Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
Annual Review of Political Science
A Framework for the
Study of Persuasion
James N. Druckman
Department of Political Science, Northwestern University, Evanston, Illinois, USA;
email: druckman@northwestern.edu
Annu. Rev. Political Sci. 2022. 25:65–88
First published as a Review in Advance on
November 18, 2021
The Annual Review of Political Science is online at
polisci.annualreviews.org
https://doi.org/10.1146/annurev-polisci-051120-
110428
Copyright © 2022 by Annual Reviews. This work is
licensed under a Creative Commons Attribution 4.0
International License, which permits unrestricted
use, distribution, and reproduction in any medium,
provided the original author and source are credited.
See credit lines of images or other third-party
material in this article for license information
Keywords
persuasion, attitude change, public opinion, political communication,
framing
Abstract
Persuasion is a vital part of politics—who wins elections and policy disputes
often depends on which side can persuade more people. Given this centrality,
the study of persuasion has a long history with an enormous number of
theories and empirical inquiries. However, the literature is fragmented, with
few generalizable findings. I unify previously disparate dimensions of this
topic by presenting a framework focusing on actors (speakers and receivers),
treatments (topics, content, media), outcomes (attitudes, behaviors, emotions,
identities), and settings (competition, space, time, process, culture).
This Generalizing Persuasion (GP) Framework organizes distinct findings
and offers researchers a structure in which to situate their work. I conclude
with a discussion of the normative implications of persuasion.
65
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
INTRODUCTION
The study of persuasion has a long and storied history dating back to Aristotle (Barnes 1984).In the
twentieth century, social scientists joined modern philosophers and rhetoricians in the pursuit of
better understanding persuasion (Habermas 1984, Booth 2004). Like the ancients, modern social
scientists focus on how speakers’ characteristics, what they say, and how recipients think determine
whether a given persuasive attempt is successful. But what, exactly, do we know about persuasion?
On the one hand, an enormous knowledge base provides insight into when persuasion may or
may not work, and there is clearly no silver bullet that always works. On the other hand, the vast
literature remains disconnected, sometimes contradictory, and without a structure to facilitate
connecting one study with another.
This article aims to remedy the situation by offering a framework for the study of persuasion:
the Generalizing Persuasion (GP) Framework. By placing their work within the GP Framework,
scholars will help the field resolve inconsistencies, identify and address open questions, and ensure
collective progress. This does not require that researchers account for the multiple dimensions I
discuss but instead asks that they be explicit about which elements they study and how. While I
focus primarily on political persuasion, the approach can be applied across domains.
POLITICAL PERSUASION
Persuasion is “a successful intentional effort at influencing another’s mental state through communication
in a circumstance in which the persuadee has some measure of freedom” (O’Keefe 2016,
p. 4). The focal approach entails exploring how characteristics of the speaker/source, message, receiver/audience,
and channel/medium determine persuasive outcomes (Lasswell 1948, McGuire
1969). Theories have integrated these dimensions by identifying psychological variations that explain
the relative influence of each factor. Most notable are the dual-process Elaboration Likelihood
Model (ELM) (Petty & Cacioppo 1986) and Heuristic-Systematic Model (Chen & Chaiken
1999). These posit that motivated and able individuals evaluate details of messages while others
rely on cues (e.g., source credibility).
Political persuasion scholarship partially owes its circuitous evolution to two books published
in 1960. First, The American Voter (Campbell et al. 1960) put forth an influential approach (i.e., the
funnel of causality) that excludes “nonpersonal”factors such as communications.Second,Klapper’s
(1960) review of the early scientific literature concluded that mass communication effects are minimal,
undermining researchers’ incentives to explore them.
Slowly but surely theoretical orientations evolved, and evidence accrued such that by the mid-
1990s, scholars studied the influence of mass communications and interpersonal exchanges. Mutz
et al. (1996) made the case that political persuasion should be a sui generis topic of study. The
subsequent quarter century has seen substantial advances, culminating in handbooks (e.g., Suhay
et al. 2020) and meta-analyses (e.g., Amsalem & Zoizner 2022). Further, the politicization of various
areas of social life led scholars to expand the study of political persuasion to consumer choices
(Gil de Zúñiga et al. 2014) and science (Bolsen & Druckman 2015).
Despite these developments,the field lacks a coherent theoretical structure.The guiding theory
continues to focus on the speaker-message-receiver-channel framework (Holbrook 2011, p. 150;
Sydnor 2019, p. 9; Lau 2020; Suhay et al. 2020, p. 3). Despite occasional calls for attention to
context, systematic study of competitive scenarios with strategic persuaders is rare. This is the case
even though these settings define much of politics (Druckman & Lupia 2006). We also know little
about how political persuaders construct their messages in the first place, even though “opinion
leader” was one of the first concepts in the field (Lazarsfeld et al. 1944).
66 Druckman
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
The lack of an overarching framework contributes to ostensibly contradictory conclusions.
Consider the following statements:
Politicians are “relatively unconstrained by public opinion and able to shape it merely by
announcing their positions” (Broockman & Butler 2017, p. 208).
“Our results indicate that participants are more likely to stick with their initial decisions
than to change them no matter which reasons are considered” (Stanley et al. 2020, p. 891).
“The results from this analysis show quite clearly that campaigns are persuasive” (Flores
2019, p. 189).
“Sizable persuasive effects from campaign activities seem very unlikely to be observed”
(Nickerson & Rogers 2020, p. 1181; also see Kalla & Broockman 2018).
Evidence “points to the fairly short duration (or rapid decay) of most persuasion effects”
(Tesler & Zaller 2017, p. 79).
“Media can have durable effects on public opinion that subsequently may influence policy”
(Markovich et al. 2020, p. 25).
Thus, persuasion is unconstrained or extremely limited by prior opinions, campaigns are persuasive
or they are not, and persuasion effects taper or endure. To be clear, the above quotations
accurately represent conclusions of exemplary studies, and the authors do not overgeneralize their
claims. Nonetheless, taken together, the statements portray a fragmented field—one without an
organizing framework that identifies the reasons why studies arrive at distinct conclusions.
THE GENERALIZING PERSUASION FRAMEWORK
Even foundational process theories such as the ELM ignore crucial components of persuasion
including competition and speaker intent. This approach makes sense for a single study, but “the
task of creating dependable generalizations from such research can be more challenging than
might appear at first look” (O’Keefe 2016, p. 184).
I propose the GP Framework (Table 1) based on what we know about external validity—that
is, confidence that a relationship holds over variation in actors, treatments, outcomes, and settings
(Shadish et al. 2002, p. 38; Druckman 2022). The GP Framework is not meant to compete with
other theories (such as the ELM) but rather to fill in two gaps. First, it allows one to consider how
individual persuasion studies connect to one another and why studies may arrive at contradictory
conclusions. Second, it highlights the sources of variations that should be studied. Progress will
come if scholars of persuasion place their findings in the context of the GP Framework. This can
be done by stating the elements under study and the aspects being held constant. In what follows,
I review each element independently. The collective power of the framework will become clear
upon revisiting the above-listed ostensibly contradictory conclusions.
ACTORS
Speakers
Speakers come in a host of forms including media, opinion leaders, friends and family, and elites
who devote themselves to politics. Most studies treat the speaker’s message as exogenous, focusing
on the audience’s reactions. This approach ignores whether the specific identity of the
speaker (e.g., an elite) has an impact (Minozzi et al. 2015) and sidesteps the motivations behind
the speaker’s messaging. The latter omission reflects the difficulty of studying communicators’
decisions.
www.annualreviews.org • The Study of Persuasion 67
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
Table 1 Generalizing Persuasion Framework
Dimension Components
Actors Speaker(s)
Types (e.g., elites, media, opinion leaders, friends/family)
Motivations in crafting messages
Receiver(s)
Assessments across weighted dimensions
Effort, motivation, prior attitudes
Treatments Topic
Persons/groups, issues, institutions, products
Variation within a topic (e.g., different policy issues)
Message content
Argument strength (and inadequacy)
Framing and evaluations
Matching to receivers’ goals
Altering receivers’ motivations (e.g., using narratives)
Medium
Alters frames, processing goals, and/or effort
Interactions with other persuasion variables
Outcomes Attitude
General evaluation of an object (where the “object” is broadly construed)
Behavior
Does not always follow from an attitude
Depends on attitude attributes, injunctive and descriptive norms, behavioral control, and emotions
Emotion
Can inform conscious evaluations or override them
Identity
A dimension of evaluation
Often activated when threatened
Settings Competition
Number of speakers
Number of receivers
Observers
Space
Attitude or behavioral change in one setting may not generalize to other settings
Time
Pretreatment effects—what happened prior to the persuasive message
Posttreatment duration—how long an effect lasts
Time between exposure and outcome measurement
Process
Threatening settings
Political (conflictual) settings versus deliberative settings
Culture
Shapes understandings of topics
Alters salience of different values
Consider Druckman & Jacobs’s (2015) study of presidential rhetoric. They obtained private
polls from three presidents (Johnson, Nixon, and Reagan) and explored how polling results informed
the presidents’ public statements. They find that presidents choose their rhetoric to
prime the salience of issues on which they are favored and to cater to issue publics (e.g., taking
positions on family value issues that align with religious voters) (also see Hager & Hilbig
68 Druckman
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
2020). Yet, elite polls are private for a reason and do not always become publicly available for
analyses.
Other researchers look at the decisions of media outlets, exploring the influence of market
pressures (Bovitz et al. 2002) and ideology (Groeling 2013). It is notoriously difficult, however, to
pinpoint the drivers of media decisions (Druckman & Parkin 2005; cf. Lelkes 2020). Looking at
readers, owners, reporters, and local politicians, Gentzkow & Shapiro (2010) find scant evidence
of supply-side incentives shaping the slant of US newspapers. Yet, there are exceptional cases. For
example, newspapers with ownership interest in 1996 television deregulation covered the issue in
a skewed fashion, which suggests that under some conditions, ownership preferences can shape
news coverage (Gilens & Hertzman 2000; also see Martin & McCrain 2019).
