Our program of research over the past two decades explores the dynamics of shared, social identities (Van Bavel & Packer, 2021). Our work is grounded in the notion that our sense of self is derived from our social environment. We flexibly form social identities with groups, from partisan to university to national affiliations, and these identities have a profound influence on how we think and act in the world—providing a lens through which we interpret the social world. The Social Identity & Morality Lab examines what happens to people psychologically when they define themselves in terms of group memberships, from our most rapid evaluations and verbal expressions to belief updating and behavior.
Drawing upon research in social psychology and cognitive neuroscience, as well as the social sciences more generally, we have been focused on a set of core principles about social identity. This has led our work to evolve from using implicit measures of social judgments and functional neuroimaging of face perception to economic cooperative games and linguistic analysis of social media posts, and, more recently, to studying the movement of millions of people and using international surveys of public health behavior during the pandemic. By triangulating across methods, we have developed a deeper understanding of how identity shapes human cognition and action across a wide range of domains.
This research has converged on what we believe is a deep truth about human nature: groups—even the most trivial of groups—are central to how people define themselves (Tajfel, 1970; see also Van Bavel & Packer, 2021; Van Bavel & Cunningham, 2010). Humans evolved in small groups, which has had a large impact, not only on how we behave, but what we believe and perceive. It is a human universal, observed in every culture (Brown, 1991), that we form these flexible coalitions to achieve a sense of social connection and pursue various goals. We have studied these processes in both international and developmental contexts, and find that our preference for group-based judgements emerges early in young children (Leshin et al., under revision; Yudkin, Van Bavel, & Rhodes, 2022) and persists through adulthood around the globe (Van Bavel et al., 2022). This fundamental feature of our psychology affords us the potential for a sense of solidarity with strangers that is unique among primates and can spark prosocial behavior and a shared sense of social reality.
Different social identities become salient at different times—and when a particular social identity is active, it can have profound effects on our thoughts, feelings, and behaviors (Turner, 1987; Van Bavel & Packer, 2021). When a specific social identity is salient, people experience the world through the lens it provides, embrace a shared reality with fellow group members, and engage in prosocial actions with fellow group members, including making economic sacrifices. But there is a flipside to this psychological tendency: cooperation and shared cognition are often bounded. Social identities can make people attend to and help members of their own groups, but it can also make them want to harm—or at least avoid helping—people who belong to other groups (Cikara, Van Bavel, Bruneau, & Saxe, 2014; Yudkin et al., 2016).
Taken together, my program of research highlights the truly social nature of cognition. It also reveals the broad significance of social identity for understanding prejudice (Van Bavel et al., 2008; 2009), the structure of beliefs (Van Bavel & Pereira, 2018; 2021), leadership (Reicher et al., 2019), social media (Brady et al., 2017; 2020; Rathje et al., 2021), the spread of misinformation (Van Bavel et al., 2020), and public health (Gollwitzer et al., 2021; Van Bavel et al., 2020). This work has also elucidated potential strategies for mitigating prejudice (Nødtvedt, Sjåstad, Skard, Thorbjørnsen, & Van Bavel, 2021) and fostering cooperation (Hackel, Wills, & Van Bavel, 2021). Jay recently summarized this program of research in a new book The Power of Us (Van Bavel & Packer, 2021), but offer a brief summary below with an emphasis on more recent research projects from the lab.
Social Identity Emerges Easily and Shapes Cognition
Perhaps one of the most important findings in the social sciences is that people favor ingroup members at the flip of a coin. Known as “minimal groups,” these arbitrary social identities provide an elegant paradigm for understanding the basic nature of identity. We have found extensive evidence that merely assigning people to a group has surprising, and important, consequences for human cognition. For instance, joining a minimal group elicits an automatic preference for ingroup members on implicit measures of evaluation (Van Bavel & Cunningham, 2009), guides rapid attentional orienting based on group membership (Brosch & Van Bavel, 2012; Park et al., 2016), and drives facial identity encoding (Van Bavel, Packer, & Van Bavel, 2011; Van Bavel & Cunningham, 2012; Van Bavel, Swencionis, O’Connor, & Cunningham, 2012). This suggests even the most trivial of identities plays a hand in shaping our basic judgments of the social world.
More striking perhaps is how our currently salient identity overrides our reactions to other social categories, like race. For instance, we have found that merely assigning people to a mixed-race group can override their initial, implicit racial biases to Black and White faces (Van Bavel & Cunningham, 2009; see Kaul, Ratner, & Van Bavel, 2014; Ratner, Kaul, & Van Bavel, 2013 for boundary conditions), shape their neural reactions in social cognitive systems (Guassi Moreira, Van Bavel, & Telzer, 2016; Van Bavel et al., 2008; 2011), and inspire richer mental state attributions about ingroup members (Cikara, Bruneau, Van Bavel, & Saxe, 2014; Hackel, Looser, & Van Bavel, 2014). This work led us to develop a theoretical framework for social cognition that centers the role of the active social identity of the perceiver, rather than features of the perceptual input (Xiao, Coppin, & Van Bavel, 2016a, 2016b; Van Bavel & Cunningham, 2011). It provides an overarching view of identity on cognition and argues that understanding the social self is key to understanding the mind.
