Core claim. Social identities satisfy two opposing needs: a need for
assimilation (inclusion, belonging) and a need for differentiation
(uniqueness, distinctiveness). People identify most strongly with groups that
reconcile these needs: groups inclusive enough to provide belonging but
exclusive enough to provide distinctiveness.
Origins. Social psychology, Brewer (1991), reviewed and extended in
Leonardelli, Pickett, and Brewer (2010), consumer applications in Chan, Berger, and Boven (2012).
Level of analysis. Individual and group.
Key constructs. Need for assimilation, need for differentiation, inclusiveness
of the social category, optimal distinctiveness, need for uniqueness.
Boundary conditions. Distinctiveness is context-specific. The optimum shifts
with the frame of reference and with the relative strength of the two needs,
which vary across cultures, individuals, and recent experience.
Related theories. Social identity theory,
Identity signaling (Section 8.5), Snob effect (Section 29.5),
Conformity.
Origins and core logic
Brewer (1991) started from a puzzle in social identity theory: if group
membership provides self-esteem, why do people not simply identify with the
largest, most inclusive, highest-status groups available? Her answer was that
social identity is the reconciliation of two opposing needs. Personal identity
is the individuated self: the characteristics that differentiate one person
from others. Social identities are categorizations of the self into more
inclusive social units that depersonalize the self-concept, so that “I” becomes
“we”. Too much individuation produces isolation. Too much deindividuation means
that self-definition is lost. People therefore seek a balance between
assimilation and differentiation: a state of optimal distinctiveness.
Brewer also noted that while some group memberships are imposed, the identities
people embrace among the groups available to them are partly a matter of choice.
Brewer formalised the theory in four assumptions (Brewer 1991, 478):
A1. “Social identification will be strongest for social groups or categories
at that level of inclusiveness which resolves the conflict between needs for
differentiation of the self and assimilation with others.”
A2. “Optimal distinctiveness is independent of the evaluative implications of
group membership, although, other things being equal, individuals will prefer
positive group identities to negative identities.”
A3. “Distinctiveness of a given social identity is context-specific. It
depends on the frame of reference within which possible social identities are
defined at a particular time, which can range from participants in a specific
social gathering to the entire human race.”
A4. “The optimal level of category distinctiveness or inclusiveness is a
function of the relative strength (steepness) of the opposing drives for
assimilation and differentiation. For any individual, the relative strength of
the two needs is determined by cultural norms, individual socialization, and
recent experience.”
A2 is a notable departure from social identity theory: Brewer argued that people
may identify strongly with small, distinctive groups even when those groups have
modest status. Figure 22.2 depicts the opposing-processes logic.
As a category becomes more inclusive, the need for differentiation is
increasingly activated. As it becomes more exclusive, the need for assimilation
is increasingly activated. The optimum lies where the two forces balance, and,
consistent with A4, it shifts toward more inclusive groups when the need to belong
is temporarily heightened: for example, after an experience of rejection.
Evidence and extensions: consumption as a marker of membership
Consumer research has made optimal distinctiveness one of its most productive
theoretical lenses, because consumption gains symbolic meaning as a marker of group
membership. The relevant literatures had long emphasised one motive or the other.
On the assimilation side, people conform to others under informational and
normative influence (Deutsch and Gerard 1955), and they make choices similar to those of
their aspiration groups to signal desired identities (Englis and Solomon 1995, Berger and Heath 2007). On the differentiation side, people have a genuine need for
uniqueness: they can experience negative emotion when they feel too similar to
others, and they seek difference through possessions, opinions, and style
(Snyder and Fromkin 1980, Lynn and Snyder 2002). Consumers’ need for uniqueness can be measured as a
stable individual difference (Tian, Bearden, and Hunter 2001), but it is also driven by situational
factors. Research on social influence in networks likewise shows that influence
can produce both convergence and divergence among individuals, depending on the
structure of relationships and the identities at stake (W. A. Mason, Conrey, and Smith 2007).
Earlier work tended to assume that one motive prevails in any given decision.
Chan, Berger, and Boven (2012) showed instead that consumers can satisfy both needs in a single
choice by assimilating on one dimension and differentiating on another.
Consumers conform to their group on attributes that signal group identity, such as
the brand, while distinguishing themselves on attributes that signal individuality,
such as colour. The desire to communicate social identity leads consumers to
conform on dimensions strongly associated with their group, particularly in
identity-relevant categories such as clothing, while higher need for uniqueness
leads them to differentiate within the group by choosing less popular options
among those associated with it. The result is a self-concept that is, in the
title’s phrase, identifiable but not identical.
A related line of work examines differentiation between groups. People diverge
from members of dissimilar outgroups: when another social group adopts a taste,
people may abandon it to avoid signalling an undesired identity and to avoid the
costs of misidentification (Berger and Heath 2007, 2008). This mechanism is
developed in the section on identity signaling (Section 8.5).
The creativity literature offers an intriguing parallel. Novel ideas tend to
balance familiarity and novelty, recombining established elements in new ways,
and averaged, prototypical forms are often judged attractive. Those ideas are
developed elsewhere in the book, in the sections on the
Geneplore model (Section 31.2) and beauty in averageness (Section 29.10) (Table 22.4).
