change was 370% higher than southern Democrats. The effect of intraregional crosspressuring is
supported only in the northern model; members from the northwest are at greater risk (116%) of
instability than northeastern members (p=.001). The electoral marginality effect, on the other
hand, is supported by the southern model but not the northern model. Incumbents with higher
victory margins in the south were more likely to reverse positions (p=.003)—the difference in
risk of change was 147% between members with victory margins one standard deviation below
and one standard deviation above the mean. The constituency interest proxies are not related to
the risk of change at statistically significant levels in either region, though the northern
antislavery dummy variable is correctly (negatively) signed and nearly reaches the .10 level of
significance. Neither section’s members experienced an increased risk of change when partisan
control of the House shifted.
[Table 3 about here]
Two additional factors prove to have a significant relationship to position change. In the
north, members were sensitive to the content of the specific gag at issue. When the gag rule
shifted from a resolution to a standing rule at the start of the 26
th
Congress, northern members
were at a 209% higher risk of position shifts than in other votes.
southerners, the length of the vote history itself has some impact on the risk of position change.
Figures 2 and 3 plot the baseline survivor rates over time, showing a member’s probability of
“surviving” (continuing to cast a consistent vote) at each point in their own vote history, when
other factors are held constant. The largest drop-offs in the survivor rates appear early in the
vote history for northerners and southerners alike—a trend that demonstrates that the vote history
gains value over time.
18
Convention