The point was debated endlessly: Was the 2004 election about mobilization or persuasion — exciting the base or bringing the swing?
In the immediate aftermath, many proclaimed proof for the mobilization hypothesis. Turnout increased by 4.8 percentage points between 2000 and 2004 and by 7.7 percentage points in the 16 battleground states.
The swing appeared much smaller. Bush defeated Kerry by 2.4 points nationwide, garnering 2.86 percentage points more than he had against Gore in 2000.
Analysts often measure swing by comparing the vote across two or more elections. This approach, however, is fundamentally flawed. Indeed, the problem is so significant and so common that statisticians have a name for it: the ecological fallacy. Simply put, it is impossible to determine individual-level behavior from aggregate-level data.
Examine an admittedly extreme illustration. Let’s say in two elections the Democrat and the Republican each got 50 percent. Zero swing. But it is of course possible that the 50 percent who voted for the Democrat in the first election all voted for the Republican in the second election and vice versa. That would be a 100 percent swing, but there is no way to know that from analyzing county or even precinct data.
The specific scenario is extremely unlikely, but the point remains the same. Mathematicians have been grappling with this problem for more than 50 years, but there is no simple solution.
The only way to know how many people switched sides is through polls that reinterview the same people at two points in time. The National Election Study did exactly that, talking to about 600 voters who cast ballots in both 2000 and 2004. By studying these voters individually, we measure the true swing.
Survey says: around 10 percent of the electorate voted for a different party in ’04 than in ’00. About 7 percent switched from Gore to Bush, while 3 percent swung from Bush to Kerry. In battleground states, the swing was a bit smaller and more balanced.
So those who said it was all about turnout were wrong, as were those who said it was just about the swing. Like most elections, 2004 was about both. But while Democrats mostly succeeded on mobilization, we failed, in relative terms, on persuasion.
Nearly two-thirds of those who moved away from us were women, and switchers were less likely to have attended college — in short, the downscale women upon whom everyone was focused.
The why would be even more revealing than the who. Despite all the talk about social issues, those who switched away from the Democrats had opinions on abortion no different from the rest of the electorate and were only somewhat less supportive of gay rights. They were less likely than average to believe they paid too much in taxes and more likely to favor affirmative action.
National security seems to have been the defining difference for switchers. While 57 percent of voters overall wanted to increase spending for the war on terrorism, 71 percent of those who moved toward Bush had that view; 66 percent overall wanted more spent on border security, compared to 82 percent of anti-Democratic switchers. Thirty-six percent of those who switched to Kerry wanted more spending on defense, while 64 percent of those who moved to Bush took the more hawkish line. Only 14 percent of those who switched to the Democrats thought the war in Iraq was worth it, but 58 percent of those who moved away from the Democrat held that view.
So Sept. 11, combined with Democrats’ now seemingly congenital weakness on national security, does more to explain ’04 than a simple focus on the mechanics of 72-hour programs or on moral revivalism.
Mellman is president of The Mellman Group and has worked for Democratic candidates and causes since 1982, including Sen. John KerryJohn KerryIran’s nuclear deal just the tip of the iceberg for Trump Trump needs to stand firm on immigration, 'religious-test' insticts Budowsky: Ellison, Kerry to DNC? MORE (D-Mass.) in 2004.