Mitigating the affective and cognitive consequences of social exclusion: an integrative data analysis of seven social disconnection interventions

Social ties serve essential self- and social-regulatory functions at the levels of affect, cognition, behavior, and physiology [33‐47], making them fundamental to well-being [48‐50]. As a highly social species, humans have inherited cognitive architecture that makes affiliation, or even the promise of it, rewarding [51].

One particularly important type of social tie is friendships. Not only do friends provide pleasure, companionship, and intimacy, they stimulate social and intellectual development [52‐54]. Correlational work finds that the number and quality of friendships are associated with greater self-reported well-being and self-esteem [55‐58]. The number of friends one has predicts health, happiness, and academic, financial, and professional success [52, 59‐61]. We propose that friendships may be particularly instrumental in protecting individuals from social disconnection through their affective and cognitive regulatory functions. Although affective and cognitive regulatory functions are typically associated with an individual’s closest, deepest attachment bonds—such as those with a primary caregiver in early life and a romantic partner in adulthood—friends serve similar functions, though with muted intensity, frequency, specificity, and persistency [53, 54, 62‐68]. Specifically, friends can restore feelings of comfort and security following threats, and can confer confidence to explore new environments. Supporting these ideas, developmental research shows that by middle childhood, children increasingly turn to their friends for emotional support, by late adolescence, peers are often preferred over parents as sources of comfort, and in adulthood, friends remain a source of comfort and security [53, 63, 64, 69, 70].

People are able to identify who they turn to for specific affective regulatory needs (e.g., who do you turn to to cheer you up when you are feeling sad?) and having a diverse social network not only facilitates this process, but is associated with well-being [71, 72]. While people are more likely to seek support from those they perceive as being closer [73], some work suggests that even unknown others (i.e., strangers) may also provide affective and cognitive benefits. Indeed, social interactions with peripheral members of social networks and strangers have the power to increase feelings of belongingness [74‐76]. Moreover, interactions with unknown others can modulate physical pain, reducing pain intensity and feelings of unpleasantness while fostering feelings of social support (see [77] for a review). This literature suggests the possibility that strangers may also confer protection against the negative consequences of social disconnection. Thus, in our research, we examine social disconnection interventions that involve friends and unknown peers. Additionally, we investigate how relationship quality with the friend may be related to the efficacy of social disconnection interventions specifically involving friends.

Previous research has found that friendly interactions can mitigate the affective and cognitive consequences of common life stressors [78]. Even the mere thought of another person can help. Thinking about an attachment figure (such as a close friend or family member) can help to regulate negative emotions and cognitions triggered by stress [79‐81]. However, the mere presence of another may not always be enough in the face of social disconnection. Research using experience sampling finds that feelings of social disconnection can actually be amplified when others are simply present [82]. In our research, we examine the nature of interpersonal engagement (e.g., engaging in a conversation vs. being reminded of an upcoming interaction vs. simply being in the presence of another) to provide clarity to the seemingly divergent findings in the existing literature.

A social disconnection intervention before social exclusion may buffer against the affective sting and cognitive reactions of exclusion. Conversely, a social disconnection intervention following social exclusion may promote recovery from the negative consequences of exclusion. With regard to buffering, perhaps surprisingly, there is mixed experimental evidence that friendships buffer against adverse social experiences [81, 83‐85]. Moreover, of the studies that do find some evidence, buffering effects are usually weak [83]. In contrast, studies using recovery interventions generally find stronger and more robust effects of social presence on reducing the affective sting and cognitive reactions of social exclusion [86, 87]. Indeed, even mundane interactions with an unknown peer can promote recovery following exclusion [88]. In our investigation, we systematically examine whether the timing of social disconnection interventions matters.

The current work aims to understand how friends, and even the mere presence of an unknown peer, may mitigate the negative consequences of being excluded or ignored in a social setting, a common form of social exclusion that may be a precursor to chronic loneliness and social isolation [14‐23]. Across five studies featuring seven distinct social disconnection interventions, we systematically examine how the affective sting and cognitive reactions of social exclusion are dampened by various forms of social presence. All studies were part of the same project and participants were restricted from participating in more than one study. We focused on social disconnection interventions that are likely to occur naturally in day-to-day life (e.g., a conversation that occurs prior to or following a negative event). We instantiated the experience of social exclusion using a modified version of Cyberball, a well-validated, virtual ball-tossing game used to experimentally manipulate exclusion that reliably induces negative affect and threatens four fundamental needs (i.e., self-esteem, feelings of belonging, meaningful existence, and control) (see Hartgerink et al., (2015) for a meta-analysis) [89].

