Week 1: Democracy between elections, and what officials actually hear

In plain words

Week 1 of the My Vote Forward seminar asks a simple question with no simple answer: between elections, how do officials find out what the public wants, and should they do what they hear? Students read two founding arguments, Federalist No. 10 and Edmund Burke’s 1774 speech to the electors of Bristol, and four empirical studies that measure how opinion and policy actually connect.

An election picks a person. It does not say which of that person’s positions voters liked. So in the years between votes, officials learn about the public from letters, calls, meetings, polls, and the people who contact them. This week asks how accurate that picture is, and whether officials should act on it even when it is accurate.

Defined terms

• Representation: acting for, standing for, or speaking for others in political decisions. Week 2 takes the concept apart. See the glossary entry.
• Responsiveness: the degree to which policy changes when public preferences change. It is an empirical relationship, not a moral verdict.
• Delegate and trustee: two classic models of a representative. A delegate follows constituents’ expressed wishes; a trustee follows their own judgment of constituents’ interests.
• Dyadic representation: the relationship between one representative and one district, as opposed to the relationship between a whole legislature and the whole public.
• Feedback channel: any route by which information about public preferences reaches officials between elections, such as polls, public comments, town halls, letters, and direct contact.

Learning objectives

By the end of week 1, a participant should be able to:

1. State the course’s central puzzle: what officials hear between elections, how accurate it is, and how much it should count.
2. Explain Madison’s argument in Federalist No. 10 for filtering popular passions through a large republic, and Burke’s argument for the representative’s independent judgment.
3. Summarize what Miller and Stokes, Page and Shapiro, Broockman and Skovron, and Butler and Nickerson each measured, and what each did not measure.
4. Distinguish a normative claim (what officials should do) from an empirical claim (what they do), and identify which kind each reading makes.
5. Explain why a more accurate picture of public opinion does not by itself settle whether officials should follow it.

Core readings

All readings below are real, published works. Each was checked on the review date against a library, DOI, or primary-document page listed in Sources.

1. James Madison, The Federalist No. 10 (1787). Full text at the Library of Congress.⁠¹
2. Edmund Burke, Speech to the Electors of Bristol (1774). Excerpt in The Founders’ Constitution, University of Chicago Press.⁠²
3. Warren E. Miller and Donald Stokes, “Constituency Influence in Congress,” American Political Science Review 57(1), 1963, pp. 45 to 56.⁠³
4. Benjamin I. Page and Robert Y. Shapiro, “Effects of Public Opinion on Policy,” American Political Science Review 77(1), 1983, pp. 175 to 190.⁠⁴
5. David Broockman and Christopher Skovron, “Bias in Perceptions of Public Opinion among Political Elites,” American Political Science Review 112(3), 2018, pp. 542 to 563.⁠⁵
6. Daniel M. Butler and David W. Nickerson, “Can Learning Constituency Opinion Affect How Legislators Vote? Results from a Field Experiment,” Quarterly Journal of Political Science 6(1), 2011, pp. 55 to 83.⁠⁶

Recommended: Christopher H. Achen and Larry M. Bartels, Democracy for Realists (Princeton University Press, 2016), chapter 1.⁠⁷

Lecture notes

The puzzle the course is built on

Democracy between elections is the subject of this seminar, and week 1 states the puzzle in its sharpest form. An election compresses years of experience, many subjects, and millions of individual judgments into a single choice among candidates. The winner learns that more voters preferred them than the alternative. The winner does not learn which of their positions voters supported, which they tolerated, and which they opposed, or what voters want done next. Every representative democracy therefore relies on a second set of signals that arrive between elections: letters and calls, public hearings, comments on proposed rules, surveys, organized campaigns, news coverage, and the people who happen to seek out an official. The site’s chapter on broken feedback between elections documents how those channels can be skewed or flooded.

The seminar asks three questions about that second set of signals, and keeps them apart because they are easily confused. First, an empirical question: what do officials actually hear, and how accurate is it? Second, a normative question: how much should what they hear count against their own judgment, their party’s program, or the rights of people who would lose? Third, a design question: if the signal is poor, could a better one be built, and what would it cost in privacy, equality, and manipulation risk? This week concentrates on the first two. The rest of the course builds toward the third, and the final weeks hold the framework this site describes, My Vote Forward, to the same standards as every other proposal.

