RESEARCH DOSSIER
Political Communication, Framing, and Media Effects
Abstract. Political communication shapes the availability and interpretation of information about institutions, public problems, and collective choices. This fundamental research monograph critically synthesizes the psychological study of agenda setting, attribute salience, news framing, responsibility attribution, issue persuasion, source credibility, elite frame competition, repetition, strategic news, perceptions of polarization, selective exposure, electoral communication, and observed civic participation. The research examines historical conceptual developments and primary experiments, field experiments, contemporary platform-intervention evidence, and the methodological boundaries of competing causal accounts. The resulting synthesis distinguishes changes in attention and interpretation from changes in policy support, candidate choice, and behavior; analyzes message–recipient–institution interactions; evaluates replication and generalization limitations; and identifies research questions concerning media fragmentation, comparative political systems, and human–artificial communication. This is a structured critical narrative review of published scholarship, not a preregistered systematic review, quantitative meta-analysis, or original experiment.
Keywords: political communication; media framing; political news; agenda setting; media effects; issue framing; political psychology; source credibility; partisan cues; public opinion; campaign persuasion; political mobilization; selective exposure; political polarization; content analysis; randomized experiments; Artificial Era
Political Communication as a Psychological Research Object
Political communication is the production, circulation, reception, and interpretation of messages concerning political authority, collective problems, institutions, competing interests, and public action. A psychological account investigates the route from available messages to attention, interpretation, judgments, affect, expressed opinions, and observable behavior. Communication occurs within institutions of journalism, government, campaigns, civic association, and ordinary conversation; its effects depend on both message properties and the social position of audiences. The present research object is the political message environment and its frame-dependent consequences. That specification preserves a common explanandum across news stories, campaign rhetoric, televised debate, and civic information while allowing multiple mechanisms and outcomes Entman (1993) Scheufele (1999).
Three outcome families organize this monograph. Agenda outcomes concern which problems people rank as important. Interpretive outcomes concern which considerations and causal explanations are made relevant when evaluating a problem. Behavioral outcomes concern measurable participation, choices, communication, or support for public policies. These outcomes may correlate without responding identically to exposure. A news message can make a topic prominent without changing policy preference, and a frame can alter which arguments citizens mention without changing their vote. Distinguishing outcomes at the beginning prevents a general term such as media influence from standing in for several distinct causal propositions McCombs & Shaw (1972) de Vreese (2004).
The review is a critical, structured narrative synthesis of landmark original studies, methodological analyses, books, and peer-reviewed reviews. It does not claim a preregistered search, systematic-review completeness, newly collected data, or a pooled effect estimate. Studies are evaluated for exposure assignment, temporal ordering, construct validity, confounding, population coverage, outcome measurement, and the distinction between a laboratory contrast and an everyday information environment. Particular attention is given to null or bounded effects because the conditions under which political communication fails to change attitudes are as theoretically informative as observed differences Chong & Druckman (2007b) Broockman & Kalla (2023).
Historical Formation of Media-Effects Research
Research on political media acquired distinctive empirical architecture through surveys connecting media content with public priorities, controlled comparisons of news presentation, and field studies of behavior. The Chapel Hill agenda-setting study by McCombs & Shaw (1972) linked the prominence of campaign issues in news to the perceived importance of issues among voters. Its historical contribution was to formulate a testable relation between media emphasis and issue salience. Because the original design was correlational, shared responsiveness to the campaign agenda and prior audience interest remain relevant alternatives to simple one-directional influence.
A different experimental tradition asked how the presentation of the same problem changes judgments. The framing experiments of Tversky & Kahneman (1981) demonstrated preference shifts when decision problems were expressed through different descriptions of outcomes. Their work concerned decision framing, which provides a psychological analogue for political message framing but should not be equated automatically with news-production practices. Political communication research subsequently developed accounts of how issue attributes, responsibility assignments, and competing principles enter audience reasoning. The experiments discussed in Iyengar (1991) made the distinction between episodes centered on individual cases and thematic reports centered on social conditions especially consequential.
The conceptual field expanded through Entman (1993), who connected selection and salience with problem definition, causal interpretation, moral evaluation, and possible remedies. Scheufele (1999) later distinguished media frames from audience frames and specified frame building, frame setting, individual processing, and feedback. This progression is important because it moved research from descriptions of news stories toward observable stages and causal mechanisms. Earlier classics remain foundational historical sources; their strongest claims require evaluation in contemporary, more selective and fragmented media environments Arceneaux & Johnson (2013).
The Concept of a Frame and Its Scientific Boundaries
A political frame is an organizing pattern in a message that makes some features of an issue more accessible, diagnostically relevant, or normatively significant than others. It can foreground economic consequences, rights, public safety, personal responsibility, institutional causes, conflict, or collective welfare. Frames are identified through explicit claims, recurring interpretive devices, source selection, metaphors, exemplars, and relations among problem, cause, evaluation, and remedy. An experimental frame becomes a manipulable stimulus only when its distinctive content and the common information held constant are documented. Entman (1993) provides the canonical conceptual vocabulary, and de Vreese (2005) explains how generic versus issue-specific frames can support comparison.
Frame is frequently used for incompatible levels of analysis. A researcher may refer to a journalist's organizing convention, a politician's rhetorical intervention, a participant's mental representation, or an analyst's coding category. Borah (2011) documented persistent conceptual variation in a broad review of published framing studies. A valid design therefore states who constructs the frame, who encounters it, what feature the frame changes, how exposure is assigned, and which response qualifies as an effect. Content prevalence in a corpus establishes message distribution; it does not independently establish that a recipient noticed, understood, or acted upon a frame.
Political framing has a special boundary with decision framing. When Tversky & Kahneman (1981) compared descriptions of equivalent choice problems, mathematically equivalent formulations could produce different preferences; political frames often juxtapose substantively distinct considerations instead. A free-speech argument and a public-order argument, for example, can be simultaneously valid considerations rather than equivalent numerical formulations. Political effects thus require studying the weighting of competing principles, persuasive credibility, and the strength of existing commitments. The model of competitive framing developed by Chong & Druckman (2007a) specifically addresses citizens encountering alternative formulations, making the social ecology part of the causal design.
Agenda Setting, Issue Salience, and Media Prominence
Agenda setting refers to relations between issue prominence in communications and issue importance attributed by audiences. Operational measures include space or airtime devoted to a topic, positional prominence, frequency across media outlets, unaided mentions of the most important public problem, and survey rankings. The familiar formulation traces to McCombs & Shaw (1972), but modern causal claims require attention to the possible direction of influence. Political actors may respond to public concerns; audiences can seek information about already salient problems; major events can cause news attention and citizen concern simultaneously. A time sequence, experimental intervention, or credible quasi-experiment helps adjudicate these alternatives.
