2020). Recent empirical research corroborates this idea and indicates that the bigger picture is usually a social one. For instance, studies have found that the more a person has contributed something of value to society the more others will regard their life as meaningful (Prinzing, De Freitas, et al., 2021). Similarly, individuals’ perceptions of the meaning in their own lives seem to be based largely on their perceptions of their own significance (Costin & Vignoles, 2020)—especially their significance to other people (Prinzing et al., 2022).

In line with these findings, other studies have found that, when asked what makes their lives meaningful, people overwhelmingly mention communities and relationships, especially with family and close friends (Delle Fave et al., 2013; Lambert et al., 2010; van Kessel et al., 2018). Studies using densely repeated nightly reports have found that days with positive social experiences feel more meaningful than days without such experiences (Machell et al., 2015; Steger & Kashdan, 2009). One study found that a sense of belonging predicts changes in perceived meaning in life over the following three weeks (Lambert et al., 2013). Another found that perceived social support predicts such changes over seven years (Krause, 2007). These longitudinal results are supported by laboratory experiments, which have found that eliciting a sense of belonging and social closeness boosts perceived meaning in life (Hicks & King, 2009; Lambert et al., 2013), whereas being socially excluded has the opposite effect (Stillman et al., 2009).

In sum, this body of research suggests that, for people to see their lives as meaningful, they typically need to see how they matter to something larger than themselves—such as relationships, groups, or communities.

Positivity Resonance Theory

The Positivity Resonance Theory of co-experienced positive affect (Fredrickson, 2013a, 2016) is an offshoot of the Broaden-and-Build Theory (Fredrickson, 2001). According to the Broaden-and-Build Theory, positive affective states broaden individuals’ mindsets in ways that enable them to incrementally build personal and social resources. In this context, “resources” are defined as stable yet mutable attributes that help individuals to effectively meet life’s challenges and take advantage of its opportunities. Social resources therefore include such things as caring relationships with family and friends, supportive communities, closely-knit neighborhoods, and so on. “Positivity resonance” refers to a subtype of positive emotion that is experienced at a collective level (in dyads or larger groups). It involves co-experienced positive affect, mutual care, and synchrony. According to Positivity Resonance Theory, and consistent with other theories of collective-level emotions (Butler, 2017; Goldenberg et al., 2020), this form of co-experienced positive affect is a unique and especially strong contributor to the accrual of social resources, more than simply the sum of positive affect and social contact, each of which predicts well-being (Choi et al., 2017; Tov & Lee, 2016).

Like emotions generally, positivity resonance is defined in terms of coordinated responses across experiential, behavioral, and physiological systems (Levenson, 2014; Mauss et al., 2005). It occurs within social interactions when multiple individuals show coordinated increases in shared positive affect (an experiential component), caring nonverbal synchrony (a behavioral component), and biological synchrony (a physiological component). Shared positive affect is a pleasant state, co-experienced in real time by multiple individuals. Caring nonverbal synchrony involves coordinated movements and gestures that indicate each person’s engagement with and concern for others. This includes mutual direct body orientation and gaze, smiles, nods, and forward leans (Gonzaga et al., 2001; Tickle-Degnen & Rosenthal, 1990; Wells et al., 2022), as well as a nonconscious, cross-personal coordination of body movements in form, tempo, and intensity (Bernieri et al., 1988; Vacharkulksemsuk & Fredrickson, 2012). These nonverbal behaviors convey engagement, support, responsivity (Sharon-David et al., 2019), and ultimately, attentive care and goodwill (Reis et al., 2004). Biological synchrony occurs when the biological processes of multiple people show parallel changes, which has been found to occur during positive social interactions (Marci et al., 2007; Stephens et al., 2010), and more so during shared positive emotions than during shared negative emotions (Chen et al., 2021). A recent study found that these collective-level experiential, behavioral, and physiological components, assessed during a 15-minute conversation between spouses (via through individual participant reports, behavioral coding, and physiological monitoring) constituted a single latent factor of positivity resonance, and that this latent factor predicted health outcomes over 13 years and longevity over 30 years (Wells et al., 2022).

Positivity Resonance as a Source of Meaning in Life

Based on this past research, we speculated that positivity resonance might be a unique source of meaning in life in two ways. First, such moments of interpersonal connection might feel inherently meaningful. Some studies have found that positive affect inductions can momentarily increase perceived meaning in life (King et al., 2006). In line with the “affect as information” theory (Clore et al., 2001), the researchers argued that people use their current mood to help them decide whether their life is meaningful (Hicks & King, 2008, 2009). Interestingly, however, subsequent research (Van Cappellen et al., 2013) has found that positive affect inductions only influence perceived meaning in life when they elicit self-transcendent positive emotions. This suggests that the boost in meaning may not come simply from pleasantness, but from changes in the degree to which individuals see themselves as connected to something or someone beyond themselves. If so, then moments of positivity resonance may feel particularly meaningful insofar as they involve mutual experiences of oneness and care between people. Such an immediate awareness of interpersonal significance will likely feel meaningful in ways that solo-experienced positive affect does not.

Positivity resonance might also uniquely support meaning in life over longer periods of time by weaving people into the social fabric. As discussed above, supportive relationships and communities are potent sources of perceived meaning in life, and positivity resonance is theorized to be distinctively well-suited to build these and other social resources. Past work indicates that positivity resonance and its components are associated with a variety of social resources, both cross-sectionally and longitudinally (Brown & Fredrickson, 2021). Self-reports of perceived positivity resonance are associated with and less loneliness cross-sectionally (Major et al., 2018), and stronger prosocial tendencies (West et al., 2021; Zhou et al., 2022) cross-sectionally and longitudinally. Observer ratings of behavioral indicators of positivity resonance are cross-sectionally associated with marital satisfaction in long-term couples (Otero et al., 2020). In teams of workers, co-experienced positive affect is cross-sectionally associated with better teamwork, coordination among members, and group cohesion (Peñalver et al., 2019). Similarly, in long-term married couples, co-experienced positive affect during conversation is cross-sectionally associated with both relationship quality (Brown et al., 2021) and synchrony between spouses’ heart rates, sweat gland activity, and peripheral vasoconstriction (Chen et al., 2021). Such physiological synchrony is, in turn, associated with higher-quality interactions and relationships, both concurrently and longitudinally (Chen et al., 2021; Feldman, 2015; Helm et al., 2014). Synchronized body movements during conversations between new acquaintances predict subsequent reports of embodied rapport (Vacharkulksemsuk & Fredrickson, 2012), and experimental manipulations of nonverbal synchrony have been found to increase compassion (Valdesolo & DeSteno, 2011), perceived emotional support (Jones & Wirtz, 2007), and affiliation (Hove & Risen, 2009) between social partners. A meta-analysis of 42 studies, encompassing 4,327 participants, found that experimentally manipulated behavioral synchrony improves social cognition, prosocial behavior, and perceived bonding (Mogan et al., 2017).

In sum, past research suggests that, as a self-transcendent state, positivity resonance may feel particularly meaningful in the moment. Moreover, it may be a key leverage point for building social resources, which in turn contribute to an enduring sense that one’s life is meaningful. Hence, this specific kind of hedonic state may be uniquely suited to support eudaimonic well-being.

The Present Research

Based on research summarized above, we hypothesized that:

• H1. Perceived positivity resonance is positively associated with perceived meaning in life, both between- and within-persons.

The between-person component of this hypothesis posits that people who experience more positivity resonance than others tend also to feel a greater sense of meaning in life. The within-person component posits that the times in a person’s life when they experience more positivity resonance than they typically do, tend to be times when they also find life more meaningful. We further hypothesized that social resources are a mechanism behind these associations:

• H2. Considering the differences between individuals, perceived social resources mediate the link between perceived positivity resonance and perceived meaning in life.

• H3. Considering change over time within individuals, changes in perceived social resources mediate the link between perceived positivity resonance and subsequent changes in perceived meaning in life.

• H4. Experimental interventions that increase individuals’ perceived positivity resonance also increase their perceived social resources, which in turn increase their perceived meaning in life.

We tested these hypotheses in four longitudinal studies (total N = 1,791), ranging from 5 weeks to 18 months. (For Study 3, we collected two samples, one of which was longitudinal, the other cross-sectional.) Longitudinal designs with repeated assessments enabled us to examine both the differences between individuals and the processes of change over time within individuals. Understanding these within-person processes is especially important for identifying avenues through which individuals can improve their lives (Curran & Bauer, 2011).

