Persistent non-response bias in sample-matched 2024 polls is quantified at ρ=-0.0030 for Trump voters, and a historical-data-informed correction estimator reduces RMSE from 0.13 to 0.05.
The Credibility Revolution in Empirical Economics: How Better Research Design Is Taking the Con out of Econometrics
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
Using decision theory, the authors conclude that ML-based AI systems are not positioned to solve missing data identification problems better than humans in prediction and treatment choice.
citing papers explorer
-
The Persistent Non-Response Bias in a Sample-Matched Poll for the 2024 U.S. Presidential Election
Persistent non-response bias in sample-matched 2024 polls is quantified at ρ=-0.0030 for Trump voters, and a historical-data-informed correction estimator reduces RMSE from 0.13 to 0.05.