Media salience, issue importance, emotional urgency, and policy preference are separate constructs. Two citizens may agree that housing affordability is the most important political problem while endorsing different interventions. A research design that records only issue rankings can identify an agenda effect but cannot infer consensus or policy persuasion. Media systems also present several agendas: reporters, political parties, local communities, social networks, and personal relationships jointly influence what is available and attended to. This complexity makes comparative and longitudinal measurement especially important. The interpretive salience addressed by Nelson & Oxley (1999) represents an additional process, in which beliefs already held can receive different decision weights.
Agenda-setting research becomes sharper when issue attention is analyzed as a finite resource. Communication reallocates scarce attention among many public problems; apparent growth in one issue's ranking can reflect declining priority for others. Repeated exposure may also make a problem seem familiar or politically consequential without increasing factual understanding. The researcher must separate recall, perceived importance, perceived severity, and willingness to act. As Scheufele (1999) emphasizes in his broader model of media effects, repeated contact with a frame can modify both the message environment and subsequent individual processing, making feedback and selection essential parts of a causal description.
Attribute Framing, Priming, and Consideration Weighting
Agenda setting concerns the question of what receives attention; attribute emphasis concerns which qualities of an issue or actor become prominent; and priming commonly denotes the altered relevance of accessible considerations during subsequent judgments. Framing theories overlap with these processes but generally include interpretive organization and possible shifts in weighting rather than only memory accessibility. Chong & Druckman (2007b) examines the theoretical relations among framing and priming. An empirical test must observe whether treatment changes retrieval, perceived importance of a consideration, factual beliefs, or the final preference; an attitude difference alone does not discriminate between mechanisms.
Nelson & Oxley (1999) used laboratory experiments to distinguish changes in belief content from changes in the subjective importance attached to beliefs. Their results support the idea that issue frames can influence opinion through more than one psychological pathway. A citizen may already know about both individual and institutional causes of a public problem; a particular story can change how much each cause matters in forming a judgment. This mechanism has different implications from teaching new facts. It predicts sensitivity to issue commitment, cognitive accessibility, and the presence of competing accounts, while leaving open the possibility that the same evidence evokes distinct judgments across audiences.
The content of thoughts is an intermediate measure when researchers wish to observe interpretive activation. In experiments by Price et al. (1997), participants read news stories containing substantially shared information but different framing material. Their reported thoughts differed in thematic focus and evaluative implications. Thought-listing tasks are useful but also reactive: asking respondents to articulate thoughts can alter processing. Investigators should prespecify whether a thought category is an outcome, a proposed mediator, or simply an exploratory diagnostic, and should distinguish reliable human coding from post-treatment overinterpretation. The empirical value lies in connecting treatment variation to a defined processing claim.
Episodic, Thematic, and Responsibility Framing
Episodic reports organize an issue around particular people or events, whereas thematic reports place the same issue in statistical, structural, or historical context. This distinction became central to Iyengar (1991) and the subsequent synthesis by Iyengar (1996), which associated episodic television framing of national problems with greater emphasis on individual responsibility. The theoretical claim is about attribution: the level at which a problem is represented changes which actors appear causally responsible and which remedies seem appropriate. Examples do more than illustrate an issue; they may define the apparent unit of explanation.
The relation between report format and responsibility judgment is affected by the facts included in a story. A thematic account may introduce policy history, prevalence statistics, and institutional actions that are absent from a case-focused report. Such differences could mediate a framing effect or change the informational treatment itself. Rigorous experimental contrasts hold as much descriptive content constant as the research question allows and separately evaluate factual learning. In field content analyses, assignment to episodic or thematic form also requires reliable rules for mixed stories, since many reports combine individual narratives with systemic context Iyengar (1996) de Vreese (2005).
Responsibility frames have implications for how citizens represent institutions, but conclusions about accountability must follow the outcome actually measured. An increase in individual blame on a survey does not by itself show decreased electoral accountability, turnout, or trust in governance. Those outcomes require additional observations and causal testing. The crucial contribution of the episodic–thematic distinction is a reproducible question: which causal unit does a message encourage an audience to consider? Its contemporary relevance extends to crisis reporting, public health, environmental policy, and economic narratives, where individual stories and structural evidence may be both important and differently salient.
Frame Building in Journalism and Political Institutions
Frame building concerns the social production of interpretive patterns before they reach an audience. Journalists work with deadlines, beats, editorial conventions, sources, institutional incentives, professional norms, and event characteristics; political actors offer explanations that align with their interests and identities. News content emerges from interaction among these actors. The process model developed by Scheufele (1999) distinguishes frame building from the subsequent question of whether a frame changes recipient interpretation. That distinction provides a methodological requirement: research on why a story adopts a frame needs evidence about news production, whereas research on audience effects needs observed exposure and outcomes.
Issues can move between science, governance, and journalism through successive struggles over interpretation. The historical content study by Nisbet et al. (2003) traced news attention, frames, and sourcing around the stem-cell controversy in major American newspapers. Such research establishes patterns in the public record and generates hypotheses about the institutional conditions that produce them. Because sources and reporters are selected rather than randomly assigned, content analysis alone cannot identify which actor caused a particular frame to dominate. Interviews, document analysis, news-production observation, and natural experiments can improve process explanation while preserving the difference between descriptive prevalence and causal influence.
Generic conflict, responsibility, economic consequences, and human-interest frames can be compared across issues, while issue-specific claims may capture political meaning with greater precision de Vreese (2005). An adequate coding scheme defines a frame as a coherent set of features, specifies units of analysis, and permits reports containing multiple frames. It treats the prominence of a quoted source as a separate variable from whether the source's claims are endorsed by the reporter. It also measures omissions and counterarguments where feasible. These distinctions matter when the scientific objective is to examine how institutional communication renders certain causal explanations ordinary and others peripheral.
The Role of Source Credibility and Political Sponsorship
A political claim is encountered together with cues about who advances it. Source identity can signal expertise, partisan membership, trustworthiness, authority, or conflict of interest. In two laboratory experiments, Druckman (2001) examined credibility as a limit on elite framing effects. The studies show why a frame's verbal content cannot be treated as its complete causal identity. A credible source may increase the probability that an otherwise familiar argument is accepted, while a distrusted source may provoke counterargument, rejection, or reinterpretation. The relevant inference is conditional on both the claim and the source.
Party sponsorship can activate motivated reasoning and group commitments. Slothuus & de Vreese (2010) embedded experiments in a nationally representative survey and found that issue framing differed according to party sponsorship and respondents' orientations. This work extends message-only designs into political environments where citizens recognize competing parties and connect new claims to prior loyalties. It also complicates the notion of political awareness as a uniformly protective resource: political knowledge can facilitate recognition of source identities and production of arguments consistent with prior commitments. Therefore investigators should measure prior partisanship, issue salience, political knowledge, and source credibility before exposure where possible.
Credibility and sponsorship must be separated analytically. A party cue does not always convey factual expertise, and a source can be recognized without being regarded as trustworthy. Experiments crossing argument content with source labels can isolate these dimensions, while observational surveys often conflate them. The dependence of effects on source credibility does not imply that all receivers are passive; it demonstrates that social inference accompanies message interpretation. Attribution of causal effects must also consider whether experimentally presented sources are plausible within the relevant country's party system, whether respondents recognize them, and whether stated attitudes persist after an intervening debate Druckman (2001) Slothuus & de Vreese (2010).