In these studies, we assessed individuals’ perceptions of positivity resonance, social resources, and meaning in life. Such measures come with pros and cons. “Objective” measures of positivity resonance would include reports from multiple interactants as well as indexes of behavioral and physiological synchrony (Wells et al., 2022). However, it is difficult to collect such data outside of laboratory contexts. Previous studies have validated a measure of perceived positivity resonance (Major et al., 2018), which can be used to study social interactions in situ. Importantly, every item on this scale refers to a collective-level phenomenon—e.g., mutuality, connection, synchrony—rather than individual-level affective experiences. Although reports of collective-level experiences from individuals are necessarily one-sided, this measure enables us to examine real-life interactions, rather than staged interactions in laboratory settings. Social resources could similarly be assessed more “objectively” as the size of a person’s social network, or the support (emotional, financial, etc.) offered by social contacts. Yet, past research indicates that perceptions of social resources are equally if not more important for health and well-being than such indexes (Cacioppo & Cacioppo, 2014; Holt-Lunstad et al., 2015; Shor & Roelfs, 2015; Wethington & Kessler, 1986).

In Study 1 we tested H1 by examining between- and within-person links between momentary perceptions of positivity resonance and meaning. Then, in Study 2, we zoomed out from the micro-level of moments within days to the macro-level of enduring attributes. We administered trait measures of perceived positivity resonance and perceived meaning in life on a quarterly basis for a year and a half. We tested H1 again at this macro-level. We also tested whether perceived social resources mediated any evident between-person links (H2), and whether changes in perceived social resources mediated any longitudinal links between perceived positivity resonance and changes in perceived meaning in life (H3). Study 3 re-tested hypotheses H1–3 in a very different context (i.e., in the early months of the COVID-19 pandemic) and with more demographically and geographically diverse samples. Finally, in Study 4 we used a randomized, controlled experiment to test H4, investigating whether interventions that increased perceived positivity resonance also indirectly increased perceived meaning in life by increasing perceived social resources. Additionally, we tested whether the links between perceived positivity resonance and perceived meaning in life are unique by controlling for overall pleasant and unpleasant emotion and (when possible) the quantity of participants’ social interaction.

It is common in well-being research for variables to display reciprocal causality—or, “upward spiral” dynamics, as they are often called in positive psychology (Fredrickson & Joiner, 2002, 2018; Kok et al., 2013; Ramsey & Gentzler, 2015; Van Cappellen et al., 2021). Hence, we also conducted exploratory analyses in each study, testing for reciprocal links between the three key constructs.

Table 1 summarizes the recruitment methods, sample sizes, and measures used in each study. All studies were secondary analyses of data collected for other purposes. We initially conducted exploratory analyses on the Study 1 data, before pre-registering analysis plans for Studies 2–4. Pre-registration forms, along with data, materials, and analytic code for all studies are available online: https://osf.io/6p9wj/.

Table 1.

Summary of Studies

Study	Design	Recruitment	N	Measures
				Positivity Resonance	Meaning in Life	Social Resources
1	Longitudinal Observational	Community advertisements	102	PPRS	2-item ad hoc	–
2	Longitudinal Observational	Community advertisements	226	Trait PPRS	4-item PIL	Pictorial entitativity scale, 2 UCLALS items
3 Student Sample	Cross-sectional Observational	Undergraduate participant pool	484	2-item PPRS	mMEMS	MHC-SF, social integration, social contribution, and positive relations items
3 Adult Sample	Longitudinal Observational	Mechanical Turk	575	2-item PPRS	mMEMS	MHC-SF, social integration, social contribution, and positive relations items
4	Longitudinal Experimental	Community advertisements	404	PPRS	PPMS	Latent variable indicated by GBS-A, UCLALS, and MHC-SF items

Note. PPRS indicates the Perceived Positivity Resonance Scale (Major et al., 2018). PIL indicates the Purpose in Life Test (Crumbaugh & Maholick, 1964). mMEMS indicates a modified version of the Multidimensional Existential Meaning Scale (Prinzing et al., 2020; for the original MEMS, see George & Park, 2017). PPMS indicates the Perceived Personal Meaning Scale (Wong, 1998). UCLALS indicates the University of California, Los Angeles Loneliness Scale (Russell, 1996). MHC-SF indicates the Mental Health Continuum – Short Form (Keyes, 2009). GBS-A indicates the Acceptance subscale of the General Belongingness Scale (Malone et al., 2012).

Study 1

In this study, we examined perceptions of positivity resonance and meaning at the episodic level—i.e., moments within individuals’ days—using data from Day Reconstruction Method (DRM) surveys (Kahneman et al., 2004). In a DRM survey, participants break down their previous day into a series of episodes and then report on what happened and how they felt during each episode. Studies have found that this method can provide comparable data to ecological momentary assessment, another method that aims to minimize self-report biases (Grube et al., 2008; Stone et al., 2006). For present purposes, the DRM has two key advantages over ecological momentary assessments. First, it does not interrupt the experiences being measured. Second, participants’ responses are not influenced by the affective states that they experienced during the episode on which they are reporting. This helps to avoid a real-time “rosy glow” effect, in which participants’ positive affect at a given moment biases how they respond to ostensibly unrelated questions about that moment.

At the episode-level, we tested whether perceived positivity resonance is associated with perceived meaning both between- and within-persons (H1). We also tested whether any evident links were independent of overall pleasant and unpleasant emotions and the presence of other people. Since these data were collected over the span of 9 weeks, we were able to aggregate each participants’ episode-level data to compute day-level scores and examine longitudinal changes in, and relations between, these constructs over time.

Method

The data were collected as a part of an experiment on the effects of different meditation practices on emotions and health-related outcomes.¹ Although they were not the focus of our analyses, we tested for longitudinal effects of experimental condition on positivity resonance and perceived meaning and controlled for condition when examining the links between these constructs.

Procedure.

The Institutional Review Board at the University of North Carolina at Chapel Hill approved all procedures for the study (#14–1938). The protocol lasted 12 weeks. During Weeks 1 and 12, participants were invited to our laboratory to provide biological samples and complete surveys. In the interim, participants were randomly assigned to learn either Loving Kindness or Mindfulness meditation. In-person classes were held during Weeks 3–8. Participants were asked to complete DRM surveys on Weeks 1, 3, 5, 7, and 9. For the present analyses, we use only the data from these DRMs.

Participants.

Using email listservs and community advertisements, we recruited two waves of participants between February 2015 and May 2016. To be eligible, participants needed to be between 30 and 69 years old, new to meditation, fluent in English, and have home internet access for online surveys. Additionally, given the purposes of the parent study, we pre-screened prospective participants for low childhood socioeconomic status. There were 113 total participants. Of these, N = 102 participants completed DRM surveys (M_age = 49.15, SD_age = 10.75; 53.3% female, 46.6% male; 4.1% Asian, 31.2% Black, 58.8% White, 4.8% mixed race, 1.1% declined to answer). These responses provided a total of 3,010 episode-level observations.

Measures.

For each DRM episode, participants indicated whether they were socializing (0 = No, 1 = Yes), which we defined as, “interacting with anyone for more than a few minutes (including by phone, text messaging, e-mail, social media etc.).” For social episodes, participants completed the Perceived Positivity Resonance Scale (Major et al., 2018). This 7-item measure asks respondents to indicate the proportion of time during the episode (from 0 to 100 percent) that they experienced shared positivity, mutual warmth, and synchrony with others (sample items: “…did you experience a mutual sense of warmth and concern toward other[s]”; “did you feel ‘in sync’ with the other[s]”). The measure displayed excellent internal reliability (coefficient ω = .99/.98 between-/within-persons). For greater correspondence with the response scales of the other measures, we divided positivity resonance scores by 20. Episodes in which participants were not socializing were coded as having 0 positivity resonance.

For all episodes, participants indicated the greatest amount (0 = “Not at all”, 4 = “Extremely”) that they experienced pleasant emotions and unpleasant emotions. We gave examples of each. For pleasant emotions, these were: amusement, awe, excitement, joy, gratitude, hope, inspiration, interest, love, pride, compassion, and serenity. For unpleasant emotions, these were: anger, contempt, disgust, embarrassment, fear, guilt, hate, sadness, shame, and stress.

Using the same 0–4 response scale, we also asked participants, “To what extent did you feel that this episode had a sense of direction or meaning to it?” and “To what extent did this episode feel significant to you?” We computed episode-level meaning scores by averaging these two items (coefficient ω = .98/.87 between-/within-persons).

Analytic Plan

Analyses involved a series of multilevel regression models. We used the lme4 and lmerTest packages in R (Bates et al., 2015; Kuznetsova et al., 2017), and computed standardized coefficients using the “standardize_parameters” function in the effectsize package (Ben-Shachar et al., 2020). To test H1 we used episode-level reports of positivity resonance and meaning, nesting episodes within days (i.e., DRM surveys) and days within participants. To disaggregate within- and between-person associations, we computed grand mean-centered person means and deviations from person means and included both as predictors (Curran & Bauer, 2011). As a robustness check, we controlled for pleasant and unpleasant emotions and whether participants were socializing.