This work evolved with the media ecosystem to study social media. A robust literature studies
the impact of social media on opinions (Settle 2018) and how users engage with one another
(Bail 2021). Again, what is missing is research on why a speaker produces a particular message.
This lacuna can be traced back to work on opinion leaders “to whom. . .the rank-and-file voters
typically ‘delegate’ the main burden of political discussion” (Berelson et al. 1954, p. 109). Scholars
explore how to identify opinion leaders based on self-reports of engagement (Ahn et al. 2014).
They also study opinion leaders who strictly engage on social media or “prosumers” (Weeks et al.
2017).
Whether people choose the prototypical opinion leader when gathering information seems
unclear, at best. Minozzi et al. (2020) use panel network data to show that people tend to incidentally
discuss politics with others who share their demographic characteristics (e.g., gender,
race/ethnicity, religion) rather than with politically interested and knowledgeable opinion leaders.
This, in turn, could affect what people learn; for instance, Carlson (2019) shows that people
learn less from nonideal informants than from “ideal informants” who resemble knowledgeable
and trustworthy opinion leaders.
Regardless of these choices, little is known about how opinion leaders or incidental discussion
partners (or other types of speakers) formulate their messages. A starting point is to recognize
that speakers have goals, such as objectively informing, persuasion/advocacy, self-expression, selfpresentation,
or mutual understanding (Cionea et al. 2017). Some recent scholarship identifies the
correlates of advocacy. For example, the likelihood of advocacy increases with confidence in one’s
opinions (Cheatham & Tormala 2015) and a belief that the receiver is open-minded (Teeny & Petty
2021). A next step would be to further unravel what generates self-expression and self-presentation
(Hillman 2010). Along these lines, Kraft et al. (2020) show that the tone of stories differs across
media: News stories that are viewed on news web sites are more negative than those chosen to
be shared on email, Twitter, and Facebook. The authors explain that this media distinction stems
from speakers being motivated by self-presentation.
How these distinct speaker motivations influence the success of persuasive efforts remains unclear.
It may be that those motivated by advocacy are less successful, as receivers often have an
aversion to those who actively pursue change (Bashir et al. 2013). Ultimately, the hope is to move
the study of persuasion from a perspective that sees a one-way transmission (speaker to receiver)
to one that sees an exchange between the actors—the reality in many modern political contexts.
Moreover, speakers may be more persuasive when they engage in interactions with receivers, as
such interactions enhance receivers’ feelings of self-efficacy (Levine 2021).
Receivers
A common starting point is characterizing the target of a persuasive attempt, classically the receiver’s
attitude about (i.e., evaluation of) an object. A useful approach, termed the summative
www.annualreviews.org • The Study of Persuasion 69
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
or expectancy value model (Chong & Druckman 2007b, Fishbein & Ajzen 2010), holds that an
individual’s attitude toward an object is the weighted sum of a series of evaluative beliefs about
that object. Specifically, Attitude = ࢣvi
∗
wi,, where vi is the evaluation of the object on attribute i
and wi is the salience weight (ࢣwi = 1) associated with that attribute. For example, one’s overall
attitude, A, toward a new housing development might consist of a combination of negative and
positive evaluations, vi, of the project on different dimensions i. An individual may believe that the
project will favor the economy (i = 1) but harm the environment (i = 2). If this individual cares
about both the economy and the environment, then v1 is positive and v2 is negative, and his or
her attitude toward the project will depend on the relative magnitudes of v1 and v2 discounted by
the relative weights (w1 and w2) assigned respectively to each attribute (Nelson & Oxley 1999).
This conceptualization applies to any object of evaluation (e.g., issues, candidates, institutions,
products) as well as behaviors (e.g., voting, donating, purchasing).
Whether a persuasive attempt succeeds depends on how the receivers assess the message. I
highlight three main factors: effort, motivation, and prior attitudes. First, receivers put forth a
certain amount of effort in assessing a communication. Dual-process models identify effort as
a key determinant of whether a receiver scrutinizes the content of the message (high effort) or
relies on a cognitively simple cue such as the perceived credibility of the speaker (low effort)
(Chaiken & Trope 1999). Effort depends on opportunity, ability, and the salience of the topic. For
example, Arceneaux & Vander Wielen (2017, pp. 98–105) exposed individuals to mixes of party
cues and arguments about whether the federal or local government should control environmental
policy. They find that those who put forth greater thought are swayed by the arguments, whereas
those who exert less effort rely on the party cues. On the more salient issue of government health
care, low-effort individuals relied on the arguments rather than the cues (also see Boudreau &
MacKenzie 2014). Even though politics remains distant for most people, many still assess message
content on salient topics.
Second, receivers have a motivation or goal: “a cognitive representation of a desired endpoint
that impacts evaluations, emotions and behaviors” (Fishbach & Ferguson 2007, p. 491). Nondirectional
goals are independent from specific conclusions, involving broad objectives such as forming
an accurate opinion in light of evidence/arguments (Lodge & Taber 2013). Directional goals, in
contrast, involve confirming a specific desired conclusion (e.g., that climate change is a natural
process over which humans have little control, or that it is not occurring at all). Many scholars
point to directional goals as a reason why persuasion efforts fail (Cotter et al. 2020).
Stanley et al.(2020) first measured people’s opinions on five low-salience policy issues (fracking,
laboratory animal testing, drone strikes, the gold standard, and standardized testing of students).
They then (randomly) asked participants to assess affirming arguments, conflicting arguments,
or a mix of arguments, after which the authors again queried their opinions. They find that “no
matter which set of reasons was evaluated, participants were more likely to stick with their initial
decisions than to change them” (Stanley et al. 2020, p. 901). Stanley and colleagues conclude
that individuals are motivated to protect their initial opinions, leading them to evaluate contrary
reasons negatively and affirming reasons positively.
Directional reasoning takes a variety of nonexclusive forms (Bayes & Druckman 2021). For instance,
Stanley et al. (2020) posit a belief consistency–seeking motivation with the goal to maintain
a standing belief, regardless of the strength of an argument put forth. One might ask, though, what
respondents would do if the contrary messages had come from their party. Would they stick to
their prior beliefs or follow their party (Mullinix 2016)? The latter would indicate a motivation of
social identity protection aimed at maintaining a feeling of identity or status within a social group
(e.g., agreeing with your party on an issue even if the argument on the issue is poor) (Leeper &
Slothuus 2014). In other situations, individuals are motivated by value affirmation. For example,
70 Druckman
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
Feinberg & Willer (2013) show that messages framing environmental issues in terms of purity
and sanctity lead conservatives who typically oppose environmental legislation to become more
supportive; this argument resonates with cherished conservative values (Feinberg & Willer 2019).
Extant work rarely theorizes motivations a priori or directly documents the presence of a
given motivation (e.g., accuracy, belief consistency, social identity protection, value affirmation).
Druckman & McGrath (2019) refer to an observational equivalence problem: Scholars infer a directional
motivation when a persuasive message fails, even though it could be that the message
simply is not compelling. For instance, evidence suggests that persuasion during high-salience
(e.g., presidential) campaigns rarely occurs (Kalla & Broockman 2018). One could conclude that
this failure stems from partisan motivated reasoning where voters seek to agree only with their
party (i.e., their social group) (Nickerson & Rogers 2020, p. 1181). Alternatively, it is possible
that voters aim for accuracy but hold strong prior beliefs, with an increasing share being “likely
to agree with their party on any issues on which they have opinions in the first place” (Kalla &
Broockman 2018, p. 150; cf. Hillygus & Shields 2008).
It matters which process occurs. In the former case successful persuasion requires altering
voters’ motivations, whereas in the latter it entails convincing voters that candidates share their
perspectives. Isolating the motivation requires manipulating it or directly measuring it (Tappin
et al. 2020). For instance, Bayes et al. (2020) distinguished between accuracy motivation, social
identity protection, and a value affirmation goal by experimentally inducing each before providing
participants with climate change messaging designed to appeal to one of the motivations. They
find that Republican participants were more likely to believe in climate change when they received
a message that facilitated the pursuit of the specifically induced motivation. For example, those
with an accuracy motivation were more persuaded by scientific information, while those with a
value goal were more persuaded by a moral value message.
The final key construct influencing receivers is their prior attitudes. Stronger prior attitudes—
i.e., those held with more certainty, confidence, importance, or extremity—are more difficult to
change (Boudreau & MacKenzie 2014, Howe & Krosnick 2017) because accuracy-motivated individuals
have less uncertainty around their beliefs and thus hold a higher threshold to change them
(Druckman & McGrath 2019). For example, Druckman & Nelson (2003) find that individuals
who hold stronger prior attitudes about campaign finance reform are significantly less likely to be
swayed by messages that frame reform as an issue of free speech or minimizing special interests.
However, just how great an impact prior attitudes have remains unclear; for instance, Guess &
Coppock (2020) offer consistent evidence of persuasion on gun control, the minimum wage, and
the death penalty regardless of attitude importance or extremity.
Beyond prior attitude strength, scholars might consider the salience of directionally relevant
constructs such as social identities, values, and normative expectations. For instance, Girvan et al.
(2010) show that information about a normative social consensus (e.g., opinions of friends, family,
celebrities) has a particularly strong influence on the vote decisions of high self-monitors—
individuals who are likely to base decisions on social cues (i.e., they value normative expectations).
That said, Coppock (2021) finds that political persuasion often occurs “in parallel,” meaning that
people respond in the same way (i.e., in the same direction, by the same amount) to persuasive information.
This is the case across categories such as partisanship, age, education, race, and gender.