Competitive Frames and the Structure of Public Debate
Citizens encounter multiple messages, often presenting incompatible grounds for judgment. The competitive-framing account of Chong & Druckman (2007a) addresses this situation directly by examining how alternative elite frames operate when exposed together. A single-frame experiment demonstrates that a particular consideration can matter when foregrounded; it does not establish durable control over public opinion in a plural media environment. Competition changes the relevant unit from an isolated message to a configuration of arguments, sources, timing, and audience commitments. The central research question becomes which considerations survive comparison and which prevail only under temporary exclusion of alternatives.
Strength is a critical feature of competitive frames. An argument connected to accepted values, relevant facts, and a coherent policy implication may be more persuasive than a weak or implausible alternative; numerical symmetry in exposure does not imply equal interpretive force. The review by Chong & Druckman (2007b) emphasizes methods for identifying and comparing frames and for understanding the psychological process behind opinions. Researchers should avoid selecting the strongest argument for one side and a caricature for another. Pilot studies can assess perceived strength without using the same respondents to calibrate and test the final contrast.
Competition can also induce deliberation, ambivalence, or stable preference even when it changes the considerations citizens articulate. Measuring only final agreement may miss these processes. Designs can compare immediate and delayed opinions, confidence, argument recall, and perceived legitimacy of alternative reasons. Yet post-treatment mediators require cautious interpretation because the message can affect several components at once. Competitive framing is a theory of communication within a social setting, and its ecological validity depends on whether real institutions actually deliver the opposing arguments in the timing, prominence, and credibility assumed by the experiment Chong & Druckman (2007a).
Repetition, Recency, and Temporal Persistence of Frames
Political communication is repeated through campaigns, commentary, and follow-up reports. A single exposure may produce a brief shift that disappears, whereas recurring messages may establish a more persistent interpretive routine. Repetition should therefore be treated as a design factor rather than an assumed multiplier. In an experiment with 1,324 participants, Lecheler & de Vreese (2013) found that repeated frames did not systematically strengthen effects on opinion, while shorter intervals between exposures sometimes mattered. Under competitive conditions the most recently encountered frame was particularly influential. These results support temporal specificity rather than a universal accumulation law.
The same study found that political knowledge moderated temporal patterns: more knowledgeable participants were less susceptible to recency effects in that design and showed greater signs of cumulative response. Such heterogeneity requires a theoretical account of attention, elaboration, and prior beliefs. Someone who can integrate new claims with extensive issue knowledge may interpret repetition differently from someone encountering the issue for the first time. Yet knowledge measures are not interchangeable with education or analytic ability, and subgroup findings should be examined for precision, multiplicity, and replication Lecheler & de Vreese (2013).
Persistence is an independent outcome. If opinion is recorded immediately after a news item, investigators have measured an immediate effect, even when a campaign may run for months. Delayed measurement introduces attrition, unobserved media exposure, and social discussion that can obscure a treatment's residual influence. Such observations are themselves informative about real-world competition. Designs that document intervening exposure, use repeated waves, and distinguish panel conditioning from substantive attitude change are better positioned to estimate durability. The research synthesis should report when a treatment altered a temporary interpretation, when it shifted a stated evaluation, and when any observable change survived later contact with competing arguments.
Strategy, Game, and Policy Frames in Political News
Political news frequently organizes campaigns around electoral competition, tactics, polls, winners, and strategic calculations. A strategy or game frame contrasts with explanations emphasizing public problems, proposed policies, institutional responsibilities, and consequences for citizens. Cappella & Jamieson (1997) developed an influential account linking strategic political coverage to cynicism toward political actors and diminished public-oriented engagement. Its importance lies in identifying a mechanism through which reports can teach people how to interpret political activity, not merely which candidate is ahead. Conclusions about generalized cynicism should nevertheless be evaluated against alternative selection and preexisting distrust.
The distinction between game and strategy framing has itself been measured inconsistently across studies. Some schemes include horse-race indicators, others emphasize politicians' motives or political maneuvering, and still others distinguish gaming from strategic negotiation. Conceptual heterogeneity makes results difficult to compare across electoral systems and news formats. A robust study should specify the coding criteria, estimate coder agreement, distinguish election coverage from ordinary policy reporting, and examine whether strategic elements coexist with substantive policy information. The prevalence of a game frame does not alone establish its effect on citizens de Vreese (2005) Borah (2011).
Effects of strategic framing may depend on prior engagement and the availability of issue-centered reporting. A viewer can simultaneously learn information about coalition formation and become more skeptical about political motives; neither outcome should automatically be classified as civic deterioration. Behavioral indicators such as participation, issue knowledge, information seeking, and institutional trust provide richer evidence than a single feeling thermometer. Research should also separate the effects of a story's political content from the tendency of already cynical individuals to select strategy-heavy coverage Arceneaux & Johnson (2013) Cappella & Jamieson (1997).
Polarization Narratives and Perceived Political Division
News can portray a society as sharply divided even when citizens' actual attitudes display a more complicated distribution. Exposure to narratives of widespread polarization may alter perceptions of fellow citizens, the expected severity of political conflict, and attitudes toward opposing groups. Levendusky & Malhotra (2016) investigated the consequences of media coverage of partisan polarization and reported that such coverage increased perceptions of an ideologically divided electorate. Their reported results also differentiated issue-position moderation from greater dislike of the opposing party. This is a decisive reminder that perceived issue distance and affective hostility are separable psychological constructs.
The mechanisms linking coverage to perceived division include salient exemplars, implied social norms, social identity, and inference from highly visible elites to ordinary voters. A news story featuring extreme representatives may encourage overgeneralization if it provides little information about the wider population. Yet the effect depends on who is represented and how explicit the report is about prevalence. Political communication researchers therefore need separate measures for estimated population polarization, perceived partisan stereotypes, personal issue positions, and social affect. Treating them as one undifferentiated scale would conceal opposing treatment responses Levendusky & Malhotra (2016) Jost et al. (2022).
The integrative account of Jost et al. (2022) places information sources, messages, and channels inside motivational dynamics involving personal commitments, group identity, and support for existing systems. It is useful for designing moderator tests, while the direct causal support for each proposed link must be sought in the underlying studies. Perceived polarization should be measured against external distributions where reliable data exist, rather than defined solely by how divided respondents feel. The possibility that coverage changes citizens' expectations of others also creates a longitudinal question: do altered perceptions influence subsequent conversations, social avoidance, and the kinds of political news people choose?
Audience Selection and the Problem of Voluntary Exposure
Audiences select news according to interest, available time, prior attitudes, trust, identity, social routines, and the alternatives in the surrounding media market. Consequently, a difference between people who watch competing political channels is simultaneously a possible consequence of messages and a possible consequence of selection. The research program developed in Arceneaux & Johnson (2013) distinguishes willingness to encounter partisan news from the effects of assigned exposure. Media choice is itself an outcome worth explaining: citizens can strengthen their existing opinions by actively seeking congenial sources, or they can avoid political news altogether. Experiments with forced exposure answer a different question from studies in which participants can opt out.