We also conducted longitudinal analyses using day-level scores. We computed day-level meaning by averaging across all episodes within each day for each participant, and day-level positivity resonance by averaging across all social episodes. We also computed an index of social interaction quantity by summing the number of social episodes each participant reported on each day. Using these day-level scores, we tested for reciprocal links between positivity resonance and meaning. Using the data from all 5 timepoints, we tested whether positivity resonance at one time, t, predicted meaning at t + 2 weeks (i.e., in the next DRM), controlling for meaning at t. In a similar but inverted model, we tested whether meaning predicted subsequent changes in positivity resonance. As another robustness check, we controlled for pleasant and unpleasant emotions and the number of social episodes reported at t. Finally, we examined the overall changes in positivity resonance and meaning over the course of the 9-week reporting period. Nesting days within participants, we tested for effects of time (coded 0–8), as well as time × condition interactions.

Results

Descriptive statistics and correlations between person means for study variables are available in Section 1a of the Supplemental Materials. Supporting H1, at the episode-level, positivity resonance was associated with meaning both between-persons, b = .49, 95% CI: [.25, .73], β = .28, p < .001, and within-persons, b = .24, 95% CI: [.18, .30], β = .31, p < .001. (See Figure 1, Study 1.) Adding pleasant emotions, unpleasant emotions, and whether participants were socializing to the model revealed that each was also significantly associated with meaning. Pleasant emotions, b = .38, 95% CI: [.32, .44], β = .33, p < .001, and unpleasant emotions, b = .06, 95% CI: [.01, .12], β = .05, p = .030, were positively associated. Socializing was negatively associated, b = –.21, 95% CI: [–.39, –.04], β = –.08, p = .019. Links between positivity resonance and meaning remained significant, both between-persons, b = .32, 95% CI: [.11, .54], β = .21, p = .004, and within-persons, b = .17, 95% CI: [.11, .23], β = .23, p < .001. This model showed no main effects of, or interactions with, randomized condition (see Supplemental Materials, Section 1b).

Figure 1.

Within- and Between-Person Links Between Perceived Positivity Resonance and Perceived Meaning – Studies 1–3

Note. Plotted lines represent the association between perceived positivity resonance and perceived meaning for individual participants.

Day-level positivity resonance scores predicted increases in meaning over the following 2 weeks, b = .11, 95% CI: [.002, .222], β = .13, p = .048. This link remained significant when controlling for overall pleasant and unpleasant emotions and social interaction quantity, b = .13, 95% CI: [.002, .267], β = .16, p = .048, none of which were significant predictors. In contrast, meaning did not predict changes in positivity resonance over the following 2 weeks, b = .09, 95% CI: [−.07, .25], β = .07, p = .261. (Full results of these models are available in Section 1c of the Supplemental Materials.)

Over the 9-week reporting period, across conditions we observed modest but significant increases in positivity resonance, b = .08, 95% CI: [.04, .12], p < .001, and meaning, b = .04, 95% CI: [.01, .07], p = .007. There was a marginally significant time × condition interaction for positivity resonance, b = –.07, 95% CI: [–.139 –.002], p = .058, suggesting that participants in the Mindfulness condition may have seen greater increases than participants in the Loving Kindness condition. There was no time × condition interaction for meaning, b = .01, 95% CI: [–.05, .07], p = .831.

Discussion

Supporting hypothesis H1, we found that perceived positivity resonance was associated with perceived meaning both between- and within-persons. These links were independent of individual-level pleasant and unpleasant emotion. In line with past work (Murphy & Bastian, 2019; Tov & Lee, 2016), both pleasant and unpleasant emotion were positively associated with perceived meaning. Interestingly, after accounting for perceived positivity resonance, participants experienced time with others as significantly less meaningful than time spent alone. This suggests that what feels meaningful is not the presence of other people, but the warmth, care, and synchrony that one co-experiences with them.

We also found that the average level of positivity resonance that participants perceived on a given day predicted increases in their perceptions of meaning two weeks later. Again, this was independent of overall pleasant and unpleasant emotions and the quantity of social interaction, none of which predicted changes in perceived meaning themselves. These results offer clear support for a key claim of Positivity Resonance Theory: this form of co-experienced positive affect is uniquely associated with positive outcomes, independently of individual-level affective states and social contact. Yet, because perceived meaning did not predict subsequent increases in perceived positivity resonance, the results are not indicative of an upward spiral dynamic.

Although the effects of the experimental intervention were not the focus of our analyses, we found that participants in both conditions showed increases in perceived positivity resonance over the 9-week reporting period and that these changes may have been larger in the Mindfulness condition than in the Loving Kindness condition. Although one might have expected loving kindness meditation to have larger effects, past research has consistently found that mindfulness practices improve social interaction quality (Adair et al., 2018) and reduce loneliness (Lindsay et al., 2019; Zhang et al., 2018). Modern technologies often distract people from what they are doing (Gazzaley & Rosen, 2016), with negative consequences for their social interactions (Misra et al., 2016). Mindfulness, by contrast, involves attending to the present moment. In a social context, this means focusing on one’s interaction partner(s), which is a precondition for positivity resonance.

Study 2

This study expanded on Study 1 by zooming out to encompass a much longer timeframe. Rather than moments in daily experience, assessed over relatively a short period, we examined more general experiences over a period of a year and a half. Because the intervals between assessments were quite long (3 months), and because we are interested in enduring changes in eudaimonic well-being, we used trait measures of each construct, rather than the episodic measures used in Study 1. We first tested for between- and within-person links between perceived positivity resonance and perceived meaning in life (H1). We then tested whether individual differences in perceived social resources mediated any evident between-person link (H2), and whether intraindividual changes in perceived social resources mediated any evident longitudinal link (H3).

Method

The data were collected as a part of an experiment on the effects of different meditation practices on emotions and health behaviors.² The Institutional Review Board at University of North Carolina at Chapel Hill approved all procedures for the study (#12–1376). As in Study 1, midlife adults were randomly assigned to learn either Mindfulness or Loving Kindness meditation. Participants completed 7 waves of surveys every 3 months over an 18-month period. Meditation classes ran during the first quarter and follow-up assessments continued for another 15 months. Although they were not the focus of our analyses, we again tested for longitudinal effects of experimental condition on positivity resonance and perceived meaning and controlled for condition when examining the links between these constructs.

Participants.

We recruited participants using email listservs and community advertisements between the summer of 2013 and fall of 2014. To be eligible, participants needed to be between 35 and 64 years old, interested in making healthy lifestyle changes, and new to meditation. The initial sample included 231 participants. As pre-registered, prior to analysis we excluded participants with excessively missing data in any given wave, leaving N = 226 participants for analyses (M_age = 48.58, SD_age = 8.86; 61.1% female, 38.1% male, < 1% other or declined to answer; 4.9% Asian, 16.8% Black, 77.0% White, < 1% other, < 1% declined to answer).

Measures.

At quarterly intervals, participants completed trait measures of perceived positivity resonance, pleasant and unpleasant emotions, perceived meaning in life, and perceived social resources.

Positivity Resonance.

Participants completed a trait version of the Perceived Positivity Resonance Scale.³ Rather than focus on specific exchanges, this measure asked participants to think of all their “experiences and encounters with other people—the people you interact with regularly on a daily basis, including family, friends, neighbors, work colleagues, customers, etc. (do not just focus on one person individually but how you feel collectively with other people, in general).” This measure showed good internal reliability at each timepoint (coefficient ωs between .88 and .92). As in Study 1, we divided positivity resonance scores by 20 for similarity with the response scales of the other measures.

Pleasant and Unpleasant Emotions.

Participants also completed the modified Differential Emotions Scale (Fredrickson, 2013b). For reasons related to the parent study, they only completed this questionnaire in waves 3–6 (660 observations from 194 participants). The instructions asked participants to indicate the greatest amount (0 = “Not at all”, 4 = “Extremely”) in the previous 2 weeks that they had experienced 20 different emotions: 10 pleasant emotions (e.g., awe, joy, and pride) and 10 unpleasant emotions (e.g., anger, sadness, and disgust). We averaged the pleasant and unpleasant emotions separately (all coefficient ωs between .86 and .94).

Meaning in Life.

Participants completed a 4-item version of the Purpose in Life Test developed by Hicks and King (2007; for original, see Crumbaugh & Maholick, 1964). Participants indicated the extent to which (1 = “Not at all”, 7 = “Extremely”) four statements were true of them (sample item: “My personal existence is very purposeful and meaningful”). The scale demonstrated good internal reliability at each timepoint (ωs between .92 and .96).

Social Resources.

