What can we Learn from Predictive Modeling?

TL;DR

This paper advocates for the integration of predictive modeling into empirical political research, highlighting its benefits, establishing benchmark criteria, and demonstrating its application in predicting interstate conflict.

Contribution

It introduces a framework for using predictive modeling as a complement to traditional association-based analysis in political science.

Findings

Predictive modeling enhances understanding of political phenomena.

Benchmark criteria for evaluating predictive models are proposed.

Application to interstate conflict prediction demonstrates practical utility.

Abstract

The large majority of inferences drawn in empirical political research follow from model-based associations (e.g. regression). Here, we articulate the benefits of predictive modeling as a complement to this approach. Predictive models aim to specify a probabilistic model that provides a good fit to testing data that were not used to estimate the model's parameters. Our goals are threefold. First, we review the central benefits of this under-utilized approach from a perspective uncommon in the existing literature: we focus on how predictive modeling can be used to complement and augment standard associational analyses. Second, we advance the state of the literature by laying out a simple set of benchmark predictive criteria. Third, we illustrate our approach through a detailed application to the prediction of interstate conflict.