Selective exposure changes the population for whom a message effect is estimated. A controlled laboratory sample can reveal what a message does to receptive or assigned viewers, but actual media audiences may consist disproportionately of people who already endorse its perspective. Field experiments should record take-up, noncompliance, alternative exposure, and the possibility that people search elsewhere after assignment. The causal estimate from an offer of exposure differs from a hypothetical estimate for viewers who would have watched it regardless. Conflating these quantities can make media effects appear either universal or negligible depending on design Arceneaux & Johnson (2013) Guess et al. (2023).
Repeated political news consumption also interacts with interpersonal communication. A recipient may hear an argument through family, coworkers, or civic organizations after initially encountering a televised frame. These social pathways can sustain or revise interpretations without being attributable solely to the original medium. Research questions should distinguish the immediate information treatment, the later diffusion of talking points, and the independently chosen channels through which people revisit political topics. Longitudinal panels can trace such processes, but without randomization they require careful assumptions about time-varying preferences and unobserved events.
Campaign Persuasion and the Difference Between Attitude and Vote
Political campaigns attempt to influence candidate evaluation, issue position, turnout, contribution, interpersonal discussion, and actual voting. These outcomes have different behavioral costs and decision structures. Kalla & Broockman (2018) combined a meta-analysis of 40 field experiments with nine additional experiments and reported that average persuasive effects of general-election campaign contact on candidate choice were approximately zero in the studied circumstances. This finding challenges any inference from memorable campaign rhetoric to widespread conversion of vote choice. It does not imply that every message, election, or response domain has an identical null effect.
The message's informational content and the audience's prior knowledge matter. In four survey experiments during the 2020 U.S. presidential election, involving 113,742 participants and 291 messages, Broockman & Kalla (2023) found evidence consistent with information-based updating: voters were more responsive when messages concerned candidates about whom they knew less, and informative messages could change evaluations and choices. This study clarifies why population averages can be small while specific subgroups or informational conditions exhibit detectable effects. Importantly, evaluations reported in surveys and verified final ballots are not interchangeable outcomes.
Campaign research needs credible treatment definition, outcome measurement, and time windows. An advertisement may change awareness, issue knowledge, interpersonal discussion, or short-lived enthusiasm without changing a durable vote choice. Conversely, effects on marginal groups may matter within a close contest even if their average magnitude is small; such electoral implications require election-specific evidence. Tests should report assignment, message reach, prior candidate familiarity, baseline preferences, attrition, and the possibility of interference across participants. This monograph treats such studies as evidence about message effects in specified settings rather than as a general manual for campaign optimization Kalla & Broockman (2018) Broockman & Kalla (2023).
Political Mobilization and Observed Participation
Political communication can seek action rather than persuasion. A person who already supports a candidate may nonetheless require information, reminders, social encouragement, or practical opportunities to vote. The field experiment by Gerber & Green (1999) tested nonpartisan face-to-face mobilization and found increased turnout in the studied population. Its methodological importance is the use of an observable behavioral outcome with a randomized field design. The experiment also illustrates why media and interpersonal communication should be distinguished when interpreting effects: personal contact can engage social expectations unavailable in a broadcast advertisement.
Mobilization outcomes include validated turnout, event attendance, contact with representatives, civic contribution, and other documented behavior. Self-reported intention or survey enthusiasm is a different variable, often susceptible to measurement and social desirability biases. An experimental comparison of canvassing, mail, and telephone contact also shows why channels cannot be assumed equivalent: interaction may change message credibility, attention, and reciprocal commitment Gerber & Green (2000). The action of voting is influenced by laws, transportation, registration, and perceived legitimacy, so communications are embedded in concrete opportunities and constraints.
For a research dossier centered on political framing, mobilization evidence provides a boundary condition rather than replacing the central object. Framing can change the meaning of political participation, but the effect of a reminder or a doorstep conversation can operate without complex argument framing. To attribute a behavioral effect to framing, an experiment would need to hold contact format and opportunity approximately constant while varying an explicitly specified interpretive message. An observed effect of an entire mobilization package instead identifies the combined intervention.
Randomized Message Experiments and Causal Identification
Randomized message experiments manipulate one or more aspects of communication and compare outcomes across assigned groups. Properly implemented randomization supports causal attribution for the treatment contrast in the experimental population. A design might hold policy facts constant while varying responsibility language, or cross a framing argument with a credible and noncredible source. Studies such as Druckman (2001), Nelson & Oxley (1999), and de Vreese (2004) illustrate the variety of mechanisms that can be isolated by assigning different message versions. The estimand must be stated precisely: the difference between specified stimuli under specified exposure conditions, not an unrestricted estimate of the power of mass media.
Treatment validity is often the major methodological challenge. If one message includes more evidence, emotional language, rhetorical force, or length, a difference in opinion may reflect those features rather than the intended frame. Experimenters should disclose stimulus materials, use suitable pretests, and measure recognition and attention without conditioning causal estimates improperly on post-treatment variables. Manipulation checks can illuminate engagement, but excluding everyone who fails a post-exposure check can introduce selection bias. Factorial designs help identify interactions, although the resulting effects may be imprecise when the sample is divided across many conditions.
The design must also anticipate demand characteristics and question wording. Participants who infer the research hypothesis may provide responses shaped by perceived expectations rather than persistent civic judgments. Order effects, interviewer cues, and survey framing can create additional treatments inside the measurement process. Delayed outcomes, behavioral validations, heterogeneous respondent samples, and transparent reporting of all tested endpoints improve inference. A well-powered null finding is itself evidence about a specified treatment; it should not be interpreted as proof that framing is absent from political communication generally de Vreese (2004) Lecheler & de Vreese (2013).
Content Analysis, Frame Detection, and Measurement Reliability
Content analysis measures how frequently and in what form frames appear in a defined communication corpus. Researchers may sample news programs, articles, speeches, institutional statements, or publicly visible messages, then apply rules for identifying frame elements. A good codebook separates topics from frames, source quotations from journalists' evaluations, and explicit blame from implicit causal suggestions. Borah (2011) shows why inconsistent conceptualizations compromise cumulative inference across framing studies. The measurement unit—article, paragraph, headline, image, claim, or exposure event—must be matched to the scientific question.
Human coding requires a sampling scheme, coder training, blinded or independent classification where possible, and reportable agreement. High overall agreement on a common category can conceal poor identification of rare frames, while agreement statistics do not establish conceptual validity. Automated text coding introduces additional decisions about labeled training data, dictionary assumptions, cross-topic performance, and drift in language use. Both human and computational classification can fail when sarcasm, mixed frames, or context-dependent political references change the meaning of a phrase. Case studies such as Nisbet et al. (2003) demonstrate the historical strength of traceable corpora for identifying changes in attention and sourcing.
