Conceptualizing a New School of Political Conflict and Terrorism Studies

Conceptualizing a New School of Political Conflict and Terrorism Studies

From Attributes to Behavior and Why Policy Makers will Care

By Will H. Moore and Stephen M. Shellman

Some interesting changes in the interchange among US government officials and social science researchers, political scientists in particular, in the nation's universities are afoot. One might date the beginning of this new relationship to then Vice President Al Gore's initiation of the State Failure Project in 1994.[1] More recently, the Defense Department's Combating Terrorism Technology Support Office has begun a Human, Social, and Cultural Behavior Modeling project, [2] and the Pentagon has recently announced the Minerva project, which seeks to "embrace eggheads and ideas" and "recruit social scientists and direct the nation's brainpower to combating security threats like the Chinese military, Iraq, terrorism and religious fundamentalism."[3]

The above mentioned projects explicitly seek the advice and input of political and other social scientists that use mathematical theoretical models, statistical tools to test hypotheses, and advanced computing resources to collect and analyze data. Many scholars and policy makers are unaware that this represents a renaissance of a similar moment in the 1970s.[4] Will this renaissance experience the same fate as its predecessor, or is there reason to believe that the scholarly community has more to offer policy makers today than we did in the 1970s? We argue that there is considerable cause for optimism and make our case by describing an important shift in the conceptual and theoretical orientation scholars are taking toward the study of dissent, repression, terror, and violent conflict within countries.

This essay emphasizes two criticisms with regard to the literature on terrorism and violent political conflict – specifically the quantitative conflict literature. First, this essay takes issue with the dominant structural and systems-oriented theoretical approaches to studying terrorism and political conflict and instead advocates an actors-based process approach to studying such phenomena. Studies of the former category emphasize specific country-attributes such as under-developed economies, the lack of democratic political outlets, and higher levels of ethnic fractionalization, that have been associated with higher levels of political conflict. Yet, many of these variables change very little over time making it difficult to explain variations in levels of violence from one point in time to the next. The approach advocated here argues that actors' decisions are based on the decisions of their opponents and this strategic inter-dependence can more effectively explain conflict processes. Second, this essay calls for more quantitative studies of terrorism to treat terror as a tactical choice within a framework of other violent and cooperative actions. Future work should model the interaction between state and nonstate actors' tactical choices (and vice-versa). Most of the previous literature treats terrorism separately from other violent political tactics and considers it as a distinct area of inquiry. More recently, other scholars have even gone further treating "suicide terror" separately from other terrorist and non-terrorist tactics (e.g., Bloom 2005, Pape 2005). Yet, terrorist tactics are often utilized within an array of other violent, non-violent, and cooperative means. This essay emphasizes that the study of terrorism should not be segregated from the larger study of political conflict and instead argues that studies focused on explanations of terror should be discussed within the broader field of contentious politics.

Structure v. Process

One way to study violent political conflict and terrorism within countries is to focus on the structural conditions that impact the chance that a country will experience such phenomena. Another way is to analyze the behavioral relationships among parties to potential conflicts, how they make decisions, how such decisions impact other parties' decisions, and how the sequences of behavioral interactions escalate and de-escalate across various thresholds of violent political conflict. Or, as Harry Eckstein put it over 25 years ago, we can distinguish between "contingency" and "inherency" approaches to the study of violent conflict.[5] The first perspective assumes that conflict is contingent on unusual or irregular conditions that cause disruptions in conventional politics. The contingent approach leads one to study the political, economic, and social attributes of countries to explain variation in their conflict experiences. The inherent perspective assumes that violent political conflict emerges out of low-level contentious interactions among a set of political players. This approach leads researchers to focus on the conditional behavior of parties to conflict and how that behavior changes over time. While Eckstein laid out these two approaches in 1980, the past 25 years have borne witness to few scholars taking the latter path.

