Swing Voters

Given that I've never heard of anyone who believes that political parties face zero-sum choices between base-mobilization and persuasion strategies, nor have I ever heard of anyone who denies that both are part of political success, I'm a bit puzzled as to why The Democratic Strategist would bother to host a "debate" on this issue in which we'll discover that liberals and centrists alike agree that both are important and both sides just claim that adopting policies they prefer on the merits is also the key to electoral success. That said, it is worth pointing out that DLC honcho Al From is trying to trick you with this chart:



From would clearly like you to believe that the combination of increased margin and higher turnout among self-identified Democrats in 2006 relative to 2004 was a smaller factor in the Democratic Party's superior election outcome than was the even larger increase in margin combined with lower turnout among self-identified independents. If you check the math, though, you'll see that this isn't true.

In 2004, 37 percent of the electorate were Democrats, and Kerry got 89 percent of their votes. Thus 32.9 percent of the electorate was Dems voting Dem. Independents were 26 percent of the electorate and 49 percent of them voted for Kerry, so 12.7 percent of the electorate was indies voting Dem. In 2006, 38 percent of the electorate were Democrats, and 93 percent of them voted Democratic. Thus 35.3 percent of the electorate was Democrats voting Democratic, an improved performance of 2.4 percentage points. Independents were 26 percent of the electorate, and 57 percent of them voted Democratic, making 14.8 percent of the electorate indies voting Democratic, an improved performance of 1.1 percentage points.

In short, contrary to From's chart, it's simply false to say that "the difference" between 2004 was that "centrist voters with loose party attachments voted Democratic in much higher numbers." The Democrats improved their performance among both groups, but the combination of turnout and vote-share factors clearly indicates that improved performance among self-IDed Dems was a more important factor than was improved performance among self-IDed Republicans. What's more, note that the numbers I used were identical to the exit polls From is using in his chart. But he presents the numbers in a such a way (using the change in margin of victory rather than showing actual vote shares) as to make it difficult to do a quick calculation of the change in performance.

I don't personally have a huge dog in this fight. As everyone agrees, both groups are important and it's not an either/or choice. But it's nice to have reviews of exit poll be done accurately.