Media Exposure

Abstract

Media exposure, which MeSH classifies under sociology, is the extent to which people encounter messages through communication media, and the study of what that contact does to thought, feeling, and behavior. This article traces the field's central ideas for a cognitive-psychology reader: the mere exposure effect, in which repeated contact alone breeds liking; cultivation and agenda-setting, in which the media diet shapes what audiences believe is real and important; social cognitive theory and the long program on media violence; the shift from uniform effects to differential susceptibility; and selective exposure with its reinforcing spirals, in which people choose the media that confirm them. It closes on the contested science of screen time and adolescent well-being. Three interactive demonstrations let the reader manipulate exposure, cultivation, and the spiral directly. The through-line is that media effects are cognitive effects.

Keywords: media exposure, mere exposure effect, cultivation, media effects, selective exposure

Media exposure is the elementary quantity of media research: how much of what content reaches a given person, through which channels, and under what conditions. It matters to the science of mind because contact with media is one of the largest sources of the second-hand experience from which people build their picture of a world they cannot observe directly. What a person takes crime, illness, wealth, and risk to be like is assembled in large part from portrayals never personally witnessed, so the distribution of media content becomes, statistically, a distribution of mental content. The field that studies this has moved a long way from its first image of an all-powerful media injecting messages into a passive audience. Its modern form treats exposure as the start of a chain that runs through selection, attention, interpretation, and memory, and that returns to shape which media a person seeks next, so that audience and content condition each other over time (Valkenburg et al., 2016). This article follows that chain from the simplest possible effect of contact to the reciprocal, individually moderated processes that current work takes to be the rule.

What Media Exposure Is

Media exposure names both a measured variable and a research tradition. As a variable it is the quantity of contact between a person and mediated content, indexed in practice by how much television is watched, how many hours are spent online, how often a genre or platform is used, or how much of a particular message a person has actually received. As a tradition it is the study of media effects: the changes in what people know, believe, feel, and do that follow from that contact. MeSH files the descriptor under sociology because exposure is measured at the level of populations and channels, but the mechanisms that turn exposure into effect are cognitive, running through attention, comprehension, memory, and the mental models that organize social knowledge (Valkenburg et al., 2016). The field's history is often told as a swing between two extremes. An early era imagined powerful and direct effects, media as a hypodermic needle delivering uniform reactions to a mass audience. A reaction to it emphasized limited effects, with media merely reinforcing what social groups already believed. The contemporary synthesis is neither: effects are real but conditional, small on average but consequential in aggregate, and shaped at every step by what the audience brings. The remainder of this article works through the major theories in roughly the order a stimulus travels, from the bare fact of repeated contact to the reciprocal dynamics of choice and change.

Figure 1

From Exposure to Effect: The Chain Modern Media Research Traces

The Mere Exposure Effect

The simplest thing repeated contact can do is make a stimulus better liked, with no argument, incentive, or even recognition involved. Robert Zajonc demonstrated this mere exposure effect in a series of experiments in which people rated novel stimuli, nonsense words, Chinese-like characters, and unfamiliar faces, more favorably the more often they had previously been shown them, with liking rising as a roughly logarithmic function of the number of exposures (Zajonc, 1968). The effect is robust, does not depend on the stimulus carrying any meaning, and is in fact often strongest when exposures are brief and the person cannot consciously recall having seen the item, which places a large part of the mechanism outside deliberate evaluation. Zajonc's own interpretation stressed that affect can be produced by exposure directly, without the prior cognitive appraisal that many theories of emotion had assumed was necessary, a claim that helped open the study of automatic and preattentive influences on preference. For media the implication is immediate and general: sheer repetition, of a brand, a face, a slogan, a melody, or a claim, tends to build positive regard for it, which is one reason advertising and political campaigns purchase frequency as much as content. The demonstration below lets the reader vary the number of exposures and the shape of the exposure-to-liking function, and watch how liking accrues quickly at first and then flattens as familiarity saturates.