Critically, across the five studies, we varied three features of the social context during which the exclusion took place: (1) whom participants interacted with (alone vs. an unknown peer vs. a friend), (2) the nature of interpersonal engagement (engaging in a face-to-face, in-person conversation vs. being reminded of an upcoming interaction), and (3) the timing of the social presence relative to the experience of social exclusion (before vs. during vs. after). Our goal was to investigate how specific features of social presence serve to either protect individuals from the negative effects of exclusion (i.e., buffering), retroactively help in the restoration of affect and fundamental needs following the effects of exclusion (i.e., recovery), or both. We utilized integrative data analysis (IDA) because the designs of all five of the studies were highly similar and examined the same dependent variable.¹

IDA is the statistical analysis of pooled data and is ideal when the original individual data from multiple studies are available. IDA approaches differ from and offer advantages over meta-analytic techniques, which also have the goal of building a cumulative knowledge base. Specifically, IDA approaches pool the original raw data, allowing researchers to examine what works, for whom, and in which contexts. This is in contrast to meta-analysis, which provide a synthesis of summary statistics drawn from multiple studies (see Curran & Hussong (2009) and Hussong, Curran, & Bauer (2013) for discussions) [90, 91]. IDA has been applied in medical research to assess treatment efficacy such as the comparison between medications and cognitive-behavioral therapy for severe depression [92]. Similarly, in the field of personality research, IDA has been utilized to explore the links between personality and life outcomes [93].

Our approach draws parallels with research in medicine. For example, in the study of skin wound healing, investigators create a small abrasion and implement interventions to observe how the body responds [94, 95]. Similarly, we use and adopt an experimental social psychological lens, manipulating social disconnection interventions before or after exposing participants to a form of social exclusion that effectively induces feelings of social disconnection.

To our knowledge, no study has systematically investigated the extent to which various forms and timing of social presence of friends influences the affective and cognitive reactions triggered by social exclusion.

Six hundred sixty-four participants (M_age = 20.29, SD_age = 2.29) were recruited to participate from Cornell University. Participants self-identified as 64.5% Female; 55.4% White, 30.8% Asian, 7.7% Black, 4.2% Latino, 4.2% Other.² See Table 2 for a breakdown of demographics by Study. Participants who completed the study with a friend reported a mean friendship length of 19.76 months (SD = 19.51; Range = 1–150) with that specific friend. Participants were compensated for their involvement, and could choose to receive either course credit or a payment of $5.

Below, we provide a broad overview of the social exclusion paradigm, the recruitment strategy, and then the key dependent variable that we used to examine affective and cognitive consequences. Then, we provide a description of each of the five studies, detailing the exact nature of each intervention and unique and shared experimental features (e.g., social presence of a close friend), in Table 3.

We aimed to replicate past work showing that one-person exclusion alone leads to negative affective and cognitive state [93, 94]. In our alone comparison conditions (top two bars in Fig. 1), we successfully replicate past work, and demonstrate the negative affective and cognitive consequences of experiencing exclusion alone as compared to inclusion alone. Experiencing social exclusion alone led to negative affective and cognitive consequences, as compared to inclusion alone, t(83) = − 4.461, p < .001, d = .97.

To examine the effect of our various social disconnection interventions in dampening the negative affective and cognitive consequences to exclusion, we first examined the effect of increasing the amount of social presence prior to experiencing social exclusion (i.e., buffering). Overall, three key findings emerge about the ingredients that give rise to effective buffering interventions.

First, the strongest buffering interventions consisted of participants having actual interactions with their friend before experiencing social exclusion. As shown in Fig. 1 (bars listed under “buffering”), having a conversation with a friend before experiencing social exclusion reduced the negative affective and cognitive consequences of social exclusion, as compared to experiencing exclusion alone, t(217.17) = 4.001, p < .001. Having a conversation with a friend before exclusion was not statistically significantly different from experiencing inclusion alone, t(184.52) = 1.652, p = .100. Moreover, the quality of the relationship predicted the affective and cognitive state for those who had a conversation with a friend before the experience of exclusion, r(33) = 0.409, p = .015.