Prior to the turn of the century, the study of intrastate conflict was much more focused on the former approach than the latter. While studying the political, economic, and social attributes of countries is a useful approach for understanding and highlighting general patterns of conflict, it is ill-suited to address conflict processes because such approaches "are essentially static 'input-output' or 'stimulus-response' type models, not dynamic models of interaction."[6] Charles Tilly argues that because "collective action is dynamic… its outcomes depend very strongly on the course of interaction."[7]

A recent wave of scholarship turns its back on the structural attributes approach and instead focuses attention on the escalation and de-escalation processes of political conflict instantiated by actors' strategic behavioral interactions.[8] A common thread running through this new generation of conflict scholarship is a shift from countries as the unit of analysis to the parties to the conflict and their behavior. This work focuses on competition between governments and various dissident groups over policy, control of the state, and—especially—the support of the population.

This shift is critically important because it means that theory becomes much more useful to policy makers: the emphasis on parties to the conflict leads this research to develop hypotheses about the conflictual behavior of dissidents in response to government behavior and vice versa. By moving away from thinking about the impact of democratic v. autocratic institutions, the size of GNP/capita, and the ethnic composition of society these scholars have begun to ask the following sorts of questions:

When does repression work? When does it backfire?

Why are some dissident groups so much more violent than others?

What explains varying levels of discrimination in targeting across time and space?

What event sequences lead to conflict escalation?

What are the effects of government countermeasures on the tactical choices of dissidents?

What explains the ebb and flow of government-dissident behavioral exchanges?

What event sequences lead to negotiated peaceful settlements?

Why can factions spoil a peace process?

Note that information about political institutions, economic output, and ethnic composition are of limited usefulness for answering these questions. Why? Because those characteristics of the country in which these conflicts unfold do not change much over time. To the extent that they change, they change rather slowly. As one recent paper put it "the factors analyzed in country level analyses are the same for a given country during war and peace, and are therefore incapable of predicting shifts from one period to the other."[9] To better understand such conflict processes, we must study the behavior of the parties to the conflict. And if we are going to study behavior, then it is surely reasonable to study it as purposive, strategic behavior that varies systematically in response to the behavior of other parties to the conflict.

To be sure, we have learned from the structural approach that characteristics of the state such as regime type, the economy, terrain, capabilities, and demographics like population and ethnicity are correlated with the level of political conflict we observe across countries. However, the structural attributes approach has not taught us much about conflict processes as they unfold over time within specific countries. For example, we know that a country with mountainous regions is more likely to experience an insurgency. Yet, knowing that El Salvador is mountainous tells us little about when we are likely to observe peace or conflict in El Salvador as those mountains change very little over time.

That said, we are not arguing that we should throw the baby out with the bathwater. Rather, the best scholarship should situate behavioral studies of conflict processes within a structural framework. And this is precisely what we see in recent studies.[10]

This essay argues that the renaissance of the policy community's interest in the scientific study of conflict is more likely to bear fruit than its predecessor because the scholarly community is conceptually and theoretically better prepared to make a contribution to the questions policymakers want answered. The shift away from the study of the correlation of national aggregates toward the study of the behavior of dissidents and states within the context of national aggregates has led to a theoretical emphasis on the strategic behavior of governments, dissident groups, and members of the population. This approach lends itself well to studying the propensity and frequency of terror attacks committed by state and nonstate actors. In addition to these new theoretical tools, exciting new data tools and statistical models have been created for the express purpose of facilitating the testing of the hypotheses implied by the new theories. We briefly describe these new developments, emphasizing what we have learned about violent dissent.

Disaggregating Violent Political Conflict

To begin, the new school points out that political conflict is not best characterized as something that countries catch (or experience): the implicit epidemiological analogy is passé. Instead the new generation of scholarship recognizes that governments often face multiple challengers fighting for the same cause and/or very different causes, and that these challenges vary across both space and time. As such, these theorists have been disaggregating the study of civil conflict across actors, tactics, space and time.

Civil conflicts often involve infighting among members or branches of the government (e.g., military coups in Nigeria) and often yield dissident group splits (e.g., the Moro Islamic Liberation Front emerged out of the Moro National Liberation Front). In other cases, multiple groups may interact with each other and even form alliances or coalitions (e.g., the Coalition Government of Democratic Kampuchea – comprised by the Khmer Rouge, FUNCINPEC and KPLNF). Further, governments and dissident groups compete for the support of the public. Stathis Kalyvas' recent book makes the importance of distinguishing among the government, the dissidents, and the public abundantly clear: he observes that death tolls in civil wars vary systematically depending on whether either the dissidents or the government can exercise authority over the town or whether the territory is actively contested by dissident and government troops.[11] Other research has disaggregated actors and their behavior to demonstrate that both diplomatic and military intervention by third countries on behalf of dissidents or governments can have strong bearing on pushing parties to the table or escalating violent activity.[12] In sum, intrastate conflict is comprised of many different parties with different motivations, who make a variety of decisions as to how to behave in both the short and long run. To answer questions like the ones listed above we need to adopt a disaggregated unit of analysis and account for the behavior of these different parties.

