581 statistical models from 65 peer-reviewed articles published between 2010 and 2022 are analyzed to identify consistent predictors of election violence. The meta-analysis, by Richard Frank, finds 13 of 44 variables are reliable predictors, highlighting election-specific factors like fraud as more significant than structural conditions. It suggests future research should focus on election-specific triggers and measurement issues.
Use Cases
- Identify robust predictors of election violence based on the meta-analysis of 581 models.
- Compare the predictive power of election-specific factors like fraud versus structural conditions like democracy.
- Analyze differences in predictor significance at national versus subnational levels.
- Guide future research design by focusing on the 13 consistently significant variables.
Strengths
- Analysis is based on 581 models from 65 peer-reviewed articles.
- Covers research published over a 12-year period from 2010 to 2022.
- Examines over 440 predictor variables to identify consistent patterns.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
Provenance
- Source
- Richard Frank Dataverse
- Collection Method
- Meta-analysis of published quantitative research.
- Time Range
- Covers research published between 2010 and 2022.
- Freshness
- Last updated 2026-06-03 04:53:42; freshness should be verified.