Figure 3.4. Executive Corruption and Invalid Vote Campaign Emergence Source: Varieties of Democracy (V-Dem), original data. Note: The figure shows results from a logistic regression analysis predicting the emergence of an invalid vote campaign using levels of executive corruption (N = 218). Year controls are included, and standard errors are clustered by country. Shaded area indicates 95% confidence intervals.

Figure 3.5. Distinct Party Platforms and Invalid Vote Campaign Emergence Source: Varieties of Democracy (V-Dem), original data collection. Note: The figure shows results from a logistic regression analysis predicting the emergence of an invalid vote campaign using levels of the distinctiveness of party platforms (N = 218). Year controls are included, and standard errors are clustered by country. Shaded area indicates 95% confidence intervals.

Figure 3.6. Antiestablishment Candidates and Invalid Vote Campaign Emergence Source: Original data collection, electoral management bodies, Carreras (2012). Note: The figure shows the predicted probability that an invalid vote campaign will emerge when an antiestablishment candidate is (not) present, using the results from a logistic regression analysis (N = 217). Year controls are included, and standard errors are clustered by country. Whiskers indicate 95% confidence intervals around estimates.

Figure 3.7. Election Intimidation and Invalid Vote Campaign Emergence Source: Varieties of Democracy (V-Dem), original data. Note: The figure shows results from a logistic regression analysis predicting the emergence of an invalid vote campaign using levels of election intimidation by the incumbent (N = 218). Year controls are included, and standard errors are clustered by country. Shaded area indicates 95% confidence intervals.

Figure 3.8. Invalid Vote Campaign Emergence Conditional on Grievances, Affective Polarization. Source: Varieties of Democracy (V-Dem), original data collection. Note: The figure shows results from logistic regression analyses predicting the emergence of an invalid vote campaign using an interaction between each measure of democratic quality and affective polarization (N = 218). Year controls are included, and standard errors are clustered by country. The figure presents 84% confidence intervals around point estimates, which is equivalent to a 95% confidence interval around the difference across lines.