In summary, the result of any political persuasion effort hinges on the receiver’s effort, motivation,
and prior attitude strength (although, as noted, there is some debate on the latter feature).1
1Some readers may note the exclusion of sophistication from the discussion. While this surely is a causal factor
in shaping attitude change, it typically operates through more proximate variables including effort (e.g., via
www.annualreviews.org • The Study of Persuasion 71
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
Existing studies rarely consider all three. Some even rely on stereotypical portraits of the public
as exerting little effort (Lau 2020), being directionally motivated (Lodge & Taber 2013), and
holding weak prior attitudes on most issues (Kinder & Kalmoe 2017). Yet, as the above examples
demonstrate, there exists substantial variance in effort, motivation, and attitude strength. Scholars
should more actively theorize about how these variables influence persuasion.
TREATMENTS
Topic
The classical framework considers what is said (message content) and how it is delivered (medium)
(Lasswell 1948). But it ignores something perhaps even more central: the topic. For example, there
exist at least four nonexclusive political topical areas.First are persons or groups such as candidates,
elected officials, media figures, parties, and racial/ethnic groups. Second are public policy issues.
Third are institutions—i.e., norms/rules or organizations—such as the government, science, and
the military. Fourth are products; the rise of political consumerism or private politics has politicized
businesses, as demonstrated when activists try to sway consumers to boycott a business due
to poor environmental practices (Abito et al. 2019).
People process distinct topics differently. For instance, attitudes about individuals are more enduring
and less vulnerable to persuasion than attitudes about policies (Druckman & Lupia 2000).
In this vein, McGraw & Dolan (2007) show that when a message personifies the state (e.g., focusing
on its leader), people form stronger attitudes than when the message characterizes the state in
terms of institutions (e.g., the parliament). Hence, topical framing affects attitude strength.
There also exists substantial variation within so-called topics. Consider the literature on the
persuasive impact of party cues. Studies suggest that when a party endorses a policy, partisans move
their opinions by 3–43% on the policy opinion scale (Bullock 2011). Variation in the policy issues
studied partially drives these inconsistent results: There is no sampling from a population of policy
issues (Druckman & Leeper 2012a, Slothuus & Bisgaard 2021a). Tappin (2020) studies the impact
of party cues across 34 distinct policy questions and finds that the impact ranges from 15% (on
whether Congress should audit the Federal Reserve) to 1% (on whether police should be required
to wear body cameras). The difference between topics may not explain all the variation, but it
clearly plays a role. Relatedly, variation across topics depends on the strength of prior opinions
(Bartels 1993), self-interest (Slothuus & Bisgaard 2021b), and, as I discuss below, timing—that is,
the extent to which people have already been exposed to arguments on the topic (Druckman &
Leeper 2012b).
Message Content
Scholarship on the content of persuasive messages is expansive and exasperating. One prominent
concept is argument quality. As mentioned, when people are able to put forth effort, they scrutinize
the argument, and it persuades them if they consider it high quality (Chaiken 1980, Petty
& Cacioppo 1986). Identifying what constitutes high quality entails an empirical exercise of asking
individuals whether they perceive the argument as strong and cogent rather than weak and
specious (Eagly & Chaiken 1993, p. 311). While this approach helps to document when content
receives attention or which message wins out in competition (Chong & Druckman 2007a),
ability), motivation (cf. Lodge & Taber 2013, Tappin et al. 2021), and attitude strength (Howe & Krosnick
2017).
72 Druckman
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
it provides scant theoretical insight. In fact, it seems to be an inadequate tautology. It is a tautology
because it involves identifying persuasive messages as persuasive, and it is inadequate because
“asking respondents about perceived or expected persuasiveness [is] no more informative about
relative actual persuasiveness than flipping a coin” (O’Keefe 2018, p. 133).
Where does that leave the study of message content? I suggest three steps as a way forward.
First, tie the study of persuasive messages to the aforementioned model of attitude structure that
includes evaluative dimensions, evaluations on those dimensions, and weights applied to each dimension.
Different messaging strategies target each of those elements. For instance, framing (or
the analogous processes of priming and agenda setting) captures altering the dimensions under
consideration or the relevant weights, as in the above example of assessing a housing development
on economic and/or environmental grounds. Social scientists have long recognized altering
the frame or scope of conflict as a crucial part of politics (Schattschneider 1960). Alternatively,
messages can focus on evaluations such as how the housing development will affect the
environment.
Second, consider the audience’s motivations in assessing a message. Recall the Bayes et al.
(2020) study showing that climate change arguments succeed when they align with the audience’s
motivation—for example,value framing (e.g.,protect the sanctity of the environment) works when
receivers seek value affirmation, while a science fact message is persuasive for accuracy-motivated
receivers. This dynamic explains the appeal of targeted messaging crafted to match the recipient’s
characteristics (Hillygus & Shields 2008). For example, Matz et al. (2017) show that messages that
comport with the audience’s level of extraversion or openness to experience enjoy more success
in affecting product purchasing behaviors than mismatched messages (but see Eckles et al. 2018).
Targeting in politics, though, may not be an ideal strategy. Hersh & Schaffner (2013) show that
targeting specific groups (e.g., born-again Christians, union members, Latinx respondents) with
specialized messages tends not to persuade the group and backfires among the broader public (also
see Ostfeld 2019). Instead of focusing on social groups, more success could come from connecting
messages to the audience’s goals (e.g., knowing whether they care most about their values, what
others are doing, or particular policies).
Third, alter the audience’s motivation (Mullinix 2016). An example of such a messaging strategy
is narrative persuasion, a message format that describes events in chronological order with
information about characters and their actions. Gooch (2018) uses two experiments to show that
a personalized narrative from a partisan elite (Joe Biden) can increase support for the discussed
issue (Social Security) among in-partisans and increase the elite’s favorability among all partisans.
Kalla & Broockman (2020) present evidence from three field experiments on the impact of the
“nonjudgmental” exchanging of narratives where the speaker offers portrayals of unauthorized
immigrants or transgender people (also see Broockman & Kalla 2016, Kalla & Broockman 2021).
They find that, relative to a control placebo (and in one case relative to arguments alone), the
inclusion of the narrative durably reduces exclusionary attitudes toward these groups. Kubin et al.
(2021) similarly show that relaying personal experiences, particularly those that are relevant to the
issue at hand and that involve harm/suffering, generates respect across group lines and perceptions
of rationality more effectively than relaying facts.
Personal narratives “transport” the receiver into the story so that the receiver becomes focused
on the world it depicts (Green & Brock 2000). Hamby et al. (2017) suggest that transportation
works via retrospective reflection, the recall of self or other relevant memories to validate and
extend the story. It consequently can shift the audience away from directional motivations to a
focus on connecting with the narrative (Carpenter 2019, p. 15). This is not the same as an accuracy
motivation but, in essence, it shifts motivations by reducing the need to be protective and allows
for persuasion on ego-threatening topics.
www.annualreviews.org • The Study of Persuasion 73
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
Medium
Much has changed since Lasswell (1948) mentioned the “channel” as one of the key persuasive
variables. The political communication environment evolved from newspapers and radio to television
to the internet to social media. Researchers have leveraged social media data in remarkably
innovative ways to study political communication patterns (Persily & Tucker 2020). Here I make
two points that have received less attention.
First, the medium by which one receives a message can alter frames, processing goals, and/or
effort. For example, in his experiment using the famous first Nixon–Kennedy debate, Druckman
(2003) shows that, compared to those who listen to the debate, those who watch it on television
rely more on personality assessments in evaluating the candidates (e.g., perceived integrity) and
less on issue position assessments. This difference alters overall evaluations (in Kennedy’s favor).
Here, the medium acts as a framing mechanism that shifts the criteria of assessment. Althaus &
Tewksbury (2002, pp. 182–83) show that reading The New York Times online instead of in print
facilitates directional goals because it “draw[s] readers immediately to those stories most likely
to fit their information preferences.” Print readers view international stories on the front-page
section and then see them as more salient, while directionally motivated online readers skip over
such stories by following links. In this case, the medium prompts distinct goals.
In another such example, Druckman et al. (2018) compare communication via media segments
to communication via in-person conversations (which I construe as variations in media, although
one could see them as variations in context). They show that exposure to an out-partisan media
segment (e.g., Democrats watch a Fox News segment supporting drilling for oil) appears to
stimulate directional partisan motivated reasoning such that the audience counterargues the story
and moves in the opposite direction (e.g., they become more opposed to drilling). In contrast,
when individuals watch partisan media and talk with a group that includes out-partisans about the
topic (e.g., Democrats talk to Republicans about drilling for oil), they seem to move away from
directional partisan motivated reasoning and become relatively persuaded (e.g., they become less
opposed to drilling). The give-and-take of the in-person format stimulates more reflecting thinking
than the media format.
With regard to motivation and social media, the results are surprising. Many researchers conclude
that misinformation on social media shapes beliefs due to directional motivations where
partisans believe anything that supports their group (Flynn et al. 2017, p. 128). Pennycook et al.
(2021) show that this is not always the case and, instead, the persuasiveness of misinformation on
social media reflects receivers’ lack of effort in processing social media (cf. Osmundsen et al. 2021).
When people put in more effort to assess social media information, they are less likely to believe
false messages and to share them. The implication is that simple effort prompts can make a difference:
Pennycook et al. (2020) show that asking people to rate the accuracy of a single headline
subsequently increases the likelihood of sharing true rather than false posts.
My second point is about the interaction between medium and other variables. For example,
considerable debate concerns how persuasive argumentation from out-group sources works on social
media (Guess et al. 2021). Bail et al. (2018) find that exposing Republicans to liberal messages
via Twitter leads them to move in a more conservative direction (a backlash effect). Democrats
exposed to conservative Twitter messages move in a liberal direction, but not statistically significantly
so. In contrast, Levy (2021) finds no effects on political opinions among partisans exposed to
counter-attitudinal messages via Facebook.2
Even though both Bail et al. (2018) and Levy (2021)
2These contrasting findings reflect a larger set of conflicting results about potential backlash effects (e.g.,
Porter & Wood 2019, Ma et al. 2019).