Participants completed a pictorial measure of group entitativity—i.e., sense of interconnection or unity between the self and others (Gaertner & Schopler, 1998). Past work has found that group entitativity is associated with outcomes such as a sense of belonging and social trust (Dang et al., 2019; Sani et al., 2008; Yzerbyt et al., 2000). Participants were shown six images of a group of circles. One circle was labeled “self.” Three unlabeled circles were depicted as increasingly overlapping with the “self” circle. The instructions read: “Select the diagram that you feel best represents you in relation to other people—the people you interact with regularly on a daily basis, including family, friends, neighbors, work colleagues, customers, etc. (do not just focus on one person individually but how you feel collectively to other people, in general. Do not select the diagram based on physical distance or personal space. Each of the ‘self’ circles represents you. Blank white circles represent other people you encounter in your daily life.”

Participants also completed two (typically reverse-scored) items from the UCLA Loneliness Scale (Russell, 1996): “How often do you feel that you are ‘in tune’ with the people around you?”; and “How often do you feel close to people?” Following a previous study (Kok & Fredrickson, 2010), these items were selected prior to data collection as a minimally burdensome measure of social closeness and integration.

Our pre-registered analytic plan involved computing a social resources variable by taking the standardized average of these three items. However, after pre-registration, we noticed that the first UCLA Loneliness item (about being “in tune” with others) is conceptually similar to positivity resonance. Hence, we instead computed the social resources variable using just the pictorial entitativity measure and the second loneliness item (“How often do you feel close to people?”). This 2-item measure showed adequate internal reliability at each timepoint (ωs between .69 and .80). For comparison, we also ran our analyses with the pre-registered 3-item measure. The results were nearly identical. We indicate two minor differences in footnotes below.

Analytic Plan

We used the same analytical techniques and R packages as in Study 1 to test H1 and to test for longitudinal effects of condition on positivity resonance and meaning in life. We tested H2–3 simultaneously with a multilevel mediation model, using the lavaan package (Rosseel, 2012) and robust (Huber-White) standard errors. At the between-persons level, this model tested whether social resources mediated the link between positivity resonance and meaning in life (H2). At the within-person level, it tested whether changes in social resources mediated the longitudinal link between positivity resonance and changes in meaning in life (H3). That is, we temporally separated the predictor, mediator, and outcome by time-lagging assessments of positivity resonance and social resources. Hence, over seven timepoints, positivity resonance at a given time, t, predicted meaning in life at t + 6 months (while controlling for meaning at t), both directly and indirectly through social resources at t + 3 months (while also controlling for social resources at t). By including the auto-regressive paths, this model tests whether positivity resonance predicts changes in social resources and meaning in life. We allowed all three variables to covary freely at time t. In robustness tests, we added pleasant and unpleasant emotions as covariates.

Finally, in an exploratory analysis, we used another multilevel path model to test for reciprocal links, indicative of “upward spirals,” between the three focal constructs. At the within-person level, the model included autoregressive and cross-lagged paths (again across all 7 timepoints). At the between-person level, and at each timepoint at the within-person level, the three constructs were allowed to covary freely.

Results

Descriptive statistics and correlations between person means of each variable are available in Section 2a of the Supplemental Materials. Over the 18-month reporting period, we observed modest but significant increases in positivity resonance, b = .02, 95% CI: [.01, .04], p < .001, and meaning in life, b = .03, 95% CI: [.003, .049], p = .026. There were no time × condition interactions for positivity resonance, b = –.00, 95% CI: [–.03 .03], p = .995, or meaning in life, b = .001, 95% CI: [–.04, .05], p = .964.

Supporting H1, positivity resonance was strongly associated with meaning in life both between-persons, b = 1.04, 95% CI: [.74, 1.34], β = .48, p < .001, and within-persons, b = .46, 95% CI: [.30, .62], β = .10, p < .001. (See Figure 1, Study 2.) These links remained significant when controlling for overall pleasant and unpleasant emotion—even though adding these covariates drastically reduced the number of available observations. (Because participants only completed the emotions measure at 4 of the 7 timepoints, only 653 observations were available for this model, versus 1,245 in the previous model.) Pleasant emotions were positively associated with perceived meaning, though unpleasant emotions were not. Full results of these models are available in Section 2b of the Supplemental Materials.

The multilevel mediation model, illustrated in Figure 2, showed good fit: χ²(2) = 14.81, p = .001; CFI = .999; TLI = .955; RMSEA = .088; within/between SRMR = .022/.000. Supporting H2, social resources significantly mediated the between-persons link between positivity resonance and meaning in life, indirect path: b = .51, 95% CI: [.25, .76], β = .28, p < .001. The direct path from positivity resonance to meaning in life was also significant, b = .42, 95% CI: [.09, .75], β = .23, p = .012.⁴ At the within-person level, positivity resonance at t significantly predicted changes in social resources between t and t + 3 months, b = .19, 95% CI: [.13, .26], β = .17, p < .001. Social resources at t + 3 months also significantly predicted changes in meaning in life between t and t + 6 months, b = .24, 95% CI: [.13, .34], β = .15, p < .001. Supporting H3, the indirect path was significant, b = .05, 95% CI: [.02, .07], β = .03, p < .001. The direct path was not, b = .02, 95% CI: [–.09, .13], β = .01, p = .700.

Figure 2.

Perceived Social Resources Mediate Between-Persons and Longitudinal Links Between Trait Perceived Positivity Resonance and Perceived Meaning in Life – Study 2

Note. * indicates p < .05; ** indicates p < .01; *** indicates p < .001. All path coefficients are fully standardized. Indirect effects are indicated below the mediator. Within-persons, all three variables were allowed to covary freely at t. For visual simplicity, these covariances are not illustrated here.

As a robustness check, we tested whether these indirect effects remained significant when including individual-level pleasant and unpleasant emotions as predictors of social resources at t + 3 months and meaning in life at t + 6 months. (Full results of this model are available in Section 2c of the Supplemental Materials.) Adding these covariates drastically reduced the number of observations (396 in this model). Between-persons, pleasant emotions were positively associated with social resources and meaning, whereas unpleasant emotions were only negatively associated with social resources. Positivity resonance was still associated with social resources (β = .57, p < .001), but social resources were no longer significantly associated with meaning in life (β = .13, p = .218). Hence, the indirect path from positivity resonance to meaning in life through social resources was also no longer significant (β = .07, p = .217). Within-persons, pleasant emotions did not predict changes in social resources or meaning in life, though unpleasant emotions predicted decreases in social resources. Positivity resonance still significantly predicted increases in social resources (β = .14, p < .001), which in turn predicted increases in meaning in life (β = .12, p = .019). However, the indirect path fell just short of significance (β = .02, p = .053).

Testing for reciprocal links (Figure 3), we found that positivity resonance predicted increases in social resources over the following 3 months, b = .14, 95% CI: [.08, .20], β = .13, p < .001. Similarly, social resources predicted increases in positivity resonance over the same period, b = .16, 95% CI: [.10, .21], β = .17, p < .001. Social resources also predicted increases in meaning in life, b = .10, 95% CI: [.02, .19], β = .06, p = .018. And meaning in life predicted increases in social resources over the same period, b = .05, 95% CI: [.02, .08], β = .08, p = .002. In contrast, although meaning in life predicted changes in positivity resonance, b = .03, 95% CI: [.003, .056], β = .05, p = .032,⁵ positivity resonance did not predict changes in meaning in life, b = .07, 95% CI: [–.02, .17], β = .04, p = .138.

Figure 3.

Reciprocal Links between Positivity Resonance, Social Resources, and Meaning in Life – Study 2

Note. * indicates p < .05; ** indicates p < .01; *** indicates p < .001. All path coefficients are fully standardized.

Discussion

Results of the pre-registered analyses supported hypotheses H1–3. Replicating our previous findings over a period of 18 months rather than 9 weeks, perceived positivity resonance and perceived meaning in life were positively associated both between- and within-persons (H1) and remained so independently of individual-level pleasant and unpleasant emotions. Individual differences in perceived social resources mediated the between-person link (H2). That is, people who experienced more positivity resonance than others tended to find their lives more meaningful because they tended to feel a stronger sense of interconnection and closeness with people in general. Similarly, the amount of positivity resonance a person typically experienced at one point in time indirectly predicted changes in their perceptions of meaning in life over the following six months via changes in perceived social resources that occurred during the intervening months (H3). Hence, results suggest that frequent co-experiences of positive affect, marked by mutual care and synchrony may support an enduring, generalized sense of closeness, connection, and integration with others, which in turn makes life feel meaningful.

Although the within- and between-person associations between perceived positivity resonance and perceived meaning in life remained significant while controlling for individual-level pleasant and unpleasant emotions, the mediating paths through perceived social resources were reduced to non-significance. What does this indicate? The link between perceived positivity resonance and perceived social resources remained significant. It was the link from perceived social resources to perceived meaning that was attenuated (see the Supplemental Materials, Section 2c). Hence, one possibility is that perceived social resources support a person’s sense of meaning in life, in part, because they are typically accompanied by pleasant moods. Another possibility is that the loss of statistical significance for the indirect paths was due to the loss of statistical power. Since individual-level pleasant and unpleasant emotions were not assessed at all seven timepoints, controlling for them meant that these models only used about a third of the original number of total observations.

