Social comparison theory is a framework in social psychology explaining how people use information about others to evaluate themselves. Proposed by Leon Festinger in 1954, it initially addressed the evaluation of opinions and abilities, especially when objective standards were unavailable. Subsequent research extended the framework to emotions, achievement, appearance, income, health, and social standing. Comparison may be deliberate, as when an employee examines colleagues’ salaries, or relatively spontaneous, as when a student’s confidence changes after hearing a classmate’s examination score.
The theory’s central insight is that self-evaluation depends partly on a social frame of reference rather than solely on absolute circumstances. The same performance can seem excellent among beginners and inadequate among experts. Yet comparison does not have a single psychological effect: a successful peer may inspire improvement, threaten self-worth, provide practical information, or produce several of these responses together. Modern research therefore examines not only whether comparisons are upward or downward, but also which targets people select, what they compare, how they interpret differences, and what opportunities for action they perceive.
Origins and the original theory
Festinger’s “A Theory of Social Comparison Processes,” published in Human Relations in 1954, proposed that people have a drive to evaluate their opinions and abilities. Where physical or otherwise nonsocial criteria are unavailable, people turn to other people. A person can establish whether a container holds a litre of water through measurement; determining whether a political judgment is reasonable or a level of artistic ability is impressive usually involves less decisive standards. Social information supplies a basis for evaluation under such conditions.
The original theory linked comparison to similarity. Festinger argued that the tendency to compare with a particular person decreases as the discrepancy between their opinions or abilities increases. Someone whose performance is vastly different may provide a poor standard for judging a person’s present capabilities. A beginning runner generally obtains more useful diagnostic information from other beginners than from an Olympic champion. Similarity, however, was not simply a claim that people always choose equally successful targets: background, experience, and other characteristics can make another person’s performance relevant.
Festinger also distinguished opinions from abilities. Opinions often invite agreement and pressures toward uniformity. Abilities frequently contain an additional “unidirectional drive upward”: in domains such as speed or accuracy, improvement has a relatively clear direction. Groups could reduce discrepancies by members changing their positions, attempting to change others, or ceasing to treat particular people as appropriate comparison partners. The theory thus connected individual evaluation with group communication and conformity, rather than treating comparison as an exclusively private calculation.
Several intellectual traditions preceded or developed alongside this account. Reference-group research in sociology examined the groups people use as standards for evaluating their circumstances. The looking-glass self, associated with Charles Horton Cooley, concerned self-understanding through imagined judgments by others. These approaches overlap with social comparison but are not identical: imagining how someone judges a person does not necessarily involve comparing that person’s standing with another’s.
Later work substantially broadened Festinger’s formulation. Stanley Schachter’s 1959 research on affiliation under anxiety helped establish the importance of other people as sources of emotional and situational information. Thomas Wills’s 1981 account of downward comparison emphasized comparisons with worse-off others under threat. Joanne Wood’s 1989 review organized an expanding literature around self-evaluation, self-improvement, and self-enhancement. The resulting field became a family of related models rather than a single unchanged set of propositions.
What is compared, with whom, and for what purpose
A social comparison contains several analytically distinct elements: a person making the comparison, a dimension, a target, and an interpretation of the relationship between self and target. Dimensions include measurable outcomes, such as income or running time, and less directly measurable attributes, such as attractiveness, competence, or moral character. Targets may be identifiable individuals, a group average, a statistical percentile, or an imagined typical person. An explicit thought such as “she earns more than this person does” is only one form; comparison may also influence a judgment without appearing in a verbal report.
Target relevance depends on the question being asked. A person assessing whether a salary is competitive may compare with workers in the same occupation, location, and career stage. The same person assessing broader social standing may compare with friends or neighbours whose occupations differ. Someone choosing a training strategy may seek a highly successful expert, whereas someone estimating current ability may prefer a peer with similar experience. Consequently, similarity is dimension-specific: sharing a profession may matter more than sharing an age, and sharing a medical condition may matter more than sharing an occupation.