Critically, the results show that having interactions with a friend, and not just simply being in the presence of a friend, might be an important component of the buffering intervention. Specifically, having a conversation with a friend before experiencing exclusion significantly reduced its negative affective and cognitive consequence as compared to simply having the mere presence of a friend, t(531.48) = 2.367, p = .018. Furthermore, having the mere presence of a friend was not significantly different than experiencing exclusion alone, t(387.21) = − 0.235, p = .814. The effect of mere presence of a friend was not predicted by relationship quality, r(132) = − 0.096, p = .272.

Second, our results highlight that conversations with friends may be a particularly effective component of buffering interventions. Having a conversation with an unknown peer did not offer conclusive evidence for buffering the affective and cognitive consequences of social exclusion as compared to experiencing exclusion alone. Our statistical model revealed a statistically significant effect, t(368.04) = 2.338, p = .020. However, when you look at the raw means (see Fig. 1), we see that the confidence intervals for having a conversation with an unknown peer and experiencing exclusion alone overlap, diminishing our confidence in the observed effect.

Third, and interestingly, having a friend in the same game while experiencing social exclusion did not significantly buffer against the negative affective and cognitive consequences of social exclusion as compared to experiencing exclusion alone, t(343.92) = 0.868, p = .386. The effect of friend in the same game was not predicted by relationship quality, r(51) = 0.208, p = .136. We discuss possible explanations for this boundary effect to buffering interventions in the Discussion.

Overall, the findings examining buffering interventions demonstrate that social interactions with friends and not just the mere presence are particularly important in ameliorating the negative consequences of social exclusion. They highlight the importance of dyadic conversations with friends in conferring buffering benefits, and that these benefits might be most pronounced when the conversations are with established social connections, especially high-quality relationships.

To examine how the social disconnection interventions may aid in the recovery from the negative consequences of social exclusion, we next examined the effect of social presence following the experience of social exclusion in Cyberball.

As shown in Fig. 1 (bars listed under “Recovery”), both recovery interventions significantly reduced the negative affective and cognitive consequence of experiencing social exclusion. Having a conversation with a friend after the experience of social exclusion reduced the negative affective and cognitive consequences of social exclusion, as compared to experiencing exclusion alone, t(244.00) = 7.359, p < .001. Having a conversation with a friend after exclusion was statistically indistinguishable from experiencing inclusion alone, t(202.18) = 1.548, p = .123. Moreover, relationship quality with the friend was not associated with the affective and cognitive state following a conversation with a friend after the experience of social exclusion, r(32) = 0.162, p = .361.

Moreover, our results demonstrate that even a symbolic representation (i.e., without direct physical engagement) of a future friend interaction facilitates the recovery following social exclusion. We find even a simple reminder of an upcoming interaction with a friend after exclusion significantly reduced the negative affective and cognitive consequence of experiencing social exclusion, as compared to experiencing exclusion alone, t(113.42) = 3.663, p < .001. Once again, the quality of the relationship was not associated with the affective and cognitive state following a reminder of an upcoming interaction with a friend after the experience of social exclusion, r(42) = 0.076, p = .626.

Why does a simple reminder of an upcoming interaction with a friend facilitate recovery? Could this effect be the result of simply reminding participants of their friend? It is noteworthy that in the presence of a friend condition, the friend was present throughout the entire experimental session, so participants were going to ostensibly reconnect with the friend at the end of the study. But the mere presence of a friend did not buffer against social exclusion, and both actual and symbolic interaction interventions significantly reduced negative affect compared to the presence of a friend condition. Participants who received a reminder of an upcoming interaction with a friend after inclusion reported higher affective and cognitive state than those who simply experienced inclusion alone in the game, t(226.25) = 3.300, p = .001.

Overall, these findings demonstrate that social interactions (real or symbolic) are particularly important in promoting recovery from negative consequences of social exclusion. This was the case whether there was an actual conversation, or the promise of an interaction, which demonstrates the power of symbolic reminders of social bonds in facilitating recovery from social exclusion. Importantly, the quality of the friendship does not seem to influence the magnitude of these effects. We discuss implications in the Discussion.