74 Druckman
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
focus on the impact of social media out-party exposure, the differences could stem from variation
in other dimensions that affect persuasion. These include treatment medium (Twitter versus
Facebook), sources (retweets from more than 4,000 political accounts versus news outlets), receiver
motivation (being paid to follow a Twitter bot/account versus being nudged to subscribe
to an outlet that could reveal new perspectives), and settings (receiving 24 messages each day versus
choosing whether to comply and view the outlets). Of those variations, the most important
may have been the relative onslaught of oppositional messages in the experiment by Bail et al.
(2018), triggering negative reactions (Bail 2021). This differs from the careful engagement likely
stimulated in the Levy (2021) study. This comparison illustrates why one may find conflicting results.
When studying messages in a particular medium, it is crucial to consider differences across
other variables. The comparison also reveals the need to consider multiple variables, as one could
mistakenly, for example, conclude that the distinct results stemmed from different social media
platforms.
OUTCOMES
Attitudes
The modal target of persuasion is the receiver’s attitude, which has been my implicit focus thus far.
Here I make a tangential point. The increasing availability of behavioral data due to the internet
and social media, equipping researchers with “big data,” has led some to minimize the importance
of studying attitudes (for discussion, see Matz & Netzer 2017, Miller 2017). This is foolhardy
for two reasons. For one, a tenet of representative democracy is the responsiveness of elected
officials to the preferences of the populace (Dahl 1971). Thus, political preferences—regardless
of actions based on those preferences—matter. Their importance need not depend on Gallup
& Rae’s (1940) idyllic vision of The Pulse of Democracy, where polls reveal genuine thoughts to
the nation’s leaders. Regardless of one’s normative stance on what makes for the best democracy,
politicians attend to opinion surveys and so, practically, they are important (Druckman & Jacobs
2015).3
Second, as I discuss in the next section, scholars have taken low correlations between
attitudes and behaviors as evidence that attitude measures may be insufficient. Yet, in the age of
social media data, one must also consider whether the behaviors studied are the right behaviors
to study. Put another way, attitude surveys are constrained by sampling issues and nonresponse
(Miller 2017), but researchers using behavioral data often ignore the sample and use whatever
social media expressions are available. These may be meaningful but are far from representing
the population of possible behaviors. In short, if attitudes and behaviors do not match, it could be
because the wrong behavior is being studied and not because attitudes do not matter (see Groves
2011, p. 870).
Behaviors
Early studies presumed that attitudes straightforwardly explain behaviors (e.g., a voter likes a candidate
and therefore votes for or donates to the candidate). But “we should not expect that a
change in general attitudes. . . will have much of an impact on any particular behavior” (Fishbein
& Ajzen 2010, p. 278). To see the implications for persuasion, consider Levine & Kline’s (2017)
experiment on climate change. They provided recipients with a generic message about taking action
to address climate change, a public health message about taking climate change action, or a
3Further, the public themselves look at attitudes of others as pieces of information on which they act (e.g.,
Moy & Rinke 2012).
www.annualreviews.org • The Study of Persuasion 75
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
food risk message about taking climate change action (e.g., to prevent food shortages). They find
that—in line with the idea that loss frames persuade—the health and food risk messages cause
people to support clean energy policies and believe climate change should be a policy priority. Yet
in an analogous experiment, these types of loss frames depressed climate-relevant behaviors such
as joining a climate advocacy group. The logic is that, relative to a gains frame, the loss frame leads
people to support preventing the loss (their attitude), but it also makes salient their own resource
constraints, causing them to be less likely to act (also see Levine & Kline 2019).
This work reveals that one must be careful in identifying the target outcome. There are at least
two possible routes for exploring behavioral outcomes (or judgmental choices). One is to downplay
or evade a focus on the attitude–behavior connection if one’s goal is to understand behaviors.
For example, Groenendyk (2019) shows that the linchpin of stimulating voter turnout and intentions
is not resolving attitudinal ambivalence but rather activating positive “gut feelings.” Altering
behaviors may depend less on attitudes than on motivations and emotions (also see Groenendyk
2016).
Another route is to identify the conditions under which attitudes shape behaviors (Fishbein
& Ajzen 2010, p. 278). For example, Druckman & Bolsen (2011) exposed people to arguments
about the use of carbon nanotubes (CNTs), an alternative energy source to coal, oil, or natural
gas. They randomized study participants to exposure to pro arguments (e.g., lower energy costs)
or con arguments (e.g., health risks). They also randomly varied whether the arguments included
facts, operationalized as a reference to a scientific study (e.g., showing that mice injected with
CNTs react as if they were injected with asbestos or showing that CNTs double the efficiency
of solar cells). They find that regardless of the facts, the frames move attitudes in the predicted
direction (e.g., a pro frame leads to more support for the use of CNTs). However, the frames only
influence behavioral intentions (e.g., likelihood of personally using CNTs) when accompanied by
the supporting facts. The accompanying facts increase the certainty with which individuals hold
their attitudes, and that certainty leads people to act on their attitudes.
This study reveals the importance of attitude strength as a driver of behavior (Howe &
Krosnick 2017). More generally, behavior or the intention to engage in a behavior (O’Keefe 2021)
depends on attitudes/attitude strength, injunctive norms (i.e., if the behavior is desired by others),
descriptive norms (i.e., if other people perform the behavior), perceived behavioral control (i.e.,
perception of the ease or difficulty of the behavior, including resource constraints), and emotions
(Fishbein & Ajzen 2010).4
Persuaders who hope to influence behaviors must consider which of
these components matter to the receiver(s) in a given situation, which components seem more
moveable, and the current intention of the receiver(s).
Emotion
Emotions—instinctive states emanating from one’s situation or mood—are not an explicit part of
the attitudinal construct presented above. Yet, emotions inform the considerations that drive attitudes
or behaviors and sometimes override conscious considerations driving behaviors (see Marcus
2002, pp. 10–11). Either path makes emotions an important target of persuasion (Albertson et al.
2020).
For example, Clifford & Jerit (2018) provided study participants with messages about an infectious
disease and varied the absence or presence of text prompting disgust (e.g., symptoms include
4These factors come from the reasoned action approach. Habit is another predictor, as is clear from research
on habitual voting (e.g., Malhotra et al. 2011).
76 Druckman
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
bloody diarrhea, boils) and/or anxiety (e.g., high likelihood of spread). Anxiety messages prompted
individuals to request more information about the disease while disgust messages lowered such
requests. Brady et al. (2017), who exposed participants to messages about gun control, same-sex
marriage,and climate change,find that messages including moral-emotional words (e.g.,hate) lead
individuals to pass them along in social media (i.e., moral contagion). Other work demonstrates
the role of negative emotions such as anger in stimulating political action (Valentino et al. 2011)
and how individuals who regulate negative emotions reduce their likelihood of political action
(e.g., protesting, donating, volunteering, online posting) (Ford et al. 2019).5
This latter finding
suggests that those interested in stimulating action may want to reframe upsetting political events
to accentuate personally relevant negative emotions (that can vitiate regulation), although this can
also lead to deleterious behaviors such as violence and moral condemnation.
Identity
People often base political decisions on their partisan or social identities (Achen & Bartels 2016).
Indeed, the implications for one’s identity group are often a crucial consideration when one forms
an attitude. It is thus not surprising that speakers often seek to activate or pander to particular
identities.6
For instance, Levendusky (2018) shows that priming a common national identity can lead opposing
partisans to move away from animus and hold more positive feelings toward one another.
Klar (2013), looking at issue attitudes, finds that when Democratic parents receive a message that
makes their partisan identity salient, they become more supportive of social service spending regardless
of the national deficit (which falls to future generations), less supportive of antiterrorism
spending, and more supportive of releasing sex offenders early to start rehabilitation. In contrast,
when the message instead accentuates their parental identity, their policy opinions move in the
opposite directions. Klar further shows that the key to activating an identity is through threat
(i.e., current policies pose a threat to one’s partisanship or parental status) (also see Wojcieszak &
Garrett 2018).
SETTINGS
Competition
Classical persuasion experiments involve a single speaker sending a message to receivers; there
is scant attention to context. Yet, many contexts involve multiple speakers trying to persuade the
audience. To see why this matters, consider a typical framing effect, such as when people hear that
a hate group rally should be thought of as a public safety concern and not be allowed (Nelson et al.
1997). Invariably, the involved group or another group will argue that the decision to allow the
rally constitutes a free speech issue. When two arguments clash, they could cancel out (Sniderman
& Theriault 2004). Chong & Druckman (2007a), however, argue that competing rhetoric prompts
effort, leading people to opt for the argument they perceive as stronger. The authors show that if
the public safety argument comes from a credible source and the free speech argument comes
from a noncredible source, then the public safety perspective sways individuals (although see
the discussion on argument strength in the Message Content subsection of the section titled
5Positive emotions also can stimulate action; Brader (2006) shows how campaign advertisements that prompt
enthusiasm stimulate campaign participation.
6Speakers also often target values, such as in the moral framing examples above (see Howat 2019 on the
relationship between identity and values).
www.annualreviews.org • The Study of Persuasion 77
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
Treatments). Yet, in general, studies of competitive persuasion remain limited in the social sciences
(Della Vigna & Gentzkow 2010, p. 665).
The number of receivers also matters. When someone hears an argument, the presence of others
can stimulate normative pressures. Levitan & Verhulst (2016) show that having others nearby
shapes how people form their attitudes on issues such as abortion, affirmative action, gay rights,
and immigration. Receivers look to see how others react and follow suit (Sinclair 2012). Alternatively,
if listeners observe a persuasive argument targeting someone else, they may learn about its
applicability to themselves. Lupia & McCubbins (1998, p. 61) provide an example where a Democratic
observer hears a Republican speaker endorse a policy to his/her Republican constituents.
The Democratic observer then takes the opposite stance (what the authors call an observer effect).
This type of dynamic partially explains the aforementioned backlash of targeted messages among
those not targeted (Hersh & Schaffner 2013).