Second, much of the scholarship to date has focused on particular tactics or events such as terror attacks, protests, war, militarized interstate disputes, negotiations, or treaties. By adopting a particularistic conceptualization of conflict behavior, we ignore the full underlying dimension of state and nonstate actors' behavior (Moore 2006) and fail to distinguish among the various actions available in an actor's choice set. For example, if behavior such as hostile rhetoric, guerrilla attacks, cease fires, and peace talks have a non-trivial effect on the probability of a terrorist event occurring and co-vary positively with independent variables used to explain terrorist events, then relationships reported in the terrorism literature are biased.[13] Aside from the bias introduced into empirical models of terrorism, the narrow focus on terrorist events and the desire to explain them alone without examining how other forms of behavior effect and intertwine with such events causes scholars to narrowly theorize about the behavior of insurgents, rebels, separatists, etc. Thus, a focus exclusively on terrorism has both negative theoretical and empirical consequences. As Moore (2006) puts it elsewhere, "Conceptualization matters; it influences our research agendas; it influences the questions we ask, and, ultimately, the policy implications we produce." At the present, political violence, nonviolence, diplomacy, accommodation, and terrorism are often treated as separate domains of inquiry, yet such phenomenon are clearly connected and intertwined. They are often tactical substitutes for state and nonstate actors engaged in political struggles. Making such connections should spawn new theories yielding larger explanations than extant theories. To analyze conflict processes effectively, we need to conceptualize the choices and choice sets of state and nonstate actors more holistically.

Third, civil conflicts rarely span an entire country's territory. Rather, they are often confined to sub-national regions based on certain geographic features that generate conditions favorable for conflict. Teams of researchers are making considerable progress pursuing this seemingly obvious point.[14] For example, mountains or jungle provide cover for rebels to hide and wage guerilla campaigns. Resource rich zones abundant in minerals (e.g., diamonds) or other valuables (oil, etc.) may also contribute to reoccurring conflict.[15] Motives may also determine conflict locations. For example, recent research claims that separatist groups tend to fight away from the capital in order to make their territory autonomous, whereas insurgents aiming to overthrow the state tend to fight nearer to the source of the power – in the capital city or province where the capital is located.[16] Moreover, international borders can often provide refuge from governmental control, and this may cause conflict events to cluster near international boundaries which, in turn, can have important implications for neighboring countries.[17] The general point is this: if we wish to investigate theories that have a geographic element we need to abandon the country level of analysis in favor of a disaggregated spatial approach. Importantly, doing so will produce research with far more rich policy implications than we have seen prior to the emergence of this new generation of research.

Fourth, the unit of time over which one aggregates is consequential.[18] Until recently most research has focused on the year as the unit of temporal aggregation. Marcellino explains that "temporal aggregation arises when the frequency of data generation is lower than that of data collection so that not all the realizations of the stochastic process… are observable."[19] Yearly aggregation obscures the actions and reactions of actors who respond to one another in much smaller units of time such as monthly, weekly, or even daily intervals. At what temporal units do dissidents and governments respond to one another? Surely the answer is: it varies. But just as surely the answer is not: annually! To focus on the kinds of questions that the new generation of civil conflict scholars is asking requires that we abandon annually aggregated data in favor of data collected over the unit of time in which they occur. In the next section we briefly describe the exciting data collection revolution that is taking place and making possible the renaissance we are seeing as Washington engages the scientific study of violent political conflict.