In an exploratory analysis, we also found evidence of upward spiral dynamics. The first was between perceived social resources and perceived meaning in life. Corroborating past research (Stavrova & Luhmann, 2016), this suggests that the sense that life is meaningful may encourage people to invest more in their relationships and become more involved in their communities, which may then further strengthen their sense of meaning in life. We found evidence of a similar reciprocal dynamic between perceived positivity resonance and perceived social resources. Moments of positive connection with others are building blocks for flourishing relationships and communities (Brown & Fredrickson, 2021). Such relationships, and communities also seem to provide social contexts in which people are more likely to experience these moments of shared positive connection. Lastly, although the model indicated that perceived meaning in life predicted changes in perceived positivity resonance, the reverse link was not evident. Past research has found that people with a stronger sense of meaning in life are seen by others as more desirable social partners (Stillman et al., 2011). If perceived meaning in life draws others in, this could lead to more frequent opportunities for positivity resonance. The lack of a direct link between perceived positivity resonance and perceived meaning in life corroborates the results of our pre-registered model. It suggests that, over intervals of several months, the link between perceived positivity resonance and subsequent increases in perceived meaning in life is indirect, operating through the accrual of social resources.

Although the effects of experimental condition were not the focus of our analyses, we again found that participants experienced increases in perceived positivity resonance and perceived meaning in life over the course of the study. We saw no evidence of interactions with randomized condition, suggesting that both mindfulness and loving kindness meditation practices may improve the affective quality of social interaction and perceived meaning in life.

This study corroborated and extended the findings from Study 1. Yet these first two studies share a potential limitation. Because both samples were composed of midlife adults from North Carolina who were learning to meditate, the generalizability of these findings remains unclear. Study 3 addresses this limitation.

Study 3

In this study, we tested hypotheses H1–3 again, only in a very different context—namely, during the COVID-19 pandemic. This health crisis affected individuals’ social lives in many ways. In the spring and summer of 2020, many local and state governments in the United States imposed lockdowns or “stay at home” orders, officially restricting citizens’ freedom of movement and ability to socialize. Even where official restrictions were not imposed, many people independently chose to limit their social activity to reduce the risk of viral transmission. Moreover, risk of exposure to the disease made socializing less comfortable for many people.

Data for this study were collected between late March and early July of 2020 as a part of a longitudinal study on mental health and community functioning during the pandemic.⁶ We recruited a cross-sectional sample of college students and a longitudinal sample of adults from across the United States. As reported in Prinzing et al. (2020), and in line with national trends, large proportions (approximately one third to one half) of each of these samples showed “moderate” to “severe” levels of negative mental health, according to recommended cut-off scores (Health Measures, 2020). This raises the question of whether positivity resonance could make life feel more meaningful even during a period of social distancing and heightened individual distress.

Method

Student Sample.

Between March 28 and April 30, we recruited undergraduate students from two public universities, one in California and one in North Carolina. The analysis sample included N = 484 participants (M_age = 20.45, SD_age = 2.89; 68.0% women, 31.2% men, < 1% other gender or declined to answer; 35.5% Asian, 3.1% Black or African American, 11.2% Hispanic or Latinx, 33.9% White or European American, 12.2% mixed race, 3.1% other, 1.0% declined to answer).

Adult Sample.

We recruited a second sample between April 1 and 8 (T1), using Amazon’s Mechanical Turk and the Cloud Research platform. The T1 analysis sample included N = 575 participants (M_age = 36.69, SD_age = 11.54; 49.9% women, 49.6% men, < 1% other, 8.3% Asian, 14.4% Black or African American, 4.2% Hispanic or Latinx, 65.2% White or European American, 6.1% mixed race, 1.2% other, < 1% declined to answer).

We invited these Mechanical Turk participants to complete follow-up assessments at one-month intervals: between April 29 and May 7 (T2), May 27 and June 4 (T3), and June 24 and July 1 (T4). The T2 analysis sample included N = 300 participants (M_age = 39.16, SD_age = 12.01; 54.0% women, 45.0% men, 1.0% other; 9.0% Asian, 11.3% Black or African American, 4.3% Hispanic or Latinx, 68.3% White or European American, 6.0% mixed race, < 1% other, < 1% declined to answer). The T3 analysis sample included N = 250 participants (M_age = 39.20, SD_age = 11.67; 53.2% women, 45.6% men, 1.2% other; 8.8% Asian, 12.8% Black or African American, 3.6% Hispanic or Latinx, 67.6% White or European American, 4.8% mixed race, 1.6% other, < 1% declined to answer). The T4 analysis sample included N = 283 participants (M_age = 38.66, SD_age = 11.93; 51.6% women, 47.3% men, 1.1% other; 8.1% Asian or Asian American, 12.4% Black or African American, 4.6% Hispanic or Latinx, 68.2% White or European American, 4.6% mixed race, 1.4% other, < 1% declined to answer).

Procedure and Measures.

The Institutional Review Board at University of North Carolina at Chapel Hill approved all procedures for the study (#20–0841). As in Study 1, we assessed perceived positivity resonance using DRM surveys. Participants broke the previous 24 hours into a series of episodes and reported on their thoughts and feelings during each episode. For each, they indicated whether they were socializing—i.e., whether they had an “encounter (including by phone, text messaging, e-mail, social media, etc.) of a few minutes or longer with another person(s) in which the participants attended to one another and adjusted their behavior in response to one another.” Students reported an average of 5.61 (SD = 2.80) social episodes. Adults reported an average of 4.66 (SD = 3.44) social episodes per DRM survey. We used the number of social episodes that participants reported as an index of social interaction quantity.

For each episode, we assessed pleasant and unpleasant emotions using the same measure as in Study 1. We computed day-level scores for pleasant emotions and unpleasant emotions by averaging across episodes within DRM surveys (i.e., for each day for each participant).

For each episode in which participants reported socializing, they also completed an abbreviated version of the Perceived Positivity Resonance Scale (PPRS). Past research has found that dropout rates in online studies increase and data quality decreases as survey length grows (Galesic & Bosnjak, 2009; Hoerger, 2010). This becomes a particular problem in DRM surveys because questions are repeated across episodes. (To illustrate, some participants reported as many as 30 social episodes in one day. Such participants would have been asked to report on their perceptions of positivity resonance 30 times.) Hence, to reduce survey length, we used two of the items from the standard 7-item PPRS. Participants indicated the proportion of time (from 0% to 100%) that they “…experienced a mutual sense of warmth and concern toward one another” and “…felt ‘in sync’ with the other(s)”. We chose these items because they encompass the key features of positivity resonance (i.e., shared positive affect, mutual care, synchrony), and because this 2-item scale was extremely highly correlated (r = .97) with the full scale in another study (Zhou et al., 2022). This episode-level measure displayed good internal reliability (coefficient ω = .92/.86 between-/within-person for both students and adults). In all analyses, we used day-level positivity resonance scores, computed by averaging across social episodes within DRM surveys and dividing by 20.

After the DRM survey, participants completed a series of additional measures, including a modified version of the Multidimensional Existential Meaning Scale (mMEMS; George & Park, 2017). The mMEMS includes subscales for three theorized facets of perceived meaning in life: comprehension, purpose, and mattering. The original mattering items repeatedly reference “the grand scheme of the universe”, which we have found to alter participants’ responses (Prinzing et al., 2022). Hence, we replaced the original mattering subscale with 5 uncontextualized items: “My life matters”; “My life is important”; “The things I do are important”; “My life is worthwhile”; “The things I do have value and significance.” We computed meaning in life scores by averaging all 15 items (ω = .96 for students and between .95 and .98 for adults).

Participants also completed the Mental Health Continuum – Short Form (MHC-SF; Keyes, 2009). This 14-item measure includes items that assess social integration (“That you belonged to a community [like a social group, or your neighborhood]”), social contribution (“That you had something important to contribute to society”), and positive relationships (“That you had warm and trusting relationships with others”). We computed a social resources variable by averaging these three items (ω = .76 for students and between .69 and .74 for adults). We selected these items a priori based on face validity and reported this decision in the pre-registered analysis plan.⁷

Analytic Plan

We first tested H1, using the same analytical techniques and R packages as in Studies 1–2. These analyses included the adult sample only, as the student sample included only one observation per participant. Next, we tested H2 using a multigroup mediation model with sample as the grouping variable, and social resources mediating the between-person link between positivity resonance and meaning in life. To test H3, we used the adult sample to run a longitudinal mediation model in which T1 positivity resonance predicted changes in meaning in life between T1 and T3, both directly and indirectly via changes in social resources between T1 and T2 (Figure 4, Panel B). We used T1–3 for this analysis to maximize sample size while keeping the temporal spacing consistent. For these mediation models, we used the lavaan package in R, bootstrapped standard errors (5,000 resamples), and a robust (Yuan-Bentler) test statistic. As robustness checks, we added pleasant emotion, unpleasant emotion, and social interaction quantity as covariates to each of these models. Finally, in an exploratory analysis, we tested for reciprocal links between positivity resonance, social resources, and meaning in life, using the same analytical techniques as in Study 2.