A high frequency of a frame in media archives is not the same as high human exposure. An article can be published but never read, seen but ignored, or understood differently by different audiences. Linking coded messages with readership, viewing behavior, attention indicators, or panel responses improves ecological relevance. Researchers should preserve the provenance of public materials, local context, dates, amendments, and platform-specific presentation. Comparisons across countries also need translation rules and measurement invariance; a responsibility frame may carry different institutional implications when the relevant actors and welfare arrangements differ.
Digital News Environments and Boundary Evidence from Field Trials
Digital distribution changes who sees political messages, how repetition is organized, and which reactions are observable. Platform interfaces can aggregate professional reporting, personal commentary, campaign statements, and social cues into a single stream. Large experiments nevertheless show that sizable changes in content exposure need not produce equally large changes in political attitudes. In research during the 2020 U.S. election, Guess et al. (2023) experimentally replaced personalized ranking with reverse-chronological feeds for consenting users and documented changes in exposure and engagement, while finding no significant change in measured issue polarization, affective polarization, and several other attitudes during the three-month study period.
Related experiments reducing exposure to politically like-minded sources changed the composition of Facebook content without measurable changes in several preregistered political attitude outcomes Nyhan et al. (2023). These studies are important boundary evidence for claims that information environments mechanically produce opinion change. Their inference is specific to the tested platforms, intervention, population, electoral period, and measured outcomes. Null estimates should be interpreted with uncertainty intervals and compliance information. A lack of detected movement in survey measures does not imply that ranking has no consequences for attention, participation, or long-term political learning.
Platform algorithm interventions constitute a neighboring research object: their primary manipulation concerns assignment or ranking of exposures, whereas this dossier examines the political messages and frames encountered. The connection is scientific because ranking affects the distribution of frames and their potential audience. A rigorous joint design would combine exposure-level records, reliable message framing, and randomized assignment where feasible. It would distinguish effects of the selection system from effects of communicative content and from interactions between the two. The general inference is that communication effects are conditional on a selection architecture, rather than being separable from the route by which a message reaches citizens Guess et al. (2023) Arceneaux & Johnson (2013).
Narratives, Exemplars, and Visual Representations
Political information is frequently organized as a narrative involving actors, motives, conflict, consequences, and an implied explanation of responsibility. A narrative can provide a sequence through which an audience understands a complex problem; a vivid exemplar can also become a misleading substitute for a population distribution. The distinction between episodic and thematic reporting studied by Iyengar (1991) and Iyengar (1996) supplies a historical framework for investigating the psychological consequences of case-centered presentation. When a report illustrates an institutional problem through one individual, the example may direct attention toward personal choices, particular harms, or identifiable sources of assistance. The outcome should be evaluated through measured attributions rather than inferred from narrative form alone.
Images and audiovisual sequences add another layer of selection. A news report may present a politician alongside approving or disapproving crowds, or use photographs of a specific incident while discussing a large-scale social trend. These choices establish a perceptual context for claims, potentially making some actors or consequences more salient. Yet a coding system that detects a photograph of a protest does not automatically establish the picture's evaluative meaning. Captions, sequencing, speech, historical context, and audience familiarity can all alter interpretation. The distinction between frame presence and audience response, formalized by Scheufele (1999), becomes especially important in multimedia research.
To distinguish narrative content from rhetorical framing, researchers may compare versions that preserve the event and its factual detail while varying causal emphasis, or compare case-centered and statistical accounts while measuring comprehension and attribution. Such designs should test whether the narrative also changes emotional involvement, perceived vividness, or recall. A significant difference in recall could be a mechanism of later opinion change, a separate outcome, or a competing explanation for the apparent treatment effect. The framing tradition supplies hypotheses about these pathways, but each pathway requires direct operationalization Price et al. (1997) Nelson & Oxley (1999).
Emotion, Appraisal, and Political Message Reception
Political messages often include affectively meaningful references to threat, injustice, solidarity, uncertainty, or hope. The psychological relevance of emotion lies in the relationship between appraisal and judgment: different interpretations of an event can change what consequences appear urgent and which actions seem appropriate. This possibility fits within the broader framing definition of Entman (1993), which includes problem definition, causal interpretation, evaluation, and treatment recommendation. It also suggests that emotion should be measured separately from opinion. An emotionally involving message can alter attention or memory without producing the predicted change in policy endorsement.
Political identities may determine whether a purported threat is accepted, dismissed, or treated as evidence against a communicator. The source experiments of Druckman (2001) and the party-sponsorship research of Slothuus & de Vreese (2010) suggest that sender credibility and partisan fit interact with evaluation. A rigorous account should assess whether emotional responses occur before or after a stated judgment and should distinguish momentary feelings from durable emotional dispositions. Survey measures collected immediately after exposure are vulnerable to common method variance; physiological or behavioral measures may offer additional evidence while introducing their own validity problems.
Emotion is also implicated when media describe the public as hostile or polarized. Levendusky & Malhotra (2016) found that reporting on partisan polarization could produce both greater perceptions of division and increased dislike of political opponents in the tested setting. Such a result does not establish a single universal emotional mechanism. Identity inference, expectations about intergroup behavior, and changes in perceived norms offer distinct explanations. Studies should therefore separately measure perceived social distance, issue disagreements, and affect toward individuals or groups, preferably before and after a precisely defined exposure.
Information Accuracy, Misleading Frames, and Epistemic Outcomes
A frame may organize accurate information, selectively emphasize true but unrepresentative facts, or carry factual errors. These states should be distinguished because the psychological effects of a frame cannot be assessed solely through agreement with its conclusion. An accurate thematic account of a public problem may change beliefs by adding new context, while an inaccurate account can generate a superficially similar attitude response through false premises. The work of Nelson & Oxley (1999) shows that framing can operate through changes in belief content as well as consideration importance; experimental designs should measure both if they wish to identify how factual and interpretive information contribute to judgment.
Political messages may also produce perceptions of public agreement that depart from actual distributions. Reporting on polarization, studied by Levendusky & Malhotra (2016), is therefore an example of the broader epistemic problem: audiences infer the characteristics of a population from selected portrayals. Researchers can compare perceived distributions with independently measured opinions, while recognizing sampling uncertainty and the possibility that respondents interpret a scale differently. When the research question concerns misinformation or truth discernment as independent cognitive processes, it belongs to a neighboring dossier. The present question is how a politically consequential frame organizes factual claims and affects interpretation.
The scientific treatment of accuracy also requires distinguishing the source's factual assertion, the analyst's assessment of that assertion, and the recipient's belief. A content analysis should preserve the original claim and its context, while any accuracy classification should rely on transparent evidence and temporal relevance. A claim that was accurate before a policy change may no longer describe current conditions. In experiments, researchers should take care that misleading treatments do not impose uncompensated harm and should consider corrective information and ethical debriefing. Credible research can distinguish the political effectiveness of a message from the reliability of the worldview it encourages Entman (1993) Borah (2011).