Space
Many political and social interactions occur in a specific space, so what happens in one context
may not affect attitudes and behaviors in another. This aligns with the prior discussion of outcome
variables but concerns the application of a particular outcome in distinct contexts. Mousa’s
(2020) study of intergroup contact shows that attitudes can change due to interpersonal exchanges,
even those not involving explicit argumentation. Mousa assigned Christian amateur soccer teams
in northern Iraq to randomly receive three additional Christian players or three additional (displaced)
Muslim players. Christian players on teams that added Muslims expressed more tolerant
attitudes toward training with Muslims or playing on a mixed team in the future. However, they
did not become more tolerant of interactions with Muslims not on the teams, in contexts (or
spaces) such as attending neighborhood social events or patronizing Muslim businesses.
Wiest et al. (2015) find a related dynamic in a study of climate change messaging. Frames
focusing on local (state) conditions alter perceptions of the severity of climate change and support
for subnational policies. However, the frames do not carry over to influence support for policy
efforts at the national or global levels. These types of findings connect to construal theory, which
holds that people construe psychologically proximal targets in more concrete terms (Liberman &
Trope 1998). The bottom line is that attitudinal and/or behavioral change in one setting does not
generalize to other spatial contexts.
Time
Timing matters. In their review of attitude change, Albarracin & Shavitt (2018) discuss time in
terms of a receiver’s development, lifespan, and generation. Here, I focus on timing as an aspect
of the setting. Political arguments occur over time; thus, understanding the outcome of persuasion
requires recognizing what came before. For example, in a study conducted among engaged
individuals (Druckman & Leeper 2012b), arguments concerning a public-funded casino—both
pro arguments focusing on economic benefits and con arguments focusing on social costs (e.g.,
addiction)—failed. They failed because these individuals previously had been exposed to the economic
argument and had formed strong opinions in favor of the casino. Hearing the economic
argument again did not matter. They had already formed strong opinions and rejected the con
argument. Such pretreatment effects make clear that one often must attend to communication
over time to grasp its effects (Gaines et al. 2007, Slothuus 2016).
Druckman et al. (2012) show how early arguments can set opinions in place, leading individuals
to seek out confirmatory information and making them more resistant to later arguments. For
instance, those exposed to an initial argument for universal health care to minimize inequality
78 Druckman
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
subsequently chose to read articles with that framing, and then when exposed later to opposing
arguments concerning economic costs, they rejected them. In other circumstances, more recent
arguments may win out, particularly among individuals who are averse to elaboration (Chong &
Druckman 2010).
The flip side of “what happened before” is “what happens after”: how long a given persuasive
effect lasts. Although the question has been considered since Hovland & Weiss’s (1951) sleeper
effect hypothesis (e.g., people initially discount a message due to a noncredible source but later
adjust their attitudes in the direction of the message), it is far from settled (Albarracin et al. 2017).
Related to duration is the amount of elapsed time between exposure to the persuasive message and
measurement of outcome—immediate or delayed. Not all studies of persuasion need to account
for prior and subsequent dynamics, but they need to situate themselves in time and recognize that
timing can alter what persuades.
Process
Different settings involve distinct decision-making processes. For instance, situations with personal
threats alter how individuals receive persuasive messages. Druckman et al. (2021) show that
those who felt more threatened by coronavirus disease 2019 (COVID-19) (e.g., more cases in
their surroundings) focused on substantive information instead of following messages from their
parties.
Political processes often involve environments rife with conflict and individualism
(Groenendyk & Krupnikov 2021). Yet, another depiction of politics posits a public endeavor of
deliberation for a common good (also see Karpowitz & Mendelberg 2014). In this latter situation,
individuals move away from directional goals and engage in open-minded (not synonymous with
accurate) assessments. Using the issue of gun control, Groenendyk & Krupnikov (2021) demonstrate
that settings privileging open-mindedness (e.g., linking it to life success) lead individuals to
evaluate arguments counter to their standing beliefs in an open-minded fashion. Individuals also
arrive at moderate opinions in response to a mix of arguments. In political settings, by comparison,
individuals think of conflict and engage in directional processing. This is particularly the case
in the presence of extreme partisan differences (Druckman et al. 2013). An intriguing question,
given these results, concerns the impact of various social media and entertainment contexts on
processing (Kim 2019): When do they prompt directional as opposed to open-minded reasoning
in response to arguments?
Culture
In recent years, persuasion scholars have come to recognize that culture shapes how arguments
are made and processed (Albarracin & Shavitt 2018). For instance, Song et al. (2020) conducted
a survey to assess what individuals viewed as an environmental issue. They find that minorities
and low-income respondents perceive several human-oriented issues as environmental. For instance,
relative to White people, Black and Hispanic individuals are significantly more likely to
identify poverty, unemployment, diabetes, and racism as environmental issues. Those with lower
household income are significantly more likely to identify unequal access to education and racism
as environmental issues. Presumably, social experiences become embedded in a cultural understanding
of what constitutes an environmental issue. This, in turn, matters for persuasion, since
it shapes the nature of conversations and surely affects what constitutes a persuasive argument on
these issues.
Stark et al. (2020) show that priming evenhandedness alters attitudes. Reminding labor activists
of the possibility of labor campaign contributions leads them to be more supportive of
www.annualreviews.org • The Study of Persuasion 79
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
business contributions (i.e., they offer more evenhanded evaluations). Yet, this effect is not nearly
as strong in countries that have collectivist cultures (e.g., Asian countries) because evenhandedness
considerations are chronically accessible. Here, culture generates what could be thought of as a
pretreatment effect.
Soroka et al. (2019) study the impact of negative news content in 17 countries across 6 continents.
They find evidence of higher attentiveness and arousal during exposure to negative stories.
They also report individual-level variations in reactions; however, these differences do not connect
to country-level contextual factors. In this case, there is less evidence of a cultural impact,
which itself is intriguing given that prior research had been concentrated in Western countries.
Clearly, understanding persuasion requires attending the intersection of multiple variables.
REVISITING CONTRADICTORY CLAIMS
To see the usefulness of the GP Framework (Table 1), consider the contradictory statements
from the bulleted list near the beginning of this review. Broockman & Butler (2017) conclude that
politicians can shape opinions with ease, while Stanley et al. (2020) hold that individuals do not
move from their initial opinions regardless of the reasons presented. The studies differ in terms
of actors (sources) and settings (timing). The former presented statements from legislators and
measured change from initial opinions after a substantial amount of time (at least a week). The
latter provided arguments without sources and assessed opinions very close in time to the initial
reporting of opinions.
Flores (2019) states that ballot initiative campaigns are persuasive in his study of television
advertisement effects on opinions about the legality of same-sex marriage. Nickerson & Rogers
(2020) state that campaigns have little to no effect. However, these studies differ in treatment and
outcome: Nickerson and Rogers’s treatment focuses on candidate elections (as opposed to initiative
campaigns), and their outcome of interest is voting behavior (as opposed to an issue attitude).
Kalla & Broockman (2018) also find scant campaign effects in general election campaigns. Yet,
they acknowledge that persuasion can occur in unusual circumstances depending on the timing
(e.g., earlier in the campaign), voters (e.g., with weaker prior opinions), targets (e.g., atypical candidates),
and messages (e.g., exploiting notable candidate blunders). These variables map onto the
GP Framework’s settings, actors, and treatments. Kalla & Broockman’s (2018) reasonable conclusions
are based on the set of extant studies and their assessment of “typical” general elections
campaigns. The challenge, though, is that no one has identified the population of campaigns and
how they vary across the GP Framework’s dimensions. If scholars were to situate their studies
within the framework and simultaneously consider variation across variables in the population of
interest (e.g., general election campaigns), it would engender more confidence about what is and
is not typical.
Tesler & Zaller (2017) find that persuasion effects are short-lived, on the basis of evidence
from election studies that focus on candidate television advertisements altering voting behavior
(Gerber et al. 2011, Hill et al. 2013). In contrast, Markovich et al. (2020) suggest that persuasion
effects endure, but they study a distinct outcome (opinions), treatment (the topic of marijuana
legalization and the medium of online news), and actor (partisan networks).
The contradictory claims thus stem from work that differs on multiple dimensions—with all
four dimensions of the GP Framework being relevant. In fact, in each contradictory pair, the works
differ along at least two of the dimensions. Understanding why studies with the same ostensible
focus arrive at contradictory conclusions requires attention to multiple possible differences. The
purpose of the GP Framework is to offer scholars a way to proceed when they study persuasion,
identifying which variables may matter and how their work connects to other research. The goal
80 Druckman
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
is to structure the literature in a way that better ensures collective progress. Authors can apply it
by being explicit about which elements they are studying and which they are not (e.g., in a table),
thereby enabling others to situate the work.
FINAL THOUGHTS: NORMATIVE CONSIDERATIONS
There is much to say about normative implications, and theorists have considered how to treat
persuasion from a host of angles (Garsten 2006, Disch 2011). There exists a long-standing tension
between idyllic conceptions of the citizen—as engaged and holding consistent preferences—and
the existence of persuasion by political elites. That elites can persuade means that the very individuals
who are supposed to respond to the authentic opinions of the citizenry have an incentive
to engage in disinformation and manipulation. This tension contributes to confusion about
salubrious persuasion (learning) versus detrimental manipulation (Druckman 2014). Adding to
the complexity is that engaged citizens are more likely to have directional motivations that make
them averse to open-minded deliberation (Druckman 2012). If one continues down this line of
reasoning, an infinite regress of trade-offs emerges. Normative criteria for a good citizenry and
the role of persuasion in promoting or hindering good citizenship are inherently subjective, often
arbitrary, and empirically challenging (e.g., how to intuit speakers’ motivations).
An alternative would be to focus on processes—such as articulating an ideal of accuracymotivated
citizens with access to competitive information. The problem here is that accuracy
may be a lot to ask for, and even if not, accuracy motivation need not lead to accurate conclusions
(Kunda 2001, p. 238). Moreover, specifying the availability of competitive information may be
doable at the extremes but ultimately can become arbitrary; it is not clear what the full range of
information includes, particularly given institutional constraints on representation.