Table 22.4: How consumers reconcile assimilation and differentiation (based on Brewer, 1991, Chan, Berger, & Van Boven, 2012, Berger & Heath, 2007).
|
Consumer strategy
|
Need served
|
Choice dimension
|
Example
|
|
Adopt the group’s brand or category
|
Assimilation
|
Identity-signalling attribute
|
Buying the brand of trainers one’s team wears
|
|
Choose a less common variant within it
|
Differentiation
|
Uniqueness attribute (colour, style, edition)
|
Choosing an unusual colourway of that brand
|
|
Customise a mainstream product
|
Both
|
Personalisation
|
Monogrammed or configurable products
|
|
Join a small, distinctive community
|
Both (optimal group size)
|
Group membership
|
A niche brand community or enthusiasts’ club
|
|
Abandon a taste adopted by an outgroup
|
Differentiation (between groups)
|
Identity-relevant domain
|
Dropping a style once a dissimilar group adopts it
|
Identifiable but not identical: designing product lines. Optimal
distinctiveness suggests that brands with strong group associations should
offer variety along dimensions that do not dilute the identity signal. A
sportswear brand can keep its logo, silhouette, and brand name constant,
the assimilation cues, while offering many colours, limited editions, and
personalisation: the differentiation cues. The theory also warns that variety
on the identity-signalling dimension itself (for example, removing the logo) may
weaken the reason consumers chose the brand. For managers, the practical question
is to identify which attributes in their category carry group meaning and which
carry individual meaning, and to design assortment accordingly (Chan, Berger, and Boven 2012).
Treating uniqueness and conformity as opposite ends of one scale. The theory’s
central claim is that both needs operate simultaneously and are satisfied on
different dimensions or at different levels of inclusiveness. Measuring a single
“conformity versus uniqueness” preference will miss the pattern. Research should
distinguish the attributes or groups on which consumers assimilate from those on
which they differentiate.
22 Social Comparison, Identity, and Influence
A household receives its monthly utility statement. Alongside the usual numbers is a small chart: the household’s own electricity use, the average of a hundred similar homes nearby, and the use of the most efficient neighbours. Nothing about the price of electricity has changed, and the chart contains no technical advice, yet households that receive such reports reliably use less energy. A hotel guest reads a card saying that most guests who stayed in this room reused their towels, and hangs hers back on the rail. A teenager wants the same brand of trainers as her friends, but in a colour none of them owns. An engineer notices a serious flaw in a project plan, looks around a meeting room in which no one else seems worried, and says nothing.
Each scene turns on the same fact: people do not evaluate themselves, form preferences, or decide how to act in isolation. They look to others: to learn what is true, to know where they stand, to belong, and to stand out. This chapter assembles the core theories of social influence that operate at the level of the individual in a social context. We begin with social comparison theory, which explains how people evaluate themselves by reference to others, and social identity theory, which explains how group memberships become part of the self. Optimal distinctiveness theory reconciles the competing needs to belong and to be unique. We then examine conformity and social norms, the processes by which groups shape individual judgement and behaviour, before turning to the effects of the mere presence of others on performance (social facilitation and distraction-conflict theory) and, finally, to the dynamics by which perceived majorities silence minorities (the spiral of silence) (Table 22.1).
22.1 Social Comparison Theory
Core claim. People have a drive to evaluate their opinions and abilities. When objective standards are unavailable, they evaluate themselves by comparison with other people, preferably similar others.
Origins. Social psychology, Festinger (1954), later developments on downward comparison (Wills 1981), upward comparison (Collins 1996), and self-evaluation maintenance (Tesser 1988), reviewed in Suls, Martin, and Wheeler (2002) and Gerber, Wheeler, and Suls (2018).
Level of analysis. Individual (the comparer) in relation to comparison targets (individuals or groups).
Key constructs. Comparison target, similarity, upward versus downward comparison, assimilation versus contrast, relevance, attainability, closeness.
Boundary conditions. Comparison is most likely when objective standards are absent or ambiguous, when the dimension matters to the self, and when information about others is salient. Its effects depend on whether the comparer sees the target as similar and the target’s standing as attainable.
Related theories. Social norms, Equity theory (Section 19.3), Social identity theory, Goal-gradient hypothesis (Section 19.5).
22.1.1 Origins and core logic
Festinger (1954) began from a simple premise: people need to know whether their opinions are correct and how good their abilities are. Where physical reality provides an answer, whether a piece of furniture will fit through a doorway, people consult it. Where it does not, whether one is a good writer, or whether one’s political views are reasonable, they turn to social reality, comparing themselves with others. Festinger further proposed that people prefer to compare with similar others, because comparison with someone very different is uninformative, and that for abilities (in Western cultures at least) there is a unidirectional drive upward: a pressure to do somewhat better than others.
Later work distinguished the direction of comparison:
Lockwood and Kunda (1997) identified a key moderator: attainability. An outstanding role model inspires when the comparer believes similar success is achievable and demoralises when it seems out of reach. Tesser (1988) added relevance and closeness: being outperformed by a close friend on a dimension central to one’s self-definition is threatening (a comparison process), whereas being outperformed on an irrelevant dimension lets one bask in the friend’s success (a reflection process). A meta-analysis covering six decades of research supported Festinger’s core idea that people actively seek comparison information while showing that the choice of targets and the consequences of comparison depend strongly on context (Gerber, Wheeler, and Suls 2018) (Table 22.2).
22.1.2 Mechanism
Figure 22.1 summarises how a comparison produces its effects. The comparison target and its direction set up the comparison. Whether the comparer assimilates or contrasts depends on attainability, similarity, and the relevance of the dimension, and the outcome is a change in self-evaluation, affect, and motivation.
Figure 22.1: Mechanism of social comparison: a comparison target produces assimilation or contrast depending on attainability, similarity, and relevance, which in turn shape self-evaluation, affect, and motivation.
22.1.3 Applications in business
Organizations create comparison information constantly: sales leaderboards, customer-satisfaction rankings, peer benchmarks in performance reviews, “customers like you” panels, and fitness apps that show where a user ranks among friends. Social comparison theory predicts that such information motivates when it highlights attainable, relevant comparisons and backfires when it exposes large, apparently unbridgeable gaps.