Cultivation and Agenda-Setting

If mere exposure is the effect of contact with a single stimulus, cultivation and agenda-setting are effects of the aggregate diet. George Gerbner's cultivation theory held that heavy television viewers, exposed over years to the medium's recurrent patterns, come to see the real world as resembling the television world, most famously in overestimating the prevalence of violence and danger (Gerbner & Gross, 1976). Because dramatic television has long portrayed far more violence than occurs in life, Gerbner argued, its heavy viewers develop a mean world syndrome, a chronic sense that people are dangerous and cannot be trusted, and inflate their estimates of the frequency of crime. The cultivation differential, the gap in such beliefs between heavy and light viewers, is the theory's characteristic measure. In cognitive terms cultivation is a story about accessibility and construct availability: the examples the medium supplies most often are the ones most readily retrieved when a person estimates how common something is, so a viewing history becomes a sampling bias in the mental frequency count. Agenda-setting, formulated by Maxwell McCombs and Donald Shaw from a study of an election campaign, makes a parallel claim about importance rather than prevalence: the media may not tell people what to think, but they are strikingly successful in telling them what to think about, so that the issues the news covers most heavily become the issues the public judges most important (McCombs & Shaw, 1972). Both theories locate the effect not in persuasion by any single message but in the cumulative priming of some contents over others. The demonstration below lets the reader set the violence saturation of a media environment and the balance of heavy and light viewers, and see the cultivation differential open up between the two groups' estimates of real-world risk.

Social Cognitive Theory and Media Violence

The most sustained and most contested body of media-effects research concerns whether exposure to violent content increases aggression, and its principal theoretical frame is Albert Bandura's social cognitive theory. Bandura's account of mass communication holds that people acquire patterns of behavior by observing models, that mediated models function much as live ones do, and that whether an observed behavior is enacted depends on the incentives, self-regulation, and outcome expectancies the observer brings (Bandura, 2001). On this view a violent portrayal can teach a script, lower inhibitions by showing aggression rewarded or unpunished, and prime aggressive constructs, without requiring that every viewer imitate what they see. Meta-analysis of the experimental and correlational literature on violent video games reported a reliable association between exposure and increases in aggressive behavior, aggressive cognition, and aggressive affect, together with decreases in prosocial behavior, though the average effect sizes are modest and their real-world interpretation remains disputed (Anderson & Bushman, 2001). A review spanning media types distinguished short-term effects, driven by priming, arousal, and imitation, from long-term effects, driven by the observational learning and desensitization that repeated exposure can produce over development (Bushman & Huesmann, 2006). These two pathways were unified in Craig Anderson and Brad Bushman's general aggression model, which treats a violent portrayal as one situational input that raises aggression in the moment by priming aggressive cognition, arousal, and affect, and across development by the rehearsal and consolidation of aggressive knowledge structures, so that the same integrative machinery explains both a transient spike and a slow shift in disposition (Anderson & Bushman, 2002). Desensitization itself has been shown directly: participants exposed to violent media were subsequently slower to help an injured person and rated the emergency as less serious, a blunting of the normal response to others' distress (Bushman & Anderson, 2009). The size and durability of these effects, and how far laboratory measures of aggression generalize, are among the most actively debated questions in the field, and the modest average effect is exactly what the individual-difference models in the next section were built to explain.

From Uniform Effects to Differential Susceptibility

The history of media theory is in large part the story of abandoning the assumption that a message does the same thing to everyone. Patti Valkenburg and Jochen Peter formalized the alternative in the differential susceptibility to media effects model, which holds that media effects are conditional on dispositional, developmental, and social susceptibility variables that determine both which media a person selects and how strongly the content moves them, and that media use and its effects are related reciprocally rather than in one direction (Valkenburg & Peter, 2013). On this account the average effect measured across a whole sample can be small precisely because it averages over people whom the content strongly affects and people it does not touch, so the scientific task is to find the moderators rather than to estimate a single number. A comprehensive review placed this individual-difference logic at the center of a general framework for media effects, organizing the field around the media, the audience factors that condition response, and the reciprocal processes that link them (Valkenburg et al., 2016). Table 1 sets the major traditions side by side, from the early direct-effects image through cultivation and agenda-setting to the conditional and reciprocal models that dominate current work. The movement across the rows is a movement from a passive audience toward an active one, and from a single message toward the accumulated environment and the individual who selects within it.

Selective Exposure and Reinforcing Spirals

People do not receive a random sample of the media environment; they choose within it, and they tend to choose what agrees with them. The active-audience premise had already been formalized in the uses and gratifications approach, which asks what needs and satisfactions lead people to select particular media rather than treating them as passive targets, and selective exposure is its sharpest case. This selective exposure, the preference for attitude-consistent over attitude-challenging content, turns the causal arrow into a loop. Michael Slater's reinforcing spirals model formalized the loop: an attitude leads a person to select congenial media, that media strengthens the attitude, the strengthened attitude sharpens the selection, and so on, so that media use and its effects are mutually reinforcing over time rather than related as one-off cause and effect (Slater, 2007). The spiral is bounded, not explosive, because competing influences and the ceiling on an attitude eventually stabilize it, but it explains how relatively small per-exposure effects can accumulate into durable polarization within self-selected audiences. The rise of high-choice media environments sharpens the dynamic. Markus Prior showed that when cable and the internet gave audiences the ability to choose entertainment over news around the clock, the politically interested consumed more news and the uninterested largely abandoned it, widening gaps in political knowledge and turnout that a low-choice broadcast era had suppressed (Prior, 2005). Social media were expected to counter such fragmentation by exposing people to diverse views, but a field experiment found the opposite: paying partisans to follow opposing-view accounts on a social platform increased rather than decreased their political polarization (Bail et al., 2018). The demonstration below lets the reader set the strength of selectivity and the amount of cross-cutting exposure, and watch an attitude spiral toward an equilibrium that the two forces jointly determine.

