Data from election studies in Sweden, the Netherlands, and the U.S. were analyzed to test the external validity of Sears' person positivity hypothesis. Although people in Sweden and the Netherlands did not rate their preferred party less favorably than that party's leader, overall the results in those two countries tended to support the person positivity hypothesis. However, the U.S. data indicated, contrary to the person positivity hypothesis, that people rated political parties more favorably than the nominees, or the people competing for the nominations, of those parties. An alternative, the principal actor hypothesis, was offered which is compatible with the data from all three countries. This hypothesis is that positivity or leniency will be extended in judgments that do not involve the principal actors in a political system. This accords with our findings that in the strong party systems of Sweden and the Netherlands, parties are judged less favorably than party leaders, while in the weak party system of the U.S., parties are judged more favorably than the individuals nominated by the parties.