According to social comparison theory, people have an innate tendency to compare themselves to others (Festinger, 1954). In doing so, they derive at various outcomes, including an evaluation of themselves (Festinger, 1954), regulation of emotions and well-being (Taylor and Brown, 1988), and aspirations to improve their skills or abilities (Wood, 1989). Upward social comparison occurs when people compare themselves to others who are better than them; although upward social comparison motivates people to become more like their comparison target, it also causes dissatisfaction and lowers self-esteem (Emmons and Diener, 1985; Taylor and Lobel, 1989; Wheeler and Miyake, 1992). In contrast, downward social comparison occurs when people compare themselves to others who are worse-off than them and such comparison often leads to more positive self-evaluation and enhanced mood (Wills, 1981; Pyszczynski et al., 1985).

People are highly selective in what they present on social media (Mendelson and Papacharissi, 2010). They carefully curate the things they upload on social media that portrays the “perfect” aspects of their lives, such as flattering photographs, expensive goods, and personal successes (Siibak, 2009; Gonzales and Hancock, 2011; Blease, 2015). People also tend to present themselves positively on social media (Vogel and Rose, 2016). They typically upload content that best represents their ideal self (Rosenberg and Egbert, 2011), or a version of themselves that they believe will be best liked by others (Madden and Smith, 2010). As such, what results is a proliferation of profiles on social media suggesting that a large number of people are doing well and lead happy and perfect lives. On top of these, the “like” button provides further information about a person’s popularity and social capital (Kim and Lee, 2011; Vitak and Ellison, 2013). Collectively, these serve as social information that people take in and compare themselves against (Fox and Moreland, 2015).

While people engage in both upward and downward social comparisons when they use social media, existing evidence suggest that upward social comparisons are engaged more frequently than downward social comparisons. Through experiential sampling, where participants were monitored across 2 weeks, Kross et al. (2013) found that Facebook use was associated with declines in subjective well-being over time. Blease (2015) also proposed that depression, resulting from Facebook use, is likely to be brought about by the conspicuous amount of positive impressions people are exposed to from their Facebook friends, which opens up opportunities for comparison and escalates risk for negative appraisals. These studies suggest that the use of social media triggers upward social comparisons, or “harmful” social comparisons (Kross et al., 2013), which underlies the declines in subjective well-being and increased likelihood for depression.

With the constant exposure to information about how perfect the lives of others are, people consistently perceive that others are better off than oneself (Chou and Edge, 2012; de Vries and Kühne, 2015; Appel et al., 2016). Consequently, the constant upward social comparison that people engage in while using social media results in lowered self-appraisals or self-esteem (Vogel et al., 2014). Existing meta-analyses show support for a negative relationship between social media use and self-esteem, evidencing that increased social media use is associated with decreased self-esteem (Liu and Baumeister, 2016; Huang, 2017; Saiphoo et al., 2020). However, the effect sizes reported for the relationship between social media use and self-esteem are often very small [r = −0.09 by Liu and Baumeister (2016); r = −0.04 by Huang (2017); and r = −0.08 by Saiphoo et al. (2020)], suggesting the presence of moderators that account for the different relationships between these variables.

Higher effect sizes for the negative association between social media use and self-esteem were found for studies that assessed problematic social media use (i.e., addictive social media use) than those that measured the frequency of general social media use (Saiphoo et al., 2020). Studies that measured social and collective self-esteem, instead of global self-esteem, reported a positive relationship between social media use and social self-esteem (Valkenburg et al., 2017; Saiphoo et al., 2020). A recent study by Valkenburg et al. (2021), which employed a 3-week experience sampling design, showed that people differed in their susceptibility toward the content on social media (e.g., not receiving many likes), which contributes to the small effect size found between social media use and self-esteem People who were less susceptible to social media content reported smaller fluctuations in their self-esteem; in contrast, people who were more susceptible to social media were likely to experience bigger fluctuations in self-esteem that would have canceled each other out across time (Valkenburg et al., 2021).

In this paper, beyond measurement artifacts and individual differences, we turn our focus to the features of social media and propose that the amount of social information uniquely afforded by social media plays a significant role in determining the relationship between social media use and self-esteem. Employing an evolutionary mismatch perspective, we argue that novel features of social media—in particular, large social network sizes—influence the social comparison process such that greater social media use may not necessarily result in self-esteem loss.