Opower and the power of position. The utility-reporting firm Opower built its business on a comparison. Its home energy reports did not try to teach customers the science of energy efficiency. They told customers where they stood relative to their neighbours: whether they were better or worse at conserving energy than similar households nearby (Cuddy, Doherty, and Bos 2010). The insight was that position relative to others is a more powerful motivator than abstract facts. The reports also illustrate the theory’s boundary conditions: the comparison group was chosen to be similar (homes of comparable size in the same area), and the reports paired the comparison with an evaluative message to prevent already-efficient households from drifting toward the average. We return to the evidence on these reports in the section on social norms.
Assuming that comparison always motivates. Ranking employees or customers can demoralise as easily as it inspires. People far below the top may contrast rather than assimilate, disengage from the dimension, or question the fairness of the metric. Designs should specify the comparison group (similar others), make improvement appear attainable (e.g., showing the next rung rather than the top), and anticipate effects on those who are already ahead.
22.2 Social Identity Theory
Core claim. Part of a person’s self-concept derives from membership in social groups. People are motivated to achieve a positive social identity, which they do by favourably distinguishing their ingroups from relevant outgroups.
Origins. Social psychology, minimal group experiments (Tajfel et al. 1971), Tajfel and Turner (1979), extended by self-categorization theory (Turner et al. 1987), reviewed in Hornsey (2008).
Level of analysis. Individual and group. Extended to organizations (Ashforth and Mael 1989) and to consumer-brand relationships (Section 27.5) (C. B. Bhattacharya and Sen 2003).
Key constructs. Social categorization, social identification, social comparison between groups, positive distinctiveness, ingroup favouritism, permeability, stability, and legitimacy (Section @ref(legitimacy (Section 7.3))) of group boundaries, depersonalization, prototypes.
Boundary conditions. Effects depend on the salience of the category in the situation and on how strongly people identify with the group. When personal identity is salient, behaviour is guided more by individual than by group considerations.
Related theories. Optimal distinctiveness, Identity signaling (Section 8.5), Social comparison, Institutional theory (Section 7.2).
22.2.1 Origins and core logic
Social identity theory began with a surprising experimental result. In the minimal group paradigm, Tajfel et al. (1971) assigned participants to groups on the basis of trivial criteria, such as a preference between paintings by two abstract artists, and asked them to allocate rewards to anonymous ingroup and outgroup members. Even though the groups had no history, no interaction, and no conflict of interest, participants favoured their own group, and in some allocation choices preferred to maximise the difference between groups rather than the absolute reward to their own group. Mere categorization was enough to produce intergroup discrimination.
Tajfel and Turner (1979) explained this with a three-step process. Social categorization sorts people, including oneself, into groups. Social identification makes group membership part of the self-concept, so that the group’s fortunes become one’s own. Social comparison between groups then motivates people to seek positive distinctiveness: a favourable difference between ingroup and relevant outgroups. When a group’s identity is unsatisfactory, members choose among strategies whose viability depends on the social structure, as summarised in Table 22.3.
Self-categorization theory extended the account by specifying when people think of themselves as individuals or as group members (Turner et al. 1987). When a social category becomes salient, people undergo depersonalization: they see themselves less as unique individuals and more as interchangeable exemplars of the group prototype, and they adopt the group’s norms as their own. The theory explains why identity-based influence is strongest from ingroup members who are seen as prototypical, and why the same person can behave very differently as a “marketing professional”, a “parent”, or a “fan” depending on which identity the situation activates (Hornsey 2008).
22.2.2 Evidence and extensions
Ashforth and Mael (1989) imported social identity theory into organization studies, defining organizational identification as the perception of oneness with or belongingness to an organization. Identification helps explain loyalty, cooperation, and support for organizational goals, and also resistance to mergers that threaten a valued identity. In marketing, C. B. Bhattacharya and Sen (2003) proposed consumer-company identification: consumers identify with companies whose identities are distinctive, prestigious, and similar to their own, and such identification fosters loyalty, advocacy, and resilience to negative information.
Illustrative scenario: post-merger identity. Two engineering firms of similar size merge. Employees of each legacy firm continue to describe themselves by their old affiliation, sit with former colleagues, and attribute problems to “them”. Social identity theory interprets this as a predictable response to a threat to valued identities rather than as mere resistance to change. Remedies suggested by the theory include creating a superordinate identity that both groups can embrace, preserving valued elements of each legacy identity, ensuring that integration is seen as fair (legitimacy), and avoiding status signals, office location, leadership appointments, systems choices, that mark one side as the “acquired” group.
Measuring membership instead of identification. Being a member of a group is not the same as identifying with it. Many studies use membership (for example, working for a firm or owning a brand) as a proxy for social identity. The theory’s predictions apply to identification and to situations in which the identity is salient. Both should be measured or manipulated.
22.3 Optimal Distinctiveness Theory
Core claim. Social identities satisfy two opposing needs: a need for assimilation (inclusion, belonging) and a need for differentiation (uniqueness, distinctiveness). People identify most strongly with groups that reconcile these needs: groups inclusive enough to provide belonging but exclusive enough to provide distinctiveness.
Origins. Social psychology, Brewer (1991), reviewed and extended in Leonardelli, Pickett, and Brewer (2010), consumer applications in Chan, Berger, and Boven (2012).
Level of analysis. Individual and group.
Key constructs. Need for assimilation, need for differentiation, inclusiveness of the social category, optimal distinctiveness, need for uniqueness.
Boundary conditions. Distinctiveness is context-specific. The optimum shifts with the frame of reference and with the relative strength of the two needs, which vary across cultures, individuals, and recent experience.
Related theories. Social identity theory, Identity signaling (Section 8.5), Snob effect (Section 29.5), Conformity.
