are substantively significant.
percent to 29.9 percent (see Table 1), and thus small increases of 1 percent misclassification or
inconsistency is actually quite substantively significant given that the two DW-NOMINATE
dimensions generally explain most roll calls, even for the most inconsistent senators.
Legislator uncertainty: independent variables of interest
As discussed above, there are five independent variables of interest that capture the
concept of legislator uncertainty about constituent preferences. Four of these variables are coded
identically in both sets of models. Amateur senator is a dummy variable coded 1 for all senators
who had no prior experience before their election to the U.S. Senate. The Carpetbagger/non-
native senator dummy is coded “0” for senators born in the state they represent and 1 for
senators who are born out of state. Democratic senator is also an indicator variable coded 1 for
all Democratic senators and 0 otherwise. These three indicator variables are all expected to lead
to more classification errors.
The Percent state population growth variable is measured as the increase in statewide
population during each congress for each senator, from the year preceding the start of the
congress to the end of the congress (e.g., for the 87
th
Congress, 1961-62, this would be the
percentage change in the total state population between 1960 and 1962). Higher values mean
census figures during census years (1970, 1980) and state population estimates from the census
for intervening years to construct this variable.
9
Very lopsided votes (e.g., 99-1, 100-0) and unanimous consent agreements are not included in the scaling of DW-
NOMINATE estimates.
10
We also estimated an alternative specification of both models with the absolute value of the change in percentage
state population over a two-year period instead of population growth/change. The results are similar to the results
presented here. States almost always show positive growth, though some show small percentages of growth, while
others exhibit substantial growth over two-year periods
18