Week 3: Who participates, and whose preferences count

In plain words

Week 3 of the My Vote Forward seminar asks whether some citizens take part in politics more than others, and whether those who take part more also get more of the policies they want. Students read the resource model of participation, Arend Lijphart’s 1996 presidential address on unequal turnout, the much-cited Gilens and Page study of policy influence, and two published critiques of it.

Voting, volunteering, donating, and contacting officials all take resources. People with more time, money, and practice at speaking and organizing tend to do more of them. That part is well supported. Whether government then ignores everyone else is harder to prove, and serious researchers disagree. This week reads both sides.

Defined terms

• Resource model of participation: the theory that time, money, and civic skills explain much of who participates, and why socioeconomic status predicts participation.
• Civic skills: communication and organizational capacities, such as running a meeting or writing a letter, that make political activity easier.
• Unequal turnout: turnout that differs systematically across groups, as opposed to merely low turnout.
• Policy influence: the independent effect of a group’s preferences on policy outcomes, after accounting for other groups’ preferences.
• Coincidental representation: a situation in which a group gets the policy it prefers because another, more influential group prefers the same thing.
• See also participation and opt-in sample.

Learning objectives

By the end of week 3, a participant should be able to:

1. Explain the resource model and why different political acts require different resources.
2. State Lijphart’s five reasons low turnout is a democratic problem and his two proposed remedies.
3. Describe the design and headline finding of Gilens and Page (2014).
4. State the specific methodological objections raised by Enns (2015) and Bashir (2015), and what each implies.
5. Explain why a channel that collects input from volunteers can reproduce the participation inequalities this week documents.

Core readings

1. Henry E. Brady, Sidney Verba, and Kay Lehman Schlozman, “Beyond SES: A Resource Model of Political Participation,” American Political Science Review 89(2), 1995, pp. 271 to 294.⁠¹
2. Sidney Verba, Kay Lehman Schlozman, and Henry E. Brady, Voice and Equality: Civic Voluntarism in American Politics (Harvard University Press, 1995), introduction.⁠²
3. Arend Lijphart, “Unequal Participation: Democracy’s Unresolved Dilemma,” American Political Science Review 91(1), 1997, pp. 1 to 14.⁠³
4. Martin Gilens and Benjamin I. Page, “Testing Theories of American Politics: Elites, Interest Groups, and Average Citizens,” Perspectives on Politics 12(3), 2014, pp. 564 to 581.⁠⁴
5. Peter K. Enns, “Relative Policy Support and Coincidental Representation,” Perspectives on Politics 13(4), 2015, pp. 1053 to 1064.⁠⁵
6. Omar S. Bashir, “Testing Inferences about American Politics: A Review of the ‘Oligarchy’ Result,” Research and Politics 2(4), 2015.⁠⁶

Lecture notes

From the realist challenge to the participation problem

Unequal participation is the subject of week 3, and it follows from week 1’s question about what officials hear. If officials learn about the public largely from the people who vote, write, call, donate, and attend, then the composition of those people shapes the picture. Week 3 asks two questions in order. Who participates, and why? And does unequal participation translate into unequal influence over policy? The first question has a strong, well-replicated answer. The second is one of the most contested findings in recent political science, and the contest is itself instructive.

The resource model

Henry Brady, Sidney Verba, and Kay Lehman Schlozman set out the resource model in 1995. The resources they consider are “time, money, and civic skills,” the last defined as “those communications and organizational capacities that are essential to political activity.” They argue that civic skills are acquired early in life and also developed in the “nonpolitical institutional settings of adult life: the workplace, organizations, and churches and synagogues.” These resources are distributed unevenly across socioeconomic groups, and the authors’ two-stage least squares analysis shows the resources “have powerful effects on overall political activity,” which explains why socioeconomic status has traditionally predicted participation so well.⁠¹

The most useful part of the model for this course is its disaggregation. The authors separate acts that require giving time, acts that require donating money, and voting, and show that “each requires a different configuration of resources resulting in different patterns of stratification across various political acts”.⁠¹ Donating money is stratified by income in a way that voting is not; contacting officials and attending meetings are stratified by time and skills. The companion book, Voice and Equality, frames the broader claim: citizens’ involvement in “family, school, work, voluntary associations, and religion has a significant impact on their participation as voters, campaigners, donors, community activists, and protesters”.⁠²

The implication for between-election channels is direct. A public hearing on a weekday evening requires time, transportation, and confidence speaking in public. An online comment portal requires internet access, literacy in the portal’s language, and the skill to write a persuasive comment. Each channel selects for a different mix of resources, and none is neutral. The site’s chapters on unequal participation, disability access, and language access document specific barriers.