22.3.1 Origins and core logic
Brewer (1991) started from a puzzle in social identity theory: if group membership provides self-esteem, why do people not simply identify with the largest, most inclusive, highest-status groups available? Her answer was that social identity is the reconciliation of two opposing needs. Personal identity is the individuated self: the characteristics that differentiate one person from others. Social identities are categorizations of the self into more inclusive social units that depersonalize the self-concept, so that “I” becomes “we”. Too much individuation produces isolation. Too much deindividuation means that self-definition is lost. People therefore seek a balance between assimilation and differentiation: a state of optimal distinctiveness. Brewer also noted that while some group memberships are imposed, the identities people embrace among the groups available to them are partly a matter of choice.
Brewer formalised the theory in four assumptions (Brewer 1991, 478):
A2 is a notable departure from social identity theory: Brewer argued that people may identify strongly with small, distinctive groups even when those groups have modest status. Figure 22.2 depicts the opposing-processes logic. As a category becomes more inclusive, the need for differentiation is increasingly activated. As it becomes more exclusive, the need for assimilation is increasingly activated. The optimum lies where the two forces balance, and, consistent with A4, it shifts toward more inclusive groups when the need to belong is temporarily heightened: for example, after an experience of rejection.
Figure 22.2: Stylised optimal distinctiveness model (after Brewer, 1991). The need for differentiation rises as a group becomes more inclusive and the need for assimilation rises as it becomes more exclusive. Identification peaks where the drives balance. A stronger assimilation drive (dashed line) shifts the optimum toward more inclusive groups.
22.3.2 Evidence and extensions: consumption as a marker of membership
Consumer research has made optimal distinctiveness one of its most productive theoretical lenses, because consumption gains symbolic meaning as a marker of group membership. The relevant literatures had long emphasised one motive or the other. On the assimilation side, people conform to others under informational and normative influence (Deutsch and Gerard 1955), and they make choices similar to those of their aspiration groups to signal desired identities (Englis and Solomon 1995, Berger and Heath 2007). On the differentiation side, people have a genuine need for uniqueness: they can experience negative emotion when they feel too similar to others, and they seek difference through possessions, opinions, and style (Snyder and Fromkin 1980, Lynn and Snyder 2002). Consumers’ need for uniqueness can be measured as a stable individual difference (Tian, Bearden, and Hunter 2001), but it is also driven by situational factors. Research on social influence in networks likewise shows that influence can produce both convergence and divergence among individuals, depending on the structure of relationships and the identities at stake (W. A. Mason, Conrey, and Smith 2007).
Earlier work tended to assume that one motive prevails in any given decision. Chan, Berger, and Boven (2012) showed instead that consumers can satisfy both needs in a single choice by assimilating on one dimension and differentiating on another. Consumers conform to their group on attributes that signal group identity, such as the brand, while distinguishing themselves on attributes that signal individuality, such as colour. The desire to communicate social identity leads consumers to conform on dimensions strongly associated with their group, particularly in identity-relevant categories such as clothing, while higher need for uniqueness leads them to differentiate within the group by choosing less popular options among those associated with it. The result is a self-concept that is, in the title’s phrase, identifiable but not identical.
A related line of work examines differentiation between groups. People diverge from members of dissimilar outgroups: when another social group adopts a taste, people may abandon it to avoid signalling an undesired identity and to avoid the costs of misidentification (Berger and Heath 2007, 2008). This mechanism is developed in the section on identity signaling (Section 8.5).
The creativity literature offers an intriguing parallel. Novel ideas tend to balance familiarity and novelty, recombining established elements in new ways, and averaged, prototypical forms are often judged attractive. Those ideas are developed elsewhere in the book, in the sections on the Geneplore model (Section 31.2) and beauty in averageness (Section 29.10) (Table 22.4).
Identifiable but not identical: designing product lines. Optimal distinctiveness suggests that brands with strong group associations should offer variety along dimensions that do not dilute the identity signal. A sportswear brand can keep its logo, silhouette, and brand name constant, the assimilation cues, while offering many colours, limited editions, and personalisation: the differentiation cues. The theory also warns that variety on the identity-signalling dimension itself (for example, removing the logo) may weaken the reason consumers chose the brand. For managers, the practical question is to identify which attributes in their category carry group meaning and which carry individual meaning, and to design assortment accordingly (Chan, Berger, and Boven 2012).
Treating uniqueness and conformity as opposite ends of one scale. The theory’s central claim is that both needs operate simultaneously and are satisfied on different dimensions or at different levels of inclusiveness. Measuring a single “conformity versus uniqueness” preference will miss the pattern. Research should distinguish the attributes or groups on which consumers assimilate from those on which they differentiate.
22.4 Conformity
Core claim. Individuals change their judgements and behaviour to match those of a group, either because they take others’ responses as evidence about reality (informational influence) or because they want to meet others’ expectations and be accepted (normative influence).
Origins. Social psychology, Sherif (1936), Asch (1956), Deutsch and Gerard (1955), reviewed in Cialdini and Goldstein (2004).
Level of analysis. Individual within a group, aggregated to markets and networks (W. A. Mason, Conrey, and Smith 2007, Salganik, Dodds, and Watts 2006).
Key constructs. Informational influence, normative influence, public compliance versus private acceptance, unanimity, group size, anonymity, social proof, herding.
Boundary conditions. Informational influence is strongest when the task is ambiguous and others seem knowledgeable. Normative influence is strongest when responses are public, the group is valued, and the majority is unanimous.
Related theories. Social norms, Optimal distinctiveness, Wisdom of the crowd (Section 30.5), Diffusion of innovations (Section 10.1).