The Events Data Revolution

In the mid 1990s Phil Schrodt revolutionized the collection of events data when he released the Kansas Events Data System KEDS computer program.[20] This program demonstrated that it is possible to use computers to code news reports to generate data about the behavior of dissidents toward governments, governments toward dissidents, government toward other governments, etc. Over the past 15 years the KEDS project has spawned a number of similar projects, and this technology has spilled over into a variety of other areas of political science as well.[21] Where hundreds of hours of human labor were required to code such reports computers are able to produce such data in mere minutes. This has radically changed the information that is available to scholars. Further, the shift in conceptual and theoretical interests described above demand just such data. To illustrate, we briefly describe Project Civil Strife (PCS), which is directed by the second author of this essay.

Project Civil Strife uses computerized coding technology to generate disaggregated data useful for testing the hypotheses advanced by the new generation of intrastate conflict researchers.[22] The project directly confronts actor, tactical, spatial, and temporal aggregation head on by collecting information on multiple actors' behavioral interactions each day in various geographic locations. With respect to actors, PCS codes the behavior of just about any dissident or government group discussed in open source media reports. It codes individual's names and offices and even tracks individuals like Norodom Sihanouk in Cambodia who at different times of the conflict served as a government official and then as a rebel leader. By disaggregating the state and the social and dissident actors, scholars and policy makers can examine interactions among multiple parties as depicted in Figure 1.


Figure 1A and 1B shows how one might go about constructing a model of the conflict processes between two rebel groups and the state (e.g., Democratic Kampuchea and the KPLNF in Cambodia). The first can help uncover different processes at work between the government and group 1 and the government and group 2. The second shows how the relationship between Group 1 and Group 2 might affect the relationships between the state and each group or how the relationship between the state and each group might affect the relationship between the two groups themselves. Figure 1C illustrates the potential relationships one could uncover by breaking up the government into its principle component parts. Such a model might be particularly useful in Pakistan, Chile, and/or Nigeria where the government and the military have competed in the past with each other.

Moreover, the project employs several different coding ontologies to disaggregate tactics and events. The project primarily uses the CAMEO and WEIS event codes to capture the various actions taking place by myriad actors in various places.[23]

With respect to spatial disaggregation, PCS also tracks the behavior of actors across disaggregated space. Figure 2 depicts a map of Cambodia and highlights the frequency of violent events (use of force, bombings, ambushes, etc.) across different regions. The two most populated regions in terms of conflict events include Phnom Penh and Pailin. Of course Phnom Penh is the capital of Cambodia and the central location of the government, while Pailin was a major stronghold and resource center for the Khmer Rouge. Pailin was an area rich in gem stones that was mined to fund the rebels. The preliminary spatial plot supports other findings in the literature. For example, conflict takes place in resource rich areas and in the mountains (Pailin and Kampot). Moreover, the abundance of conflict events seems to be greater near the Vietnam border (Prey Veaeng, Sbaay Rieng, Kampong Chaam, and Kracheh).


To summarize, the PCS data can be used to study multiple-party interactions over disaggregated time and space. By doing so it encourages researchers to address the types of questions identified above by studying the behavior of parties to intrastate conflicts. For example, PCS puts terrorism in the context of other violent, non-violent, and cooperative tactics such that we can theorize and empirically model a more holistic conceptualization of political behavior.[24] Such a conceptualization reduces bias and yields theories that explain more than their predecessors. It also allows for the examination of how low intensity conflict escalates into "civil war".[25] In sum the project and others like it bring together process oriented theories, holistic conceptualizations of tactical choices, and disaggregated data to better understand the causes and consequences of violent political conflict.


Paul Collier's influential World Bank project on the co-variates of civil war is an archetypical example of the structural attributes approach to the conceptualization and analysis of violent political conflict.[26] We find it instructive that the final product from that project abandoned the structural attributes approach in favor of case studies that focused attention on historical description of the behavior of the parties to the conflict.[27] Sidney Tarrow argues that this shift from the statistical analysis of structural aggregates organized in country-year data containers to descriptive-historical analyses of cases demonstrates the weakness of large-N data analysis: it can illuminate a little bit, but to really understand conflict processes we must engage in descriptive case studies.[28] While this essay shares Tarrow's concerns about the weakness of the structural aggregates approach to the study of violent political conflict, it also demonstrates the poverty of Tarrow's diagnosis: the weakness is not inherent in its methodological choices for hypothesis testing. Rather, it is a conceptual weakness about the units of analysis one chooses. One irony of Tarrow's essay is that he also reviews some of the new generation scholarship we celebrate here. Yet, rather than recognize its value for having abandoned the overly-aggregated country-year in favor of a focus on the behavior of parties to the conflicts within the context of institutions, economies, and societies, Tarrow flogs the dying horse of his generation's intellectual battleground: the so-called qualitative—quantitative divide that grew out of the behavioral revolution in political science.