Survey Wording, Response Scales, and Measurement Invariance
Research on political communication frequently measures effects through survey questions, but the survey itself is a communication encounter. The precise wording of a candidate evaluation, policy description, or perceived problem can foreground considerations similar to those manipulated in the treatment. The framing experiments of Tversky & Kahneman (1981) show the general sensitivity of judgment to presentation, while Nelson & Oxley (1999) demonstrates the importance of measuring beliefs and their relative weights separately. To identify a media-treatment effect, researchers must ensure that the outcome questionnaire does not inadvertently introduce an additional, asymmetric frame across experimental groups.
Measurement design should distinguish issue knowledge, frame recognition, perceived argument strength, emotional response, opinion, confidence, and behavior. A single agreement item often blends substantive preference with interpretation of the statement and social desirability. Multi-item measures can improve reliability but may still combine concepts that respond differently to political news. Surveys administered before treatment can sensitize respondents to the focal issue, while repeated post-treatment surveys can encourage elaboration. Experimental reports should document question order, response options, missing data, exclusion criteria, and the handling of undecided respondents. Such transparency makes the reported result interpretable as a change in a specific measurement rather than an undefined shift in public opinion.
Cross-cultural measurement adds translation, party-system context, and institutional referents. A question about trust in government does not necessarily represent the same underlying judgment when different governmental structures or histories are salient. Likewise, a generic responsibility frame may be interpreted through different causal expectations. Even when a framing treatment is identical in wording, its meaning may vary by audience. Comparative studies should therefore use cognitive interviews, measurement-equivalence checks where possible, and locally validated scales rather than assuming textual sameness ensures construct sameness de Vreese (2005) Scheufele (1999).
Causal Pathways, Mediation, and Interpretation of Null Effects
A political message can cause more than one downstream response: attention, recall, perceived source credibility, changed issue considerations, emotional appraisal, and declared preference may all be affected. These variables often influence one another, but their temporal order and conditional relations are not established by collecting them in a single post-treatment questionnaire. The mechanism-focused experiments of Nelson & Oxley (1999) are valuable because they distinguish changes in belief importance from changes in belief content. More complex mediation analysis requires assumptions about confounding between mediator and outcome, and such assumptions may remain untestable even when treatment assignment is randomized.
Null effects also require precision. The television-news study by de Vreese (2004) reported differences in issue interpretation but no corresponding policy-support difference in its experimental comparison. The campaign evidence synthesized by Kalla & Broockman (2018) indicates limited average effects on candidate choice in the general-election settings studied. Contemporary platform interventions Guess et al. (2023) Nyhan et al. (2023) likewise report substantial changes in exposure without detectable movement in several measures of political attitudes. These results define boundaries on certain treatment–outcome relations. They do not justify a sweeping declaration that news framing, electoral communication, or information distribution cannot matter.
A rigorous null-result appraisal examines confidence intervals, the smallest effect of scientific interest, treatment adherence, baseline stability, sample composition, and follow-up period. An underpowered study can fail to distinguish a meaningful effect from zero; a precisely estimated small effect can impose much stronger constraints on theory. Moderators should be analyzed through appropriately powered interactions and independent tests. The causal estimand must also account for whether the outcome is a temporary survey answer or an actual civic behavior. Clear separation of these claims creates cumulative knowledge across studies that otherwise seem contradictory.
Unresolved Research Questions and an Integrative Agenda
Several questions remain unresolved in the study of political communication. One concerns persistence: how long does an induced interpretive emphasis survive contact with competing messages, interpersonal discussion, and subsequent events? Another concerns transportability: whether a frame effect found in one issue, language, institutional setting, or population remains similar when any of those dimensions change. The experimental findings of Chong & Druckman (2007a) on competition and Lecheler & de Vreese (2013) on repetition suggest that these questions must be answered through multi-wave, multi-message designs rather than extrapolation from isolated single-exposure comparisons.
A second set of questions concerns the distribution of agency in political message environments. Journalists and political sources construct frames, audiences select or resist them, and technical systems alter which messages are encountered. The framework of Scheufele (1999) provides a useful separation between building, setting, and processing; field research on platform feeds Guess et al. (2023) makes the allocation of exposure experimentally observable. Joint studies should record message provenance, delivery conditions, source cues, content features, and actual attention. Such data could help identify when distribution influences opinion through repeated framing and when it primarily changes what people see without moving their reported beliefs.
A final methodological priority is cross-design convergence. Researchers should compare laboratory message manipulations with field experiments, historical content analysis, panel observations, and validated behavioral outcomes while keeping their estimands distinct. Pre-registered hypotheses, archived stimuli, transparent frame codebooks, independent replication, and clear reporting of null results would make theories more cumulative. Studies should also examine the civic and ethical implications of audience privacy, contested factual claims, artificial text production, and the interaction of institutional power with personal interpretation. The central question remains explanatory: what conditions transform political messages into changes in attention, interpretation, knowledge, and action, and what evidence identifies each step?
Individual Differences, Prior Beliefs, and Effect Heterogeneity
The same political frame can affect different recipients in different ways because audiences bring distinct knowledge, values, partisan identities, issue commitments, and expectations about sources. A useful model identifies whether heterogeneity is caused by message comprehension, consideration accessibility, motivated evaluation, or differences in the outcome's baseline ceiling and floor. Slothuus & de Vreese (2010) connects party sponsorship and motivated reasoning, while Lecheler & de Vreese (2013) demonstrates that political knowledge can modify recency and cumulative effects. Neither finding implies a fixed psychological trait that determines susceptibility in all settings. The frame, issue, source, and measurement window jointly define the expected variation.
Statistical subgroup differences should be treated as hypotheses that require adequate precision and correction for multiple comparisons. When investigators examine many identities and issue categories, some differences will arise by chance. Strong designs prespecify theoretically motivated interactions and report uncertainty for both conditional estimates and contrasts between them. The population average is important for societal description, while conditional effects can disclose how communication changes the structure of disagreement. Treating a statistically significant result in one subgroup and an insignificant result in another as proof of an interaction is a familiar error; the interaction itself requires a direct test.
The distinction between understanding and accepting a message is especially important. Knowledgeable respondents may recognize the intended frame and reject its implication; others may agree with its factual claim but continue to prioritize different values. The integrative synthesis of Jost et al. (2022) situates these patterns within identity and motivational processes. Such accounts should generate falsifiable predictions: what observed source, framing, timing, and recipient characteristics would distinguish an identity-protective explanation from ordinary uncertainty reduction? Evidence from Broockman & Kalla (2023) suggests that responsiveness to unfamiliar candidate information can be reconciled with stable partisan commitments without assuming that all resistance is motivated rejection.
Comparative Contexts, Institutions, and Cross-Cultural Generalization
Political communication occurs within different electoral systems, media ownership structures, professional journalism practices, rights regimes, party systems, and histories of civic association. These institutional contexts shape whether responsibility claims are intelligible, which source identities are trusted, and how often competing frames are accessible. The European Union television-news experiment by de Vreese (2004) provides a concrete example of studying political framing outside a purely American campaign setting. Its reported finding—that conflict and economic-consequence frames guided interpretation without producing different policy support in that design—also cautions against treating opinion change as an inevitable outcome of interpretive variation.