Many scholars sidestep normative debates for fear that imposing their own values onto citizens
could be elitist (Lupia 2006). Alas, knowingly or not, this is what much scholarship has done.
Narratives lead to less exclusionary attitudes even though they can be far from fact-based (Van
Bavel et al. 2021), moral framing generates support for climate change policy despite including no
scientific information, and identity priming can depolarize without any mention of the need for
compromise in politics. In all of these cases, scholars have no ethical issue in studying particular
persuasive approaches that could alter opinions because they lead to, respectively, more tolerant,
evidence-consistent, and politically functional outcomes. These each invariably constitute valued
democratic collective goods to which governments contribute, and thus, scholars need not back
away from endorsing them and exploring persuasion as a means to these ends.
DISCLOSURE STATEMENT
The author is not aware of any affiliations, memberships, funding, or financial holdings that might
be perceived as affecting the objectivity of this review.
ACKNOWLEDGMENTS
I thank Eric Groenendyk, Yanna Krupnikov, Matt Levendusky, Adam Levine, and Eric Schickler
for extremely insightful comments on an earlier draft of this review. I also thank Alex Coppock,
Dan O’Keefe, and Rune Slothuus for helpful advice.
LITERATURE CITED
Abito JM, Besanko D, Diermeier D. 2019. Corporate Reputation and Social Activism. New York: Oxford Univ.
Press
www.annualreviews.org • The Study of Persuasion 81
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
Achen CH, Bartels LM. 2016. Democracy for Realists: Why Elections Do Not Produce Responsive Government.
Princeton, NJ: Princeton Univ. Press
Ahn TK, Huckfeldt R, Ryan JB. 2014. Experts, Activists, and Democratic Politics: Are Electorates Self-Educating?
New York: Cambridge Univ. Press
Albarracin DG, Kumkale T, Poyner-Del Vento P. 2017. How people can become persuaded by weak messages
presented by credible communicators: Not all sleeper effects are created equal. J. Exp. Psychol. 68:171–80
Albarracin DG, Shavitt S. 2018. Attitudes and attitude change. Annu. Rev. Psychol. 69:299–327
Albertson B, Dun L, Gadarian SK. 2020. The emotional aspects of political persuasion. See Suhay et al. 2020,
pp. 169–83
Althaus SL, Tewksbury D. 2002. Agenda setting and the “new” news: patterns of issue importance among
readers of the paper and online versions of The New York Times. Commun. Res. 29(2):180–207
Amsalem E, Zoizner A. 2022. Real, but limited: a meta-analytic assessment of framing effects in the political
domain. Br. J. Political Sci. 52(1):221–37
Arceneaux K, Vander Wielen RJ. 2017. Taming Intuition: How Reflection Minimizes Partisan Reasoning and
Promotes Democratic Accountability. Cambridge, UK: Cambridge Univ. Press
Bail C. 2021. Breaking the Social Media Prism: How to Make Our Platforms Less Polarizing. Princeton, NJ:
Princeton Univ. Press
Bail C, Argyle LP, Brown TW, Bumpus JP, Chen H, et al. 2018. Exposure to opposing views on social media
can increase political polarization. PNAS 115(37):9216–21
Barnes J, ed. 1984. The Complete Works of Aristotle: The Revised Oxford Translation. Princeton, NJ: Princeton
Univ. Press
Bartels LM. 1993. Messages received: the political impact of media exposure. Am. Political Sci. Rev. 87(2):267–
85
Bashir NY, Lockwood P, Chasteen AL, Nadolny D, Noyes I. 2013. The ironic impact of activists: Negative
stereotypes reduce social change influence. Eur. J. Soc. Psychol. 43(7):614–26
Bayes R, Druckman JN. 2021. Motivated reasoning and climate change. Curr. Opin. Behav. Sci. 42:27–35
Bayes R, Druckman JN, Goods A, Molden DC. 2020. When and how different motives can drive motivated
political reasoning. Political Psychol. 41(5):1031–52
Berelson BR,Lazarsfeld PF,McPhee WN.1954.Voting: A Study of Opinion Formation in a Presidential Campaign.
Chicago: Univ. Chicago Press
Bolsen T, Druckman JN. 2015. Counteracting the politicization of science. J. Commun. 65(5):745–69
Booth W. 2004. The Rhetoric of Rhetoric: The Quest for Effective Communication. Oxford, UK: Blackwell
Boudreau C, MacKenzie SA. 2014. Informing the electorate? How party cues and policy information affect
public opinion about initiatives. Am. J. Political Sci. 58(1):48–62
Bovitz GL, Druckman JN, Lupia A. 2002. When can a news organization lead public opinion? Ideology versus
market forces in decisions to make news. Public Choice 113:127–55
Brader T. 2006. Campaigning for Hearts and Minds: How Emotional Appeals in Political Ads Work. Chicago: Univ.
Chicago Press
Brady WJ, Wills JA, Jost JT, Tucker JA, Van Bavel JJ. 2017. Emotion shapes the diffusion of moralized content
in social networks. PNAS 114(28):7313–18
Broockman DE, Butler DM. 2017. The causal effects of elite position-taking on voter attitudes: field experiments
with elite communication. Am. J. Political Sci. 61(1):208–21
Broockman DE, Kalla JL. 2016. Durably reducing transphobia: a field experiment on door-to-door canvassing.
Science 352(6282):220–24
Bullock JG.2011.Elite influence on public opinion in an informed electorate.Am.Political Sci.Rev.105(3):496–
515
Campbell A, Converse PE, Miller WE, Stokes DE. 1960. The American Voter. Chicago: Univ. Chicago Press
Carlson TN. 2019. Through the grapevine: informational consequences of interpersonal political communication.
Am. Political Sci. Rev. 113(2):325–39
Carpenter CJ. 2019. Cognitive dissonance, ego-involvement, and motivated reasoning. Ann. Int. Commun.
Assoc. 43(1):1–23
Chaiken S. 1980. Heuristic versus systematic information processing and the use of source versus message
cues in persuasion. J. Pers. Soc. Psychol. 39(5):752–66
82 Druckman
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
Chaiken S, Trope Y, eds. 1999. Dual-Process Theories in Social Psychology. New York: Guilford
Cheatham L, Tormala ZL. 2015. Attitude certainty and attitudinal advocacy: the unique roles of clarity and
correctness. Pers. Soc. Psychol. B 41(11):1537–50
Chen S, Chaiken S. 1999. The heuristic-systematic model in its broader context. In Dual-Process Theories in
Social Psychology, ed. S Chaiken, Y Trope, pp. 73–96. New York: Guilford
Chong D, Druckman JN. 2007a. Framing public opinion in competitive democracies. Am. Political Sci. Rev.
101(4):637–55
Chong D, Druckman JN. 2007b. Framing theory. Annu. Rev. Political Sci. 10:103–26
Chong D, Druckman JN. 2010. Dynamic public opinion: communication effects over time. Am. Political Sci.
Rev. 104(4):663–80
Cionea IA, Hoelscher CS, Ile¸s IA. 2017. Arguing goals: an initial assessment of a new measurement instrument.
Commun. Rep. 301(1):51–65
Clifford S, Jerit J. 2018. Disgust, anxiety, and political learning in the face of threat: disgust and political
learning. Am. J. Political Sci. 62(2):266–79
Coppock A. 2021. Persuasion in Parallel. Chicago: Univ. Chicago Press
Cotter RG, Lodge M, Vidigal R. 2020. When, how, and why persuasion fails. See Suhay et al. 2020, pp. 51–65
Dahl RA. 1971. Polyarchy: Participation and Opposition. New Haven, CT: Yale Univ. Press
Della Vigna S, Gentzkow M. 2010. Persuasion: empirical evidence. Annu. Rev. Econ. 2:643–69
Disch L. 2011. Toward a mobilization conception of democratic representation. Am. Political Sci. Rev.
105(1):100–14
Druckman JN. 2003. The power of television images: the first Kennedy-Nixon debate revisited. J. Politics
65(2):559–71
Druckman JN. 2012. The politics of motivation. Crit. Rev. 24(2):199–216
Druckman JN. 2014. Pathologies of studying public opinion, political communication, and democratic responsiveness.
Political Commun. 31(3):467–92
Druckman JN. 2022. Experimental Thinking: A Primer on Social Science Experiments. New York: Cambridge
Univ. Press
Druckman JN, Bolsen T. 2011. Framing, motivated reasoning, and opinions about emergent technologies.
J. Commun. 61(4):659–88
Druckman JN, Fein J, Leeper TJ. 2012. A source of bias in public opinion stability. Am. Political Sci. Rev.
106(2):430–54
Druckman JN, Jacobs LR. 2015. Who Governs? Presidents, Public Opinion, and Manipulation. Chicago: Univ.
Chicago Press
Druckman JN, Klar S, Krupnikov Y, Levendusky M, Ryan JB. 2021. Affective polarization, local contexts, and
public opinion in America. Nat. Human Behav. 5:28–38
Druckman JN, Leeper TJ. 2012a. Is public opinion stable? Resolving the micro/macro disconnect in studies
of public opinion. Daedalus 141(4):50–68
Druckman JN, Leeper TJ. 2012b. Learning more from political communication experiments: pretreatment
and its effects. Am. J. Political Sci. 56(4):875–96
Druckman JN, Levendusky MS, McLain A. 2018. No need to watch: how the effects of partisan media can
spread via inter-personal discussions. Am. J. Political Sci. 62(1):99–112
Druckman JN, Lupia A. 2000. Preference formation. Annu. Rev. Political Sci. 3:1–24
Druckman JN, Lupia A. 2006. Mind, will, and choice: lessons from experiments in contextual variation. In The
Oxford Handbook of Contextual Political Analysis, ed. RE Goodin, C Tilly, pp. 97–113. Oxford, UK: Oxford
Univ. Press
Druckman JN, McGrath MC. 2019. The evidence for motivated reasoning in climate change preference formation.
