and care only about winning). However, the candidates—even taking the same position—might
not converge to the “true” median voter as there is uncertainty about its exact location. And if
the candidates care about personal policy preferences as well, they will diverge from one another
(Adams and Merrill 2006; Adams, Merrill, and Grofman 2005; Calvert 1985; Morton 1993).
While the equilibrium prediction of spatial models of position-taking imply that the two
candidates will take the same position (if they only care about winning) or may diverge (if they
care about winning and have policy preferences), the dynamics underlying that equilibrium
involve search strategies for the “best guess” about the median (for a review of this theoretical
literature see Grofman 2004; Hinich and Munger 1997). Moreover, models concerned with the
specific dynamics of legislator positioning under uncertainty (Kollman et al. 1992; 1998)
theorize that parties (or legislators) are adaptive, searching for optimal positions.
The theoretical logic of these models can be applied to incumbent members of Congress
and their representational decisions. If a senator is uncertain about the location of the
constituency median, and s/he notices that his/her opponent (who is also uncertain) takes a
position distinct from his/her own, then the senator is likely to update his/her belief about the
location of the median. Further, even when reelection is a few years away, legislators may be
aware that their constituency is changing (e.g., due to massive black enfranchisement), but
highly unsure of the location of the median voter in the constituency. In this scenario, senators
may try a variety of positions—voting inconsistently—in order to gauge reactions of
1
Before proceeding further, we would like to dispatch with a theoretical term that may sound similar to legislative
roll call inconsistency. The theoretical concept of inconsistency sounds similar to the strategy of ambiguity. While
perhaps related, we view these as analytically distinct concepts. Ambiguity implies a lack of information presented
to voters for their electoral consumption. To quote Page (1976, 745), ambiguity occurs when legislators “generally
avoid specific issues altogether.” While ambiguity may at first blush appear similar to inconsistency, it is not. This
theoretical concept of ideological inconsistency is what we investigate in this paper.
4