18
Finally, party cohesion mattered. In both models, party selection was statistically
significant and correlated with the standard deviation of the party’s W-NOMINATE
median. Thus, deputies did not select more cohesive parties, preferring less cohesive
ones. This indicates that preference heterogeneity was preferred to preference
homogeneity. If deputies considered cohesive parties, not simply their closeness to the
party, then we would have expected a different finding. Instead, it appears that
preference homogeneity was not an important resource.
Before turning to individual-specific results from these models, I want to make a
few points about the goodness-of-fit statistics for both models. The corrected model has
a higher McFadden’s R
2
, Adjusted McFadden’s R
2
, and Count R
2
than does the
uncorrected model. Moreover, in each case the improvement in specificity of the
corrected model is striking. These results lend credence to the argument that correcting
for selection produces more reliable results.
Table 4 About Here
The results of the individual-specific variables are presented in tables 4-6. Due to
the number of party alternatives, 28, the complete results of the individual-specific results
are quite cumbersome. Consequently, I present the results in a simplified form indicating
statistical significance and direction of correlation. The full results can be found in the
appendix. Table 4 presents the simplified results for all party choices on the single-
member district deputy variable. The results show the correlation between being a
single-member district deputy and selecting this particular party in comparison to
selecting the reference group, the SDPU
(o)
.
Convention