Search results

Zones of peace, conflict, and democracy -- A regional approach to conflict, integration, and democratization -- Empirical data, measurement, and methods -- Do zones of democracy and peace coevolve? -- Deutschian integration and the democratic peace -- Wealth, conflict, and the diffusion of democracy -- Conclusions and implications -- Appendixes : A. Independent states since 1816 -- B. Armed conflicts from 1816 to 1996 -- C.A primer on generalized additive models -- D. Polity democracy scale and modifications -- E. Symbols and notation

Abstract
The COVID 19 pandemic has generated much interest in the relationship between research and policy. It has drawn new attention to the limitations of a linear model, where policy is based on first observing prior scientific research and then designed in response to this. Conflict researchers often motivate the importance of their work by claiming that their "research has important policy implications", but the proposals offered are often at best incomplete. I identify a number of common limitations in claims about policy implications, including a lack of discussion of objectives and priorities, stating objectives themselves as if they were policies, claims about targeting factors without discussing the effectiveness of possible interventions, and a failure to consider uncertainty and potential tensions with other objectives or unintended effects. Research can potentially inform policy discussions and improve decisions, but the incentives in academic research are very different from policy decisions, and the latter often calls for very different evidence than what is offered by the former. Rather than attempting to offer policy prescriptions as an afterthought to academic articles, research can be more helpful to policy by trying to inform debates, focusing on what we know from the cumulative body of research than individual manuscripts, and providing new data and empirical material that allow for better problem description and analysis.

Salient events such as the COVID-19 pandemic and the Ukrainian crisis and Russian invasion lead to interest and debate on how research can inform policy responses. Research can provide important evidence on the potential consequences of different actions with regard to specific objectives. However, research conclusions also remain uncertain and typically depend on many assumptions. I argue that since policy prescriptions entail claims about future consequences, they must be seen as predictions. Although prediction is difficult, especially about the future, we can have better and more informed discussions about policy consequences if we pay attention to what we have learned about predictions and predicting better. Moreover, beyond whether predictions are ultimately correct or not, it is useful to consider what we can learn from them. In some cases, it is more helpful to understand how specific inputs influence predictions than to focus only on approaches that maximize overall fit. Predictions may not be intended to influence policy, but research is more likely to be useful to non-academics if presented in a clear and accessible manner. Finally, since the outcomes we try to predict depend on policy choices, we can potentially improve predictions by thinking about how decisions are made.

Debates over controversial articles often highlight important issues regarding academic freedom, transparency, and how to handle disagreements in publishing. I argue that a response outlining criticism is generally a more productive course of action than calling for retraction. However, there are a number of constraints that impede meaningful debates, and a problematic divergence between our common ideals of open research and free debate and the actual practices that we see in academic publishing, where our current practices often undermine transparency, replication, and scientific debate. I argue that research can benefit from more explicit recognition of politics and preferences in how we evaluate research as well greater opportunities for post-publication debate. The successful initiatives to promote data replicability over the past decade provide useful lessons for what improved post-publication transparency may look like.

A recent report by the United Nations and the World Bank argues that the world has seen "a surge in violent conflicts in recent years," with a sharp increase in reported battle-related and terrorist attacks, and calls for preventive action to ensure that increasing conflict does not undermine the UN Sustainable Development Goals. I will argue that this is a far too pessimistic assessment of conflict trends and not borne out of the available evidence, which indicates a decline in violent conflict since the end of the Cold War. Alarmist warnings may seem helpful to call for action, but they detract attention from what we can learn about the causes for why conflict has declined. For example, resort to violence has become less frequent where factors that can motivate resort to violence such as political and ethnic exclusion have decreased. Moreover, conflict of interest does not imply violence, and the space for nonviolent alternatives has increased. The future of conflict and peace depends on our confidence in whether positive changes will continue and our understanding of the possible challenges.

AbstractThis note provides a retrospective on lessons learned in research on conflict forecasting, motivated by reflections around the retirement of Professor Michael D. Ward from Duke University. I argue that an excessive focus on "black swans" or surprising events that are hard to forecast detracts from considering the more frequent "white swans" or regularities in conflict. It is often more useful to focus on modal conflicts than exceptions, and substantial progress has been made in recent research. I identify some key lessons learned and highlight the need for researchers to distinguish between features that are more or less difficult to forecast.

Barbieri, Keshk, and Pollins (2009) introduce a new Correlates of War trade data set and express various opinions on issues related to analyzing data on trade and how these may impact our inferences on the relationship between trade and conflict. Since there are 33 references to my name in their paper and the authors believe the expanded trade that I have generated to be highly problematic, I respond to some of the issues raised, clarify the specific sources of disagreement, and outline my thoughts on how trade data can be improved and best used in empirical analyses. Barbieri et al. (2009) could be interpreted as suggesting that inferences on the pacifying effect of trade on conflict are fragile and sensitive to decisions on the data. However, their paper in my view provides few new empirical results and does not change my own interpretation of the evidence that dyadic trade reduces the risk of conflict. [Reprinted by permission; copyright Sage Publications Ltd.]