🧾 The Puzzle

Surveys often misstate how frequently people engage in socially desirable behaviors, and the American National Election Study (ANES) has long overstated voter turnout. The face-to-face component of the 2012 ANES produced a turnout estimate at least 13 percentage points higher than the benchmark voting-eligible population turnout rate, but the reasons for this persistent bias have been unclear.

📊 How Turnout Was Checked Against Records

• The face-to-face 2012 ANES turnout estimate was compared to the voting-eligible population benchmark.
• Turnout data supplied by voter-file vendors were used to validate self-reports and measure multiple bias sources simultaneously — enabling these phenomena to be measured for the first time in a single survey.

🔎 Three Explanations Tested

• Nonresponse bias (differences between respondents and nonrespondents).
• Over-reporting (respondents falsely claiming they voted).
• Inadvertent mobilization (the survey acting as a treatment that increased turnout among respondents).

📈 Key Findings

• The ANES overestimated turnout by at least 13 percentage points in the 2012 face-to-face component.
• Breakdown of contributors to that overestimate:
• Over-reporting: 6 percentage points (the largest single contributor).
• Nonresponse bias: 4 percentage points.
• Mobilization (survey-induced turnout): 3 percentage points.
• These three phenomena were directly measured within the same survey by linking responses to voter-file turnout data.

💡 Why It Matters

These results clarify why a flagship survey consistently overstates turnout and show that corrective strategies must address both misreporting and selection effects, as well as potential survey-induced behavior. Comparing survey reports with administrative turnout records is essential for accurate measurement of electoral participation.