Screens, Social Media, and Well-Being

The most publicly visible current question about media exposure is whether the heavy screen and social-media use of adolescents is harming their mental health. Jean Twenge and colleagues reported that increases in depressive symptoms, suicide-related outcomes, and suicide rates among United States adolescents after 2010 coincided with the rise of smartphones and social media, and that adolescents who spent more time on new-media screens were more likely to report mental-health problems (Twenge et al., 2018). The interpretation of such correlations is fiercely contested. Amy Orben and Andrew Przybylski applied specification curve analysis, which computes the effect across the thousands of defensible ways the data could be analyzed rather than a single chosen path, and found that the association between adolescent well-being and digital-technology use is real but tiny, small enough that its practical significance is doubtful and easily dwarfed by mundane comparisons (Orben & Przybylski, 2019). A narrative review of the evidence reached a similarly cautious verdict, stressing that effect sizes are small, causal direction is unresolved, and measurement of both screen use and well-being is poor (Orben, 2020). The engagement side of the ledger is likewise mixed: a meta-analysis of social-media use and political and civic participation found a positive but small overall relationship, consistent with modest mobilization rather than the transformation sometimes claimed (Boulianne, 2015). The pattern across these literatures is the recurring one of modern media research: aggregate effects that are statistically real, generally small, and heavily conditioned by who is using the medium and how.

Worked Example

The reinforcing-spiral idea, that selective exposure and media effect feed back on each other toward a bounded equilibrium, can be made exact with a simple dynamical model, the same one the demonstration animates. Let a person's attitude strength be a number A between 0 and 1. Each cycle, selective exposure to congenial media pushes the attitude up by an amount proportional to the current attitude and to the room left below the ceiling, r times A times the quantity one minus A, where r combines the selectivity of the person and the per-exposure effect of the media. Working against it, exposure to cross-cutting content pulls the attitude down by an amount d times A, where d is the strength of counter-attitudinal exposure. The attitude settles where the two forces balance, that is where r times A times one minus A equals d times A. Dividing both sides by A, the equilibrium satisfies r times one minus A-star equals d, so A-star equals one minus d divided by r. Take a selectivity-and-effect strength of r equal to 0.40 and a cross-cutting exposure of d equal to 0.10. Then A-star equals one minus 0.10 divided by 0.40, which is one minus 0.25, or 0.75. The attitude spirals up toward a strength of 0.75 rather than running away to its maximum, because the cross-cutting exposure grows with the attitude and eventually cancels the reinforcing push. Now cut the cross-cutting exposure in half, to d equal to 0.05, as a high-choice environment lets the person avoid opposing views: the equilibrium becomes one minus 0.05 divided by 0.40, or one minus 0.125, which is 0.875. Halving exposure to the other side moves the resting attitude from 0.75 to 0.875, a one-eighth shift of the entire scale, produced not by any change in the media's per-exposure force but purely by the reduction in cross-cutting contact (Slater, 2007). This is the formal core of the claim that fragmented, high-choice media can polarize an audience even when no single message becomes more persuasive.

Discussion

Media exposure begins as the plainest of variables, the amount of contact between a person and content, and ends as a window onto how minds build a model of a world they mostly cannot see. The theories surveyed here converge on a cognitive account of how that happens. Mere exposure shows that contact alone leaves an affective residue, biasing preference beneath awareness (Zajonc, 1968). Cultivation and agenda-setting show that the aggregate diet biases the mental sample from which judgments of prevalence and importance are drawn, an accessibility effect writ across a viewing history (Gerbner & Gross, 1976; McCombs & Shaw, 1972). Social cognitive theory shows that mediated models supply behavioral scripts whose enactment is gated by the observer's expectancies and self-regulation, which is why violent content raises aggression on average yet moves individuals so unequally (Bandura, 2001; Anderson & Bushman, 2001). And the reciprocal models show that the audience is not a target but an agent, selecting the media that confirm it and being confirmed in turn, so that effect and exposure spiral together (Slater, 2007; Valkenburg & Peter, 2013). The recurring empirical lesson is that average media effects are usually small, which is neither a refutation nor a triviality: a small per-person effect distributed across a population and accumulated over years can matter greatly in aggregate, while the same small average can hide large effects on the susceptible. The open problems are correspondingly about conditionality and measurement, above all in the screen-time debate, where the science has been limited less by the size of effects than by the difficulty of measuring exposure well and establishing its causal direction (Orben, 2020). What the field offers cognitive psychology is a running demonstration that a large part of the content of thought is furnished from outside, by the media a person happens, and chooses, to encounter.