Three motives recur throughout the literature. Self-evaluation concerns obtaining an accurate estimate of a quality or circumstance. Self-improvement concerns discovering how to reach a desired standard. Self-enhancement concerns maintaining or restoring a favourable view of the self. These motives can coexist. A student may examine a high-scoring classmate’s work both to identify useful techniques and to decide whether their own result is disappointing. The comparison’s informational value does not prevent it from having emotional consequences.
Uncertainty often increases the usefulness of social information, but uncertainty is not a necessary condition for comparison. People compare objectively measured outcomes even when their own score is unambiguous. Knowing that a race took 25 minutes answers a measurement question; knowing how that time ranks among comparable runners answers a different, socially evaluative question. Standards can also be supplied involuntarily through public rankings, performance dashboards, or casual conversation.
Comparison must therefore be distinguished from simple exposure to information about others. Seeing a successful colleague does not establish that a person used the colleague as a standard. Likewise, competition is not required: people can compare while cooperating, and competitors can sometimes focus entirely on an absolute goal. Researchers increasingly distinguish target selection, comparative processing, and consequences, because an observed preference for certain information does not by itself reveal how that information changes self-evaluation.
Upward and downward comparison: contrast and identification
An upward comparison involves a target judged better off or more capable on the relevant dimension; a downward comparison involves someone judged worse off or less capable. A lateral comparison involves someone at a similar level. These directions describe perceived positions, not fixed types of people. One colleague can be an upward target for earnings, a lateral target for experience, and a downward target for work–life balance. The dimension and the observer’s interpretation determine the comparison’s direction.
Direction alone does not determine the outcome. In contrast, the person emphasizes the distance between self and target. An upward contrast can lower perceived ability: an accomplished musician makes a learner feel unskilled. A downward contrast can improve relative self-evaluation: the learner feels competent beside a novice. In assimilation, the target draws the self-evaluation toward the target’s position. Identification with a successful musician may increase confidence, while identification with someone struggling may increase concern about one’s own future.
The target’s attainability is particularly important. A successful peer whose progress appears achievable can offer a model for improvement. The same performance may be discouraging if it seems to depend on inaccessible resources or exceptional talent. Perceived similarity can therefore have different effects depending on what it implies. Similarity may make success seem transferable, but it can also make an unfavourable discrepancy harder to dismiss. Dissimilarity can protect self-evaluation while reducing the target’s instructional usefulness.
The selective accessibility model developed by Thomas Mussweiler and colleagues explains some of these outcomes through the information brought to mind during comparison. Testing whether self and target are similar makes similarity-consistent knowledge more accessible and can support assimilation. Testing how they differ highlights difference-consistent knowledge and can support contrast. The model treats comparative judgment as an active process of retrieving and interpreting information, not merely subtracting one score from another.
Downward comparison is similarly ambiguous. A person may feel relief because someone else faces greater difficulty, but may instead identify with that person and fear a similar outcome. Upward comparison can combine admiration, dissatisfaction, and determination. Experimental studies therefore often separate a target’s objective standing from instructions that encourage identification or differentiation. This distinction explains why evidence does not support simple rules such as “upward comparisons are harmful” or “downward comparisons improve well-being.” Effects depend on the comparison process, domain, perceived control, and outcome being measured.
Self-concept, relationships, and comparative emotions
Social comparisons contribute to self-concept, the organized set of beliefs people hold about their characteristics. They can influence relatively specific judgments, such as mathematical competence, as well as broader evaluations associated with self-esteem. These levels should not be conflated. A poor comparison in an unimportant activity may have little effect on general self-worth, whereas a modest discrepancy in a personally central domain can be consequential. The importance assigned to a domain is therefore part of the explanatory mechanism.
Abraham Tesser’s self-evaluation maintenance model, developed during the 1980s, focused on the interaction of another person’s performance, psychological closeness, and the domain’s relevance to the self. When a close other excels in a domain central to one’s identity, comparison can threaten self-evaluation. When the domain has little personal relevance, the same success can produce reflected pride. A sibling’s musical achievement may be enjoyable for someone who does not identify as musical but difficult for another aspiring musician.
