Journal of Conflict Resolution

The extent to which theories in the social sciences are accepted or rejected on the basis of empirical tests can be shown only by a detailed analysis of specific cases. The author examines the reception by social scientists in the 1970s and early 1980s of T. R. Gurr's theory of collective violence based on the concept of relative deprivation. The history of this theory may be considered an example of definite progress in social science: a hypothesis widely accepted at one time has been tested and rejected, thus making room for the development of alternative hypotheses. But although Gurr and other advocates of the theory have abandoned it in its original form following the mostly negative results of empirical tests, many social scientists (especially psychologists) have continued to cite it favorably. Slightly less than half of the unfavorable citations have been supported by references to empirical evidence.

What conditions determine whether a civil war ends in a negotiated settlement or a military victory? The authors address this question by developing an expected utility model of the choice between seeking an immediate settlement or continuing to fight in anticipation of eventual victory. The model implies that the likelihood of a settlement varies with estimates of the probability of winning, the time required to win, the rate at which the costs of conflict accrue, and the payoffs for victory versus settlement. Logistic regression results suggest that a settlement becomes less likely the larger the government's army and more likely the longer the civil war lasts. Payoff and cost variables had no effect on the likelihood of a settlement.

The authors develop a characterization of subgame perfect equilibrium strategies in discounted repeated games that highlights a class of strategies called countervailing. When using countervailing strategies, the players focus on manipulating the other side's payoff to induce his cooperation. These strategies do not require one player to entertain specific expectations about the strategic choice of the other. For this reason, countervailing strategies are promising vehicles for tacit bargaining. The authors test their relevance using a case of international trade—Japan's penetration of the U.S. automobile market in the early 1980s. The empirical section provides evidence of the tacit bargaining behavior that the model predicts.

Fearon (1994a) concludes that democracies should be less likely to back down in crises and thus be able signal resolve more effectively than autocratic states. The authors evaluate this argument against the population of conflicts identified by the SHERFACS phase-disaggregated conflict management data set. Their findings indicate that domestic characteristics of democracies enable them to communicate intentions effectively. Between 1945 and 1984, conflicts between pairs of democratic states have been found to have fewer phases than those involving at least one autocratic state. This suggests that the information conveyed by the regime type of a state may help resolve the security dilemma by providing communication between democracies. This finding also supports an integration of existing explanations of the democratic peace.

Does the nature of a nation's political institutions influence the types of countries with which it allies? Some previous research has suggested that democracies tend to ally with other democracies. This study reexamines alliance patterns by assessing the broader linkage between regime type and alliance partnership. The authors present a refinement of previous research designs, using new data from Polity III and the updated correlates of war (COW) alliance data sets to analyze all alliances from 1815 to 1992. The bipolar alliance structures of the cold war (NATO and the Warsaw Pact) appear to be aberrations in their strong ideological content. In general, there is very little correlation between alliance dyads and regime type. Surprisingly, democracies are less likely to ally with one another than highly autocratic regimes. Regimes of most types seem to prefer to ally with partners of dissimilar type. The authors conclude that this is due to so-called gains from trade within alliance dyads.

Current consensus in the field of democratic peace research holds that democratic states go to war in general no less than nondemocratic states. The author challenges this consensus by reevaluating the main empirical studies on which it rests, using information that previous studies ignored and statistical techniques unused or even unknown at the time. The results indicate that from 1960 to 1980, democratic nations were less involved in military conflict than other regime types. Estimates of this relationship are robust to different operational definitions of both war and democracy, to the addition of control variables for other possible correlates of war, and to the application of different statistical techniques. This indicates that lack of previous significant findings have less to do with the data than with the methods used to analyze them.

Models of international conflict often omit the option of offering rewards as a means of exerting influence. The author has developed a revised version of Bueno de Mesquita and Lalman's (1992) international interaction game that allows states either to offer rewards or threaten military force once they have made political demands and their adversary has rejected them. After laying out the game's logic, the author derives predictions about the conditions under which war occurs and rewards are offered and accepted with complete information. The costs of using force, relative military capabilities (i.e., the probability of victory in war), and the value of the rewards drive these predictions. The author also relaxes the complete information assumption and shows how uncertainty can affect the choices of both the maker of the demand and the recipient. Finally, the model is illustrated by applying it to two crises: the United States-Japan crisis leading to the Pearl Harbor attack and the Cuban missile crisis. The variables in the model help to explain both the outbreak of war in the first case and the avoidance of war in the second.

In this study, the performance of individual and team negotiators is compared. From an intergroup relations perspective, teams may be impeded by their competitive orientation and low levels of cooperation and trust. However, from a cognitive perspective, advantages should accrue to teams because more people are present to generate additional ideas, bring a greater number of perspectives to the problem, and perform strategically important roles during the negotiation. The relative effects of these opposing forces were tested by comparing the perceptions and outcomes of parties in intergroup, interindividual, and mixed (team vs. individual) negotiations in a laboratory experiment. Results indicate that the presence of teams increased competitiveness and decreased cooperativeness and trust between negotiating parties. In mixed negotiations, teams outperformed individual opponents and were perceived as having more power and more ideas for a solution. Teams also affected integrative outcomes when negotiators' level of training was considered, such that the presence of teams negatively influenced performance for novices but positively influenced integrative outcomes for trained negotiators. The implications of these results for theories of intergroup relations and negotiations are discussed.