18
competitive chambers, including both houses of Congress. We can see that there is clear
separation between the parties on the first dimension, and there is considerable
heterogeneity with the Democratic delegation on that dimension. The APRE increment
for the second dimension is notable, but not out of line, and the over time correlations of
the dimensions now accord with the general pattern (.95 for the first dimension and .41
for the second).
[Figure 8 about here]
We have done some preliminary work on what underlies the dimensions,
especially the second dimension of the Illinois House. An examination of the content of
the bills separated by cutline angles close to 90 degrees (largely party line votes) and
those that cut across the parties is suggestive, but not conclusive. Those bills that split
the parties appear to deal with the topics one would expect: health, labor, education, the
poor, as well as elections and public assistance. The bills with cutline angles closer to 135
or greater degrees deal with issues such as fees, taxes and revenues. Issues of the lottery
and regulation of liquor sales and giving localities the power to change what they charge
for services all are part of this second dimension as well. The mean angle for a set of 118
bills with these topics as part of their short summaries is 111.6. We will return to the
implications of this initial analysis in our conclusions.
Most important for understanding the overall patterns among chambers is that the
Illinois House is not an institution that clearly defies the general pattern of the impact of
parties on the shape of voting coalitions across the legislatures of the US. We can see
this nicely when we re-do Figure 9 in the next section, replacing Illinois House values for
dimensionality (APRE 1) and party Likeness with calculations based on the non-agreed
Convention