The evolutionary mismatch perspective posits that our evolved psychological mechanisms, which are designed to be adaptive in ancestral environments, are not well-suited to handle novel elements within the modern context (see Li et al., 2018, 2020). A classic example of the evolutionary mismatch concerns our evolved preference for sweet and fatty foods. As sweet and fatty foods were higher in calories, the preference for these foods were adaptive in the ancestral environment where such caloric-rich food were scarce. However, in modern environments where there is an abundance of over-processed food and food that contain large amounts of manufactured sugar, this food preference leads people to overconsume sweet and fatty foods, more than what our physiological systems are designed to handle. Because our mechanisms did not evolve to process the unnaturally high levels of fats and sugar found in modern contexts, health conditions such as obesity and diabetes ensues (Gluckman and Hanson, 2006).

Similarly, social media is a modern feature that contains several evolutionarily novel elements that can potentially influence the functioning of our evolved psychological mechanisms. Of particular focus in this paper is its affordance for an evolutionarily novel large social network size. Most popular social media platforms allow registered members to create personal profiles and interact with other users. Registered members can seek other users out via a search engine, browse their profiles, and befriend them (Blease, 2015). This ease of befriending others contributes to the large “friend” networks people have on social media. The average adult Facebook user has 338 “friends”; beyond people who they actually are friends with in real life, this social network also comprises of people who are not close friends and people they have never met (Osman, 2021). However, humans have evolved to handle only a limited number of relationships (Tooby and Cosmides, 1996). Specifically, humans have evolved a neocortex size to maintain a network size of 150 individuals (Dunbar, 1998). This introduces a mismatch situation, which carries important implications for the psychological mechanisms governing social comparison.

As people are exposed to the “perfect” lives of others on social media, the evolved tendency to take in the social information and compare themselves to others results in self-esteem loss. Typically, the more one uses social media, the more social comparisons are engaged, and the more one feels worse about themselves. Moreover, on the surface, we might expect this to be even more true for networks with a greater vs. lesser number of people. Just as how our preference for sweet and fatty foods is hijacked by the modern environment, the social comparison process is hijacked by the large amount of social information introduced by large social network sizes, such that people are drawn into more social comparisons within larger networks. As such, on one hand, larger networks increase the occurrence for comparative social evaluation, which escalates the likelihood of one feeling more depressed and greater loss of self-esteem (e.g., Blease, 2015). Yet, a key evolutionary principal suggests otherwise. That is, given the natural limitations on humans’ ability to process network sizes, when social networks are beyond the size of 150 individuals, the enormous amount of available social information may be increasingly difficult for psychological mechanisms underlying social comparison to process. As such, on the other hand, for evolutionarily novel social network sizes that exceed 150 individuals, greater use of social media may not lead to greater loss of self-esteem.

We began our research with the aim of understanding the negative but weak relationship between social media use and self-esteem. Using principles from social comparison theory and the evolutionary mismatch theory, we explore how social network size influences the relationship between social media use and self-esteem. Specifically, we predict that greater social media use is likely to be associated with lower levels of self-esteem when one’s social network size is within 150 individuals, the number of relationships we have evolved to handle. When social network sizes are larger than 150 individuals, we test the competing predictions: on one hand, with more targets for social comparison, greater use of social media is likely to result in greater self-esteem loss; on the other hand, the huge, evolutionarily novel amount of social information makes it difficult for psychological mechanisms underlying social comparison to process such that greater use of social media is not associated with low self-esteem.

A total of 123 participants were recruited through an Australian university’s subject pool system (106 females, M_age = 22.78, SD_age = 7.92). All participants indicated that they engage in at least one social media platform (M = 2.76, SD = 0.82), with Facebook (N = 111) and Instagram (N = 104) being the most used social media platforms. Participants reported having a mean of 1,186 friends (SD = 1,601) across all social media platforms that they engaged in.¹

Upon providing informed consent, participants completed a series of questionnaires that measured their social media usage and self-esteem. Participants were also required to provide the number of friends they have across all the social media platforms they use. Finally, participants provided demographic details before completing the study.