Fortunately, a growing number of policy analysts in Washington are disinterested in these academic "tempests in a teapot" and recognize the potential value in leveraging the new school's insights to inform policy makers. This renaissance is welcome and it will be interesting to see how it unfolds. Political scientists are on the cusp of being able to say very interesting things about the violent behavior of dissidents and governments as the outcome of dynamic processes that vary over time and space across the countries in which such conflicts take place. We invite the reader to stay tuned.

About the Authors:

Will H. Moore is Associate Professor of Political Science at the Florida State University with research interests in violent political conflict within and between countries.

Stephen M. Shellman is a Research Scientist within the Institute for the Theory and Practice of International Relations at The College of William & Mary and is Director of the Violent Intranational Political Conflict and Terrorism (VIPCAT) Research Laboratory.


[1] D.C. Esty, J. Goldstone, T.R. Gurr, B. Harff, P.T. Surko, A.N. Unger, and R. Chen. "The State Failure Project: Early Warning Research for US Foreign Policy Planning." Preventive Measures: Building Risk Assessment and Crisis Early Warning Systems. 1988, Vol. xx, pp. 27-38.

[2] Office of the Secretary of Defense. Broad Agency Announcement 08-Q-3028 Packet, available online at:

[3] See for the story and for the announcement.

[4] SJ Andriole and GW Hopple. "The Rise and Fall of Event Data: From Basic Research to Applied Use in the US Department Of Defense." International Interactions, 1984. Vol. 11, pp. 293-309.

[5] Harry Eckstein. "Theoretical Approaches to Explaining Collective Political Violence." In T.R. Gurr (ed.) Handbook of Political Conflict, pp. 135-166. 1980. New York: The Free Press.

[6] Will H. Moore. "Action-Reaction or Rational Expectations? Reciprocity and the Domestic-International Conflict Nexus During the 'Rhodesia Problem'." Journal of Conflict Resolution. 1995. Vol. 39, No. 1, p.132.

[7] Charles Tilly. "Models and Realities of Popular Collective Action." Social Research. 1985. Vol. 52, No. 4, p. 1.

[8] Recent reviews of this literature can be found in Christian Davenport. "State Repression and Political Order." Annual Review of Political Science. 2007. Vol. 10, pp. 1–23, and Mark I. Lichbach. "Internal Wars Over the State: Rational-Choice Institutionalism and Contentious Politics." M.I. Midlarsky (ed.) Handbook of War Studies, Volume III. Forthcoming. University of Michigan Press.

[9] C. Butler, S. Gates and M. Leiby. "To Fight or Not to Fight: Contest Success Functions and Civil Conflict." Paper presented at the American Political Science Association Annual Meeting. 2005. P. 18. See, also, Will H. Moore, Ronny Lindström and Valerie O'Regan. "Land Reform, Political Violence, and the Economic Inequality—Political Violence Nexus: A Longitudinal Analysis." International Interactions. 1996. Vol. 21, No. 4, pp. 335-363.

[10] For example, see MaCartan Humphreys. "Natural Resources, Conflict, and Conflict Resolution: Uncovering the Mechanisms." Journal of Conflict Resolution. 2005. Vol. 49, Num. 4, pp. 508-537; Stathis Kalyvas. The Logic of Violence in Civil War. 2006. New York: Cambridge University Press; and Jeremy Weinstein. Inside Rebellion: The Politics of Insurgent Violence. 2007. New York: Cambridge University Press.

[11] Kalyvvas, 2006, op cit.

[12] Moore, 1995, op cit; Will H. Moore and David R. Davis. "Ties That Bind? Domestic and International Conflict Behavior in Zaire." Comparative Political Studies. 1998. Vol. 31, No. 1, pp. 45-71; Kristian Skrede Gleditsch and Kyle Beardsley. "Nosy Neighbors." Journal of Conflict Resolution. 2004. Vol. 48, No. 3, 379-402; and Clayton L. Thyne. "Cheap Signals with Costly Consequences: The Effect of Interstate Relations on Civil War." Journal of Conflict Resolution. 2006. Vol. 50, No. 6, pp. 937-961.

