This article—one of only two on the “top twenty” list published later than 1986—undertook two tasks. The first was to examine whether the “democratic peace” was a spurious relationship that could be disconfirmed by several possible alternative explanations. The second was to provide a stepping stone—to be revised and improved by others—for converting the democratic peace from a theoretically loose, substantially intuition-driven project into a research program in which theory and empirics reinforced each other.

Our collaboration stemmed from submission of a manuscript by Maoz and Abdolali (1989) to the journal edited by Russett. That was then the most systematic analysis of the democratic peace, using Polity regime data, scales for violent disputes (not just war), and a long time period. It found evidence that democracies rarely fight one another, but it did not support the idea that democracies are more peaceful in general. Russett, considering the accumulating evidence for a democratic peace but troubled that all the analyses had been solely or nearly bivariate, knew that few scholars would take it seriously until it had passed various multivariate tests that controlled for influences like wealth, relative power, alliances, and contiguity. He was far from confident that it would pass, but felt that this required a more thorough set of empirical tests. When he was a Fulbright scholar in Israel during spring 1989, he met Maoz in Tel Aviv and asked him to collaborate. The first outline of their collaborative research was written on a napkin in a restaurant. The collaboration was done mostly by e-mail after Russett's return home.

At this point Maoz made the important theoretical innovation of insisting that what were becoming known as the normative and institutional explanations should be treated as alternative hypotheses for their relative contributions, and he articulated the theoretical propositions more carefully. We established a claim to priority with a simple version using contingency tables (Maoz and Russett 1992), and then published the more sophisticated analysis in the Review article and in parts of Russett (1993).

We confronted several fundamental problems in working on this study. One was to outline in a more coherent manner the underlying assumptions of both explanations in a manner that would allow deducing seemingly inconsistent behavior by democracies (that they are pacific with respect to each other, but sometimes belligerent toward nondemocratic states). Another problem entailed differentiating the normative explanation from the structural one on both a theoretical and an empirical level. We also struggled a great deal in attempting to find empirical measures of democratic norms. Finally, we spent much effort in trying to develop critical tests for these explanations.

Subsequently this work became an expanding democratic peace research program to include the effects of trade (e.g., Oneal, Oneal, Maoz, and Russett 1996) and international organizations (e.g., Russett and Oneal 2001; Pevehouse and Russett 2006) over a much longer time period (1885–2001) into what we now call the Kantian Peace project. Maoz also wrote some of the most detailed and scientifically compelling defenses of the democratic peace (e.g., Maoz 1997, 1998). Since 1993 the project has continued with at least 18 co-authors and many hundreds of other scholars who built on, complemented, or criticized our work to make the research program central to understanding the effects of internal and transnational influences on national foreign policy decisions. It has also been influential in policy circles, although in some cases the findings of this program have been misused and misinterpreted by politicians.