This paper considers the direct impact of political parties, interest groups, and SMOs on policy-that is, the impact of each controlling for the impact of the others and of other relevant factors. It tests a "core" hypothesis and three others that refine or qualify it.

The core hypothesis is that all three types of organizations will have substantial impacts on policy. The other three: (1) that when public opinion is taken into account, the political organizations will not have such an impact; (2) that parties will have a greater impact than interest groups and SMOs, and interest groups a greater impact than SMOs; (3) that interest groups and SMOs will affect policy only to the extent that their activities provide elected officials with information and resources relevant to their election campaigns.

The source of data is articles published in major sociology and political science journals during the 1990s, systematically coded to record the impact of organizations on policy. The major findings include: our understanding of the impact of organizations on policy rests on a narrow base, in that most studies focus on only a few issues in a few countries; contrary to the core hypothesis, political organizations have an impact on policy only about half the time; parties may have a very slightly greater impact than interest groups and SMOs, but there is no difference between the latter; there is some evidence that organizational activities that respond to the electoral concerns of elected officials are especially likely to have an impact