Posted by Terry Bouricius on September 10, 2012
This study was prompted by unusual results of a standard analysis of the ratio of party votes received compared to legislative seats won. Typically, in winner take all voting systems such as Vermont’s, there is a deviation from a one to one ratio, which overrepresents the largest party, and may be exaggerated further as the result of gerrymandering. Vermont’s Senatorial Districts are less prone to gerrymandering, being based largely on preexisting county boundaries. A superficial look at the Vermont 2000 election for governor raises a question.
Since 60% of the voters selected pro-civil union candidates for governor (Democrat Howard Dean or Progressive Anthony Pollina), why did control of the House switch from Democrat to Republican, and why did the Democratic majority in the Senate shrink? We focused our analysis on the Senate elections, simply because the amount of data entry is a fraction of that required for a House analysis. We anticipated finding that, although there were more overall Democratic votes, they would be more concentrated in certain districts, and that a roughly 40% Republican minority vote translated into 14 seats (as opposed to the 12 seats that a straight proportional vote would work out to) as a result of a more dispersed geographic distribution of Republican voters. We were surprised to discover that, in fact, there were actually more votes cast for Republican State Senate candidates statewide (356,051) than for Democratic ones (343,190). This seemed counter-intuitive considering the statewide results in the governor’s race.