Madison: a large republic as a filter

James Madison’s Federalist No. 10 is usually read as an argument about factions, groups “united and actuated by some common impulse of passion, or of interest,” whose aims run against other citizens’ rights or the community’s interests. Madison argues that the causes of faction cannot be removed without destroying liberty, so a republic must control their effects. His answer is scale and representation. In a small community, he writes, a majority is more easily found and can more easily “concert and execute their plans of oppression.” His remedy is direct: “Extend the sphere, and you take in a greater variety of parties and interests; you make it less probable that a majority of the whole will have a common motive to invade the rights of other citizens”.⁠¹

For this course, Federalist No. 10 matters because it treats some distance between public sentiment and public decisions as a feature. Representatives are supposed to “refine and enlarge the public views” by passing them through a chosen body. A reader who wants officials to hear the public more clearly has to answer Madison: what keeps a clearer signal from also being a faster channel for a passing majority? The site takes up that worry in tyranny of the majority and does a majority decide rights?.

The strongest reply to Madison, which later weeks develop, is that the filter he described does not filter evenly. If the people who reach officials between elections are mostly the organized, the wealthy, or the angry, the filter refines some voices and drops others. Week 3 examines that claim with evidence on who participates.

Burke: judgment as a duty

Edmund Burke’s 1774 speech to the electors of Bristol gives the classic statement of the trustee model. Burke accepts that a representative owes constituents close attention and should “prefer their interest to his own.” But he refuses to treat their instructions as binding: “Your representative owes you, not his industry only, but his judgment; and he betrays, instead of serving you, if he sacrifices it to your opinion”.⁠² Government and legislation, he argues, are “matters of reason and judgment, and not of inclination.”

Burke’s position is often caricatured as elitism. The strongest version is narrower. It holds that a representative sees information constituents do not, hears arguments from other districts, and is accountable at the next election for the judgment exercised. On that view, better information about constituent opinion is welcome as an input but should never become an instruction. The delegate model replies that a representative who ignores clear, informed, stable constituent preferences on a matter that affects them has stopped representing them. The course does not settle this argument. It asks students to notice which model is assumed whenever someone proposes a new feedback channel.

Miller and Stokes: measuring the link

Warren Miller and Donald Stokes moved the delegate and trustee argument from theory to measurement. Their 1963 article starts from the observation that substantial constituency influence over the House of Representatives is “commonly thought to be both a normative principle and a factual truth,” yet “much of the evidence of constituency control rests on inference”.⁠³ They frame constituency control as one pole of the debate Burke started, and they ask what a reasonable definition of control would require: that representatives know their constituents’ views, that those views shape representatives’ votes, and that constituents can tell the difference.

The lasting contribution is methodological. Miller and Stokes showed that “representation” can be broken into measurable links, each of which can fail separately. A representative can vote with constituent opinion by coincidence, through party, or through shared background, without any information flowing from the district. That warning recurs throughout the course: congruence between opinion and policy is not proof that opinion caused policy.

Page and Shapiro: opinion and policy over time

Benjamin Page and Robert Shapiro took the question to the national level. Examining United States public opinion and policy data from 1935 to 1979, they report “considerable congruence between changes in preferences and in policies, especially for large, stable opinion changes on salient issues,” and they present evidence that opinion is often a proximate cause of policy, affecting policy more than policy affects opinion. They end with a caution: one should not conclude “that democratic responsiveness pervades American politics”.⁠⁴

Two features of the finding matter for this course. Responsiveness was strongest where opinion change was large, stable, and on salient issues. That leaves a large territory, including small changes, unstable opinion, and low-salience subjects such as many local budget lines, where the evidence for responsiveness is weaker. The site’s chapter on public budget blindness takes up that territory, and week 7 asks who decides which issues become salient in the first place.

Broockman and Skovron: officials can misread the public

If officials respond to what they believe constituents want, the accuracy of that belief becomes a central variable. David Broockman and Christopher Skovron surveyed 3,765 state legislative politicians about their perceptions of constituency opinion on nine issues in 2012 and 2014. They report that politicians from both parties “dramatically overestimated their constituents’ support for conservative policies on these issues,” that Republican politicians drove much of the overestimation, and that the pattern held across methods, districts, and states. They suggest one cause: biases in who contacts politicians, since in those years Republican citizens were especially likely to contact legislators.⁠⁵

Students should read this finding carefully and without partisan spin. The authors’ general claim is that “politicians can systematically misperceive what their constituents want,” and that misperception can arise from who chooses to speak up. The mechanism is not partisan by nature. Any group that contacts officials more than others can skew an official’s picture, in any direction, on any issue. The study measured one period and one set of issues; it does not show that the same direction of bias holds everywhere or always. What it does show is that an official’s sense of “what I hear from my district” is itself a measurement, and it can be biased.