Our work is far from conclusive and poses questions for future work. Within the evolutionary framework, general intelligence may have evolved to solve evolutionarily novel problems (Kanazawa, 2010). This implies that the evolutionary constraints on the human brain proposed by the mismatch theory may be less strong among more intelligent than less intelligent individuals as they are more able to comprehend and deal with evolutionarily novel entities and situations (e.g., Kanazawa and Li, 2015). Given the evolutionarily novel nature of social media and large social networks, it is likely that intelligence may play an important role in influencing the effects of large social network sizes. For instance, more intelligent people may be more likely to be able to process the inputs of group sizes larger than the evolutionarily familiar limit of 150, and hence, engage in social comparisons as they would for network sizes of 150 individuals. However, it is also possible that they may be more able to perceive that these social networks consist of people who are not real (i.e., virtual friends) and not have their self-esteem affected in the first place. As such, future directions should examine the effects of intelligence on social media use.

Similar to existing studies that had examined the effect of social media use on subjective well-being (Kross et al., 2013) and depression (Blease, 2015), we proposed that social comparisons underlie the relationship between social media use and self-esteem, and that upward social comparisons tend to be made when people are engaged in social media, which would account the negative relationship observed between social media use and self-esteem (Liu and Baumeister, 2016; Huang, 2017; Saiphoo et al., 2020). While our findings show support for this negative relationship, social comparisons were not directly measured in this paper—we are only able to infer the social comparisons that could have taken place based on self-esteem, which would not accurately elucidate the social comparison process. Moreover, the type of social comparison influences self-esteem differently; while upward social comparisons result in low self-esteem, downward social comparison can boost self-esteem (Vogel et al., 2014). Future studies should explicitly assess the type and frequency of social comparisons people make when they engage in social media. Examining the social comparison process would not only provide evidence for the type of social comparisons people make when using social media, it would also shed light on how exactly large social networks affect the functioning of psychological mechanism governing social comparisons.

The analytic approach employed in this paper allows us to draw inferences about the association between naturally occurring levels of social media use and self-esteem, but it is not conclusive of the definitive causal relations between them. As such, an alternative interpretation to the findings in this paper is that self-esteem is also likely to influence social media use. Social media typically offers users opportunities for self-disclosure, feedback validation, and relationship development (Stern, 2004; Boyd, 2008; Reich et al., 2012). With these opportunities, social media was deemed to be particularly useful for individuals with lower levels of self-esteem who face difficulties in social situations in real life. Social media allows them to compensate their need for social interactions by allowing them to expand their social capital (Forest and Wood, 2012; Błachnio et al., 2013, 2016) and social snack through photos, representational reminders of social connections, and parasocial relationships (see Gardner et al., 2005). Thus, experiments manipulating social media use, and the number of friends one has, would be required to establish causal relationship between social media use and self-esteem. That said, these experiments would have to be carefully set up as participants should still feel socially connected after the possible manipulations (e.g., manipulated social profiles) in order to elicit meaningful social comparisons.

Having used a university’s psychology subject pool system, we recognize some of the shortcomings that accompany this sample. Participants in our participants were predominantly women who may be more influenced by the cues on social media than men. Compared to men, they tend to internalize media-promoted ideals to a higher degree (Knauss et al., 2007) and are more oriented to the activities of others when using social media (Steinsbekk et al., 2021). Studies also show that women were more likely to have negative emotional responses and experience depressive symptoms than men when using social media (Fleuriet et al., 2014; Kelly et al., 2018). As such, the effect of social media use on self-esteem is likely to be more pronounced in the current sample than if it was from a more gender-balanced sample. To this end, future studies may want to consider including non-binary gender measures to derive at more accurate conclusions for the effects of social media use (Cameron and Stinson, 2019). Additionally, participants were categorized into groups according to the quartiles of the number of friends they reported they had across all their social media platforms. While we observed significant findings for the interaction term and simple slopes, the number of participants in each group is considered small (roughly 30 per group). Hence, a larger sample size in future studies would provide greater confidence to the findings of this paper.

Furthermore, the amount of social information one is exposed to on social media is inferred from the number of friends one has in this study, which may not be a nuanced enough measure. The exposure to social information could be different depending on specific behaviors and the types of activities people engage in on social media. For instance, they may spend more time curating their profiles and working on their own posts than reading and interacting with those of others, and this implies that they would be less exposed to social information regardless of the number of friends they have on social media. As such, it would be beneficial for future studies to breakdown social information exposure through the different ways people spend their time on social media.