Saturday, November 17, 2012

The misleading party ID #s in exit polls

The exit polls are showing a big advantage for Democrats that may not be as big as it seems. They are 38%D/32%R, D+6.

In 2004, the exit polls had John Kerry winning independents by 1 and Bush won nationally by 2. In 2008, they had Barack Obama winning independents by 8 and he won by 7. This year Barack Obama won by 3, but Mitt Romney won independents by 5. The speculation is that the exit polls are capturing people previously identified as Republicans in the independent groups.

By designating them independents, instead of Republicans, you're putting them in a different place than they were 4 years ago, making comparisons difficult.

If we take 2% of the independents and assume they vote like Republicans, Romney 93%-6%, we end up with a 38%D/34% electorate and Barack Obama wins independents by 3%, 49%-46%. In a 38%D/35%R breakdown Obama wins independents by 5%.

Why is it better to use this breakdown instead of the exit poll breakdown? When many of us try to figure out who is going to win an election we look at the spread in party breakdown. When Obama wins by only 3% in a D+6 environment it means Republicans have an inherent advantage. Greater turnout doesn't mean a Democratic win.

In 2008, turn-out was D+7. Obama won by 7%. To compare the two elections we need to have both of them designating voters the same way.

No comments:

Post a Comment