In the present work, we varied whom participants interacted with (a friend vs. an unknown peer vs. alone), the nature of interpersonal engagement (having a face-to-face conversation vs. being reminded of an upcoming interaction vs. merely being in the presence of another), the timing of the social disconnection intervention relative to the experience of social exclusion (before vs. during vs. after). We find that not all social disconnection interventions are equally effective.

First, in our studies, whom the person interacted with emerged as a critical factor. As compared to being alone, having a conversation with a friend both before and after experiencing exclusion, regardless of the quality of the relationship, mitigated the affective and cognitive consequences of social exclusion. Notably, the benefit of interactions before experiencing social exclusion (i.e., buffering interventions) was related to relationship quality with said friend; the better the relationship, the better the outcome. However, the benefit of interactions after experiencing social exclusion was not related to the quality of the relationship. We speculate that when the conversation occurs after a threat, the interaction provides relief to a concrete and immediate social threat (i.e., the experience of social exclusion), and does so regardless of relationship quality. In contrast, when the conversation occurs before a threat, individuals may be uncertain about the nature and severity of potential threats (if any). Naturally, when relationship quality is lower, the confidence of the quality of future support remains in question. The perceived and received benefits of social support may be inconsistent or uncertain due to past experiences or ambiguity. This ambiguity can extend to the kind of support that might be needed in case of a threat, making it difficult to anticipate the level of support that will be available or necessary.

Were these benefits due to simply having a conversation with anyone? Our data suggest that the answer to this question is no. Specifically, our findings did not provide conclusive support that interacting with a previously unknown peer was effective in mitigating the affective and cognitive consequences of social exclusion. We note that the conversations between the friends and the unknown peers differed. Researchers have found that conversations between strangers are mostly spent finding common ground, whereas conversations between friends are mostly spent exploring new ground (see Speer, Mwilambwe-Tshilobo, Tsoi, Burns, Flak, & Tamir (2023) for a discussion) [113]. Thus, the conversations between the unknown peers were structured with the goal of facilitating connection, and used “getting to know you” type questions (e.g., “What is your favorite class at Cornell?”). In contrast, the conversations between friends were structured with the goal of prompting exploration, with the goal of deepening the bond, and incorporated questions from Aron’s “Fast Friends” procedure (e.g., “What would constitute a perfect day for you?”). While ecologically valid, this has implications for the interpretation of results. One possibility is that if the conversation methods were the same, a conversation with an unknown peer before might have buffered against the negative consequences of social exclusion. However, a more likely possibility is that, even if the conversation methods were the same, a conversation with an unknown peer would still not buffer. We suspect this is more likely due to uncertainty regarding the unknown peer’s capability and availability to provide future social support. The unknown peer is still a relative stranger who may never be seen again (for comparison, friends had a mean relationship length of nearly 20 months).

Second, the nature of interpersonal engagement of the social disconnection intervention also mattered. Having an interaction with a friend before (buffering) or after (recovery) mitigated the negative affective and cognitive consequences of social exclusion. Remarkably, even a simple reminder of an upcoming interaction with a friend after the experience of social exclusion was also powerful. Such effects may reflect the immediate consequences of activating the symbolic representation of the friend, such as positive affect, but may also reflect the anticipation of the future interaction. Critically, the mere presence of a friend or an unknown peer in the same room did not mitigate the negative consequences. This finding echoes previous research showing that merely being around others does not alleviate feelings of social disconnection [82].

Nonetheless, our present results extend previous research on how actual and imagined interactions regulate emotions and cognitions by demonstrating it in the context of social exclusion and underscore the importance of interactions (both real and imagined) in mitigating the negative affective and cognitive consequences of social disconnection. Future research should address the absence of recovery interventions that involved unknown peers. It remains to be seen whether conversations with unknown peers and reminders of an upcoming interaction with an unknown peer after the experience of social exclusion would be effective. Some research suggests that momentary vulnerabilities of social disconnection motivate individuals to form new bonds in the service of social connection [114]. Yet at the same time, there is mixed evidence on how short-term, momentary vulnerabilities of social disconnection either assist or hinder social processing. For example, some research suggests that after experiencing social exclusion, individuals are better at distinguishing between Duchenne (“real”) and non-Duchenne (“fake”) smiles, and that the experience increases the cone of direct gaze [115, 116]. Yet, other research has found that individuals are less likely to categorize faces as “happy” faces and less accurate in perceiving gaze direction after the experience of social exclusion [117, 118].

