[js1138]

Your chart omits a lot of relevant data, such as the fact that in years when the vote broke for the challenger, it was a blowout year. Stevenson, McGovern, Ford and Dole were wiped out. The polls were just regressing towards the mean.

This year is likely to be an under two percent year. It's also likely to be a high turnout year. I think the polls are meaningless at this point. It will be won by the party that gets its vote out.

[swilhelm73]

I'd hardly call the 1976 election a blowout with its 2.1% margin of victory. Anyway, the chart disproves your original assertion

The pattern has been clear since the 1930s. The only exception was Reagan vs Carter, in which it broke for Reagan. Undecideds break for the incumbent, or the incumbent party.

, though if you wish to qualify it now as "Undecideds break for the incumbent, or the incumbent party, except when there is a blowout by either the incumbent or challenger, or the polls are regressing to a nominal mean". Then sure, I guess the data supports that.

The data set itself is awefully small to begin with though, and taking out you exceptions leaves us with only a scant 8 elections to work from.

I still contend that my original statement is the most accurate analysis of the data.