18
medians. Interestingly, the names in Table 5, for the most part, are who a student of
Congress might expect.
TABLES 3, 4, AND 5 ABOUT HERE
The ideal point estimates in Table 3 appear reasonably accurate on their face,
based on knowledge of state ideologies garnered via other measures. For the most part,
the states scaled with liberal (conservative) ideal points are coded as liberal
(conservative) by other state measures (e.g., Berry et al.). One key difference between
the state ideal points presented here and existing measures of state ideology is the ability
to compare senators and states on the same scale. Another difference between these
measures and other scores is that we present ideal point estimates of the state median
voter. Most other state ideology measures use a mean value of some existing measure
(e.g., U.S. House data aggregated to the state level, survey data at the state level). Thus,
in some instances, the estimates in Table 3 are slightly different than might be expected
based on previous scores, though are improvements for those interested in measuring the
median of a state.
For instance, Montana’s ideal point estimate is a perhaps surprisingly moderate
0.136, while Florida’s is a similar 0.123. Conventional wisdom based on the 2000
election results might suggest that Montana is substantially more conservative than
Florida.
9
However, a closer look at the revealed preferences of Montana’s median voter
and Florida’s median voter preferences for president and the Senate in the last few
elections indicate quite a bit of similarity. Floridians just barely favored George W. Bush
over Al Gore, while a substantial majority of Montanans also voted for George W. Bush.
9
Montana gave Bush 64 percent of the two-party vote, while Florida gave Bush just over 50 percent of the
two-party vote.
Convention