22.4.1 Origins and core logic
Sherif (1936) showed that when people judged the apparent movement of a point of light in a dark room, an inherently ambiguous stimulus, their estimates converged on a group norm, which persisted when they later judged alone. Here conformity looks like rational learning: when reality is uncertain, others’ judgements are informative.
Asch (1956) asked a sharper question: would people conform even when the right answer was obvious? Participants judged which of three comparison lines matched a standard line, a task on which people working alone almost never erred. In the critical trials, a unanimous group of confederates gave the same wrong answer before the participant responded. Roughly a third of participants’ responses on these trials conformed to the erroneous majority, and most participants conformed at least once, although a substantial minority never did. Variations showed that a single dissenting ally sharply reduced conformity, highlighting the power of unanimity.
Deutsch and Gerard (1955) distinguished the two forces at work. Informational influence is the tendency to accept information from others as evidence about reality. Normative influence is the tendency to conform to the positive expectations of others. Their experiments showed that conformity in an Asch-type task declined when participants responded anonymously, evidence of a normative component, but did not disappear, consistent with informational influence. Cialdini and Goldstein (2004) integrate the literature by linking conformity to three goals: accuracy (informational), affiliation (normative), and the maintenance of a positive self-concept (Table 22.5).
22.4.2 Evidence and extensions
A meta-analysis of Asch-type studies found that conformity levels were higher in collectivist than in individualist cultures and had declined in the United States since the 1950s (R. Bond and Smith 1996), suggesting that conformity is shaped by cultural values (see Hofstede (Section 18.2)). In markets, conformity aggregates into herding and social proof. In a large online experiment, Salganik, Dodds, and Watts (2006) created parallel “worlds” in which participants could download songs. When participants could see how often others had downloaded each song, popularity became more unequal and less predictable across worlds than when they could not. A randomized experiment on a social news site found that giving a comment an artificial positive vote made subsequent positive votes more likely and raised the comment’s final rating, whereas artificial negative votes tended to be corrected by later users (Muchnik, Aral, and Taylor 2013). And in a field experiment in a restaurant chain, displaying the most popular dishes increased demand for those dishes, consistent with observational learning (Cai, Chen, and Fang 2009).
Social proof in e-commerce. Online retailers and platforms routinely display “best-seller” badges, review counts, “N people are viewing this item”, and “customers also bought” panels. Each exploits informational influence: under uncertainty about quality, others’ choices are a signal. The research above suggests two managerial cautions. First, social proof amplifies early random advantages, so displayed popularity may reflect luck as much as quality (Salganik, Dodds, and Watts 2006). Platforms that care about matching customers to the best products should guard against self-reinforcing rankings, for example by mixing popularity with quality signals or by randomising exposure for new items. Second, herding is asymmetric: positive signals snowball more readily than negative ones (Muchnik, Aral, and Taylor 2013), which makes rating systems vulnerable to early manipulation.
Interpreting all similarity of behaviour as conformity. People who behave alike may have been influenced by one another, may share unobserved preferences (homophily (Section @ref(homophily (Section 30.2)))), or may face common shocks. Inferring social influence from correlated behaviour in observational data is notoriously difficult. Randomised manipulations of what others are seen to do, as in the studies above, provide the clearest evidence.
22.5 Social Norms
Core claim. Norms influence behaviour mainly when they are salient. Two kinds of norm must be distinguished: descriptive norms (what most people do) and injunctive norms (what most people approve or disapprove of). Messages that activate the wrong norm can backfire.
Origins. Social psychology, focus theory of normative conduct (Cialdini, Reno, and Kallgren 1990), field applications in energy (Schultz et al. 2007, Allcott 2011), hotels (N. J. Goldstein, Cialdini, and Griskevicius 2008), and taxation (Hallsworth et al. 2017).
Level of analysis. Individual within a community. Designed interventions at the level of firms and public agencies.
Key constructs. Descriptive norm, injunctive norm, norm salience (focus), reference group, boomerang effect, pluralistic ignorance.
Boundary conditions. Norm messages work best when the reference group is similar and identifiable, the norm is credible, and the behaviour is feasible. Descriptive norms can backfire for people who are already doing better than the norm or when they reveal that an undesirable behaviour is common.
Related theories. Conformity, Social comparison, Social identity theory, Spiral of silence.
22.5.1 Core logic
Cialdini, Reno, and Kallgren (1990) argued that the concept of norms had lost explanatory power because it was invoked after the fact to explain almost anything. Their focus theory of normative conduct restores predictive content by distinguishing two kinds of norm and by holding that a norm guides behaviour only when attention is focused on it. In field experiments on littering, people were least likely to litter when they saw someone drop litter into a clean environment, an act that drew attention to the descriptive norm that people here do not litter, and most likely to litter when the environment was already littered. Seeing someone pick up litter, which focused attention on the injunctive norm, reduced littering regardless of the state of the environment (Table 22.6).
Norm perceptions are often inaccurate. Prentice and Miller (1993) showed that university students systematically overestimated their peers’ comfort with heavy drinking: a case of pluralistic ignorance, in which most members of a group privately reject a norm but believe that most others accept it. Correcting such misperceptions is one of the main levers of norm-based interventions. A striking feature of normative influence is that people underestimate it: in a study of energy conservation, residents rated information about their neighbours’ behaviour as the least motivating reason to conserve, yet it produced larger changes in behaviour than appeals to environmental protection, social responsibility, or saving money (Nolan et al. 2008).
