The data on Senate campaign content examined in the previous chapter indicates that some dialogue does indeed take place. These instances of dialogue make it possible to test ideas as to when dialogue will occur. In particular, these relatively rare instances of dialogue can be used to build a statistical model, which explains dialogue's occurrence. This analysis will require additional data, specifically information on the candidates' positions and other background factors. By developing this explanation, we will be able to generate predictions as to when to expect dialogue. Thus, this final analysis should complete our understanding of the dynamics surrounding campaign dialogue.

I will begin with a reexamination of the game-theoretic model that was originally presented in chapter three. This reexamination will concentrate on developing hypotheses to explain dialogue's appearances. I will develop hypotheses related to the newspaper's editorial policies, characteristics of dimensions and the candidates' positions (among others) that will predict and explain the occurrence of dialogue. The new data necessary to test these hypotheses will be introduced and the appropriate tests conducted. By the end of this chapter, we will have an understanding derived from the model as to when circumstances will lead to more substantive electoral campaigns.

SUMMARY OF THE MODEL

The concept of dialogue served as this investigation's point of departure. I framed the decision to dialogue or not as a strategic choice, made in light of the candidates' expectations as to what responding to or ignoring a particular campaign theme would do to their electoral prospects.