Lijphart: unequal turnout as a dilemma

Arend Lijphart’s 1996 presidential address to the American Political Science Association, published in 1997, makes the normative case. He gives five reasons low voter turnout is “a serious democratic problem”: it means “unequal turnout that is systematically biased against less well-to-do citizens”; unequal turnout “spells unequal political influence”; United States turnout is especially low, though turnout measured against voting-age population is also relatively low in many other countries; turnout in “midterm, regional, local, and supranational elections” tends to be especially poor; and turnout “appears to be declining everywhere”.⁠³

His remedies come in two forms. One is a package of institutional choices that raise turnout: “voter-friendly registration rules, proportional representation, infrequent elections, weekend voting, and holding less salient elections concurrently with the most important national elections.” The other, which he argues can maximize turnout by itself, is compulsory voting, whose advantages he says “far outweigh the normative and practical objections to it”.⁠³ The site’s case study of compulsory voting in Australia examines the practice, and the low turnout in local elections chapter examines Lijphart’s fourth point.

The strongest objections to Lijphart deserve equal time. Some argue that compulsion violates a freedom not to participate, and that a vote cast under penalty carries less information than a vote cast by choice. Others argue that voters who would otherwise abstain may know less about the choices and add noise rather than signal. A third objection, relevant to this course, is that raising turnout in elections does nothing for the much more stratified acts between elections, such as contacting officials, which the resource model shows are where inequality is sharpest.

Gilens and Page: the influence test

Martin Gilens and Benjamin Page’s 2014 article moved the debate from participation to influence. They set four theoretical traditions against one another: “Majoritarian Electoral Democracy, Economic-Elite Domination, and two types of interest-group pluralism, Majoritarian Pluralism and Biased Pluralism.” Using “a unique data set that includes measures of the key variables for 1,779 policy issues,” they report that “economic elites and organized groups representing business interests have substantial independent impacts on U.S. government policy, while average citizens and mass-based interest groups have little or no independent influence.” Their results, they write, provide “substantial support” for Economic-Elite Domination and Biased Pluralism, but not for Majoritarian Electoral Democracy or Majoritarian Pluralism.⁠⁴

The article drew enormous public attention, often summarized more strongly than the authors stated it. The site’s chapter on money in politics and the objection page can wealthy groups dominate? discuss the finding in context. For this seminar, the important feature is the design: survey questions about proposed policy changes, with preferences estimated for citizens at different income levels, including the 50th income percentile that Bashir’s critique below examines, compared against whether the policy change happened. That design is powerful and also vulnerable, because the preferences of the middle and the affluent are highly correlated, and separating their independent effects is statistically demanding.

Enns: coincidental representation

Peter Enns offers what he calls a “cautionary note.” He demonstrates that “even on those issues for which the preferences of the wealthy and those in the middle diverge, policy ends up about where we would expect if policymakers represented the middle class and ignored the affluent.” This happens, he argues, because even when the two groups express different levels of support for a policy, “the policies that receive the most (least) support among the middle typically receive the most (least) support among the affluent.” He concludes that “the opportunity of unequal representation of the ‘average citizen’ is much less than previously thought,” while also finding that “substantial opportunity exists for unequal representation of strong partisan preferences”.⁠⁵

Enns does not claim the affluent lack influence. His point is that when two groups rank policies in the same order, it is hard to tell from outcomes which group is being followed, and the average citizen often gets what they want whether or not they are the cause.

Bashir: the test underestimates the middle

Omar Bashir’s critique targets the statistical test directly. He argues that the model on which the original study is based “is prone to underestimating the impact of citizens at the 50th income percentile by a wide margin,” and that a descriptive analysis of the authors’ own dataset shows “average Americans have received their preferred policy outcome roughly as often as elites have when the two groups have disagreed with each other.” He concludes that evidence that average citizens are effectively ignored “may not be as strong as is suggested by the authors”.⁠⁶

Reading the dispute fairly

Students sometimes treat a published critique as a refutation. That is a mistake in both directions. Gilens and Page did not claim average citizens never get what they want; they claimed that, once elite and organized-interest preferences are accounted for, average citizens’ preferences add little independent explanatory power. Enns and Bashir do not show that economic elites lack influence; they show that the specific inference about average citizens is fragile under different analyses. Students who want to follow the exchange should search the journal’s later issues for any replies and read them in full rather than relying on summaries.