22.5.2 Mechanism and the boomerang effect
Schultz et al. (2007) provided the canonical demonstration of how descriptive norms can backfire. Households in a Californian community received feedback comparing their electricity use to the average in their neighbourhood. Households that had been using more than average subsequently reduced consumption, as intended. But households that had been using less than average increased their consumption: a boomerang effect, as the descriptive norm pulled them toward the average. When the feedback was accompanied by an injunctive cue, a smiling emoticon for below-average users and a frowning emoticon for above-average users, the boomerang effect among low users disappeared. Figure 22.3 reproduces the pattern with simulated data.
Figure 22.3: Stylised boomerang effect of descriptive-norm feedback. Households above the neighbourhood average reduce use under both message types. Households below the average drift upward when they receive descriptive feedback alone, but not when an injunctive cue (approval) is added. Simulated data. Bars show means with 95% confidence intervals.
Hotel guests show the same sensitivity to descriptive norms. N. J. Goldstein, Cialdini, and Griskevicius (2008) found that a card stating that most guests reused their towels produced more towel reuse than the hotel industry’s standard environmental appeal, and that a “provincial” norm referring to guests who had stayed in the same room was more effective still, even though that reference group is not more meaningful in any objective sense.
22.5.3 Applications in business and policy
Opower home energy reports. Opower, a software firm working with utilities, scaled the Schultz et al. design to millions of households. Its reports compared each home’s energy use with that of similar neighbours and of the most efficient neighbours, added an injunctive element (smiling faces for good performance), and offered conservation tips (Cuddy, Doherty, and Bos 2010). Because utilities randomly assigned which households received reports, the programme became one of the best-evaluated behavioural interventions in the world. Analysing randomised trials covering hundreds of thousands of households, Allcott (2011) estimated that the reports reduced energy consumption by about 2 percent on average, with larger effects among households that had used more energy at baseline. Follow-up research found that households responded to each report and then partially backslid, that continued reports produced more persistent changes in behaviour, and that a meaningful share of the effect persisted after reports were discontinued (Allcott and Rogers 2014). The case illustrates both the power of norms at scale and the importance of the injunctive element in preventing boomerang effects among already-efficient households.
Social-norm letters and tax compliance. In two large natural field experiments, the UK tax authority varied the wording of reminder letters sent to taxpayers who had not yet paid what they owed (Hallsworth et al. 2017). Adding a descriptive-norm statement, that the great majority of people pay their tax on time, increased payment rates relative to a standard letter, and messages that made the norm more specific to the recipient’s situation tended to be more effective. Because the intervention cost almost nothing to implement, the additional revenue collected represented a very high return. The study has become a template for norm-based communications in public administration and, by extension, in business collections, compliance, and renewal reminders.
Norms as policy: scale, persistence, and ethics. Norm-based nudges are cheap, scalable, and well-evidenced in field experiments, but three debates remain. First, effects are typically modest in size, a few percent in energy use, so their value depends on scale and low cost rather than on transforming behaviour. Second, persistence varies: effects can decay when messages stop, although repeated messaging can build habits (Allcott and Rogers 2014). Third, the ethics of “managing” people through social pressure is contested, especially when the norm statements are selectively framed. Transparency about the source and accuracy of normative claims is a practical safeguard: a norm message that turns out to be misleading destroys the credibility that makes such messages work.
Sample hypotheses.
Designs. Randomised field experiments varying message content (the Opower and tax-letter designs are templates). Survey measures of perceived descriptive and injunctive norms with experimental corrections. Difference-in-differences analyses of staggered roll-outs.
Measures. Behavioural outcomes (consumption, payment, reuse, reporting). Perceived norms (estimated percentage of peers who perform a behaviour. Perceived approval). Heterogeneity by baseline position relative to the norm.
22.6 Social Facilitation and Inhibition
Core claim. The presence of others, as an audience or as co-actors, improves performance on simple or well-learned tasks (social facilitation) and impairs performance on complex or novel tasks (social inhibition).
Origins. Social psychology, Triplett (1898), the term “social facilitation” from Allport (1920), the drive theory synthesis of Zajonc (1965), evaluation apprehension (Cottrell et al. 1968), meta-analysis in C. F. Bond and Titus (1983).
Level of analysis. Individual performing in the presence of others.
Key constructs. Audience effect, coaction effect, dominant response, drive or arousal, task complexity or familiarity, mere presence, evaluation apprehension.
Boundary conditions. Effects depend on the task: facilitation requires that the correct response be dominant (well-learned). Inhibition arises when the dominant response is incorrect. Effects are generally small in magnitude.
Related theories. Distraction-conflict theory, Social comparison, Goal-setting theory (Section 19.4).
22.6.1 Origins and intellectual history
Social facilitation is often described as the topic of the first experiment in social psychology. Triplett (1898) observed from cycling records that racers rode faster when paced or in competition than when riding alone against the clock, and he followed this with a laboratory study in which children wound fishing reels either alone or alongside another child. He reported that many children worked faster in competition. A modern reanalysis of Triplett’s data concluded that, by contemporary statistical standards, the evidence in the original experiment was considerably weaker than textbook accounts suggest (Strube 2005): a useful reminder that founding studies are not always the strongest.
Allport (1920) coined the term social facilitation. Working on “the influence of the group upon association and thought”, he found that people working alongside others produced more free associations than people working alone, but that the quality of their reasoning did not improve in the same way. Two research streams developed:
By the 1960s the literature appeared contradictory: the presence of others sometimes helped and sometimes hurt.
22.6.2 Core logic: drive theory
Zajonc (1965) resolved the contradiction with an elegant argument from learning theory. The presence of members of one’s own species increases generalized drive (arousal). Drive increases the likelihood of dominant responses: the responses that are most strongly established in a given situation. On simple or well-learned tasks the dominant response is usually correct, so performance improves. On complex or novel tasks the dominant response is often incorrect, so performance suffers. In short, an audience facilitates the emission of well-learned responses and impairs the acquisition of new ones. Zajonc and Sales (1966) provided direct support: after participants had practised nonsense words at different frequencies, the presence of an audience increased their tendency to give the most-practised (dominant) words in a later task (Table 22.7).