Third, timing mattered—a proposition consistent with other work on emotion regulation [119‐121]. Although social disconnection interventions both before and following social exclusion aided in dampening the negative affective and cognitive consequences, recovery interventions were generally more robust than buffering interventions. More specifically, the mean affective and cognitive state of participants in the two recovery interventions were equal to and higher than inclusion alone. Earlier, we speculated that, for buffering interventions, the confidence of the quality of future support might not change when relationship quality is lower, whereas for recovery interventions, there is no question about receiving support. These findings are consistent with past work that has generally found more robust effects for recovery as compared to weak and inconsistent effects for buffering [81, 83‐85]. Given our findings, it will be essential to investigate the nuanced role of timing in social disconnection interventions, such as examining the optimal points for inducing buffering and recovery effects. For example, future research could explore questions such as how close to the social exclusion event does the timing of interventions need to be.

The efficacy of interventions can be understood within an attachment framework, where a primary function of attachment figures is to restore affective and cognitive equilibrium after a threat has been encountered [53, 54, 62‐67]. The realization of these benefits can be more challenging when individuals lack awareness or the ability to foresee future threats. These findings highlight the significance of implementing well-timed interventions in addressing the affective and cognitive toll of loneliness and social isolation.

Addressing public health concerns arising from social disconnection requires a consideration of the structure (how relationships are organized, the frequency of interactions), function (what roles relationships and interactions serve, such as social support), and quality (the level of satisfaction and fulfillment, which can vary across situations and relationships) [122]. In particular, a strong structural foundation, such as the proximity, contact, and presence of others, has downstream consequences on the function and quality of social relationships [20].

Building on previous work that suggests that social disconnection interventions should focus on strategies that increase opportunities for social contact and on cognitions that lead to social avoidance and withdrawal [29, 30], we manipulated factors that affect structure, function, and quality: (1) whom participants interacted with (alone vs. an unknown peer vs. a friend), influencing structure and quality, (2) the nature of interpersonal engagement (a face-to-face conversation vs. a reminder of an upcoming interaction vs. mere presence), influencing function and quality, and (3) the timing (before vs. during vs. after), influencing structure, function, and quality. Our research highlights the potential benefits of tailoring social disconnection interventions to leverage existing friendships and underscores the significance of fostering, encouraging, and nurturing relationships. Furthermore, our findings hold particular relevance for various organizations, including schools, colleges, and medical and occupational institutions. Implementing frequent, positive, and structured interactions in these settings may effectively promote social connection and reduce individuals’ vulnerability to social slights.

Our findings suggest social disconnection interventions that involve direct or symbolic interactions with friends is crucial for mitigating the detrimental effects of social exclusion. Specifically, having individuals engage in face-to-face conversations with friends before or after the threat of social exclusion, and activating symbolic representations of friends through reminders of upcoming interactions after, mitigate the negative affective and cognitive consequences of social exclusion. Integrating these interventions into public health interventions, practices, and policies can provide a proactive approach to primary prevention by targeting momentary vulnerabilities of social disconnection prior to the onset of loneliness and social isolation.

An advantage of the social disconnection interventions described here is that they have a low barrier to entry and are focused on momentary vulnerabilities of social disconnection. Part of the low barrier to entry is that the interventions do not require direct visible support that may be more difficult to provide. Direct visible support (e.g., “My advice to you is to...”) as compared to indirect invisible support (e.g., cooking dinner, hanging out) often has unintended consequences [123]. Bolger et al. have theorized that invisible or indirect social support, which recipients may not notice, or even interpret as support, leads to better outcomes. This is partly because direct social support is often associated with feelings of indebtedness or inequity [124].

Our current work primarily centers on young people, a group who is “at a high risk of feelings of loneliness but are least likely to take action” [11]. However, it will be equally important to examine how the effects of these social disconnection interventions vary across different groups of people who face loneliness and social isolation, including older populations with smaller but denser social networks, individuals from diverse racial and ethnic backgrounds, and the intersection of these factors [125‐128]. These considerations will be essential for a comprehensive understanding of social disconnection interventions and their potential impact on public health.