What all sides agree on is narrower and still important. Preferences of different income groups are often similar, influence is hard to separate statistically when they are, and the cases where groups disagree are where the stakes are highest and the evidence thinnest. The policy debate about money, lobbying, and access should be argued on that shared ground.

What the week means for feedback channels

The resource model predicts that any voluntary channel will overrepresent people with time, skills, and motivation. The site’s research chapters show the same pattern in survey data: opt-in samples tend to overrepresent politically and civically engaged people, a result discussed in measurement and taken up in week 6. A channel meant to correct unequal influence can therefore reproduce it, unless it uses methods such as probability sampling, deliberate outreach, or random selection that break the link between resources and voice. Weeks 6, 8, and 10 examine those methods and their costs.

Common misreadings this week guards against

Three misreadings recur in public discussion of these studies. The first treats unequal participation as proof of unequal influence. The resource model establishes the first; the influence debate shows the second is much harder to demonstrate. The second treats a published critique as a refutation, or a famous finding as settled, when both sides of the Gilens and Page exchange agree the evidence is contested at its most important point. The third treats higher turnout as a cure for all participation inequality, when the resource model shows that acts such as contacting officials and donating are more stratified than voting. Students who avoid these three errors will read most political claims about “who has a voice” more accurately.

Discussion questions

1. Which political acts in your community require the most civic skills? Who has had the chance to develop them, and where?
2. Lijphart argues that compulsory voting’s advantages outweigh its objections. State the strongest version of the opposing view. Which side does the resource model help more?
3. Gilens and Page and their critics agree that middle and affluent preferences often coincide. Why does that make influence hard to measure? Can you design a better test?
4. Does it matter whether the average citizen gets their preferred policy because they caused it, or by coincidence? Why?
5. If a new channel attracted mostly retirees, or mostly young people, or mostly people with strong partisan identities, how would that change what it reports?
6. Should a channel that cannot achieve representative participation publish its results at all?

Assignment

Replication memo (1,500 words). Read Gilens and Page and one critic (Enns or Bashir) in full. Write a memo, addressed to a nonspecialist legislative staff member, that states (a) what the original study measured, (b) the critic’s specific objection, (c) what each side would predict about a case where middle-income and affluent preferences diverge, and (d) what kind of evidence would settle the dispute. Do not take a side unless the evidence clearly supports one; explaining honestly why it is unsettled is a full-credit answer.

Evidence

Each description of a reading’s argument or finding above is drawn from its published abstract or, for Voice and Equality, the publisher’s description, opened on the review date. The figure of 1,779 policy issues is from the Gilens and Page abstract. The remarks about the Gilens and Page design summarize the article’s approach as described in the critiques; students should verify them against the article.

Competing interpretations

On participation, the resource model is widely accepted, but scholars weigh other factors, such as mobilization by parties and groups, and psychological engagement, differently. On influence, three positions remain defensible: that economic elites and business groups dominate policy independent of the average citizen (Gilens and Page); that the average citizen is often well represented, partly by coincidence, with partisan preferences the bigger source of inequality (Enns); and that the original test understates the middle’s influence (Bashir). On remedies, Lijphart’s case for compulsory voting is contested by those who value a freedom to abstain and those who doubt that compelled votes improve outcomes.

Application to this framework

My Vote Forward would collect priorities from people who choose to take part. This week’s readings predict that such participation will be stratified by time, skill, and motivation, unless the design breaks that link. The framework’s priorities and published releases chapters must therefore report who took part alongside what they said, and must not present an opt-in aggregate as the voice of the public.

Unresolved questions

• Can any between-election channel achieve participation that is less stratified than voting?
• If influence is hard to measure when groups agree, should research focus on the cases where they disagree, even though those cases are fewer?
• Would weighting or sampling solve the participation problem for a priorities channel, or only hide it?