Figure 22.4 shows the resulting interaction with illustrative data, and Figure 22.5 depicts the mechanism.
Figure 22.4: Illustrative social facilitation and inhibition: the presence of an audience improves performance on a simple, well-learned task and impairs performance on a complex, novel task. Simulated data. Points are means with 95% confidence intervals.
Figure 22.5: Mechanism of social facilitation and inhibition (drive theory, after Zajonc, 1965): the presence of others raises arousal, which strengthens dominant responses. Task complexity determines whether the dominant response is correct. Evaluation apprehension and distraction-conflict are alternative accounts of why presence raises arousal.
22.6.3 Evidence and extensions: why does presence arouse?
Drive theory left open why the presence of others should raise drive. Three answers have been proposed (Table 22.8).
A meta-analysis of 241 studies concluded that the presence of others has small effects overall: it slightly increases the speed of simple performance and impairs both the speed and the accuracy of complex performance, and these effects were not fully explained by evaluation apprehension (C. F. Bond and Titus 1983).
Social facilitation should be distinguished from social loafing (Section 30.9), the tendency to exert less effort when working collectively than individually (Latané, Williams, and Harkins 1979, Karau and Williams 1993). Facilitation arises when individual performance is visible to others. Loafing arises when individual contributions are pooled and cannot be identified. The same team can show either, depending on how work and evaluation are designed.
22.6.4 Applications in business
Open-plan offices, monitoring, and peer visibility. Many organizations have redesigned workspaces to increase visibility and interaction. Social facilitation theory predicts that visibility will help routine, well-practised work and hinder complex, novel work that demands concentration. Field evidence is consistent with a cautious view. In two studies of corporate headquarters that moved to open-plan layouts, face-to-face interaction fell sharply while electronic communication rose: the opposite of the intended effect (Bernstein and Turban 2018). In a laboratory study of electronic performance monitoring, monitoring raised the productivity of highly skilled participants but lowered that of less skilled participants, and monitored participants reported more stress (Aiello and Kolb 1995): the facilitation-inhibition pattern applied to skill rather than task. Visibility to peers can also raise effort: among supermarket cashiers, productivity rose when a highly productive co-worker joined the shift, especially when the cashier was in that co-worker’s line of sight (Mas and Moretti 2009). The managerial implication is to match visibility to task: open, observable settings for practised, routine work. Quiet, protected settings for learning and complex problem solving.
22.7 Distraction-Conflict Theory
Core claim. Others (and other distractors) draw attention away from a task. When a person feels pulled between attending to the task and attending to the distraction, the resulting attentional conflict raises arousal and narrows attention, producing facilitation on simple tasks and impairment on complex ones.
Origins. Social psychology, Sanders, Baron, and Moore (1978), reviewed in R. S. Baron (1986).
Level of analysis. Individual.
Key constructs. Distraction, attentional conflict, drive, attentional overload, narrowing of attention (cue utilization).
Boundary conditions. Effects require that the distractor be salient and that the person feel pressure to attend to both the task and the distractor. Distractors that can be ignored easily produce little conflict.
Related theories. Social facilitation, Information processing model (Section 31.1), Social comparison.
22.7.1 Core logic
Distraction-conflict theory offers a mechanism for social facilitation that does not depend on anything uniquely social. Sanders, Baron, and Moore (1978) proposed that others are distracting, people glance at them, wonder what they think, and compare their own progress with that of co-actors, and that distraction creates a conflict between two attentional demands. In their experiments, co-actors produced the classic pattern of facilitation on simple tasks and impairment on complex tasks mainly when they were working on the same task, making them relevant targets for social comparison and thus more distracting. Participants in those conditions also reported greater distraction.
R. S. Baron (1986) developed the theory in two directions. First, attentional conflict can raise drive, producing Zajonc’s dominant-response effects. Second, conflict can create attentional overload, which leads people to narrow their attention to a smaller set of cues. Narrowing helps on simple tasks that require attention to only a few central cues, irrelevant cues are screened out, but harms complex tasks that require attention to many cues. Because the account is about attention rather than about other people as such, it predicts that non-social distractions such as noise or flashing lights can produce effects similar to those of an audience, and Baron reviewed evidence consistent with that prediction. Figure 22.6 summarises the mechanism.
Figure 22.6: Mechanism of distraction-conflict theory (after Sanders, Baron, & Moore, 1978, Baron, 1986): a salient distractor creates attentional conflict with the task, which raises drive and produces attentional overload and narrowing. Task complexity determines whether narrowing helps or hurts.
Illustrative scenario: dashboards and interruptions in the contact centre. A customer-service centre installs wall-mounted screens showing each agent’s live call statistics and ranking. Distraction-conflict theory predicts mixed effects. For experienced agents handling routine queries, the screens may narrow attention onto the core script and raise throughput. For newer agents, or for complex complaints that require attention to many cues, the customer’s tone, account history, policy exceptions, the competing pull of the leaderboard may divide attention and increase errors. Similar logic applies to knowledge workers facing frequent notifications. A testable design would randomise the visibility of real-time metrics across teams and examine effects separately for routine and complex cases and for experienced and novice staff.
Sample hypotheses.
Designs. Laboratory experiments crossing presence (alone, audience, blindfolded audience, non-social distractor) with task complexity. Field experiments that vary monitoring or workspace openness. Natural experiments from office redesigns or remote-work transitions.
Measures. Speed and accuracy separately, physiological arousal, self-reported distraction, peripheral-cue recall as an index of attentional narrowing.
