Chapter 22: Key Takeaways

The headline points from this chapter, organized around the chapter's learning objectives.

Turnout

  • Three denominators, three different numbers. VAP (all adults) is the dirtiest. Registered voters is the smallest. VEP (voting-eligible population) is the cleanest measure of "Americans who could legally vote" and is the standard for serious turnout analysis.
  • Long-term U.S. presidential turnout has hovered between 55-65% of VEP since 1972. 2020 was 67%, the highest since 1900. 2024 was ~64%, the second-highest in over a century. The post-2016 turnout floor appears to have shifted upward from the pre-2016 norm.
  • Midterm turnout ranges from 36-50% of VEP, with 2018 (50%) and 2022 (46%) the highest midterms in a century.
  • Off-year, special, and primary elections typically see much lower turnout — often 5-25% — magnifying the influence of high-engagement voters.
  • International comparison: When U.S. turnout reaches 2020 levels, it is competitive with peer voluntary-voting democracies (Canada, U.K., Germany). The very high-turnout countries (Australia, Belgium) achieve their numbers through compulsory voting.

Demographic predictors of turnout

  • Age. ~70-75% of voters 65+ vote in presidential years; ~45-50% of 18-29 do. Driven by life-stage effects and habit formation.
  • Education. The strongest single predictor. ~78% of college graduates vote in presidential years vs. ~38% of those with less than a high school education — a 40-point gap. Education provides civic skills, social networks, civic knowledge.
  • Income. Top-quartile turnout exceeds bottom-quartile turnout by ~33 points, partly reflecting education correlation but with independent effect.
  • Race/ethnicity. White > Black > Asian American > Hispanic in presidential turnout. The Black-white gap has narrowed substantially since 2008; the Hispanic-white gap remains larger.
  • Gender. Women have voted at slightly higher rates than men since 1980 (3-5 points).
  • Marital status, religious attendance, civic engagement. All predict turnout, indexing community embeddedness more than any single trait.

Partisan demographic correlates (post-2016)

  • Education realignment is the dominant electoral fact of the post-2016 era. College-educated voters now lean Democratic; non-college voters now lean Republican. The gap was ~0 in 2008 and ~14 points in 2024.
  • Race. Black voters remain ~85-90% Democratic. Hispanic voters moved from ~71% Democratic in 2012 to ~53% in 2024 — one of the largest single-decade demographic shifts in modern polling history. Asian American voters also moved modestly toward Republicans.
  • Religion. White evangelicals ~80% Republican. Catholics roughly even (white Catholics R-leaning, Hispanic Catholics D-leaning). Jewish voters ~70% Democratic. "Nones" ~60% Democratic.
  • Geography. A density gradient: urban Democratic, rural Republican, suburbs the contested middle. Density now predicts vote choice more powerfully than income alone.
  • Gender. Women lean Democratic by 7-12 points; the gender gap has been remarkably stable since 1980. The gap interacts with marriage — single voters lean more Democratic than married voters.
  • Generation. Younger generations lean Democratic on aggregate, but Gen Z's pattern is more complicated than Millennials' was. The 18-29 gender gap in 2024 was the largest ever measured.

Why people vote (or don't)

  • The rational-choice paradox. A single vote has essentially zero probability of being decisive. Theoretical answers — civic-duty value, expressive value, group-conformity value, investment value — collectively explain why people vote, though none alone is fully satisfying.
  • Resource model (Verba/Schlozman/Brady): voting requires time, money, and civic skills, all unequally distributed.
  • Mobilization matters but with sharply diminishing returns. Door-knocking is most effective (~7 points). Phone-banking is next (~3 points). Robo-calls and generic email are nearly worthless.
  • Habit. First-time voters are substantially more likely to vote in subsequent elections. Voting is habit-forming; non-voting is also habit-forming.

How people choose

  • Party identification is the strongest single predictor of vote choice. Acquired in childhood/early adulthood, stable thereafter. ~90% of self-identified Democrats vote Democratic; ~90% of Republicans vote Republican.
  • Issue positions matter at the margin, especially for the ~10-15% who are genuinely persuadable.
  • Candidate evaluation has independent effects, especially in primaries.
  • Retrospective voting rewards incumbents in good times, punishes in bad. Achen & Bartels argue this is closer to arbitrary blame-assignment than rational evaluation; Healy & Lenz argue it is clumsy but real.
  • Economic voting is increasingly partisan-filtered: voters perceive the economy through their partisan lens, except when conditions are extreme enough to break through (the 2022-2024 inflation episode).

Voter registration and the franchise

  • NVRA (1993, "Motor Voter") required states to allow registration through DMVs and public-assistance offices.
  • HAVA (2002) modernized voting infrastructure post-2000-Florida.
  • Online registration is available in ~42 states; same-day registration in ~22 states; automatic voter registration in ~22 states. Each has modest positive effects on registration and turnout.
  • Voter ID is required in ~36 states with varying strictness.

Mail and early voting

  • The 2020 transformation. Pandemic-driven mail-vote expansion brought ~65% of votes to be cast before election day, up from ~35% in 2016. Mail voting alone accounted for ~39% of 2020 votes.
  • Partisan asymmetry in 2020 was substantial (Democrats voted by mail much more), narrowed in 2024 as Republican Party leadership reversed course on early/mail voting.
  • Permanent mail-ballot states (CA, CO, OR, WA, UT, VT, HI, NV) have higher turnout on average than otherwise-similar states with traditional polling-place voting.

Structural barriers and the suppression debate

  • Progressive critique: Many post-2010 voting laws were enacted by Republican legislatures in states with shifting demographics; even race-neutral laws have demonstrable disparate impacts; some laws (NC 2013) have been struck down for being targeted with "almost surgical precision."
  • Conservative defense: Election integrity has independent value; voter ID is normal in most advanced democracies; the empirical evidence on suppression effects is contested; universal mail balloting and unrestricted ballot collection carry real administrative risks.
  • The empirical evidence: The most rigorous studies find that effects on aggregate turnout are smaller than activists on either side claim, but effects can be consequential at the margin and can have disparate impacts across demographic groups.
  • The honest reading: The U.S. has a state-by-state administrative patchwork messier than peer democracies. Reform proposals from both directions are responding to real concerns even when their advocates overstate their case.

The registration cliff

  • About 25% of voting-eligible Americans (~60 million people) are not registered. They are disproportionately younger, more mobile, lower-income, and non-white.
  • They are the voters who would most affect election outcomes if they participated, and they are the hardest to reach.
  • American elections are not decided by the median voter alone. They are decided by the median voter who shows up.

Recurring themes from this chapter

  • Power flows to those who show up. Older, more educated, higher-income voters disproportionately decide American elections — not because they are more informed, but because they vote.
  • Data beats anecdote. Demographic and partisan patterns are knowable from rigorous data, and they consistently surprise the unanecdotal expectations.
  • The American system was designed for disagreement. Including disagreement about the rules of voting itself.
  • Every political question has at least two honest sides. Including the voter-ID question, where both progressive concern about disparate impact and conservative concern about election integrity have genuine empirical and normative grounding.

The next chapter (Identity and Politics) takes up a deeper question: how political identity is formed, and how identity-based mobilization differs from other mobilization strategies. It builds on the patterns this chapter has described.