Monday, November 13, 2006

Who Voted?

One thing that I haven't heard many people talk about in the wake of Deval Patrick's historic victory was the fact that exit polls showed more Democrats voting in Massachusetts than Independents this year. Here are the percentage breakdowns from CNN, with the results from 2002 (PDF) and 1998 for comparison:


As Christy Mihos was quick to remind everyone this campaign season, unenrolled voters make up half of the registered voters in Massachusetts. That only tells part of the story, however, since this chart shows that they have never quite been half of the subset of people who actually make it out to the polls in a gubernatorial year. In fact, what we see now is that the number of independents has been decreasing as a share of the voting population since 1998, to the point now where Democrats outnumber them.

Using that information, and the raw voter registration numbers (PDF) from the Secretary of state, we can estimate turnout in each subgroup for 2006. The number of voters by party should be the percentage from the exit poll multiplied by the total number of voters who voted on Tuesday. If we divide that the number of registered voters, we can come up with our best guess of turnout by party.


I rounded the estimated voters to the nearest thousand to emphasize that these are estimates, and not exact numbers because they're based on polling information. In any case, we don't require that much precision to get the estimated turnout.

Compare those numbers above to what I posted in February for 2002 and 1998. The percentage of independents who turned out this year (44%) was down from the levels in 1998 (45%) and 2002 (47%). In contrast, the number of Democrats shot up from 52% in 1998 to 57% in 2002 to 62% now. It would appear that the number of Democrats who came out to vote for governor has not been higher in at least the last two elections. Note also that year after year, unenrolled voters are the least likely to turn out in a gubernatorial election. Fewer than half of unenrolled voters make it to the polls on election day.

Not only did more Democrats turn out than expected, but they were unusually loyal to the Democratic ticket. Here is the vote by party from those exit polls:


Eighty-five percent of Democrats reported a vote for Deval Patrick. That's ten points higher than the percentage who voted for Scott Harshbarger in 1998. It's amazing to me that only 15% of Democrats crossed party lines, and only 3/5ths of those voters could actually bring themselves to vote for the Republican. Notice also that while independents did break for Patrick, it was not a huge swing, only a four point difference. Deval Patrick's landslide was because more Democrats came out than in years past and those who did stuck with their party.

Late in the campaign, we were asked to stop making voter ID calls to Democrats. The campaign had seen what the exit polls ended up showing -- Democrats were sticking with their candidate at a rate we haven't seen in Massachusetts for many years. When we made get-out-the-vote calls, we were told to assume that every Democrat we hadn't IDed was already with us. Some of us, myself included, were very resistant to this, but it turned out to be a sound strategy. If there was an 85% chance that any given Democrat would fill in the oval for Deval, that would be worth the 15% who would come out for one of the other candidates. Getting Democrats of any kind to the polls became just as important as getting independents we had already identified as supporters.

Back in February, I calculated that if 70% of Democrats, or just about a million, came out to vote, then the Democratic candidate would be guaranteed a victory assuming that turnout and partisanship levels stayed the same as in 2002. The Democrats fell short of that, but because of their unprecedented loyalty to their candidate, Patrick was still able to score a resounding win.