Dear Armchair Pundits:

When you post the latest poll numbers and provide us with your analysis, what you say might mean less than you would like. This is especially true when the poll you cite has a difference between Obama and Romney that is equal to or within the margin of error of that poll. Romney now leads Obama 50% to 47%, which seems significant until you read that the margin of error is +/- 3 percentage points. Is this good news for Romney? Absolutely. But does it predict the future? Absolutely not.

This could be a close race unless really big shifts happen. No one (not even paid media pundits) will be able to say with 100% confidence what is really going on until the first returns come in November 6th. Polling is half science, half heuristics. Please keep that in mind. About half of the pre-election predictions will come true, so of course someone is going to "nail it dead center" in their prediction each cycle. That much is assured.

We humans discount the Unknown and the Unpredictable: big shifts and swings can still happen, gaffes and mistakes can be made by either candidate, negative media saturation can make people stay home, and unleashed super PACs could throw some surprise punches as well. Indeed, we may have surprises that come from GOTV operations (or, the flipside, voter suppression efforts). As The Doors song goes, "the future's uncertain and the end is always near."  

When you read or repost polls, keep in mind that the methodology that made one polling firm "accurate" last cycle is a badge only earned in hindsight. That same firm’s methodology might not be any more accurate than another polling firm's methods this time around. Please know that each poll you cite is a weighted (i.e. massaged) piece of data. The pollsters have to make educated guesses and slight corrections to their results to make sure their small sample reflects the predicted weights of voter turnout amongst different voter segments. This accounts for much of the variation you see between polls that sample from the very same population at the same time. Further, given that many key demographics have different rates of cell phone use (which is harder to poll than more stationary folks who have landlines), quality data is harder and harder to find without additional effort and cost.

Ironically, the best thing you could do for your candidate is to say that the other side is winning. Please hear this contrarian out. Calls that "we're doing this, y'all" or "we're going to win" are more likely to give your social media friends an attitude that it's ok if they don't stand in line at the polling place two weeks from now. In contrast, fear of narrowly losing by a tragically small handful of votes will probably do more to your side's essential GOTV efforts than confident assertions that "we're ahead."  Sometimes when a potential voter says "oh, my team's ahead," their next thought is "maybe I'll stay at work/home." If you really want to help your side, say "this is going to be close! Every single vote counts! Victory is threatened!"

I will end with one cautionary story. In the 2004 election, I was a paid worker on a congressional campaign that was based out of the Democrat's coordinated campaign office in North Carolina. I watched as a growing number of campaign workers and volunteers start to flirt with the possibility that Kerry/Edwards might actually carry North Carolina in an upset surprise. "NC is going Tar Heel Blue," I recall one true believer shouting. This was on Election Day, in a place where people had easy access to field and polling data from different allied campaigns.

Kerry/Edwards lost North Carolina by a whopping 12.4% margin. Some polls on the eve of the election showed this margin to be half that size, leading some people to suggest that North Carolina might actually be "in play" which is wonk speech for "not ignored by the top of the ticket." Even people with the best information and intentions can still get things wrong.

Remember that the future is always uncertain, which underlies the immediacy of now. This is what makes politics exciting for spectators and participants alike.

Regards,

Bjorn Philip Beer

Bjorn is a former campaign worker and former pollster for a major DC-based firm (where he worked on projects such as Berlusconi, Blair, Hillary Clinton, Merck, Ford, Altria Group, Peabody Energy, and various smaller campaigns). He now lives in Charlottesville, VA.