The House of Representatives will vote on a financial bailout bill for the second time in a week this afternoon, and backers of the measure are desperately hoping to corral at least 13 more vote than they did on Monday, when the bill failed by a 228-205 margin.
In order to have any hope of wrangling those 13 votes, party leaders need to understand the unusual ingredients that made up the defeating coalition. Various Trailhead readers noted that the extremes of both parties voted against the measure, while donations from Freddie Mac and Fannie Mae correlated with support for the bill.
Catholic University politics professor Matthew Green took that analysis a step further. He ran the roll call-vote against 14 factors that might have affected a representative’s vote: Data points like freshman status, membership in the various House caucuses, vulnerability in the upcoming election, and so forth.
Using a standard logistic regression model , Green discovered several factors that reliably predicted a “no” vote, and another few that reliably predicted a “yes” vote.
Members of the Congressional Hispanic Caucus , the Blue Dog Coalition of fiscally conservative Democrats, the Congressional Black Caucus , and those on the ideological extremes of their party were likely to vote against the bill with a high statistical significant (p < .05, if that means anything to you).
Meanwhile, party leaders, those not running for re-election, and members of the New York delegation were significantly likely to support the bill. Members of the Financial Services Committee also trended toward supporting the bill, but with a slightly weaker correlation.
This model essentially studies each factor in a vacuum, holding other factors constant in order to study its effects in isolation. So a lawmaker with multiple, conflicting traits—take for examples, Rep. James Clyburn , D-S.C., a member of the Congressional Black Caucus and a party leader—predicting the vote gets tricky.