Chapter 14 Key Takeaways

Core Concepts

The Voting Paradox Standard rational choice theory predicts that individuals should not vote, since the probability that any single ballot is decisive is effectively zero and voting has real costs. This paradox is resolved by incorporating expressive utility (D term), social pressure effects, and civic identity into models of participation. Campaigns leverage all three levers in effective GOTV messaging.

U.S. Turnout in Comparative Perspective The United States consistently ranks among the lowest in voter turnout among OECD democracies. This gap is attributable to a combination of structural factors — opt-in registration, weekday voting, winner-take-all electoral rules — and cultural factors. International comparisons suggest structural reforms could meaningfully increase participation without changing culture.

Registration as a Barrier Individual-responsibility voter registration substantially suppresses participation, particularly among lower-income, younger, and more mobile voters. Automatic Voter Registration reduces this barrier and increases the size of the registered electorate. Roll purges and exact-match policies impose additional barriers that research suggests fall disproportionately on minority communities.

The GOTV Evidence Hierarchy Three decades of field experiments have established a clear evidence hierarchy for GOTV interventions: - Highest effect: Personal canvassing (3–9 pp), social pressure mailers (1–3 pp with ethical concerns), P2P texting (1–3 pp) - Moderate effect: Live phone calls by trained callers (3–5 pp) - Weak or null: Robocalls, generic direct mail, most digital advertising

Voting as Habit Participation in one election significantly increases the probability of participation in subsequent elections, controlling for underlying propensity. This finding supports long-term investment in first-time voter mobilization, particularly among young voters who have the most elections ahead of them.

Differential Turnout and Electoral Outcomes Because different demographic groups have different voting rates and different political preferences, the composition of the actual voting electorate systematically differs from the eligible electorate. In competitive elections, the relative turnout rates of each party's coalition often matter more than persuasion at the margin.

Turnout Modeling Modern turnout propensity scores combine vote history (the dominant predictor) with demographic and consumer data attributes. The distinction between calibration (probabilities are accurate) and discrimination (rankings are correct) matters for different use cases. Models degrade in novel electoral environments and require validation against recent analogous elections.

Key Terms

  • Paradox of participation: The puzzle that rational actors vote despite near-zero probability of being decisive
  • Expressive voting: Casting a ballot to affirm identity or duty rather than to influence outcomes
  • D term: Intrinsic civic satisfaction from voting, added to the Downs utility equation to resolve the paradox
  • Automatic Voter Registration (AVR): Systems that register eligible citizens by default through government interactions, reversing the opt-in burden
  • GOTV (Get Out the Vote): Organized campaign efforts to mobilize supporters to vote
  • Turnout propensity score: A voter-level probability estimate for casting a ballot in a given election
  • Habit formation: The process by which voting in one election increases the probability of voting in subsequent elections
  • Surge and decline: The pattern of higher turnout in presidential versus midterm elections, with compositional differences
  • Compositional effect: The difference in demographic makeup between the eligible voting population and the actual voting electorate
  • Calibration: Whether a model's predicted probabilities match observed rates; distinguished from discrimination (correct rank-ordering)

Analytical Skills Developed

  1. Applying the Downs voting utility formula and understanding its limitations
  2. Evaluating GOTV interventions based on evidence quality and effect size
  3. Building simple turnout propensity scores from vote history and demographic data
  4. Distinguishing calibration from discrimination in predictive models
  5. Calculating the vote-generation impact of different resource allocation strategies
  6. Analyzing the electoral implications of differential turnout across demographic groups

Common Misconceptions

"High turnout always helps Democrats." This was more consistently true in earlier eras but is less automatic now. Higher turnout in some populations (e.g., non-college white voters) benefits Republicans; higher turnout among young voters and voters of color benefits Democrats. What matters is which groups' turnout increases, not total turnout per se.

"Door-knocking is inefficient — digital is cheaper." Cost per contact is lower for digital, but effect per contact is dramatically lower. The relevant comparison is cost per vote generated, and door-to-door canvassing typically wins this comparison for high-priority targets, though digital may be superior for very low-cost, high-volume outreach to low-priority targets.

"If you're in a safe state, your vote doesn't matter." This overlooks state and local elections, primary elections, and the aggregate effects of turnout on party organizational health, fundraising, and national narratives. Turnout in "safe" states also affects the popular vote margin, which carries political weight even without electoral college consequences.

The Nadia-Jake Lesson

The tension between Nadia's model-driven resource allocation and Jake's intuition-based base prioritization is a microcosm of the broader challenge in applied analytics: formal models optimize well-defined objectives but miss poorly-specified variables. Jake's instinct about base communities reflects real organizational dynamics — morale, network effects, volunteer recruitment — that are genuinely important but hard to quantify. The most effective campaigns integrate both: formal models for the marginal allocation decisions, human judgment for the organizational and relational dimensions that models underweight.