decrease as the majority’s floor advantage increases. If a committee is relatively small,
this leads to even greater ideological stacking. Clearly the intercept of Figure 3 is
positive as well. The correlation of majority floor advantage with ideological stacking is
–0.79. Excluding the four recent Republican-controlled sessions, the correlation drops
only slightly to –0.69.
[Figure 3 about here]
For the years that were examined here (1975-2001), average ideological stacking never
reports a negative value. This is not always the case for individual committees, though.
Figures 4, 5, and 6 show the scatterplots for Appropriations, Rules, and the Veterans’
Committee. These committees appear to offer strong support for the contention that
ideological outlying committees can be caused by the inclusion of redundant components.
Interestingly, if one were to examine each of these committees for outlying
characteristics, the answer one receives is largely contingent upon the year examined. If
a closely divided congress were examined, then the committee might be classified as an
outlier in the favor of the majority party. If a year were examined in which the majority
party had a very large floor majority, then this same committee might be classified as an
outlier favoring the minority party. And of course there are years in between these
extremes which could produce the very common null findings. What is important to note
here, however, is that each of these three results appears to be the result of the same data
generation process.
[Figures 4, 5, and 6 about here]
Some scholars have detected these “contradictory outliers” in which a committee
switches its outlying ideological direction over time. Such a finding has commonly
Convention