Learning from the Past and Stepping into the Future: Toward a New Generation of Conflict Prediction

Abstract

Developing political forecasting models not only increases the ability of political scientists to inform public policy decisions, but is also relevant for scientific advancement. This article argues for and demonstrates the utility of creating forecasting models for predicting political conflicts in a diverse range of country settings. Apart from the benefit of making actual predictions, we argue that predictive heuristics are one gold standard of model development in the field of conflict studies. As such, they shed light on an array of important components of the political science literature on conflict dynamics. We develop and present conflict predictions that have been highly accurate for past and subsequent events, exhibiting few false-negative and false-positive categorizations. Our predictions are made at the monthly level for 6-month periods into the future, taking into account the social-spatial context of each individual country. The model has a high degree of accuracy in reproducing historical data measured monthly over the past 10 years and has approximately equal accuracy in making forecasts. Thus, forecasting in political science is increasingly accurate. At the same time, by providing a gold standard that separates model construction from model evaluation, we can defeat observational research designs and use true prediction as a way to evaluate theories. We suggest that progress in the modeling of International Studies Review (2013) 15, 473-490 conflict research depends on the use of prediction as a gold standard of heuristic evaluation.

Prediction and its discontents: guidance for Australia from the debate over social science forecasting

“…Most statistical models will assign a very high probability of peace for a given country in a given year and, most of the time, they will be correct. However, it is the small number of times when this is not correct that strategists really need to know about (Ward et al 2013).…”

“…The Vietnam War is often thought of as the archetypal case of naive and arrogant quantitative analysis run amok (Chomsky 1968). In fact, Stanford's Jeremy Milstein was already using predictions derived from his regression analyses in 1968 to suggest that the USA's bombing campaign would fail to prevent further North Vietnamese incursions into the South, and that the commitment of further US ground troops to South Vietnam would not bolster confidence in the South Vietnamese regime (Milstein 1974;Milstein and Mitchell 1968;Ward et al 2013). Moving forward in time, Caulkins, Kleiman, and Kulick's (2010) deductive microeconomic model of the Afghan drug trade allowed them to forecast that global production of opiates would remain concentrated in Afghanistan for the foreseeable future; that eradication efforts would serve simultaneously to make the trade more profitable and to concentrate it in Taliban-run areas; and, most shockingly, that alternative livelihood programs would end up funding the insurgency through de facto taxes levied by the Taliban on local farmers.…”

What Drives State-Sponsored Violence?: Evidence from Extreme Bounds Analysis and Ensemble Learning Models

“…we follow the literature on predictive political science and use the area under the ROC curve (AUC) as our model evaluation metric (e.g., Hill and Jones 2014;Ward et al 2010Ward et al , 2013. Models with higher AUC values are considered more accurate.…”

Conflict Modelling: Spatial Interaction as Threat

“…A number of studies are beginning to exploit this information and make predictions concerning conflict around the world. Ward et al (2013), for example, use online event data similar to GDELT within a mixed-effects logistic regression model to determine the likelihood of the occurrence of civil conflict at the monthly level between 1997 and 2011. Using a wide range of predictive performance metrics, the authors demonstrate high predictive capability, suggesting that the use of online data can indeed be used as an early warning signal for the onset of civil wars (another example is Hegre et al, 2013).…”