The model also identifies ways relationships and self-definitions can change. A person may improve performance, reduce the importance of the domain, or alter psychological closeness to the successful other. These are theoretical possibilities, not inevitable reactions. Relationships involve commitments and emotions that exceed comparative evaluation, and close others can simultaneously supply encouragement, collaboration, and standards of achievement. Comparisons within a peer group are especially informative because shared circumstances can make performance differences appear personally relevant.
Comparative emotions include admiration, envy, pride, shame, relief, and resentment. Research often distinguishes forms of envy associated with efforts to improve one’s own position from those associated with hostility toward an advantaged person. Perceived deservingness and control help explain the difference. A hard-earned, attainable success can elicit a different response from an advantage interpreted as unfair and unchangeable. These emotional categories are not perfectly discrete, and a single encounter may produce mixed feelings.
Social comparison also intersects with relative deprivation, the perception of being unfairly disadvantaged relative to a standard. Disadvantage alone is not sufficient: interpretations of entitlement and injustice distinguish deprivation from a neutral recognition that another person has more. Group-level comparisons may concern wages, recognition, or political rights and may contribute to collective action. This extends beyond Festinger’s original focus on evaluating individual opinions and abilities. Personal disappointment and perceived group injustice can occur together, but they involve different reference standards and do not necessarily predict the same behaviour.
Individual differences, development, and cultural context
People differ in how frequently they attend to comparative information and how strongly it affects them. Frederick Gibbons and Bram Buunk introduced the Iowa–Netherlands Comparison Orientation Measure in 1999 to assess social comparison orientation. Its 11 items concern tendencies to compare abilities and opinions with those of others. The measure captures a relatively general disposition rather than a preference for upward or downward targets alone. A high score does not establish that a person always compares unfavourably or experiences harmful consequences.
Individual differences interact with situations. Uncertainty, an unfamiliar environment, or evaluation by others can increase comparison even among people who report relatively little habitual comparison. Conversely, a frequent comparer may ignore social standards in a domain of little interest. Self-esteem, perceived competence, and emotional vulnerability can influence how targets are interpreted, but these relationships are not reducible to one-way effects. Repeated threatening comparisons may alter self-evaluation, while existing self-evaluations can shape which comparisons are noticed and remembered.
Research in developmental psychology examines how children acquire the ability to interpret performance relative to peers. During the school years, children increasingly differentiate effort, ability, and achievement, and comparative information becomes more important to academic self-perceptions. Classroom organization matters: public recognition, ability grouping, grades, and competitive tasks make some differences more visible than others. Adolescence adds changing peer relationships and identity concerns, but comparison is not confined to that period. Adults compare occupational progress, parenting, health, and retirement circumstances.
Cultural context influences which dimensions are valued, which targets are relevant, and how comparison is expressed. Research comparing national or cultural groups has examined self-improvement motives, modest self-presentation, relational obligations, and responses to outstanding performers. Broad contrasts between “individualistic” and “collectivistic” societies can obscure substantial variation within populations. A person’s occupational setting, social class, generation, and immediate relationships may change the meaning of a comparison as much as a national label does.
Material opportunities also constrain target selection. Segregated schools, neighbourhoods, and workplaces expose people to different distributions of achievement and resources. Through socialization, individuals learn which attributes signal success and which groups count as relevant standards. The result is that identical absolute circumstances can support different subjective evaluations across social settings. This does not imply that material conditions are unimportant; rather, social circumstances affect both the resources people possess and the reference points through which those resources are interpreted.
Research designs and measurement
Social comparison research uses laboratory experiments, questionnaires, field studies, longitudinal designs, and experience sampling. In a typical experiment, participants complete a task and receive information suggesting that another participant performed better, worse, or similarly. Researchers then measure self-evaluation, affect, persistence, or later performance. Random assignment helps isolate the effect of the supplied comparison information. However, invented scores and brief tasks may not reproduce the importance or complexity of comparisons involving longstanding relationships and consequential outcomes.
