Disentangling Mode and Model: Machine Learning and Multilevel Poststratification for Electoral Inference in Mexico

• In Event: Methodological Advances in the Study of Elections

Brief Overview

We compare machine learning and MRP frameworks for estimating vote intention using three national surveys from Mexico’s 2024 election. Our goal is to identify where modeling or survey mode drives accuracy gains in electoral estimation.

Authors

• Yamil Nares, Defoe, Experts on Social Reporting

• Alfredo Chaparro, Defoe, Experts on Social Reporting