
Though I haven’t gone back and looked and actually counted at the number of state poll posts that I have written up, my guess would be somewhere in the 225 to 250 range since August. But for Hawaii and the District of Columbia, I have written up at least one for each state. There were 18 for Ohio alone and over ten apiece for Pennsylvania, Virginia, Colorado, Florida and North Carolina. And I probably looked at another 200 polls that I didn’t write up as posts. My spreadsheets are filled with numbers and data that lead me to conclude a comfortable Obama win on Tuesday. Trouble is there is much I don’t know and that naturally makes one cautious.
In my career as a Wall Street equity analyst where my job was to predict the behaviour of stocks, I learned to fear the unknown. What am I missing is the question I would ask myself before making a call or changing one of my stock ratings. Even if I couldn’t answer the question, I would at least to provide hedges and probabilities to variation from my predictive course of action.
I tend to believe the polls not so much for their top line, that is Obama X% and McCain Y%, but rather for the trends that can be discerned underneath in the internals. Demographics is political destiny and it is a political hard truth that people vote in blocs. Hence the term voting bloc. The reason that I have paid attention to all these states polls is that in my view the state polls matter more because due to the infinite wisdom of James Madison in the United States one can win the Presidency without winning the popular vote. And if Senator McCain does manage to win on Tuesday, I am convinced that he won’t win the popular vote.
The state polls do have one risk compared to the national polls. The state polls generally have a smaller sample size and thus a larger margin of error and the sample breakout varies greatly by state. Getting those right is of the utmost importance. Survey USA’s polls frankly worry me, they may have overpolled Democrats in some cases. Still my measure of comparison has been to use the respected state-specific pollsters such the Field Poll in California, Siena in New York, the Ohio News Consortium in Ohio, the Cronkite Poll in Arizona among others. These polls are effectively my control. Comparison across polls can point to deficiencies in a poll. And of course all polls have confidence levels, generally 95%. So 5 times out of a 100, a poll is simply wrong.
So what I am missing? Well to answer that question, I’ll point to the UK General Election of 1992 when the Conservatives led by the then bland Prime Minister John Major defeated the energetic Neil Kinnock and the British Labour Party. The polls all had predicted a Labour win. The British electorate of 1992 was especially volatile and those supporting the Tories were reticent in telling pollsters their intentions. The electoral outcome left pollsters wondering where they had gone wrong and added a new term to the political lexicon, a “shy Tory.”
Neil Kinnock was not the only surprise loser at the last general election. The other was the polling industry. Never before had so much egg landed on its collective face. The final polls, published on election day and conducted over the two previous days, showed the Conservative and Labour parties level. NOP and MORI put Labour ahead, ICM declared a dead heat and Gallup put the Conservatives ahead by a mere half per cent; the average of the four was a Labour lead of 0.9 per cent. Everything pointed to a hung parliament and a coalition government. In the event the Conservatives won with an overall majority of 21 seats, 7.6 per cent ahead of Labour in the popular vote. The polls were out by over 8 points, having underestimated the actual Conservative vote by 4 per cent and overestimated the actual Labour vote by 4 per cent.
So what went wrong? (more…)