Butler and Nickerson: information can change votes

The final core reading is a field experiment. Daniel Butler and David Nickerson surveyed 10,690 New Mexicans about the governor’s spending proposals for a 2008 special session, then shared district-specific results with a randomly selected half of the legislature. Legislators who received their district’s results “were much more likely to vote in line with constituent opinion than those who did not.” The authors conclude that legislators “want to be more responsive to public opinion than they are in their natural state and can be if given solid information about constituent beliefs”.⁠⁶

This is the most hopeful result in the week, and it should be read with its limits. It is one state, one session, and one set of spending questions. It shows that information can move votes; it does not show that moving votes toward survey results produces better policy, protects minorities, or survives once interest groups learn to shape the survey. The site’s chapter on question design and week 6 show how easily survey results depend on wording, sampling, and nonresponse.

The realist challenge

Christopher Achen and Larry Bartels, the recommended reading, challenge the premise behind the whole week. They argue that voters, even well-informed ones, “mostly choose parties and candidates on the basis of social identities and partisan loyalties, not political issues,” and that democratic theory should be founded on identity groups and parties rather than on the preferences of individual voters.⁠⁷ If they are right, a better channel for individual policy preferences may measure something thin or unstable, or something that follows from group loyalty rather than independent judgment. The course returns to this challenge in weeks 6 and 8. It is the strongest objection to any project, including My Vote Forward, that treats collected preferences as a guide.

How this week frames the course

The five readings set out the terms the rest of the seminar uses. Madison and Burke supply normative reasons for distance between opinion and decision. Miller and Stokes supply a method for testing the links. Page and Shapiro show responsiveness that is real but uneven. Broockman and Skovron show that officials’ beliefs about opinion can be wrong in systematic ways, and Butler and Nickerson show that better information can change behavior. Achen and Bartels ask whether the preferences being measured are the right object at all. Every later week adds a mechanism, a failure, or a design to this frame.

Discussion questions

1. Madison treats distance between opinion and decision as protection. Under what conditions does that distance protect rights, and under what conditions does it protect incumbents or organized interests?
2. Burke says a representative betrays constituents by sacrificing judgment to their opinion. Is there any subject on which you think a representative should simply follow a clear district majority? What distinguishes it?
3. Broockman and Skovron attribute part of the misperception to who contacts officials. Design a feedback channel that would not have this problem. What new problem does your design create?
4. Butler and Nickerson show that information moved votes. Is a legislator who changes a vote after seeing a survey being more democratic, or less independent? Does your answer depend on the survey’s quality?
5. If Achen and Bartels are right that issue preferences follow group identity, what, if anything, would better feedback between elections accomplish?
6. Which channel do officials in your own community rely on most between elections? Who is missing from it?

Assignment

Feedback channel audit (1,500 words). Choose one public body you can observe: a city council, school board, county commission, state legislative committee, or a comparable body in another country. Identify every channel through which it receives public input between elections, using its own published materials (meeting rules, comment portals, agendas, minutes). For each channel, state who can use it, what it costs a participant in time and skill, how input is recorded, and whether the body is required to respond. Then apply one empirical reading from this week to predict how that channel might distort the body’s picture of public opinion, and one normative reading to argue how much weight the input should carry. Cite every factual claim to a document you opened. Do not contact officials or survey members of the public for this assignment.

Evidence

Every description of a reading’s argument or finding above is drawn from the reading itself or its published abstract, opened on the review date and cited. The Federalist and Burke quotations are taken from the primary texts linked in Sources. No result from classroom use of this module exists; none is claimed.

Competing interpretations

The week presents three positions that remain in live dispute. The responsiveness view, which Butler and Nickerson’s experiment supports in one setting, holds that better information makes representation better. The trustee view, from Burke and in a different form from Madison, holds that information should inform judgment but never bind it. The realist view of Achen and Bartels holds that individual policy preferences are a weak foundation for democratic theory at all. A reader can accept parts of each. Page and Shapiro’s own caution, that responsiveness is real but not pervasive, is a reasonable starting point for all three.

Application to this framework

My Vote Forward proposes collecting priorities and optional opinions between elections and keeping them separate from any private vote. Week 1 supplies the first test of that proposal: a new channel is only an improvement if it is less skewed than the channels it supplements, and even an accurate channel leaves open how much officials should weigh it. The framework’s chapter on representation states its position on that second question; students should test it against Burke and Madison.

Unresolved questions

• Would officials’ misperceptions shrink if a widely used, representative feedback channel existed, or would officials discount it as they discount polls they dislike?
• Butler and Nickerson studied one legislature in one session. Would the effect persist once organized groups learned to shape the survey?
• Can any feedback channel be designed that does not overweight the people with the most time and motivation?