• by Philip Rosenstein , Staff Writer, November 6, 2015

Should polling data this early weigh heavily on campaign decisions? If you look at the accuracy of recent pre-election polls, probably not. Donald Trump’s team seems to think differently; they went straight back to Iowa when he slipped into second behind Ben Carson.

When Scott Walker’s poll numbers fell below 5% in mid-September, dropping out of the race seemed inevitable. As recently as August 14, it was reported that the Koch brothers had picked Walker to be their $900 million surrogate. Barely a month later, his sagging polls caused him to leave the race.

Jeb Bush has had his fair share of polling woes. With polling in the mid-single digits, donors have pressured him to reshuffle staff and cut pay. Why do the media and big donors care so much about polls when they have proven so often to be inaccurate?

It has been a disappointing year for pollsters around the globe.

Nate Silver, who successfully anticipated the results of the 2008, 2010 and 2012 elections, got the British elections wrong earlier this year. Elections in Turkey, Canada and Israel have all gone surprisingly awry when compared to what polls had predicted.

Closer to home, the off-year gubernatorial election in Kentucky was a polling snafu. Democrat Jack Conway had a healthy leads in polls leading up to the Nov. 3 election, despite one last Internet poll showing a very close race. Conway ended up losing handily to Republican Matt Bevin.

Nate Silver wrote earlier this year: “There may be more difficult times ahead for the polling industry.” He points to increased problems in trying to reach entire groups, due to a move away from traditional landlines and questionable online polling standards.

Still, though the numbers we have today may be completely different from the real numbers we see in February, if you’re at the top of the polls now, your chance of being there in three months looks better. And your donors will stick around.

Rousing, traffic-driving statements tend to lead to higher polling numbers. Higher polling numbers get candidates more airtime, which in turn, ups their profile and helps disseminate their campaign messages — a positive feedback loop fueled by our obsession with polls.

Favorability polls and specific polling questions may give us a more complete understanding of where the candidates could end up. In a Quinnipiac poll this week on the question — Is there anyone whom you would definitely not vote for — over a quarter of respondents answered Trump and just under a quarter for Bush.

In contrast, only 3% of Republican primary voters said they would definitely not vote for Ben Carson.