POLS 1140

Influence and Persuasion

Updated Apr 13, 2026

Why influence and persuasion matter

Democracy assumes citizens can update their views in response to argument and evidence
- If minds never change, democratic deliberation is theater
- If minds change too easily, democracy is vulnerable to manipulation
Two central tensions in the literature:
- Minimal effects vs. real persuasion: Can campaigns, media, and conversations actually move opinions?
- Deliberation vs. participation: Does exposure to opposing views produce better or worse democratic citizens?

Questions for this week

Can political opinions be durably changed?
What conditions make persuasion more or less likely?
Does exposure to cross-cutting views help or hurt democracy?
What mechanisms underlie successful persuasion — and what does that imply for politics?

The Minimal Effects Tradition

The Columbia Studies

Lazarsfeld, Berelson, and Gaudet (1944) studied the 1940 presidential election:

Expected to find large media effects on vote choice
Instead found that most voters decided early and did not change their minds
Mass media primarily reinforced existing predispositions rather than converting voters
Coined the term “minimal effects”

Why Are Effects “Minimal”?

Several mechanisms limit persuasion:

Selective exposure: People consume media that confirms their views (Stroud 2011)
Motivated reasoning: People process new information to defend prior beliefs (Lodge and Taber 2013)
Social reinforcement: Peer networks are often politically homogeneous
Partisan identity: Strong partisans are especially resistant to cross-cutting information (Zaller 1992)

Zaller’s RAS Model

Zaller (1992) offers a systematic account of opinion change:

Citizens Receive information from elites
They Accept it if it is consistent with their predispositions
They Sample from considerations when forming a response

Political awareness increases exposure to messages, but also increases the ability to resist counter-attitudinal ones

High-awareness partisans: most exposed, most resistant
Low-awareness citizens: most persuadable, least reached

Are Minimal Effects Really Minimal?

A later generation of scholars pushed back:

Erikson and Wlezien (2012): campaigns matter at the margins — and margins decide elections
Gerber, Green, and Shachar (2003): get-out-the-vote contact causally increases turnout
Kalla and Broockman (2018): across hundreds of experiments, campaign persuasion effects average near zero — but exceptions exist
The debate is not resolved — but the field has moved toward asking when and how persuasion occurs, not just whether

Deliberative Democracy and Cross-Cutting Networks

The Deliberative Ideal

A core normative tradition in democratic theory holds that democracy requires deliberation:

Citizens should be exposed to diverse viewpoints
Through reasoned argument, they should update beliefs and develop mutual understanding
Habermas (1984): the “ideal speech situation” — open, equal, reason-giving discourse
Gutmann and Thompson (1996): deliberation produces legitimate democratic outcomes

What Are Cross-Cutting Networks?

A cross-cutting network is a social environment where people regularly interact with those who hold different political views
- Workplace colleagues with different party affiliations
- Mixed neighborhoods or religious communities
- Families with political disagreement
Contrast: reinforcing networks, where social ties share political views
- Common in partisan media environments and geographically sorted communities

The Democratic Expectation

Deliberative theorists predict that cross-cutting exposure should:

Deepen citizens’ understanding of their own views by requiring them to defend and articulate their positions
Increase awareness of legitimate rationales for opposing views
Foster political tolerance — willingness to extend rights to those with whom one disagrees
Ultimately produce more reflective, informed citizens

Mutz (2002): Testing Democratic Theory

Mutz (2002), “Cross-Cutting Social Networks: Testing Democratic Theory in Practice,” APSR

National telephone survey (Spencer Foundation, fall 1996, N ≈ 780) with detailed discussion network items
Supplemented by a laboratory experiment (N = 82 students)
Central question: does cross-cutting exposure actually produce the benefits democratic theorists claim?
Tests three hypothesized benefits: awareness of own-view rationales, awareness of opposing-view rationales, and political tolerance

Measuring Cross-Cutting Exposure

Respondents name up to three political discussants and report:
- How frequently they discuss politics with each person
- Whether that person’s views are similar to or different from their own
Exposure to dissonant views: frequency-weighted index of discussion with those holding different political views
Exposure to consonant views: frequency-weighted index of discussion with like-minded others
Key: these are not zero-sum — more cross-cutting contact does not require less consonant contact
Only about 11–14% of discussants in the sample hold clearly different views — cross-cutting exposure is rare

Hypothesis 1: Does Cross-Cutting Exposure Deepen Your Own-View Knowledge?

Prediction: confronting opposing views forces you to articulate and defend your own positions, sharpening your reasoning
Measurement: open-ended questions asking respondents to give reasons for their own positions on three issues (welfare federalism, affirmative action, presidential candidate)
Finding: No effect. Cross-cutting exposure is unrelated to the number of rationales respondents can give for their own views
The null result is informative — it rules out a simple “greater political engagement” explanation for the positive findings below

Hypothesis 2: Does Cross-Cutting Exposure Build Understanding of Opposing Views?

