Empiricism is the only way to understand the world---proves the K doesn’t turn the case.

First and most obviously, a good theory should be logically consistent and empirically valid, because a logical explanation that is consistent with the available evidence is more likely to provide an accurate guide to the causal connections that shape events.

Second, a good theory is complete; it does not leave us wondering about the causal relationships at work (Van Evera 1997). For example, a theory stating that “national leaders go to war when the expected utility of doing so outweighs the expected utility of all alternative choices” (Bueno de Mesquita & Lalman 1992) may be logically impeccable, but it does not tell us when leaders will reach this judgment. Similarly, a theory is unsatisfying when it identifies an important causal factor but not the factor(s) most responsible for determining outcomes. To say that “human nature causes war,” or even that “oxygen causes war,” is true in the sense that war as we know it cannot occur in the absence of these elements. But such information does not help us understand what we want to know, namely, when is war more or less likely? Completeness also implies that the theory has no “debilitating gaps,” such as an omitted variable that either makes its predictions unacceptably imprecise or leads to biased inferences about other factors (Nincic & Lepgold 2000, p. 28).

A third desideratum is explanatory power. A theory’s explanatory power is its ability to account for phenomena that would otherwise seem mystifying. Theories are especially valuable when they illuminate a diverse array of behavior that previously seemed unrelated and perplexing, and they are most useful when they make apparently odd or surprising events seem comprehensible (Rapaport 1972). In physics, it seems contrary to common sense to think that light would be bent by gravity. Yet Einstein’s theory of relativity explains why this is so. In economics, it might seem counterintuitive to think that nations would be richer if they abolished barriers to trade and did not try to hoard specie (as mercantilist doctrines prescribed). The Smith/Ricardo theory of free trade tells us why, but it took several centuries before the argument was widely accepted (Irwin 1996). In international politics, it seems odd to believe that a country would be safer if it were unable to threaten its opponent’s nuclear forces, but deterrence theory explains why mutual vulnerability may be preferable to either side having a large capacity to threaten the other side’s forces (Wohlstetter 1957, Schelling 1960, Glaser 1990, Jervis 1990). This is what we mean by a powerful theory: Once we understand it, previously unconnected or baffling phenomena make sense.

Fourth, at the risk of stating the obvious, we prefer theories that explain an important phenomenon (i.e., something that is likely to affect the fates of many people). Individual scholars may disagree about the relative importance of different issues, but a theory that deals with a problem of some magnitude is likely to garner greater attention and/or respect than a theory that successfully addresses a puzzle of little intrinsic interest. Thus, a compelling yet flawed explanation for great power war or genocide is likely to command a larger place in the field than an impeccable theory that explains the musical characteristics of national anthems.

Fifth, a theory is more useful when it is prescriptively rich, i.e., when it yields useful recommendations (Van Evera 1997). For this reason, George advises scholars to “include in their research designs variables over which policymakers have some leverage” (George 2000, p. xiv; also Glaser & Strauss 1967, Stein 2000). Yet a theory that does not include manipulable variables may still be useful to policy makers. For example, a theory that explained why a given policy objective was impossible might be very useful if it convinced a policy maker not to pursue such an elusive goal. Similarly, a theory that accurately forecast the risk of war might provide a useful warning to policy makers even if the variables in the theory were not subject to manipulation.

Finally, theories are more valuable when they are stated clearly. Ceteris paribus, a theory that is hard to understand is less useful simply because it takes more time for potential users to master it. Although academics often like to be obscure (because incomprehensibility can both make scholarship seem more profound and make it harder to tell when a particular argument is wrong), opacity impedes scientific progress and is not a virtue in theoreticalwork. An obscure and impenetrable theory is also less likely to influence busy policy makers.