This is an archived article that was published on sltrib.com in 2016, and information in the article may be outdated. It is provided only for personal research purposes and may not be reprinted.
You can't open a newspaper these days without seeing another political poll. One has a candidate up, another has the same candidate down. How is anyone expected to make sense of it, or perhaps the better question is, should we even care?
Let's get some perspective on polls. Polls are conducted by a whole variety of pollsters all with different methods and methodology. It's tempting to read a poll and assume they know what they're talking about … but not so fast.
Even legitimate polling firms have to rely on a number of elements to reach their conclusions. Things like margin of error, that little understood, barely mentioned figure plays a huge factor in the accuracy of a poll.
Polls only talk to a very small portion of the voting population, sometimes as small as a few hundred. They rely on various lists to determine who they are calling and many exclude use a mix of landline and cell phone numbers one of the more controversial elements considering the significant increase in cell phones and almost exclusive use by millennials.
More advanced pollsters also apply a weighting to certain demographics to attempt to better match the population they are sampling, for example, in Utah, they may weight Republican respondents higher since, statistically, there are more registered Republicans than Democrats so they are should have a higher proportional weight or value.
All of this, as well as more detailed statistical elements, factor in to create the margin of error. It can be a lot to understand, but as a general rule, the smaller the margin of error, the closer and potentially more accurate a poll is. I say potentially because even with a small margin of error, polling isn't perfect and still relies heavily on assumptions about the respondents and the pollster's weighting methodology.
There are a lot of examples of polls that were way off, all of these examples published right here in the Salt Lake Tribune.
2006 • A poll gave Salt Lake County Sheriff Jim Winder a very slight edge of 51 percent of the vote. He trounced his opponent with 63.7 percent.
2007 • A poll predicted a narrow win for school vouchers. The bill ended up losing spectacularly with 67.7 percent voting in opposition.
2007 • A poll on the Salt Lake City mayoral race put the primary election at a statistical three-way tie. Ralph Becker won by a 12-point margin.
2008 • A poll predicted Councilman Randy Horiuchi behind by 4.5 points with only 37 percent of the vote. He won with an 11 percent margin earning 55 percent of the vote.
2012 • A poll showed Jim Matheson trailing Mia Love by 12 points. Matheson prevailed.
During the 2012 election, I was the Executive Director of Alliance for a Better UTAH, a good-government advocacy organization. We conducted a poll on the Salt Lake County Parks Bond and as part of our questions, also polled on the mayoral race, showing Ben McAdams ahead by 3.3 percent. We had no intention of releasing our poll. It was intended for internal use only to track the bond initiative, but when the Salt Lake Tribune released its poll showing Ben McAdams down 10 points, we decided it was important for voters to balance the two polls rather than rely on information we didn't believe was accurate. McAdams went on to win by 5 points. And the Tribune retracted its poll.
Polls can certainly be helpful to campaigns, giving them an idea of where they need to focus their efforts, what messages that resonate with voters, etc., but there is a downside to published polls impacting voter turnout. Utah already suffers from abysmal voter turnout, some of the lowest in the nation. If potential voters see, through a published poll, that a certain candidate is already winning or is far behind, they may be less inclined to vote, believing their vote won't matter. The election is decided.
Utah voters need to understand the value of polls without putting too much stock in their results. Don't let a poll dictate who or even if, you vote. The only poll that matters is the election!
Maryann Martindale is a consultant to the Catherine Kanter campaign for Salt Lake County Council.