Current Directions

The most active current work concerns method as much as theory, driven by the recognition that the screen-time debate had been distorted by selective analysis and weak measurement. Specification curve analysis, which reports an effect across every defensible analytic path rather than a single one, has reset expectations for the size of digital-media associations and become a template for credible analysis of large observational datasets (Orben & Przybylski, 2019). Alongside it, a wave of careful reviews has insisted that self-reported screen time is a poor proxy for actual use and that the field needs objective, content-sensitive exposure measures before its causal questions can be answered (Orben, 2020). A second front investigates the political consequences of algorithmically curated, high-choice media, where the surprising finding that cross-cutting exposure can deepen rather than soften partisanship has prompted a search for the conditions under which contact with opposing views backfires (Bail et al., 2018). A third continues to refine the conditional and reciprocal models, seeking the specific dispositional and developmental moderators that differential susceptibility predicts and that would convert small average effects into identifiable strong ones (Valkenburg & Peter, 2013). Across all three the connection to core cognition is tightening, as work on selective attention and on social cognition supplies the processing-level mechanisms that population-level exposure measures leave unspecified.

Common Misconceptions

A small average media effect means media do not matter.

A per-person effect that is small on average can be large in aggregate once it is distributed across a mass audience and accumulated over years, and the same small average can conceal strong effects on susceptible subgroups, which is the whole point of differential-susceptibility models (Valkenburg & Peter, 2013). Size and importance are not the same quantity.

Cultivation and agenda-setting mean the media tell people what to think.

Both theories are about the salience and accessibility of contents, not about persuading people to hold particular opinions: agenda-setting shifts what issues seem important, and cultivation shifts estimates of how common something is, by biasing what comes readily to mind rather than by winning an argument (McCombs & Shaw, 1972).

Exposing people to the other side's views will reduce polarization.

The intuitive contact remedy can backfire: a field experiment that paid partisans to follow opposing-view accounts found their polarization increased rather than decreased, and selective-exposure dynamics can turn added cross-cutting contact into entrenchment (Bail et al., 2018). Whether contact softens or hardens an attitude is itself conditional.

Glossary

Agenda-setting.

The tendency of the issues most heavily covered by the news media to become the issues the public judges most important, an effect on salience rather than opinion.

Cultivation differential.

The gap in beliefs about the real world, such as estimates of the prevalence of violence, between heavy and light viewers of television, the characteristic measure of cultivation.

Cultivation theory.

The theory that long-term, heavy exposure to television shifts viewers' conceptions of social reality toward the recurrent patterns of the television world.

Desensitization.

The blunting, through repeated exposure to violent media, of the normal cognitive, emotional, and behavioral responses to real violence and to others' distress.

Differential susceptibility to media effects.

The model that media effects are conditional on dispositional, developmental, and social variables that shape both media selection and the strength of response.

General aggression model.

An integrative framework in which media violence raises aggression through short-term priming, arousal, and affect and through the long-term learning of aggressive knowledge structures.

Mean world syndrome.

The chronic sense, attributed by cultivation theory to heavy television viewing, that the world is more dangerous and people less trustworthy than they in fact are.

Media effects.

The changes in what people know, believe, feel, or do that follow from exposure to media content, the general object of the research tradition.

Media exposure.

The extent and kind of contact a person has with mediated messages and content, measured by amount, channel, genre, or receipt of a specific message.

Mere exposure effect.

The increase in liking for a stimulus produced by repeated contact alone, without argument or reward, often strongest when the exposures are not consciously recalled.

Reinforcing spiral.

The mutually reinforcing loop in which an attitude drives selection of congenial media, which strengthens the attitude, which sharpens the selection, bounded by competing forces.

Selective exposure.

The tendency to choose media content that agrees with one's existing attitudes and to avoid content that challenges them.

Social cognitive theory.

Bandura's account of how people acquire behavior by observing models, including mediated ones, and enact it conditionally on incentives, expectancies, and self-regulation.

Specification curve analysis.