Figure 4.

Perceived Social Resources Mediate Links between Perceived Positivity Resonance and Perceived Meaning in Life During the COVID-19 Pandemic

Note. * indicates p < .05; ** indicates p < .01; *** indicates p < .001. All path coefficients are fully standardized. Indirect effects are indicated next to the mediator. In Panel A, path coefficients for the student sample are stacked on top of coefficients for the adult sample. Although the raw coefficients for the b and c paths are constrained to equality, the standardized coefficients vary slightly. In Panel B, all constructs were allowed to covary freely at T1. For visual simplicity, these covariances are not illustrated here.

Results

Descriptive statistics and correlations between person means of each variable are available in Section 3a of the Supplemental Materials. The adults reported more positivity resonance than the students, t(1,018) = 5.37, p < .001, d = .34, 95% CI: [.21, .46]. However, meaning in life and social resources did not differ significantly across samples.

Supporting H1, in the adult sample positivity resonance was associated with meaning in life both between-persons, b = .43, 95% CI: [.36, .51], β = .43, p < .001, and within-persons, b = .05, 95% CI: [.01, .09], β = .03, p = .026. (See Figure 1, Study 3.) In the robustness check, pleasant emotions were significantly, positively associated with meaning in life, unpleasant emotions were significantly, negatively associated, and social interaction quantity was not significant. After adding these covariates, the hypothesized between-person link remained significant, b = .36, 95% CI: [.29, .44], β = .36, p < .001. However, the within-person link did not, b = –.00, 95% CI: [–.05, .05], β = .00, p = .993. (See Section 3b of Supplemental Materials for the full results of this model.)

The results of the multigroup mediation model are presented in Figure 3, Panel A. A likelihood ratio test indicated that constraining all paths to equality across samples significantly reduced model fit, χ²(3) = 12.06, p = .007. However, constraining just the paths from social resources to meaning in life and positivity resonance to meaning did not, χ²(2) = 4.06, p = .132. Hence, we allowed the path from positivity resonance to social resources to vary across samples. This partially constrained model showed excellent fit: χ²(2) = 4.06, p = .132; CFI = .999; TLI = .996; RMSEA = .045; SRMR = .028. Supporting H2, social resources significantly mediated the link between positivity resonance and meaning in life, both for students, b = .12, 95% CI: [.07, .17], β = .12, p < .001, and adults, b = .21, 95% CI: [.17, .27], β = .23, p < .001. The direct path from positivity resonance to meaning in life was also significant and positive, b = .15, 95% CI: [.10, .20], β = .16/.17 for students/adults, p < .001. Adding pleasant and unpleasant emotion and social interaction quantity as covariates, the indirect path remained significant in the adult sample, b = .12, 95% CI: [.07, .17], β = .13, p < .001. However, it was no longer significant in the student sample, b = .04, 95% CI: [–.01, .09], β = .04, p = .100, as the path from positivity resonance to social resources was no longer significant among students, b = .09, 95% CI: [–.02, .19], β = .08, p = .098. (See Section 3c of the Supplemental Materials for the full results of this model.)

The results of the longitudinal mediation model are presented In Figure 4, Panel B. The model showed good fit: χ²(2) = 8.36, p = .012; CFI = .990; TLI = .948; RMSEA = .077; SRMR = .029. Positivity resonance at T1 significantly predicted changes in social resources between T1 and T2, b = .13, 95% CI: [.05, .22], β = .13, p = .002. Changes in social resources at T2, in turn, predicted changes in meaning in life between T1 and T3, b = .22, 95% CI: [.10, .34], β = .24, p < .001. Supporting H3, the indirect path was significant, b = .03, 95% CI: [.01, .06], β = .03, p = .028. The direct path was not, b = .06, 95% CI: [–.01, .12], β = .06, p = .069. Including pleasant emotions, unpleasant emotions, and social interaction quantity as predictors of T2 social resources and T3 meaning in life, revealed that none of these paths were significant (ps ≥ .283). More importantly, their addition to the model did not affect the original pattern of results. (See Section 3d of the Supplemental Materials for the full results of this model.)

Testing for reciprocal links (Figure 5), we found that positivity resonance at t predicted changes in social resources over the following month, b = .05, 95% CI: [.001, .095], β = .05, p = .045. Social resources predicted changes in positivity resonance over the same period, b = .11, 95% CI: [.04, .18], β = .11, p = .003. Social resources at t also predicted changes in meaning in life over the following month, b = .09, 95% CI: [.04, .15], β = 10, p < .001. And meaning in life predicted changes in social resources over the same period, b = .20, 95% CI: [.13, .28], β = .18, p < .001. In contrast, although positivity resonance predicted changes in meaning in life, b = .05, 95% CI: [.01, .08], β = .06, p = .006, the reciprocal link from meaning to changes in positivity resonance was not significant, b = .04, 95% CI: [–.05, .13], β = .03, p = .400.

Figure 5.

Reciprocal Links between Perceived Positivity Resonance, Perceived Social Resources, and Perceived Meaning in Life – Study 3

Note. * indicates p < .05; ** indicates p < .01; *** indicates p < .001. All path coefficients are fully standardized.

Discussion

Overall, results suggest that—even when a global pandemic required social distancing and created fear of viral spread—creating moments of positive connection with others was one way to experience more meaning in life. Supporting H1, we again found that perceived positivity resonance was associated with perceived meaning in life, both between- and within-persons. In contrast to Studies 1–2, the within-person link was no longer significant when controlling for individual-level pleasant and unpleasant emotion and quantity of social interaction. This difference in results could be attributed to the pandemic context. This was a time when people had very few interaction partners, and individuals’ social lives were far more restricted. Alternatively, it may simply reflect the limits of these data. The adult sample included an average of just over two day-level positivity resonance observations per participant. By contrast, Study 2 had about five and a half average quarterly observations per participant, and Study 1 had an average of almost twenty-nine episode-level observations per participant. Hence, in this study, the non-significance of the within-person link, after controlling for individual-level pleasant and unpleasant emotions and social interaction quantity, may result from limited within-person variance in positivity resonance.

Supporting H2, individual differences in perceived social resources mediated the between-person link between perceived positivity resonance and perceived meaning in life. We observed differences between the student and nationwide adult samples, however. Students showed lower levels of perceived positivity resonance and weaker links between perceived positivity resonance and perceived social resources. Moreover, for the adults but not the students, that link remained significant when controlling for overall pleasant and unpleasant emotions and social interaction quantity. We speculate that this may reflect the unique social situation of university students at that time. When these data were collected, most of these students had recently been forced to leave campus (because of the risks of outbreaks in high-density residential communities) and return to their families. The affective quality of students’ interactions at home may therefore have had relatively little influence on their sense of social integration and contribution, which may have been based more on their perceptions of their disrupted social lives at university.

Supporting H3, perceived positivity resonance at the start of the study predicted changes in perceived meaning in life over the following months, and this link was mediated by changes in perceived social resources during the intervening period. This mediating path remained significant when controlling for pleasant and unpleasant emotions and social interaction quantity. Indeed, these other variables did not significantly predict changes in either perceived social resources or perceived meaning in life. This provides further support for the key claim of Positivity Resonance Theory: this specific form of co-experienced positive affect comes with benefits that individual-level positive affective states do not. Similarly, these results again suggest that the quantity of a person’s social interactions is not as important as the affective quality of those interactions.

Finally, in an exploratory analysis, we again found evidence for upward spiral dynamics. As in Study 2, we observed reciprocal links between perceived positivity resonance and perceived social resources and between perceived social resources and perceived meaning in life. However, in contrast to Study 2, perceived positivity resonance predicted increases in perceived meaning in life, whereas the reverse link did not emerge. This difference in results could perhaps be explained by the shorter interval between assessments (1 month vs. 3 months in Study 2), or the pandemic context of this study. During a period of increased isolation and “social distancing,” moments of positive connection with others may have felt especially meaningful, independently of whether they affected one’s relationships or community involvement.

Because these data come from participants of different age groups, across the United States, and during the COVID-19 pandemic, the results help to extend the generalizability of our previous findings. However, like Studies 1–2, this observational study cannot support causal conclusions. We aimed to address this limitation in Study 4.

Study 4

In this study, we tested whether experimental interventions that produce upward shifts in people’s day-to-day experiences of positivity resonance also lead to increases in perceived social resources, which in turn increase in perceived meaning in life (H4). We used a randomized, controlled experiment with two distinct experimental interventions, each of which boosted participant’s daily experiences of positivity resonance.⁸ This study also improved upon Studies 1–3 by including a wider range of social resources measures, which we used to construct a latent variable.