Cross-national comparison requires more than using the same translated survey. Researchers must ask whether a term such as political trust, public responsibility, or party identification names the same latent construct in different settings. Frame coding rules may also be culture-specific when a metaphor or institutional reference invokes a distinctive history. The ideal design documents translations, local validity studies, measurement invariance, and external indicators of political institutions. Sampling should represent the intended population rather than defaulting to readily accessible students or platform users. The meaning of an average frame effect is therefore conditional on both the message and the sampled political context de Vreese (2005) Borah (2011).
Historical timing generates further variation. A frame introduced during an acute public crisis or close election can meet unusually high audience attention; the same formulation during a low-salience period may be forgotten. Comparisons must record events that alter factual information and social stakes between measurement waves. Even within one country, geographic locality, class position, media availability, political efficacy, and civic networks can modify reception. A comparative account should integrate these as measurable contextual variables and preserve the distinction between plausible general mechanisms and empirically demonstrated transportability.
Evidence Quality, Replication, and Rival Explanations
Political framing evidence includes several complementary designs whose inferential limits differ. Controlled experiments provide stronger causal attribution for assigned treatment differences but may simplify the news environment and use short-term outcomes. Field experiments improve real-world contact and behavioral observation while often sacrificing control over what individuals notice and discuss. Survey panels track trajectories across time but are vulnerable to selective exposure, unmeasured events, and reciprocal causation. Content analyses establish distributions of messages and historical change but generally cannot demonstrate audience-level effects. A serious synthesis resists ranking all studies on a single scale; it matches each claim to the specific design that can support it.
Three empirical results illustrate why competing claims must be held together. Nelson & Oxley (1999) found that issue frames affected political opinions through changes in both beliefs and their importance in controlled laboratory contrasts. de Vreese (2004) found interpretation changes without corresponding differences in policy support in a television-news experiment. Kalla & Broockman (2018) estimated very small average candidate-choice persuasion effects from campaign contact in general-election field settings. These findings are compatible when the outcomes, treatments, populations, and time horizons are specified. They reject any one-dimensional statement that frames always work, never work, or produce a constant magnitude.
Publication bias, researcher degrees of freedom, weak construct definitions, and selective reporting can distort cumulative judgments. Evidence appraisal should examine preregistration when available, treatment materials, exact endpoints, confidence intervals, attrition, excluded cases, multiplicity, and independently conducted replications. The methodological critique of Borah (2011) shows that framing research has faced significant conceptual inconsistency, while competitive and temporal studies expose moderators omitted by simple isolated-message experiments Chong & Druckman (2007a) Lecheler & de Vreese (2013). A reproducible program should archive stimuli, sampling and coding decisions, analysis scripts, and deidentified data where lawful and ethical.
Public Knowledge, Democratic Accountability, and the Ethics of Research
Political communication affects knowledge and judgment under circumstances in which different citizens retain legitimate interests and values. Research therefore needs to distinguish factual accuracy from interpretive disagreement. A responsible experimental analysis can assess whether a frame highlights an accurate causal relation, omits relevant evidence, or appeals to a recognizable civic principle. A claim about democratic improvement requires a stated normative criterion—such as factual understanding, transparency, public accountability, or informed participation—plus measurements appropriate to that criterion. An increase in message effectiveness is not itself evidence of public benefit Entman (1993) Cappella & Jamieson (1997).
Experiments also create ethical responsibilities. Studies involving real elections, stigmatized populations, or contentious issues may influence emotional distress, perceptions of fellow citizens, or political participation. Investigators should consider consent, independent ethics oversight, debriefing when appropriate, data minimization, and protections against identification through granular media-use records. The possibility of meaningful spillovers must be considered when treatments circulate beyond enrolled participants. These questions are intensified by fine-grained digital tracking, but they also apply to conventional field mobilization and persuasive-message studies Gerber & Green (1999) Guess et al. (2023).
A public-facing scientific monograph should make room for conflicting results without arranging them into a partisan narrative. The findings of Levendusky & Malhotra (2016) on perceived division, issue moderation, and opposing-party dislike show that a single communication intervention can move several public outcomes in different directions. The broad psychological synthesis of Jost et al. (2022) likewise describes interacting motivational pathways rather than a single politically desirable endpoint. Reporting what was measured, what changed, what did not change, and which uncertainties remain is the foundation for independent assessment by readers across political commitments.
Political Communication in the Artificial Era
Artificial systems are becoming participants in the production, translation, summarization, classification, and circulation of political communication. Their psychological significance for this dossier arises where they affect the availability or interpretation of political frames, the credibility assigned to a message, or the experience of repeated exposure. The basic scientific questions remain empirically tractable: what message did a person encounter, what interpretation did it encourage, which source did the person infer, and what outcome followed? Existing theories of source credibility and frame competition offer testable starting points Druckman (2001) Chong & Druckman (2007b).
A distinction is needed between media-generated text and relational interaction with an artificial conversational agent. A summary inserted into a news feed changes a mediated message; an agent conducting a sustained dialogue introduces additional variables such as apparent social presence, personalization, conversational memory, and trust. The latter belongs to the separate research object of human–artificial political communication. Likewise, algorithmic selection belongs to the dossier on digital political exposure. Their connections to political framing can be studied through crossed interventions that vary message content independently of source representation and distribution Guess et al. (2023).
The Artificial Era enlarges the historical range of communication sources and the speed at which messages can be generated and recombined. Describing that transition as an empirical phenomenon requires observations of actual systems, specified populations, and counterfactual comparisons. Research must test whether artificial provenance changes perceived expertise, whether summaries preserve competing considerations, whether users identify generated framing devices, and whether longitudinal interaction modifies attitudes or civic participation. The philosophical horizon From Homo to Artificial gives this inquiry an interdisciplinary setting; it does not substitute philosophical attribution for behavioral evidence.
Conclusion
Political communication is a structured process in which institutions and actors create messages, audiences select and interpret them, and public judgments emerge through repeated encounters with information and other people. Its scientifically distinguishable processes include agenda setting, issue framing, source-based inference, responsibility attribution, consideration weighting, competitive interpretation, temporal dynamics, and mobilization. The historical studies of McCombs & Shaw (1972), Iyengar (1991), Entman (1993), and Scheufele (1999) established enduring conceptual questions; later experiments identify important conditions under which those mechanisms operate or fail to produce expected outcomes.
The assembled evidence supports several bounded conclusions. Framing manipulations can alter interpretation and sometimes opinion in experimental contexts Nelson & Oxley (1999); policy support can remain unchanged when interpretive focus moves de Vreese (2004). Party identity and credible sources modify response Druckman (2001) Slothuus & de Vreese (2010). Repetition is not a universal accumulation mechanism Lecheler & de Vreese (2013). Campaign communication can yield small average effects on candidate choices while informative messages produce conditional responsiveness Kalla & Broockman (2018) Broockman & Kalla (2023). Field experiments in digital political information environments show that large exposure changes do not necessarily imply measurable attitude change over the observed period Guess et al. (2023) Nyhan et al. (2023).