Target-selection studies examine which information participants choose to inspect. Participants might request feedback from a stronger or weaker performer, select a discussion partner, or choose between accounts of others’ experiences. These designs assess preferences more directly than studies that assign a target. Their interpretation still requires care: choosing a high performer’s account may reflect a desire for practical instruction, curiosity, or evaluative comparison. Selecting information and using it as a self-evaluative standard are related but distinct events.
Questionnaires can measure general comparison orientation, domain-specific comparison, or reactions to particular targets. Retrospective reports are vulnerable to imperfect memory and differences in what respondents understand by “comparison.” Experience-sampling methods ask people to report current or recent comparisons repeatedly during daily life. Such designs can distinguish variation between people from variation within the same person across situations. Their limitations include participant burden, incomplete reporting, and the possibility that repeated questions make comparison unusually salient.
Establishing causation is especially difficult in natural settings. An association between frequent comparison and lower subjective well-being may reflect an effect of comparison, a tendency for dissatisfied people to compare more, or shared influences such as stress and insecurity. Longitudinal data clarify temporal ordering but do not automatically remove confounding. Experiments offer stronger control, while field designs provide greater realism; persuasive explanations often depend on convergence across methods rather than a single design.
Several measurement distinctions are particularly important. Studies should separate comparison frequency from direction, target selection from interpretation, and immediate mood from enduring self-concept. They should also distinguish an actual target’s attributes from the participant’s perception of those attributes. A comparison can be psychologically influential even when the target’s supposed success is exaggerated. Reviews and meta-analyses must therefore consider whether studies have operationalized the same process. Combining all upward-comparison exposures into one category can conceal differences between inspiring role models, threatening competitors, and irrelevant examples.
Education, performance, and organizational life
Educational research provides a prominent demonstration that self-evaluation depends on context. The big-fish–little-pond effect, developed in work by Herbert Marsh and colleagues from the 1980s onward, describes a tendency for students of comparable achievement to report lower academic self-concept in higher-achieving educational settings. A student’s own achievement generally supports academic self-concept, while the average achievement of the surrounding group can have an opposing contextual association. Being strong relative to classmates and attending a high-achieving school are therefore not psychologically equivalent.
This effect concerns academic self-concept, not a general claim that selective schools reduce learning or damage all students. Achievement, self-perception, aspiration, and later educational choices are distinct outcomes. High-achieving environments may provide advanced instruction and capable peers while also raising the standard against which students judge themselves. Prestige or identification with a successful institution can produce positive reflected appraisals, potentially operating alongside unfavourable local comparisons.
Related work distinguishes social comparison from dimensional comparison. A student judging mathematics ability against classmates is making a social comparison; judging mathematics ability against their own language performance is making a dimensional comparison. Temporal comparison, such as evaluating current work against last year’s work, uses the past self as a standard. These processes can yield different conclusions from the same record of achievement. A high mathematics score may appear disappointing beside an even higher language score but excellent relative to classmates.
In organizations, salary, promotion, recognition, workload, and access to opportunities provide recurring comparison dimensions. A worker may evaluate compensation against colleagues in the same firm, an occupational labour market, or personal expectations. Organizational justice research distinguishes the fairness of outcomes from the fairness of procedures and interpersonal treatment. Comparative disadvantage can contribute to dissatisfaction, but judgments also depend on whether differences appear justified by experience, responsibility, contribution, or legitimate procedures.
Public rankings and performance dashboards make comparisons salient and can change incentives. A rank may motivate effort, but it may also redirect attention toward narrowly measured activities or make cooperation less attractive. Competitive effects depend on whether success is attainable and whether rewards are fixed or jointly expandable. An informative comparison standard need not be a good learning model: the highest performer may use methods unsuitable for a novice, whereas a moderately more advanced peer may provide more transferable examples. Social comparison theory thus helps explain why identical performance feedback can generate different responses across classrooms and workplaces.
Health, adjustment, and material well-being
Health-related comparison includes evaluating symptoms, functional limitations, treatment experiences, and recovery against those of other people. Social comparison is particularly relevant when outcomes are uncertain or individual experiences vary widely. Another person’s account can help someone interpret an unfamiliar situation, but similarity in one characteristic does not establish clinical comparability. A shared diagnosis, for example, does not mean that disease stage, treatment response, or prognosis is the same.
