election data show it is true that the candidate who spends more money in a campaign usually wins. But is money the cause of the victory? It might seem logical to think so, much as it might have seemed logical that a booming 1990s economy helped reduce crime. But just because two things are correlated does not mean that one causes the other. A correlation simply means that a relationship exists between two factors-lets call them X and Y-but it tells you nothing about the direction of that relationship. Its possible that X causes Y; its also possible that Y causes X; and it may be that X and Y are both being caused by some other factor, Z. Think about this correlation: cities with a lot of murders also tend to have a lot of police officers. Consider now the police/murder corre- lation in a pair of real cities. Denver and Washington, D.C., have about the same population-but Washington has nearly three times as many police as Denver, and it also has eight times the number of murders. Unless you have more information, however, its hard to say whats causing what. Someone who didnt know better might con- template these figures and conclude that it is all those extra police in Washington who are causing the extra murders. Such wayward think- ing, which has a long history, generally provokes a wayward response. Consider the folktale of the czar who learned that the most disease- ridden province in his empire was also the province with the most doctors. His solution? He promptly ordered all the doctors shot dead. Now, returning to the issue of campaign spending: in order to fig- ure out the relationship between money and elections, it helps to con- sider the incentives at play in campaign finance. Lets say you are the kind of person who might contribute $1,000 to a candidate. Chances are youll give the money in one of two situations: a close race, in     which you think the money will influence the outcome; or a cam-