Ignoring task complexity. The most common error in applying social facilitation and distraction-conflict theory is to predict a main effect of observation, monitoring, or open workspaces on “performance”. The theories predict an interaction: the same social presence helps on simple, practised tasks and hurts on complex, novel ones. Studies and workplace policies that average across task types will find small or inconsistent effects.
22.8 Spiral of Silence
Core claim. Because people fear social isolation, they continuously monitor the climate of opinion. Those who believe their view is in the minority, or losing ground, become less willing to express it, which makes the view appear even weaker and further silences its holders, in a self-reinforcing spiral.
Origins. Political science and mass communication, Noelle-Neumann (1974), meta-analytic assessment in Glynn, Hayes, and Shanahan (1997).
Level of analysis. Individual expression aggregated to public opinion. Analogues in organizations (employee voice and silence).
Key constructs. Fear of isolation, quasi-statistical sense of the climate of opinion, perceived current and future majority, willingness to speak out, mass media as a source of climate perceptions, the “hard core” who speak out regardless.
Boundary conditions. The spiral is most likely for morally loaded or controversial issues, in public settings, and when people rely on media or other indirect cues to judge opinion. It is weaker among those with strong convictions and in settings where dissent is protected.
Related theories. Conformity, Social norms, Agenda setting (Section 26.5), Pack journalism (Section 30.7).
22.8.1 Core logic
Noelle-Neumann (1974) proposed that people possess a quasi-statistical sense with which they estimate the distribution of opinion in their environment and whether it is shifting. Driven by fear of isolation, people who perceive that their opinion is shared by the majority, or is gaining ground, speak out more confidently. Those who perceive their view to be in the minority, or losing ground, fall silent. Because public expression is itself one of the main cues that others use to judge the climate of opinion, silence makes the minority view seem even rarer than it is, which further suppresses its expression. The mass media play a central role: they are a principal source of information about what “most people” think, so media emphasis can shape perceptions of the dominant opinion and of the dominant culture. One implication is that a person’s perceived social environment may differ substantially from reality. Noelle-Neumann also noted the existence of a hard core who continue to express their views regardless of the perceived climate (Figure 22.7).
Figure 22.7: The spiral of silence (after Noelle-Neumann, 1974): perceptions of the climate of opinion, shaped by media and by others’ expression, combine with fear of isolation to reduce minority expression, which in turn reinforces the impression that the view is rare.
22.8.2 Evidence and extensions
The spiral of silence has been tested mainly with surveys that measure people’s perceptions of majority opinion and their stated willingness to express their own views in hypothetical public situations. A meta-analysis of such studies found a positive but small relationship between perceived support for one’s opinion and willingness to speak out (Glynn, Hayes, and Shanahan 1997). The theory thus captures a real tendency but not a deterministic law, and its effect is easily swamped by individual differences in conviction, interest, and personality. The closely related concept of pluralistic ignorance (Prentice and Miller 1993), widespread private rejection of a norm that everyone believes others accept, describes the end state that a spiral of silence can produce.
Does the spiral survive the internet? Online platforms both weaken and strengthen the conditions for a spiral of silence. On one hand, anonymity and the ease of finding like-minded others should reduce fear of isolation and allow minority views to be expressed. On the other, visible metrics of approval, algorithmic amplification of popular content, and the permanence of online statements may make the climate of opinion more salient and the costs of deviation higher. Researchers also debate whether “willingness to speak out” in hypothetical survey scenarios captures actual expression. The theory’s core mechanisms, perceived climate, fear of isolation, and expression as a signal, offer a framework for these questions, but the size and direction of effects in digital environments must be established empirically rather than assumed.
22.8.3 Applications in organizations: employee voice and silence
The organizational analogue of the spiral of silence is organizational silence: a collective-level phenomenon in which employees withhold information about problems and concerns (Morrison and Milliken 2000). Morrison and Milliken (2000) argued that silence develops when managers fear negative feedback and hold implicit beliefs, for example, that employees are self-interested or that management knows best, which shape structures and behaviours that discourage upward communication. Employees then form shared perceptions that speaking up is futile or dangerous. In interviews, most employees in one study reported having remained silent about issues that concerned them at work, frequently citing fear of being labelled negatively or damaging relationships (Milliken, Morrison, and Hewlin 2003). Leadership matters: managerial openness has been linked to higher employee voice through its effect on employees’ sense that speaking up is safe (Detert and Burris 2007), and team psychological safety (Section 14.8), a shared belief that the team is safe for interpersonal risk-taking, supports learning behaviour in work teams (A. Edmondson 1999).
Breaking the spiral in meetings and organizations. The engineer who notices a flaw but sees no one else worried is caught in a micro-spiral of silence: the silence of others is read as agreement, and the engineer’s own silence becomes part of the cue for the next person. Organizations use several practices that map directly onto the theory’s mechanisms. Anonymous or simultaneous input (written pre-reads, independent estimates before discussion, anonymous polling) prevents early speakers from defining the climate of opinion. Designated dissent (red teams, pre-mortems, a rotating “devil’s advocate”) lowers the isolation cost of disagreement by making it a role. Leader behaviour, senior people speaking last and visibly rewarding bad news, changes the perceived climate. And measurement through confidential surveys can reveal pluralistic ignorance, showing employees that concerns they believed were idiosyncratic are widely shared.
Equating silence with agreement. Managers and analysts often infer consensus from the absence of objections, and researchers sometimes infer public opinion from public expression. The spiral of silence predicts that both inferences are biased in favour of the perceived majority. Measure private opinions separately from public expression, and treat the gap between them as an outcome of interest.
22.9 Chapter Summary
Table 22.9 presents the comparison of the chapter’s theories.
22.10 Discussion Questions
22.11 Further Reading