Nat. Climate Change 9:111–19
Druckman JN, Nelson KR. 2003. Framing and deliberation: how citizens’ conversations limit elite influence.
Am. J. Political Sci. 47(4):729–45
Druckman JN, Parkin M. 2005. The impact of media bias: how editorial slant affects voters. J. Politics
67(4):1030–49
Druckman JN, Peterson E, Slothuus R. 2013. How elite partisan polarization affects public opinion formation.
Am. Political Sci. Rev. 170(1):57–79
www.annualreviews.org • The Study of Persuasion 83
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
Eagly AH, Chaiken S. 1993. The Psychology of Attitudes. Fort Worth, TX: Harcourt Brace Jovanovich College
Publ.
Eckles D, Gordon RB, Johnson GA. 2018. Field studies of psychologically targeted ads face threats to internal
validity. PNAS 115(23):E5254–55
Feinberg M, Willer R. 2013. The moral roots of environmental attitudes. Psychol. Sci. 24(1):56–62
Feinberg M, Willer R. 2019. Moral reframing: a technique for effective and persuasive communication across
political divides. Soc. Pers. Psychol. Compass 13:e12501
Fishbach A, Ferguson MJ. 2007. The goal construct in social psychology. In Social Psychology: Handbook of Basic
Principles, ed. AW Kruglanski, ET Higgins, pp. 490–515. New York: Guilford
Fishbein M, Ajzen I. 2010. Predicting and Changing Behavior: The Reasoned Action Approach. New York: Taylor
& Francis
Flores AR. 2019. Persuasion and ballot initiatives: How persuasive were the televised campaign ads on samesex
marriage? Politics Groups Identities 7(1):177–93
Flynn DJ, Nyhan B, Reifler J. 2017. The nature and origins of misperceptions: understanding false and unsupported
beliefs about politics. Adv. Political Psychol. 38(S1):127–50
Ford BQ, Feinberg M, Lam P, Mauss IB, John OP. 2019. Using reappraisal to regulate negative emotion after
the 2016 U.S. presidential election: Does emotion regulation trump political action? J. Pers. Soc. Psychol.
117(5):998–1015
Gaines BJ, Kuklinski JH, Quirk PJ. 2007. The logic of the survey experiment reexamined. Political Anal.
15(1):1–20
Gallup G, Rae SF. 1940. The Pulse of Democracy: The Public-Opinion Poll and How It Works. New York: Simon &
Schuster
Garsten B. 2006. Saving Persuasion: A Defense of Rhetoric and Judgment. Cambridge, MA: Harvard Univ. Press
Gentzkow M, Shapiro JM. 2010. What drives media slant? Evidence from daily newspapers. Econometrica
78(1):35–71
Gerber AS, Gimpel JG, Green DP, Shaw DR. 2011. How large and long-lasting are the persuasive effects of
televised campaign ads? Results from a randomized field experiment. Am. Political Sci. Rev. 105(1):135–50
Gil de Zúñiga H, Copeland L, Bimber B. 2014. Political consumerism: civic engagement and the social media
connection. New Media Soc. 16(3):488–506
Gilens M, Hertzman C. 2000. Corporate ownership and news bias: coverage of the 1996 Telecommunications
Act. J. Polit. 62(2):369–86
Girvan EJ, Weaver J, Snyder M. 2010. Elevating norm over substance: self-monitoring as a predictor of decision
criteria and decision time among independent voters. Anal. Soc. Iss. Public Policy 10(1):321–36
Gooch A. 2018. Ripping yarn: experiments on storytelling by partisan elites. Political Commun. 35(2):220–38
Green MC, Brock TC. 2000. The role of transportation in the persuasiveness of public narratives. J. Pers. Soc.
Psychol. 79(5):701–21
Groeling T. 2013. Media bias by the numbers: challenges and opportunities in the empirical study of partisan
news. Annu. Rev. Political Sci. 16(1):129–51
Groenendyk E. 2016. The anxious and ambivalent partisan: the effect of incidental anxiety on partisan motivated
recall and ambivalence. Public Opin. Q. 80(2):460–79
Groenendyk E. 2019. Two minds, but one heart: a good “gut” feeling moderates the effect of ambivalence on
attitude formation and turnout. Am. J. Political Sci. 63(2):368–84
Groenendyk E, Krupnikov Y. 2021. What motivates reasoning? A theory of goal-dependent political evaluation.
Am. J. Political Sci. 65(1):180–96
Groves RM. 2011. Three eras of survey research. Public Opin. Q. 75(5):861–71
Guess AM, Barberá P, Munzert S, Yang JH. 2021. The consequences of online partisan media. PNAS
118(14):e2013464118
Guess A, Coppock A. 2020. Does counter-attitudinal information cause backlash? Results from three large
survey experiments. Br. J. Political Sci. 50(4):1497–515
Habermas J. 1984. The Theory of Communicative Action. Boston: Beacon
Hager A, Hilbig H. 2020. Does public opinion affect political speech? Am. J. Political Sci. 64(4):921–37
Hamby A, Brinberg D, Daniloski K. 2017. Reflecting on the journey: mechanisms in narrative persuasion.
J. Consum. Psychol. 27(1):11–22
84 Druckman
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
Hersh ED, Schaffner BF. 2013. Targeted campaign appeals and the value of ambiguity. J. Politics 75(2):520–34
Hill SJ, Lo J, Vavreck L, Zaller J. 2013. How quickly we forget: the duration of persuasion effects from mass
communication. Political Commun. 30(4):521–47
Hillman AL. 2010. Expressive behavior in economics and politics. Eur. J. Political Econ. 26(4):403–18
Hillygus DS, Shields TG. 2008. The Persuadable Voter: Wedge Issues in Presidential Campaigns. Princeton, NJ:
Princeton Univ. Press
Holbrook AL. 2011. Attitude change experiments in political science. In The Cambridge Handbook of Experimental
Political Science, ed. JN Druckman, DG Green, JH Kuklinski, A Lupia, pp. 141–54. New York:
Cambridge Univ. Press
Hovland CI, Weiss W. 1951. The influence of source credibility on communication effectiveness. Public Opin.
Q. 15(4):635–50
Howat AJ. 2019. The role of value perceptions in intergroup conflict and cooperation. Politics Groups Identities
9(4):657–80
Howe LC, Krosnick JA. 2017. Attitude strength. Annu. Rev. Psychol. 68:327–51
Kalla JL, Broockman DE. 2018. The minimal persuasive effects of campaign contact in general elections:
evidence from 49 field experiments. Am. Political Sci. Rev. 112(1):148–66
Kalla JL, Broockman DE. 2020. Reducing exclusionary attitudes through interpersonal conversation: evidence
from three field experiments. Am. Political Sci. Rev. 114(2):410–25
Kalla JL, Broockman DE. 2021. Which narrative strategies durably reduce prejudice? Evidence from field and
survey experiments supporting the efficacy of perspective-getting. Am. J. Political Sci. https://doi.org/
10.1111/ajps.12657
Karpowitz C, Mendelberg T. 2014. The Silent Sex: Gender, Deliberation, & Institutions. Princeton, NJ: Princeton
Univ. Press
Kim E. 2019. Entertaining beliefs in economic mobility. PhD Diss., Univ. Penn., Philadelphia, PA
Kinder DR, Kalmoe NP. 2017. Neither Liberal nor Conservative: Ideological Innocence in the American Public.
Chicago: Univ. Chicago Press
Klapper JT. 1960. The Effects of Mass Communication. Glencoe, IL: Free Press
Klar S. 2013. The influence of competing identity primes on political preferences. J. Politics 75(4):1108–24
Kraft PW, Krupnikov Y, Milita K, Ryan JB, Soroka S. 2020. Social media and the changing information
environment: sentiment differences in read versus recirculated news content. Public Opin. Q. 84(S1):195–
215
Kubin E, Puryear C, Schein C, Gray K. 2021. Personal experiences bridge moral and political divides better
than facts. PNAS 118(6):e2008389118
Kunda Z. 2001. Social Cognition: Making Sense of People. Cambridge, MA: MIT Press. 4th ed.
Lasswell HD. 1948. The structure and function of communication in society. In The Communication of Ideas,
ed. L Bryson, pp. 37–51. New York: Harper
Lau RR. 2020. Classic models of persuasion. See Suhay et al. 2020, pp. 29–50
Lazarsfeld PF,Berelson B,Gaudet H.1944.The People’s Choice: How the Voter Makes Up His Mind in a Presidential
Campaign. New York: Duell, Sloan & Pearce
Leeper TJ,Slothuus R.2014.Political parties,motivated reasoning,and public opinion formation.Adv.Political
Psychol. 35(S1):129–56
Lelkes Y. 2020. National and cross-national perspectives on political media bias. See Suhay et al. 2020,
pp. 572–89
Levendusky M. 2018. Americans, not partisans: Can priming American national identity reduce affective polarization?
J. Politics 80(1):59–70
Levine AS. 2021. Single conversations expand practitioners’ use of research: evidence from a field experiment.
PS Political Sci. Polit. 54(3):432–37
Levine AS, Kline R. 2017. A new approach for evaluating climate change communication. Climatic Change
142(1–2):301–9
Levine AS, Kline R. 2019. Loss-framed arguments can stifle political activism. J. Exp. Political Sci. 6(3):171–79
Levitan LC, Verhulst B. 2016. Conformity in groups: the effects of others’ views on expressed attitudes and
attitude change. Political Behav. 38(2):277–315
www.annualreviews.org • The Study of Persuasion 85
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
Levy R. 2021. Social media, news consumption, and polarization: evidence from a field experiment. Am. Econ.