Shared Reality is Driven by Social Identity
One of those most important features of groups is how they help us construct a shared sense of social reality. Whether in the context of cults, organizations, or nations, group members frequently develop a common understanding that is used to coordinate collective action (Van Bavel & Packer, 2021). Although minimal groups provide a powerful tool for conducting incisive experimental work on cognitive processes in the lab, they are a mere proxy for the richer psychological experience of real-world identities where a shared sense of reality emerges. For the past few years our work has explored how a wide variety of social identities shape our judgments of the physical world, such as our mental sense of distance towards outgroups (Xiao & Van Bavel, 2012; Xiao, Wohl & Van Bavel, 2016), our smell and taste of group symbols (Coppin et al., 2016; Hackel, Coppin, Wohl & Van Bavel, 2018), and neural synchrony with fellow group members (Dikker et al., 2017; Reinero Dikker, & Van Bavel, 2021). This work led us to propose a theoretical model linking identity to perceptual judgments and explaining how this relation might have important consequences for intergroup relations (e.g., foster discrimination towards outgroups who smell differently or seem too close for comfort; Xiao, Coppin & Van Bavel, 2016a, 2016b).
To better explain why people might be susceptible for fake news and other forms of (mis)information, we proposed the Identity-Based Model of Political Belief (Van Bavel & Pereira, 2018; see also Harris et al, 2021; Van Bavel., 2021; Sternisko, Cichocka, & Van Bavel, 2020; Van Bavel, Harris, Parnamets, Rathje, Doell, & Tucker, 2021). The model articulates at the psychological and neural level why people might be motivated to believe (mis)information that aligns with their partisan identities unless they have strong accuracy goals. When people place a great value on affirming a salient identity (e.g., because it confers a sense of belonging or social status), they will be more inclined to believe and spread (mis)information. In line with the model, we have found that both Democrats and Republicans are more likely to believe misinformation that flatters their ingroup (Periera, Harris, Van Bavel, 2021). We also found that this psychological process operates at the level of national identity across 54 nations (Sternisko, Cichocka, Cislak, & Van Bavel, 2021). However, partisan bias is mitigated when people are motivated to generate accurate beliefs (Rathje, Van Bavel, & van der Linden, under revision). Importantly, these motives are not evenly distributed between political parties—and our recent experiments and meta-analysis find that Republicans are less susceptible to accuracy nudges than Democrats (Pretus et al., under review; Rathje et al., 2022).
This work has also inspired us to explore how partisan content is spread online. Nearly 4 billion people are now on social media and the platform affords a powerful tool for finding like-minded people and expressing their social identity (Brady, Crockett, & Van Bavel, 2020; Van Bavel, Sternisko, Harris, Rathje, & Robertson, 2021). For example, our analysis of millions of messages from regular citizens as well as political leaders and media outlets on Twitter and Facebook has found that moral emotional expressions online—the language of moral outrage—is associated with greater dissemination rates (Brady, Wills, Jost, Tucker, & Van Bavel, 2017; Brady, Wills, Burkart, Jost, & Van Bavel, 2019; Rathje, Van Bavel, & van der Linden, 2021). These words appear to capture attention when people are scrolling through their newsfeed (Brady, Gantman, & Van Bavel, 2020). However, this type of rhetoric is also associated with polarization, leading people to share messages from partisan ingroup members and disengage from outgroup members (Brady et al., 2017; Brady et al., under revision). We have conducted several, massive pre-registered replications and a meta-analysis showing this effect is highly robust (Brady & Van Bavel, under review) as well as several lab experiments to provide causal evidence of this effect—and clarify the role of social identity in moral contagion (Brady & Van Bavel, invited revision).
Social Identity Impacts Action
Understanding the link between social identity, beliefs, and action has been the central goal of our lab over the past few years. As the pandemic began spreading in 2020, the President of the World Health Organization issued an expression of concern about the potential “Infodemic” of misinformation. To better understand the pandemic, and how scholars and policy makers might respond, Jay spearheaded a review of the literature with 40 other co-authors across the social and behavioral sciences (Van Bavel et al., 2020). Our paper integrated several lines of my research over the past decade, explaining how group identity, polarization, misinformation, and leadership are likely to play a role in shaping collective behavior during the pandemic. This paper became a landmark contribution, cited over 3000 times and being used by numerous governments and organizations (including the WHO) in considering behavioral policy responses during the pandemic.