Method

Procedure.

The Institutional Review Board at University of North Carolina at Chapel Hill approved all procedures for the study (#18–2810). Participants completed trait measures at baseline (T1), nightly reports on their social experiences for 35 days, and the same trait measures again at follow-up (T2). At T1, participants were randomly assigned to one of four conditions. Participants in two Social Connectedness Conditions watched a video (approximately 11 minutes) on the value positive connections with others and were encouraged to create more such moments over the 35-day reporting period. In one of these two conditions, participants were encouraged to create moments of connection with other people generally (Social Connectedness – General). In the other, participants were encouraged to do so specifically with strangers and acquaintances (Social Connectedness – Weak Ties). Participants in a Mindfulness condition watched a video (approximately 10 minutes) on the value of mindfulness and were encouraged to cultivate more mindful moments over the 35-day reporting period. Participants randomized to these experimental interventions also received regular emails with reminders of their respective behavioral goals. Participants randomized to a passive monitoring condition (Control) were not asked to change their behavior and simply completed the assessments.

Previously published analyses of these data (Zhou et al., 2022) reported that the two Social Connectedness conditions and the Mindfulness condition each showed significant increases in positivity resonance over the 35-day reporting period, whereas the Control condition showed no change. The Mindfulness condition was originally included in the study to serve as an active control group. Yet, as shown in Studies 1–2, mindfulness meditation practices can boost positivity resonance. Although in (Zhou et al., 2022) we had hypothesized that instructing participants to focus specifically on creating moments of positive social connection would have a larger effect on positivity resonance, the effect of this intervention did not differ significantly from that of the Mindfulness intervention. Hence, for present purposes, and in accordance with our pre-registered analysis plan, we collapsed the Mindfulness and Social Connectedness conditions into a single Experimental group (n = 303). Results of a less parsimonious model, in which we treated the Mindfulness and Social Connectedness conditions as separate, are available in Section 4a of the Supplemental Materials. The pattern of results in that model is identical to the results reported below.

Participants.

We recruited participants using flyers and email listservs between March and November of 2019. To be eligible participants needed to be between 20 and 65 years old, working part- or full-time, and not currently an undergraduate student. We initially recruited 416 participants. Participants who did not complete any of the nightly reports (n = 11) were excluded from analysis, leaving an analysis sample of N = 405 participants (M_age = 33.90, SD_age = 11.22; 80.2% women, 18.8% men, < 1% declined to answer; 7.7% Asian, 11.9% Black, 7.2% Hispanic or Latinx, 64.7% White, 2.7% other, 5.9% declined to answer).

Measures.

In the nightly reports we assessed participants’ overall daily pleasant and unpleasant emotions using the same measures as in Studies 1 and 3. Participants also reported on select social experiences from their day using the Event Reconstruction Method (ERM), which is like a DRM except that it focuses on a single episode (Grube et al., 2008). Participants were asked to think back to the single longest interaction that they had with one or more “strangers or acquaintances,” and then “mentally relive this event, including how the event unfolded, what time of day it was, and what it was like.” They then completed the 7-item Perceived Positivity Resonance Scale for that episode. This procedure was repeated with “close friends or loved ones” in place of “strangers and acquaintances.” We computed daily positivity resonance scores by averaging all 14 items and dividing by 20 (coefficient ω = .99/.95 between-/within-persons).

At T1 and T2, we assessed perceived meaning in life with the Perceived Personal Meaning Scale (Wong, 1998). Participants used a 7-point Likert scale (1 = “Not at all”, 7 = “A great deal”) to indicate the extent to which eight statements characterize their lives (sample item: “My life as a whole has meaning”). This measure displayed good internal reliability (ω = .91/.92 at T1/T2).

At T1 and T2 participants also completed several measures of perceived social resources. One was the General Belongingness Scale (Malone et al., 2012), which asked participants to indicate their agreement (1 = “Strongly disagree”, 7 = “Strongly agree”) with 12 statements about their everyday life. This measure has two separately scored subscales: Acceptance/Inclusion (e.g., “I have a place at the table with others”; ω = .92/.92 at T1/T2) and Rejection/Exclusion (e.g., “I feel like an outsider”; ω = .90/.90 at T1/T2). Participants also completed the 8-item version of the UCLA Loneliness Scale (Hays & DiMatteo, 1987), which asked how often (1= “Never”, 4 = “Often”) they felt socially isolated (e.g., “I lack companionship”; ω = .87/.85 at T1/T2). As in Study 3, participants also completed the Mental Health Continuum – Short Form (Keyes, 2009), and we took the average of the same three items as in Study 3 (ω = .73/.75 at T1/T2).

Analytic Plan

Replicating previously reported analyses of these data, we verified that experimental condition led to longitudinal changes in positivity resonance, using a multilevel regression model and testing for a time (i.e., study day, coded 0–34) × condition interaction, allowing random intercepts and slopes. Unlike previous analyses, however, we also ran parallel models with pleasant emotions and unpleasant emotions as the dependent variables.

We tested H4 with a structural equation model, using the lavaan package in R, Full Information Maximum Likelihood, bootstrapped standard errors (5,000 resamples), and a scaled (Yuan-Bentler) test statistic. In this model, experimental condition predicted participants’ changes in positivity resonance over the 35-day reporting period (i.e., random effects of time extracted from a multilevel model). These changes in positivity resonance, in turn, predicted changes in meaning in life between T1 and T2, both directly and indirectly via changes in social resources (see Figure 6). Social resources and meaning in life were allowed to covary freely at T1. For thoroughness, we compared the fit of this model with that of an alternative model in which the path between T2 social resources and T2 meaning in life was reversed (i.e., meaning in life mediates the link between changes in positivity resonance and social resources).

Figure 6.

Indirect Effects of Experimental Condition on Perceived Meaning – Study 4

Note. * indicates p < .05; ** indicates p < .01; *** indicates p < .001. All path coefficients are fully standardized. For visual simplicity, the indicators of the social resources latent variable are not illustrated here (but see Section 4b of the Supplemental Materials).

As pre-registered, we tested the fit of multiple social resources measurement models. The measurement model that we chose was a “partial” strong invariance model, including Acceptance/Inclusion, Loneliness, and the averaged MHC-SF items as indicators. Results of this measurement model and documentation of our pretesting process are available in Section 4b of the Supplemental Materials.

Results

Descriptive statistics and correlations between person means of each variable are available in Section 4c of the Supplemental Materials. Replicating previous analyses of these data, we found that participants experienced increases in positivity resonance in all conditions apart from Control (see Section 4d of the Supplemental Materials). In contrast, we found no effects of time, and no time × condition interactions, for either pleasant emotion or unpleasant emotion (see Section 4e of the Supplemental Materials).

Comparing the fit of our hypothesized model with that of the alternative model (in which change in meaning in life mediates the effect of change in positivity resonance on change in social resources) revealed that our hypothesized model showed significantly better fit, Vuong’s Z = 2.75, p = .003. Indeed, the hypothesized model, illustrated in Figure 6, showed excellent fit: χ²(29) = 57.48, p = .001; CFI = .996; TLI = .993; RMSEA = .049; SRMR = .037. All hypothesized paths were significant. Relative to the Control group, the Experimental group showed greater increases in positivity resonance, b = .004, 95% CI: [.001, .007], β = .14, p = .006. These changes in positivity resonance, in turn, predicted changes in social resources, b = 5.79, 95% CI: [1.20, 10.41], β = .09, p = .014, which themselves predicted changes in meaning in life, b = .54, 95% CI: [.40, .66], β = .39, p < .001. Hence, each hypothesized path in the model was significant. Moreover, the indirect path from positivity resonance slope to change in meaning in life via change in social resources was significant, b = 3.11, 95% CI: [.63, 5.85], β = .03, p = .020. However, the 3-step path—from condition to positivity resonance slope to change in social resources to change in meaning in life—fell just short of significance, b = .01, 95% CI: [.002, .030], β = .01, p = .073. The direct paths from experimental condition to T2 social resources and meaning in life were each non-significant, (ps = .311, and .992 respectively), as was the path from T1 meaning to positivity resonance slope (p = .367). The path from T1 social resources to positivity resonance slope was not significant, b = .002, 95% CI: [–.000, .005], β = .15, p = .074.

