397 US voters were surveyed in 2017 regarding their party identity and stances on 20 policy issues. Created by Philip Moniz, this dataset compares these self-reported responses against commercial vendor predictions to assess the accuracy of political microtargeting.
Use Cases
- Measuring the gap between self-reported policy positions and vendor-predicted stances
- Analyzing the impact of cross-pressured voter status on predictive accuracy
- Statistical validation of commercial voter file reliability across 20 issue categories
Strengths
- N=397 ground-truth survey responses
- Covers 20 distinct policy issues
- Direct comparison with commercial vendor predictions
Limitations
- Small sample size of 397 records
- Temporal staleness as data was collected in 2017
- Geographic bias limited to the United States
Provenance
- Source
- Research & Politics Dataverse
- Collection Method
- Survey and commercial data matching
- Time Range
- 2017
- Freshness
- Last updated February 2026; survey conducted in 2017.
- Geography
- United States