11/5: On Election Eve…
When it comes to public opinion polls, this election cycle has had more shoot the messenger reactions than ever before. There’s little doubt that pollsters are in season for October and November.
Maybe this results from the growing twitter-sphere. I can’t recall the number of times I’ve had to explain that we don’t weight by party, can’t weight by party, and shouldn’t weight by party. Party identification is a variable that moves from election to election and from poll to poll. If you had used the ’04 exit polls as a guide for ’08, McCain would have been elected.
Ironically, I’ve never been asked why we might be undercounting young people or overcounting conservatives. I guess the criticisms of poll data are motivated by the political cliche: “Where you stand depends upon where you sit.”
Pollsters can adjust data when there are population parameters but not for attitudes. By way of example, pollsters can weight by age because it is a known number, but not by whether you consider yourself to be young, middle-aged, or old, an attitude.
Then, there’s the issue of pollsters “cooking the numbers” to create some pre-desired result. This criticism is often tied to the “weighting by party” argument.” If you have any worries about pollsters forming a conspiracy, you should attend a professional gathering of number crunchers and watch them try to figure out where to go to lunch. There isn’t a scientifically based public pollster I’ve ever come across in more than three decades of polling who isn’t motivated exclusively by the desire to be accurate and informative.
Then, there’s the matter of track record. A couple of facts about The Marist Poll. In the presidential election of 2008, we polled five of the current battleground states. We called every one right. The average difference between our final estimates and the Election Day results was 2%. And, we underestimated Obama in each case. We are sufficiently humble enough to understand that you are only as good as your last election cycle. And, the battleground states this time are very close.
We are firmly committed to transparency, and make all of our numbers available to the public. Unfortunately, it is our belief in sharing all of our internal numbers that frequently creates the misuse of our polls. But, we will continue to provide the numbers nonetheless because so many people find them valuable and informative.
Finally, I’ve never been convinced that voters are waiting for the next poll to decide who to support. It’s really the other way around. Public polls measure what voters think based upon what the candidates and their campaigns are doing.
Now, I’ve been asked my take on who will win the battleground states. The ‘ol perfessor Casey Stengel used to say, “It’s very difficult making predictions, especially about the future.” Nonetheless, affix your bayonets… Here goes.
Obama has a slight advantage in Ohio and Iowa. Nevada and Wisconsin are leaning his way. The remaining swing states: Florida, Virginia, Colorado, New Hampshire, and North Carolina are simply too close to call.
If (and, it’s still a big “IF”) you give Obama Ohio, Iowa, Nevada, and Wisconsin, and Romney the remaining five states, then Obama ends up with 277 electoral votes to Romney’s 261. The assumption here is that Pennsylvania, Michigan, and Minnesota all remain blue states (another ”IF,” even if not as big). We have not polled these states.
Also, recognize these states are all within single digits and most are within the margin of error. Late movement among undecided voters and get out the vote efforts can still have a big impact on all the contested states. Why? Because it’s very close!
Have a good Election Day. My thanks and gratitude to the more than 100,000 voters who have taken their time to share their views with us this election year.