Rev. 111(3):831–70
Liberman N, Trope Y. 1998. The role of feasibility and desirability considerations in near and distant future
decisions: a test of temporal construal theory. J. Pers. Soc. Psychol. 75(1):5–18
Lodge M, Taber CS. 2013. The Rationalizing Voter. New York: Cambridge Univ. Press
Lupia A. 2006. How elitism undermines the study of voter competence. Crit. Rev. 18(1–2):217–32
Lupia A, McCubbins MD. 1998. The Democratic Dilemma: Can Citizens Learn What They Need to Know? New
York: Cambridge Univ. Press
McGraw K, Dolan TM. 2007. Personifying the state: consequences for attitude formation. Political Psychol.
28(3):299–327
McGuire W. 1969. The nature of attitudes and attitude change. In The Handbook of Social Psychology, ed.
G Lindzey, E Aronson, pp. 136–314. Reading, MA: Addison-Wesley
Ma Y, Dixon G, Hmielowski JD. 2019. Psychological reactance from reading basic facts on climate change:
the role of prior views and political identification. Envir. Commun. 13(1):71–86
Malhotra N, Michelson MR, Rogers T, Valenzuela AA. 2011. Text messages as mobilization tools: the conditional
effect of habitual voting and election salience. Am. Political Res. 39(4):664–81
Marcus GE. 2002. The Sentimental Citizen: Emotion in Democratic Politics. Univ. Park, PA: Pa. State Univ. Press
Markovich Z, Baum MA, Berinsky AJ, de Benedictis-Kessner J, Yamamoto T. 2020. Dynamic persuasion: decay
and accumulation of partisan media persuasion. Paper presented at the Annual Meeting of the Southern
Political Science Association, Jan. 9–11, San Juan, PR
Martin G, McCrain J. 2019. Local news and national politics. Am. Political Sci. Rev. 113(2):372–84
Matz SC, Kosinski M, Nave G, Stillwell DJ. 2017. Psychological targeting as an effective approach to digital
mass persuasion. PNAS 114(48):12714–19
Matz SC, Netzer O. 2017. Using big data as a window into consumers’ psychology. Curr. Opin. Behav. Sci.
18(Dec.):7–12
Miller PV. 2017. Is there a future for surveys? Public Opin. Q. 81(S1):205–12
Minozzi W, Neblo MA, Esterling KM, Lazer DMJ. 2015. Field experiment evidence of substantive, attributional,
and behavioral persuasion by members of Congress in online town halls. PNAS 112(13):3937–42
Minozzi W, Song H, Lazer DMJ, Neblo MA, Ognyanova K. 2020. The incidental pundit: Who talks politics
with whom, and why? Am. J. Political Sci. 64(1):135–51
Mousa S. 2020. Building social cohesion between Christians and Muslims through soccer in post-ISIS Iraq.
Science 369(6505):866–70
Moy P, Rinke EM. 2012. Attitudinal and behavioral consequences of published opinion polls. In Opinion Polls
and the Media: Reflecting and Shaping Public Opinion,ed.C Holtz-Bacha, J Strömbäck,pp.225–45.London:
Palgrave Macmillan UK
Mullinix KJ. 2016. Partisanship and preference formation: competing motivations, elite polarization, and issue
importance. Political Behav. 38(2):383–411
Mutz DC, Sniderman PM, Brody RA. 1996. Political persuasion: the birth of a field of study. In Political
Persuasion and Attitude Change, ed. DC Mutz, PM Sniderman, RA Brody, pp. 1–16. Ann Arbor: Univ.
Mich. Press
Nelson TE, Clawson RA, Oxley ZA. 1997. Media framing of a civil liberties conflict and its effect on tolerance.
Am. Political Sci. Rev. 91(3):567–83
Nelson TE, Oxley ZM. 1999. Issue framing effects on belief importance and opinion. J. Politics 61(4):1040–67
Nickerson DW, Rogers T. 2020. Campaigns influence election outcomes less than you think. Science
369(6508):1181–82
O’Keefe DJ. 2016. Persuasion: Theory and Practice. Los Angeles: Sage. 3rd ed.
O’Keefe DJ. 2018. Message pretesting using assessments of expected or perceived persuasiveness: evidence
about diagnosticity of relative actual persuasiveness. J. Commun. 68(1):120–42
O’Keefe DJ. 2021. Persuasive message pretesting using non-behavioral outcomes: differences in attitudinal
and intention effects as diagnostic of differences in behavioral effects. J. Commun. 71(4):623–45
Osmundsen M,Bor A,Vahlstrup PB,Bechmann A,Bang Petersen M.2021.Partisan polarization is the primary
psychological motivation behind political fake news sharing on Twitter. Am. Political Sci. Rev. 115(3):999–
1015
86 Druckman
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
Ostfeld MC. 2019. The new white flight? The effects of political appeals to Latinos on white Democrats.
Political Behav. 41(3):561–82
Pennycook G, Epstein Z, Mosleh M, Arechar AA, Eckles D, Rand DG. 2021. Shifting attention to accuracy
can reduce misinformation online. Nature 592:590–95
Pennycook G, McPhetres J, Zhang Y, Lu JG, Rand DG. 2020. Fighting COVID-19 misinformation on social
media: experimental evidence for a scalable accuracy-nudge intervention. Psychol. Sci. 31(7):770–80
Persily N, Tucker JA, eds. 2020. Social Media and Democracy: The State of the Field, Prospects for Reform. New
York: Cambridge Univ. Press
Petty R, Cacioppo J. 1986. The elaboration likelihood model of persuasion. Adv. Exp. Soc. Psychol. 19:123–205
Porter E, Wood TJ. 2019. False Alarm: The Truth about Political Mistruths in the Trump Era. New York:
Cambridge Univ. Press
Schattschneider EE. 1960. The Semisovereign People: A Realist’s View of Democracy in America. New York: Holt,
Rinehart & Winston
Settle JE. 2018. Frenemies: How Social Media Polarizes America. New York: Cambridge Univ. Press
Shadish WR, Cook TD, Campbell DT. 2002. Experimental and Quasi-Experimental Designs for Generalized
Causal Inferences. Boston: Houghton Mifflin
Sinclair B. 2012. The Social Citizen: Peer Networks and Political Behavior. Chicago: Univ. Chicago Press
Slothuus R. 2016. Assessing the influence of political parties on public opinion: the challenge from pretreatment
effects. Political Commun. 33(2):302–27
Slothuus R, Bisgaard M. 2021a. How political parties shape public opinion in the real world. Am. J. Political
Sci. 65(4):896–911
Slothuus R, Bisgaard M. 2021b. Party over pocketbook? How party cues influence opinion when citizens have
a stake in policy. Am. Political Sci. Rev. 115(3):1090–96
Sniderman PM, Theriault SM. 2004. The structure of political argument and the logic of issue framing. In
Studies in Public Opinion, ed. WE Saris, PM Sniderman, pp. 133–65. Princeton, NJ: Princeton Univ. Press
Song H, Lewis NA, Ballew MT, Bravo M, Davydova J, et al. 2020. What counts as an “environmental” issue?
Differences in issue conceptualization by race, ethnicity, and socioeconomic status. J. Env. Psychol.
68:101404
Soroka S, Fournier P, Nir L. 2019. Cross-national evidence of a negativity bias in psychophysiological reactions
to news. PNAS 116(38):18888–92
Stanley ML, Henne P, Yang BW, De Brigard F. 2020. Resistance to position change, motivated reasoning, and
polarization. Political Behav. 42(3):891–913
Stark TH, Silber H, Krosnick JA, Blom AG, Aoyagi M, et al. 2020. Generalization of classic question order
effects across cultures. Sociol. Method. Res. 49(3):567–602
Suhay E, Grofman B, Trechsel AH, eds. 2020. The Oxford Handbook of Electoral Persuasion. New York: Oxford
Univ. Press
Sydnor E. 2019. Disrespectful Democracy: The Psychology of Political Incivility. New York: Columbia Univ. Press
Tappin BM. 2020. Estimating the between-issue variation in party elite cue effects. Work. Pap., Mass. Inst. Technol.,
Cambridge, MA. https://doi.org/10.31234/osf.io/p48zb
Tappin BM,Pennycook G,Rand DG.2020.Bayesian or biased? Analytic thinking and political belief updating.
Cognition 204:104375
Tappin BM, Pennycook G, Rand DG. 2021. Rethinking the link between cognitive sophistication and politically
motivated reasoning. J. Exp. Psychol. Gen. 150(6):1095–114
Teeny JD, Petty RE. 2021. The perception of others’ attitude bases and position as antecedents of inferred openmindedness:
consequences for attitudinal advocacy.Paper presented at Psychology Department Speaker Series,
Northwestern Univ., Evanston, IL
Tesler M, Zaller J. 2017. The power of political communication. In The Oxford Handbook of Political Communication,
ed. K Kenski, KH Jamieson, pp. 69–84. New York: Oxford Univ. Press
Valentino NA, Brader T, Groenendyk EW, Gregorowicz K, Hutchings VL. 2011. Election night’s alright for
fighting: the role of emotions in political participation. J. Politics 73(1):156–70
Van Bavel JJ, Reinero DA, Spring V, Harris EA, Duke A. 2021. Speaking my truth: why personal experiences
can bridge divides but mislead. PNAS 118:e2100280118
www.annualreviews.org • The Study of Persuasion 87
Downloadedfromwww.annualreviews.org.MasarykUniversity(ar-393718)IP:147.251.15.49On:Wed,25Sept202411:15:51
Weeks BE, Ardèvol-Abreu A, Gil de Zúñiga H. 2017. Online influence? Social media use, opinion leadership,
and political persuasion. Int. J. Public Opin. Res. 29(2):214–39
Wiest SL, Raymond L, Clawson RA. 2015. Framing, partisan predispositions, and public opinion on climate
change. Glob. Envir. Change 31:187–98
Wojcieszak M, Garrett RK. 2018. Social identity, selective exposure, and affective polarization: how priming
national identity shapes attitudes toward immigrants via news selection. Human Commun. Res. 44(3):247–
73
88 Druckman