Our paper also laid the theoretical groundwork for countless studies. For instance, we analyzed the role of political partisanship on risk perceptions (Sjåstad & Van Bavel, under revision) and spatial distancing behavior during the early stages of the pandemic in the US (Gollwitzer, Martel, Brady, Knowles, & Van Bavel, 2020). By analyzing the movement of over 15 million Americans (using publicly available and anonymized smartphone data) we found that people in Republican states and counties were far less likely to reduce their physical movement during the pandemic than people in Democratic states and counties. This pattern was almost fully explained by exposure to Fox News, which amplified comments from Donald Trump and downplayed the risks of COVID-19 throughout the pandemic. Moreover, partisanship was one of the single best predictors of spatial distancing and the partisan gap only grew larger as the pandemic spread throughout the US. We found that a lack of spatial distancing at the county level predicted subsequent COVID-19 infections and mortality, suggesting that this partisan response by Republicans might have led to thousands of deaths. We have new work underway finding an almost identical pattern of national vaccination rates, with Republican regions dramatically trailing more Democratic parts of the country and this too is linked to subsequent mortality in Republican areas (Parnamets et al, in prep).
To determine if national identity was potentially related to greater compliance with public health behavior, we launched a global study of over 50,000 people in 68 nations around the globe (Van Bavel et al., 2022). Along with over 240 collaborators, we found that national identification was robustly related to numerous public health behaviors (i.e., handwashing, support for public policy, avoiding crowds) in almost every country we studied. Moreover, this pattern of national identity was much stronger than a more malignant form of national narcissism and political ideology (which tended to be only weakly related). We later replicated this pattern in 42 additional countries and used Google movement data—finding that the link with national identity persisted with real aggregate measures of behavior. This underscores the potential function of social identity in a crisis and the need for effective leadership to mobilize collective action. One open question is whether this catastrophe could have been prevented under a more competent leader. Our prior work suggests that superior national leadership—like that demonstrated by New Zealand Prime Minister Jacinda Ardern—could have harnessed social identity in a more productive and healthy fashion (see Haslam, Reicher, & Van Bavel, 2019; Reicher, Van Bavel, & Haslam, 2020).
Mitigating Bias and Promoting Cooperation
One of the primary challenges facing most organizations—as well as society—is understanding how to mitigate intergroup bias and for more effective, fair, and harmonious intergroup relations. Since beginning down this line of work in 2004, we have conducted numerous studies designed to implement and evaluate different strategies for reducing bias. One of the most robust strategies for reducing implicit bias (see Van Bavel & Cunningham, 2008; Van Bavel et al., 2009) and gaps in intergroup empathy (Cikara et al., 2014) turns out to be fostering a more inclusive sense of identity—where people a sense of shared purpose with outgroup members. Another effective strategy is giving people the time and cognitive resources to regulate their biases (Yudkin, Van Bavel & Rhodes, 2020). However, these psychological strategies are often hard to sustain in the face of other situational pressures.
We have found that generating more inclusive and cooperative social norms or implementing institutional or platform design features provides more robust solutions to promoting cooperation. For instance, exposing people to groups of people with cooperative norms dramatically increases cooperation (Hackel et al., 2021)—by increasing the value of cooperation (see Parnamets, Shuster, Reinero, & Van Bavel, 2021). Likewise, sharing norms around accuracy can reduce the spread of misinformation among hyper partisans (Pretus & Van Bavel, in progress) and reducing the impact of political bias in the publication system (Reinero, Wills, Brady, Mende-Siedlecki, Crawford, & Van Bavel. 2020). We have also found that platform design features can reduce intergroup bias and promote economic transactions with racial minorities in the online sharing economy (e.g., increasing the willingness to rent an AirBnB apartment). For instance, when positive reputational information is prominently displayed on these platforms it can significantly reduce discrimination (Nødtvedt et al., 2021). This work offers several psychological informed strategies for promoting more prosocial behavior across multiple domains, from economic decisions to the spread of misinformation.
Our research suggests that collective concerns can tune cognition and shape action in surprising ways. Our multi-level approach has already shed light on several basic topics in psychology, including person perception, implicit evaluation, moral judgment, belief updating, political partisanship, and the determinants of health. The fact that collective concerns shift from one situation to another and shape our actions and behaviors represents a challenge to models throughout the field of psychology and underscores the social nature of human cognition (Packer & Van Bavel, 2014; Van Bavel & Packer, 2021). It is our hope this work on social identity will not only inform basic theories of human social cognition, but also offer important insights into issues ranging from social conflict to public health. We are currently expanding this work to study important topics, like climate change (Doell, Pärnamets, Harris, Hackel, & Van Bavel, 2021), and spearheading a series of global collaborations to study effective interventions.