If you get the feeling that my commentary has been as much about my friends in political science as about party politics in Congress, you would be right. I hope they are still my friends. This set of essays would be as long as their collective works if I reviewed everything that I have learned from them. The state of theory about congressional parties is much improved for their efforts. Nevertheless, I must confess that I have not always liked the tone of the debate, the care taken in drawing inferences, or the treatment of the previous literature. We can do better. More important, even after years of effort, the study of congressional parties remains somewhat disordered and incomplete. My purpose has been to sort through the theoretical propositions and empirical claims to identify the strengths and weaknesses of the most influential scholarship. In this short chapter, I conclude with some extensions of the themes noted but not developed in previous discussion. These extensions suggest that there is a substantial research agenda for scholars of legislative parties to pursue.

THEME 1. THE CIRCUMSTANTIAL EVIDENCE OF PARTY INFLUENCE IS STRONG

Let us add it up. Here are well-established empirical regularities:

• The correlation between party and long-term policy positions varies across policy areas (Clausen 1973).

• The correlation between party and roll-call voting varies across types of votes (Froman and Ripley 1965; Roberts and Smith 2003; Rohde 1991; Sinclair 2002).

• Party has an independent effect on roll-call voting scores controlling for legislators' scales scores from a survey of their policy positions (Ansolabehere et al. 2001).

• […]