[13] Moore (2006, 10) makes a similar argument about studying war only

[14] See the Conference on "Disaggregating the Study of Civil War and Transnational Violence." University of California, San Diego, 2005. Program available online at: and the Armed Conflict Location and Event Data (ACLED) project at the Center for the Study of Civil War, Peace Research Institute Oslo (PRIO). Available online at:

[15] See Humphreys, 2005, op cit. and Michael Ross. "A Closer Look At Oil, Diamonds, And Civil War." Annual Review of Political Science. 2006. Vol. 9, pp. 265-300.

[16] See Halvard Buhaug and Scott Gates. "The Geography of Civil War." Journal of Peace Research. 2002. Vol. 39, No. 4, pp. 417-433, Halvard Buhaug and Päivi Lujala. "Accounting for scale: Measuring geography in quantitative studies of civil war." Political Geography. Vol. 24, No. 4, May 2005, pp. 399-418, and Jan Ketil Rǿd and Halvard Buhaug. "Using disaggregated grid cells for a study on the onset of African Civil Wars." 2007. Working Paper available online at:

[17] Idean Salehyan and Kristian Skrede Gleditsch. "Refugees and the Spread of Civil War." International Organization. 2006. Vol. 60, No. 2, pp. 335-366.

[18] Pitirim Sorokin. Social and Cultural Dynamics: Fluctuation of Social Relationships, War, and Revolution, Volume 3, 1937, New York: American Book Company, pp. 383-408, and Stephen M. Shellman. "Time Series Intervals and Statistical Inference: The Effects of Temporal Aggregation on Event Data Analysis." Political Analysis. 2004. Vol. 12, No. 1, pp. 97-104.

[19] Massimiliano Marcellino. "Some Consequences of Temporal Aggregation in Empirical Analysis.'' Journal of Business and Economic Statistics. 1999. Vol. 17, p. 129.

[20] Philip A. Schrodt, Shannon G. Davis and Judith L. Weddle. "Political Science: KEDS—A Program for the Machine Coding of Event Data." Social Science Computer Review. 1994. Vol. 12, No. 4, pp. 561-587.

[21] See, for example, Paul Pennings and Hans Keman. "Towards a New Methodology of Estimating Party Policy Positions." Quality and Quantity. 2002. Vol 36, No. 1, pp. 55-79, and Burt L. Monroe, Kevin Quinn, Michael Colaresi, Dragomir Radev, Steven Abney, Michael Crespin, Yael Shomer, Akitaka Matsuo and Sarah Hobolt The Dynamics of Political Rhetoric and Political Representation. Poster presented at the 2007 Human and Social Dynamics meeting, National Science Foundation. Available online at:, pp. 195-196.

[22] Internet links will be provided for the project, papers, publications, etc if published.

[23] See for WEIS codes and for the CAMEO scale.

[24] Author citation – deleted for anonymity.

[25] Andrew M. Reeves and Brandon M. Stewart. "Forecasting Escalation in Civil War: A Quantitative Case Study of the Aceh Conflict in Indonesia." Prepared for the 2007 Annual Meeting of the International Studies Association. Available online at:

[26] Paul Collier. Economic Causes of Civil Conflict and Their Implications for Policy. Washington, D.C.: World Bank. 2000, and Paul Collier and Anke Hoeffler. "Greed and grievance in civil war." Oxford Economic Papers, 2004. Vol. 56, pp. 563-595.

[27] Paul Collier and Nicholas Sambanis. Volumes I and II. Understanding Civil War: Evidence and Analysis, Washington, D.C.: World Bank Publications. 2005.

[28] Sidney Tarrow. "Inside Insurgencies: Politics and Violence in an Age of Civil War." Perspectives in Politics. 2007. Vol. 5, No. 3, pp. 587-600.

Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.


Perspectives on Terrorism is  a journal of the Terrorism Research Initiative and the Center for Terrorism and Security Studies

ISSN  2334-3745 (Online)

Disclaimer, Terms and Conditions