Discussion

This experiment provides suggestive but inconclusive evidence for the hypothesis (H4) that experimental interventions that increase perceived positivity resonance also increase perceived meaning in life by increasing perceived social resources. We found that encouraging participants either to practice mindfulness or to cultivate positive social connections each increased participants’ perceptions of positivity resonance in their daily lives over the following 5 weeks. Moreover, this effect was specific. Neither intervention significantly altered participants’ overall daily experiences of pleasant or unpleasant emotions. The observed increases in perceived positivity resonance predicted changes (between the start and end of the study) in perceived social resources, which in turn predicted changes in perceived meaning in life. That is, each hypothesized path emerged as significant. Moreover, the mediating path from change in perceived positivity resonance to change in perceived meaning in life via change in perceived social resources was also significant. Although this pattern of results resembles the results of the models that tested H3 in Studies 2–3, we underscore that, in this study, rather than perceived positivity resonance at one time predicting changes in the other variables at later times, we found that change in perceived positivity resonance predicted concurrent change in perceived social resources and thereby concurrent change in perceived meaning in life. Despite the overall consistency of these results with our hypotheses, we must emphasize the 3-step indirect path from experimental condition to change in perceived meaning in life fell just short of statistical significance. Overall, these results constitute suggestive—if inconclusive—evidence for causal links between engaging in intentional positive activities to boost perceived positivity resonance and an increased sense of meaning in life. Future randomized experiments are needed to establish the hypothesized causal process more decisively.

Regarding upward spirals, perceived meaning in life and perceived social resources at the start of the study did not predict change in perceived positivity resonance over the 5-week reporting period. Hence, these results to not offer clear evidence for the upward spiral dynamics observed in Studies 2–3.

General Discussion

We conducted four studies to test four hypotheses. Studies 1–3 provided strong and consistent support for the hypothesis (H1) that people who experience more shared positivity, warmth and synchrony in their day-to-day social encounters (i.e., positivity resonance) tend to feel a greater sense of meaning in life than people who experience less, and the times when a given person experiences more shared positivity, warmth and synchrony with others tend to be the times when they feel a stronger sense of meaning in life. In Study 1 we found these links at the level of moments within days. In Study 2 we found them at the level of general experiences over a period of months. And in Study 3 we found them during the tumultuous early months of the COVID-19 pandemic.

Moreover, arguing against a mere “rosy glow” effect of general positivity, these links were independent of individual-level pleasant and unpleasant emotion. (Both were independent in Studies 1–2. Though, only the between-person link was independent in Study 3.) In the two studies (1 and 3) that enabled us to test for independence from the quantity of participants’ social interactions, we established the unique predictive power of the collective affective quality of participants’ social interactions. In support of a central claim of Positivity Resonance Theory, people seem to experience additional positive outcomes when they experience pleasant states collectively and simultaneously with other people, versus alone. Moreover, the sense of meaningfulness previously found to come from interacting with others appears to depend largely, if not entirely, on the quality of the affective connection between people. Per theory, the holistic construct of positivity resonance appears to reflect a synergy between pleasant affect and social contact, rather than merely their sum.

Considering differences between individuals, results supported the hypothesis (H2) that people who typically perceive more positivity resonance also perceive themselves as closer with others, and more socially integrated and valued. These social resources in turn explain why these individuals tend to experience their lives as more meaningful. We observed this mediating pathway in Study 2, as well as in Study 3, which included samples of undergraduate students and adults from across the United States.

Considering processes of change over time, results also supported the hypothesis (H3) that perceived positivity resonance predicts subsequent changes in perceived meaning in life and that these links are mediated by changes in perceived social resources. In Study 1, the positivity resonance that participants perceived on a given day predicted increases in perceived meaning over the following two weeks. In Study 2, we found that perceived positivity resonance assessed at one time predicted changes in perceived meaning in life over the following six months, and that this link was accounted for by changes in perceived interpersonal closeness and connection during the intervening months. This indirect effect emerged again in Study 3, during the social upheaval that characterized the early months of the COVID-19 pandemic.

We found mixed results when testing whether these between- and within-person mediating paths were independent of individual-level pleasant and unpleasant emotions. The mediating paths remained significant in two of five of these robustness tests. Puzzlingly, when controlling for individual-level emotions reduced mediating paths to non-significance, it was for different reasons in different cases. In Study 2, the non-significant path was between perceived social resources and perceived meaning. In the student sample of Study 3, the non-significant path was between perceived positivity resonance and perceived social resources. Given this inconsistency, it’s unclear what might explain those results. What does seem clear from the overall pattern of results, however, is that the hypothesized link between perceived positivity resonance and perceived meaning in life is independent of individual-level pleasant and unpleasant emotions, and that the hypothesized mediating path through perceived social resources might be independent as well, though further evidence is needed.

We also found suggestive evidence for our causal hypothesis (H4) that experimental interventions that increase perceived positivity resonance also boost perceived social resources and, in turn, perceived meaning in life. In Study 4, experimentally induced increases in perceived positivity resonance significantly predicted improvements in felt belonging, loneliness, social contribution, and positive relationships. These newly accrued social resources, in turn, predicted increases in perceived meaning in life. Each individual step in the path from experimental condition to changes in perceived meaning in life was significant. However, the 3-step indirect pathway fell short of significance. Hence, although Study 4 offers preliminary evidence for the hypothesized causal process, further randomized longitudinal experiments are needed.

Finally, in exploratory analyses, we tested for reciprocal links, or “upward spiral” dynamics among the three key constructs. We found consistent evidence for upward spirals between perceived positivity resonance and perceived social resources. That is, the supportive relationships and communities that people build partly by creating moments of shared positivity, warmth, and synchrony appear, in turn, to help create social contexts in which such moments become more frequent. A parallel dynamic emerged between perceived social resources and perceived meaning in life. As observed in past research (Stavrova & Luhmann, 2016), the sense that life is meaningful seems to encourage people to invest more in their relationships and become more involved in their communities, which may then further strengthen their sense of meaning. Although we repeatedly found, across studies, that perceived positivity resonance predicted increases in perceived meaning—primarily indirectly through change in social resources—evidence for the inverse link (from perceived meaning to increases in perceived positivity resonance) only emerged in Study 2. Hence, it remains unclear whether a sense of meaning in life helps facilitate experiences of positivity resonance.

Strengths, Limitations, and Future Directions

A key strength of this work is the use of longitudinal designs, lasting between 5 weeks and 18 months, which enabled us to examine within-person associations over a range of timescales. Longitudinal analyses are particularly valuable for revealing how individuals change over time and, consequently, point to what they might do differently to cultivate a stronger sense of meaning in life. Another strength is the diversity of the five independent samples, which included 1,791 participants. We collected samples of community adults (including one selected for low childhood SES, which showed appreciable racial diversity), online samples of American adults, and an undergraduate sample. We also examined these processes in very different social contexts: prior to and during the COVID-19 pandemic.

One key limitation of the present work is the observational character of our primary analyses in the first three studies. Study 4 employed a randomized experimental design with a passive monitoring control. However, particularly given the marginal significance of the 3-step indirect path, more longitudinal experiments are needed. Another limitation of these studies arises from the use of self-report measures. As noted in the introduction, positivity resonance would ideally be measured using reports from multiple interactants paired with behavioral and physiological data. Collecting such data about real-life interactions would be very difficult. Nevertheless, it would be an important step forward for future work. One strategy might be to recruit romantic couples, work teams, or other groups that could report on their regular interactions with each other. Wearable fitness trackers might be used to collect physiological data, and smartphones to record audio (for behavioral coding) and geolocation (to estimate frequency of social contact).

Another promising avenue for future research would be to investigate whether the processes examined in these studies vary across social interaction partners. That is, does it matter who one experiences positivity resonance with? Intuitively, positivity resonance with close ties (e.g., family members, close friends, or romantic partners) might have stronger effects on perceived social resources and meaning in life. Yet research (Sandstrom & Dunn, 2014) has found that interactions with weak ties (e.g., coworkers or acquaintances) have surprisingly strong links with well-being. It may be that weak ties are treated as psychological stand-ins for human beings in general. Hence, interactions with weak ties may form the basis for individuals’ beliefs about humanity or about their significance within society.

Conclusion

These studies have implications for the broader literature on the hedonic and eudaimonic aspects of well-being. Although pleasure and meaning are sometimes portrayed as antagonistic pursuits, past research has found that they are often positively associated, and that the pursuit of meaning leads to pleasant experiences. Here, we found evidence that pleasant states, when co-experienced with others and marked by mutual caring and synchrony, feel especially meaningful. They also appear to build social resources that become potent sources of meaning in life. This body of research thus offers evidence that hedonia and eudaimonia are mutually reinforcing aspects of psychological health and, more generally, a well-lived life.

One appealing aspect of this route to a more meaningful life is the fact that positivity resonance is relatively accessible. These moments don’t cost anything and people can take steps each day to cultivate them, both with strong and weak social ties. However, this does not imply that connecting with others is always easy. Positivity Resonance Theory posits several preconditions for such moments of connection, including perceived safety and real-time sensory connection (Fredrickson, 2013a, 2016). These preconditions are not always or equitably satisfied. Hence, for those seeking more meaning in life, a practical piece of advice could be to make conscious efforts to create moments of shared positivity in daily interactions and to seek out conditions under which such moments are easier or more frequent.