The strongest future research will integrate message-content measurement, realistic exposure assignment, direct measures of interpretation, durable judgments, observable behavior, cultural context, and transparent estimation of uncertainty. It will examine interactions among rhetorical content, source credibility, competition, and distribution without collapsing them into a single claim about media power. The independent subject of this dossier is therefore defined by a precise causal question: under which communicative conditions do political messages reorganize what people consider, how they interpret collective problems, and what they subsequently believe or do? The answer is a family of conditional, measurable effects whose limits are part of the scientific account itself.
References
- Arceneaux, K., & Johnson, M. (2013). Changing minds or changing channels? Partisan news in an age of choice. University of Chicago Press. https://doi.org/10.7208/chicago/9780226047447.001.0001
- Borah, P. (2011). Conceptual issues in framing theory: A systematic examination of a decade's literature. Journal of Communication, 61(2), 246–263. https://doi.org/10.1111/j.1460-2466.2011.01539.x
- Broockman, D. E., & Kalla, J. L. (2023). When and why are campaigns’ persuasive effects small? Evidence from the 2020 U.S. presidential election. American Journal of Political Science, 67(4), 833–849. https://doi.org/10.1111/ajps.12724
- Cappella, J. N., & Jamieson, K. H. (1997). Spiral of cynicism: The press and the public good. Oxford University Press. https://doi.org/10.1093/oso/9780195090635.001.0001
- Chong, D., & Druckman, J. N. (2007a). A theory of framing and opinion formation in competitive elite environments. Journal of Communication, 57(1), 99–118. https://doi.org/10.1111/j.1460-2466.2006.00331.x
- Chong, D., & Druckman, J. N. (2007b). Framing theory. Annual Review of Political Science, 10, 103–126. https://doi.org/10.1146/annurev.polisci.10.072805.103054
- de Vreese, C. H. (2004). The effects of frames in political television news on issue interpretation and frame salience. Journalism & Mass Communication Quarterly, 81(1), 36–52. https://doi.org/10.1177/107769900408100104
- de Vreese, C. H. (2005). News framing: Theory and typology. Information Design Journal, 13(1), 51–62. https://doi.org/10.1075/idjdd.13.1.06vre
- Druckman, J. N. (2001). On the limits of framing effects: Who can frame? The Journal of Politics, 63(4), 1041–1066. https://doi.org/10.1111/0022-3816.00100
- Entman, R. M. (1993). Framing: Toward clarification of a fractured paradigm. Journal of Communication, 43(4), 51–58. https://doi.org/10.1111/j.1460-2466.1993.tb01304.x
- Gerber, A. S., & Green, D. P. (1999). Does canvassing increase voter turnout? A field experiment. Proceedings of the National Academy of Sciences, 96(19), 10939–10942. https://doi.org/10.1073/pnas.96.19.10939
- Gerber, A. S., & Green, D. P. (2000). The effects of canvassing, telephone calls, and direct mail on voter turnout: A field experiment. American Political Science Review, 94(3), 653–663. https://isps.yale.edu/resource/the-effects-of-canvassing-telephone-calls-and-direct-mail-on-voter-turnout-a-field
- Guess, A. M., Malhotra, N., Pan, J., Barberá, P., Allcott, H., Brown, T., Crespo-Tenorio, A., Dimmery, D., Freelon, D., Gentzkow, M., González-Bailón, S., Kennedy, E., Kim, Y. M., Lazer, D., Moehler, D., Nyhan, B., Velasco Rivera, C., Settle, J., Thomas, D. R., ... Tucker, J. A. (2023). How do social media feed algorithms affect attitudes and behavior in an election campaign? Science, 381(6656), 398–404. https://doi.org/10.1126/science.abp9364
- Iyengar, S. (1991). Is anyone responsible? How television frames political issues. University of Chicago Press. https://doi.org/10.7208/chicago/9780226388533.001.0001
- Iyengar, S. (1996). Framing responsibility for political issues. The ANNALS of the American Academy of Political and Social Science, 546(1), 59–70. https://doi.org/10.1177/0002716296546001006
- Jost, J. T., Baldassarri, D. S., & Druckman, J. N. (2022). Cognitive–motivational mechanisms of political polarization in social-communicative contexts. Nature Reviews Psychology, 1, 560–576. https://doi.org/10.1038/s44159-022-00093-5
- Kalla, J. L., & Broockman, D. E. (2018). The minimal persuasive effects of campaign contact in general elections: Evidence from 49 field experiments. American Political Science Review, 112(1), 148–166. https://doi.org/10.1017/S0003055417000363
- Lecheler, S., & de Vreese, C. H. (2013). What a difference a day makes? The effects of repetitive and competitive news framing over time. Communication Research, 40(2), 147–175. https://doi.org/10.1177/0093650212470688
- Levendusky, M., & Malhotra, N. (2016). Does media coverage of partisan polarization affect political attitudes? Political Communication, 33(2), 283–301. https://doi.org/10.1080/10584609.2015.1038455
- McCombs, M. E., & Shaw, D. L. (1972). The agenda-setting function of mass media. Public Opinion Quarterly, 36(2), 176–187. https://doi.org/10.1086/267990
- Nelson, T. E., & Oxley, Z. M. (1999). Issue framing effects on belief importance and opinion. The Journal of Politics, 61(4), 1040–1067. https://doi.org/10.2307/2647553
- Nisbet, M. C., Brossard, D., & Kroepsch, A. (2003). Framing science: The stem cell controversy in an age of press/politics. Harvard International Journal of Press/Politics, 8(2), 36–70. https://doi.org/10.1177/1081180X02251047
- Nyhan, B., Settle, J., Thorson, E., Wojcieszak, M., Barberá, P., Chen, A. Y., Allcott, H., Brown, T., Crespo-Tenorio, A., Dimmery, D., Freelon, D., Gentzkow, M., González-Bailón, S., Guess, A. M., Kennedy, E., Kim, Y. M., Lazer, D., Malhotra, N., Moehler, D., ... Tucker, J. A. (2023). Like-minded sources on Facebook are prevalent but not polarizing. Nature, 620, 137–144. https://doi.org/10.1038/s41586-023-06297-w
- Price, V., Tewksbury, D., & Powers, E. (1997). Switching trains of thought: The impact of news frames on readers’ cognitive responses. Communication Research, 24(5), 481–506. https://doi.org/10.1177/009365097024005002
- Scheufele, D. A. (1999). Framing as a theory of media effects. Journal of Communication, 49(1), 103–122. https://doi.org/10.1111/j.1460-2466.1999.tb02784.x
- Slothuus, R., & de Vreese, C. H. (2010). Political parties, motivated reasoning, and issue framing effects. The Journal of Politics, 72(3), 630–645. https://doi.org/10.1017/S002238161000006X
- Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481), 453–458. https://doi.org/10.1126/science.7455683