Prediction: exposure to different views imparts information about legitimate rationales for positions one does not hold
Measurement: “Regardless of your own views, what reasons can you think of for [opposing position]?”
Finding: Yes, significantly. Cross-cutting exposure substantially increases awareness of legitimate rationales for opposing views
Magnitude: those with the highest cross-cutting exposure can generate approximately two additional rationales for opposing views, compared to those in fully homogeneous networks
Effect is specific to dissonant contact — consonant contact has a negative coefficient

Two Pathways to Political Tolerance

Mutz identifies two mechanisms linking cross-cutting exposure to political tolerance:

Mechanism	Path
Cognitive	Cross-cutting → awareness of opposing rationales → tolerance
Affective	Cross-cutting → intimacy within dissonant dyads → tolerance

The Role of Perspective-Taking: Experimental Evidence

Mutz’s lab experiment randomly assigns subjects to receive rationales for dissonant, consonant, or no political views on three issues

Overall group differences in tolerance trend in the right direction but are not significant
Key moderator: perspective-taking ability — the cognitive capacity to entertain others’ viewpoints (Davis 1983)
- Among those high in perspective-taking: dissonant exposure increases tolerance by approximately 14% on the scale
- Among those low in perspective-taking: no benefit — may actually counter-argue more intensely
Cross-cutting exposure benefits tolerance most for citizens already inclined to see things through others’ eyes

Group Discussion

Can you think of a time when talking with someone who disagreed with you changed how you understood their position — even if not your own?
Does knowing the reasons behind an opposing view make you more likely to support others’ right to hold it?
Mutz finds cross-cutting contact is rare. What structures in contemporary life make it rarer or more common?

Mechanisms of Persuasion

How Does Persuasion Work?

Even if effects are small on average, persuasion clearly occurs under some conditions. What drives it?

Information provision: New facts change beliefs (Gilens 2001)
Elite cueing: Partisan cues shift opinions regardless of argument content (Zaller 1992)
Framing: How an issue is described shapes how people evaluate it (Chong and Druckman 2007)
Social pressure and norms: Conformity to perceived group expectations (Gerber, Green, and Larimer 2008)
Moral reframing: Connecting an issue to the values a person already holds

Information vs. Identity

A key debate in the literature:

Are opinions driven by information (what citizens know)?
Or by identity (who citizens are — partisans, racial group members, etc.)?
Evidence increasingly favors identity:
- Providing factual corrections often fails to change beliefs — and can backfire (Nyhan and Reifler 2010)
- Partisan cues can override substantive argument content (Broockman and Butler 2017)

Framing Effects

Chong and Druckman (2007) distinguish:

Equivalence framing: logically identical information presented differently (classic Kahneman and Tversky)
Emphasis framing: highlighting different considerations or values relevant to an issue — more common in political life
- “Welfare” vs. “assistance to the poor”
- “Estate tax” vs. “death tax”
- “Undocumented immigrants” vs. “illegal aliens”

Contact and Perspective-Taking

Extending Allport’s (1954) contact hypothesis to political persuasion:

Sustained, equal-status contact with outgroup members reduces prejudice
Conditions that facilitate contact effects:
- Absence of hierarchy between groups
- Cooperative interdependence
- Opportunities for personal acquaintance
Political implications: does personal contact with political opponents — or with members of stigmatized groups — reduce hostility?

Enter Personalized Moral Reframing

Traditional canvassing experiments show modest, short-lived attitude change

An extension — personalized moral reframing — goes further:

Canvassers first listen to identify which moral values a voter holds
They then reframe the persuasive appeal in terms of those specific values
Grounded in moral foundations theory (Graham, Haidt, and Nosek 2009)
Kalla, Levine, and Broockman (2022) test this approach on a highly contested issue: abortion access in Maine

Kalla, Levine, and Broockman (2022)

Review

Take a few minutes to review your notes on Kalla, Levine, and Broockman (2022)

What is the research question?
What is the theoretical framework?
What is the design?
What are the results?

Background: The Challenge of Persuasion

Organizations seeking to change citizens’ minds or mobilize them face substantial barriers in the contemporary United States
Many have given up on persuasion, focusing instead on turning out voters who already agree with them
Moral reframing (Feinberg and Willer 2015): framing a position as consistent with individuals’ own moral values can strengthen or change those views
- Example: environmental conservation framed around “purity/sanctity” is more persuasive to those who hold that value
Core problem: enormous heterogeneity in citizens’ values — it is difficult to predict from available data which moral foundations matter to whom

The Solution: Personalized Moral Reframing

Kalla, Levine, and Broockman (2022) test a strategy that resolves this problem:

Listen first: during an interpersonal conversation, identify which moral values the person actually expresses
Tailor the appeal: reframe the cause in terms of those specific values
Interpersonal conversations — door-to-door canvassing, phone calls, workplace discussions — are well-suited to this strategy because persuaders can hear the values before speaking
Neither moral reframing of any kind in a field context, nor personalized moral reframing specifically, had been tested before this study

Research Question

Can personalized moral reframing in door-to-door canvassing:

Increase support for abortion access (policy attitudes)?
Increase interest in taking pro-abortion political action (action-taking)?
Decrease stigma toward women who have abortions?