An analytic method that estimates an effect across all defensible ways of specifying the analysis, rather than a single chosen path, to gauge how robust the effect is.

Uses and gratifications.

The audience-centered approach that asks what needs and satisfactions lead people to select particular media, treating the user as an active chooser rather than a passive target.

Key Researchers

Craig A. Anderson. Psychologist at Iowa State University; he co-developed the general aggression model and led influential meta-analyses of the effects of violent video games on aggression. ORCID - Google Scholar - Wikipedia

Albert Bandura (1925-2021). Psychologist at Stanford University; his social learning and social cognitive theory established the observational learning of behavior from models, including mediated ones, and framed decades of media-violence research. Wikipedia

Brad J. Bushman. Communication and psychology scholar at The Ohio State University; his experimental and meta-analytic work on violent media, aggression, and desensitization is among the most cited in the field. ORCID - Google Scholar - Wikipedia

George Gerbner (1919-2005). Communication scholar and long-time dean of the Annenberg School for Communication at the University of Pennsylvania; he founded cultivation theory and the Cultural Indicators project, and introduced the mean world syndrome. Wikipedia

Amy Orben. Experimental psychologist at the MRC Cognition and Brain Sciences Unit, University of Cambridge; she brought specification curve analysis to the study of digital technology and adolescent well-being, resetting the field's estimate of effect size. ORCID - Google Scholar - Faculty Page

Jean M. Twenge. Psychologist at San Diego State University; her work links generational increases in adolescent depression and suicide-related outcomes to the rise of smartphones and social media. ORCID - Google Scholar - Wikipedia

Patti M. Valkenburg. Media psychologist at the University of Amsterdam; she developed the differential susceptibility to media effects model and leads research on conditional and reciprocal media effects in adolescence. ORCID - Google Scholar - Wikipedia

Robert B. Zajonc (1923-2008). Social psychologist at the University of Michigan and later Stanford University; he demonstrated the mere exposure effect and argued that affect can arise from exposure without prior cognitive appraisal. Wikipedia

Frequently Asked Questions

What is media exposure?
Media exposure is the extent and kind of contact a person has with mediated messages and content, measured in practice by how much media is used, through which channels, and how much of a particular message is received. In research it is the entry point for studying media effects, the changes in cognition, emotion, and behavior that follow from that contact (Valkenburg et al., 2016).

What is the mere exposure effect?
The mere exposure effect is the increase in liking for a stimulus that results from repeated contact alone, without any argument or reward. Zajonc showed that people rate novel words, symbols, and faces more favorably the more often they have seen them, and that the effect can occur even when the exposures are not consciously remembered (Zajonc, 1968).

What is cultivation theory?
Cultivation theory, developed by George Gerbner, holds that heavy television viewers come over time to see the real world as resembling the television world, for example by overestimating the prevalence of violence and developing a mean world syndrome. The gap between heavy and light viewers is called the cultivation differential (Gerbner & Gross, 1976).

What is agenda-setting?
Agenda-setting is the finding that the issues the news media cover most heavily become the issues the public judges most important. As McCombs and Shaw put it, the media may not tell people what to think, but they succeed in telling them what to think about, an effect on issue salience rather than opinion (McCombs & Shaw, 1972).

Does violent media cause aggression?
Meta-analyses report a reliable association between violent-media exposure and increases in aggressive behavior, cognition, and affect, alongside decreases in prosocial behavior, though the average effects are modest and their real-world meaning is debated (Anderson & Bushman, 2001). Social cognitive theory explains why the same content moves individuals unequally, through differences in incentives and self-regulation (Bandura, 2001).

What is selective exposure and a reinforcing spiral?
Selective exposure is the tendency to choose media that agree with one's attitudes and avoid those that challenge them. Slater's reinforcing spirals model describes how this creates a loop: an attitude drives media selection, the media strengthen the attitude, and the process repeats, producing durable polarization within self-selected audiences (Slater, 2007).

Is social media harming adolescent mental health?
The evidence is genuinely contested. Some analyses link post-2010 increases in adolescent depression and suicide to smartphones and social media (Twenge et al., 2018), but specification curve analysis finds the association between well-being and digital-technology use to be real but very small, and reviews stress that causal direction is unresolved and measurement is poor (Orben & Przybylski, 2019; Orben, 2020).

Why are media effects usually reported as small?
Average effects are small because they are computed across whole samples that mix people the content strongly affects with people it barely touches, and because per-exposure effects are genuinely modest. A small average can still matter greatly once accumulated over a mass audience and years, and can hide strong effects on susceptible subgroups (Valkenburg & Peter, 2013).

References