Research on adjustment to serious illness has examined both upward and downward comparisons. Work associated with Shelley Taylor and colleagues during the 1980s showed how people may construct favourable self-evaluations despite substantial adversity. Upward targets can offer hope or information, while downward targets can help place difficulties in perspective. Neither direction reliably produces one emotional response. A person doing well may appear encouraging or intimidating; a person doing poorly may evoke relief, compassion, identification, or fear.
This area illustrates why contrast and identification require separate measurement. Someone may contrast current functioning with a worse-off patient and feel relatively fortunate while simultaneously identifying with that patient’s possible future. Similarly, another person’s recovery can be interpreted as evidence of what is possible without being treated as a reliable prediction. Studies of health-related quality of life therefore distinguish perceived standing, emotional adjustment, and physical outcomes. Comparative satisfaction is not evidence that a medical condition has improved.
Social comparison also contributes to research on income and life satisfaction. Income has purchasing power, but it can additionally signal relative position. Richard Easterlin’s 1974 work on economic growth and happiness helped stimulate debate about absolute income, relative income, and changing aspirations. Later research has not established that absolute resources are irrelevant. Material deprivation, economic security, public services, and the comparison group all affect how the relationship is interpreted.
A central difficulty is identifying the actual reference group. Geographic neighbours, coworkers, former classmates, and people shown in the media may provide different standards. Researchers sometimes assign regional average income as a proxy, but that does not prove that individuals use that average psychologically. Reference groups can also change after migration, promotion, or entry into higher education. Consequently, improved absolute circumstances can coexist with unchanged or reduced relative satisfaction. This pattern concerns the interaction between resources and evaluative standards, not a demonstration that improvement in living conditions lacks value.
Digital environments and the boundaries of the theory
Social media make comparative information unusually abundant and visible. Profiles, photographs, follower counts, reactions, and public achievements present potential standards for appearance, relationships, status, and success. Recommendation systems can repeatedly expose users to selected high-performing or highly engaging targets. These conditions differ from face-to-face comparison in scale and persistence, but the underlying interpretive questions remain familiar: Is the target relevant, similar, attainable, authentic, or representative?
Online presentation is selective. Users may encounter an accumulation of others’ celebrations, favourable photographs, and achievements without seeing corresponding failures or ordinary routines. This can distort the apparent distribution of success even when individual posts are factually accurate. Editing, sponsorship, and selective disclosure introduce additional uncertainty. Comparisons can therefore be based on real people while still relying on incomplete information about their circumstances.
Studies linking social media with body image, envy, or well-being often identify comparison as a possible mechanism, but “social media use” is too broad to specify that mechanism on its own. Messaging a close friend, viewing appearance-focused content, sharing creative work, and monitoring a professional network involve different activities. Time spent on a platform does not establish the frequency or direction of comparison. Existing dissatisfaction can also influence content selection, making reciprocal effects plausible.
Digital environments can support constructive comparisons as well. A learner may encounter attainable role models, a person with an uncommon experience may find relevant peers, and a professional community may make previously hidden pathways visible. Whether these encounters provide useful information or threatening standards depends on interpretation and context. Likewise, the common distinction between active posting and passive viewing does not fully predict outcomes: active participation may involve intense evaluation of audience reactions, while viewing may provide instruction without adverse self-evaluation.
The framework’s boundaries remain important. Comparison is not a complete explanation of identity, inequality, motivation, or platform effects. Social approval, belonging, material incentives, discrimination, and algorithmic selection can influence outcomes through mechanisms not reducible to comparative judgment. Research must distinguish comparison from adjacent processes such as learning by observation and attempts to infer how others evaluate the self. A particularly demanding question is whether a study has demonstrated that a social standard altered judgment, or has merely documented exposure to other people. Resolving that distinction requires measures of what participants attended to, how they interpreted it, and which subsequent judgments or actions changed.