Abortion is explicitly chosen as a hard test: attitudes are deeply crystallized, partisan disagreement has been high for decades, and most citizens see their position as tied to a core moral value

Design: Field Experiment in Maine

Partnership with Planned Parenthood of Northern New England (PPNNE), summer 2018 — when advocates feared a new Supreme Court majority might overturn Roe v. Wade
Registered voters in three Maine counties recruited via mail (N = 112,010) to an ostensibly unrelated online survey
Baseline respondents (N = 3,348) randomly assigned to treatment or placebo
Treatment: personalized moral reframing canvass (~11.8 minutes on average)
Placebo: canvassing conversation about Medicaid expansion in Maine
Individuals who answered the door (N = 1,034) recruited to follow-up surveys at 1 week, 1 month, and 3 months

Moral Foundations Theory

Canvassers trained in moral foundations theory (Graham, Haidt, and Nosek 2009):

Foundation	Core Concern
Care/Harm	Protection from suffering
Fairness/Cheating	Equal treatment, justice
Loyalty/Betrayal	Commitment to family and group
Authority/Subversion	Deference to tradition, hierarchy
Sanctity/Degradation	Purity, bodily integrity

After the study, canvassers reported that voters most often expressed the loyalty and care foundations
The goal: identify the bridge between the voter’s expressed values and the case for abortion access

How Personalized Moral Reframing Works in Practice

Canvasser asks the voter to share experiences that shaped their views on abortion
As the voter speaks, canvasser listens for which moral foundation is being expressed
Canvasser shares their own story, explicitly naming the shared moral value (e.g., “Just like you, I value loyalty to my family…”)
Canvasser asks what the experience should be like for a woman who decides to end a pregnancy
Canvasser argues for abortion access using the voter’s own moral language (e.g., “Women who decide to end their pregnancies should be treated fairly, like any other medical decision”)
Voter invited to reflect; those who express support are asked to contact their elected officials

Outcomes Measured

Three preregistered outcome indexes, all standardized (mean 0, SD 1):

Action-taking index: willingness to take five specific actions in the next month (accompany someone to an abortion clinic, write to Congress to support access, protest outside a clinic [reverse-coded], etc.)
Policy attitudes index: eleven questions about when abortion should be allowed during pregnancy and Maine laws governing abortion procedures and insurance coverage
Stigma index: five questions about the morality of abortion and a feeling thermometer toward women who have had abortions

Results: Action-Taking

The strongest and most consistent effects are on action-taking interest:

1 week post-canvass: d = 0.19 (p < .001) — large, significant effect
1 month post-canvass: d = 0.11 (p = .04) — effect persists
3 months post-canvass: d = 0.08 (p = .13) — decays to nonsignificance
Effects consistent across items, and especially strong for:
- Willingness to accompany someone to an abortion clinic
- Willingness to write to Congress in support of abortion access
Effects do not differ significantly depending on whether the canvasser had personally had an abortion

Results: Policy Attitudes and Stigma

Policy attitudes: barely significant at 1 week (d = 0.061, p = .054); effect decays in follow-up waves
- Stands in contrast to the precise null results from Broockman et al. (2017)
Stigma: no significant effect; point estimates are positive but indistinguishable from zero
Planned Parenthood favorability: significant increase (d = 0.11, p = .001), though also appears to decay

Summary of Results

Outcome	Effect at 1 Week	Durability
Action-taking interest	Strong (d = 0.19)	Significant at 1 month; decays at 3 months
Policy attitudes	Marginal (d = 0.061, p = .054)	Short-lived
Stigma toward women	Null	—
PP favorability	Moderate (d = 0.11)	Short-lived

The clearest success is in motivating supporters to act, not in converting opponents
Even on a highly crystallized issue, moral reframing can shift behavioral intentions

Why Does Action-Taking Respond More Than Policy Attitudes?

Several possible interpretations:

Policy attitudes on abortion are among the most entrenched positions in American politics — accumulated over years of partisan signaling and personal moral commitment
Action-taking interest is a lower-commitment behavioral intention — it requires aligning with existing values, not abandoning them
Moral reframing may work best at activating latent sympathies rather than reversing settled positions
Effect decay after three months suggests the value salience fades once the conversation ends — attitude change is not durably encoded

Caveats and Limitations

No treatment condition without personalized moral reframing — the causal mechanism cannot be definitively established, only inferred by comparison with Broockman et al. (2017)
Maine is 95% non-Hispanic white — limits generalizability, particularly on a highly gendered issue where race may interact with attitudes and stigma
Effects decay after three months, perhaps as the salience of the named values for subjects’ abortion attitudes recedes
Future research should: replicate on other issues, collect behavioral (not just survey) outcomes, and record which moral frames voters actually articulate

Group Discussion

Why might it be easier to shift action-taking interest than policy attitudes?
What does it mean for democratic persuasion that the most effective approach requires listening before speaking?
What are the ethical implications of identifying a voter’s moral values and then deliberately using them to advance a political cause?