Well, aside from the soft hum of the kegerator chilling beer for post-5:00 p.m. consumption, a new poll by Quinnipiac was published yesterday and our client Marco Rubio is in second place!
This is fantastic news!*
*…well, wait. Sort of.
This certainly isn’t bad news. We’d much rather see our client’s name near the left side of that grid than the right, however, clicking on the “Quinnipiac” button takes you to this document published by Quinnipiac. (http://www.quinnipiac.edu/images/polling/us/us12022015_U45hkpp.pdf)
On page 2 of the document, it indicates that the margin of error for this question is +/- 3.8 percentage points. That is, we are pretty confident (about 95% confident) that the true proportion of voters who would vote for Rubio if the election were held today is between 13.2% and 20.8%. That’s a pretty big range – and we’re only 95% confident about that! We’re also about 95% confident that Cruz and Carson are between 12.2% and 19.8%. There’s a lot of overlap there where Carson and/or Cruz could be ahead of Rubio.
We also want to make sure our sampling frame (the group from which we sample) matches up with our population of interest. Quinnipiac used 672 registered Republican voters for this sample. The problem with this is that many people who are registered simply won’t vote for one reason or another. For example, in 2012 only 5% of Iowans who were registered to caucus actually did caucus. The people who are registered to vote but don’t end up voting likely have different beliefs than the people who actually will turn out to vote. For example, maybe the people who said they would support Trump are less likely to show up to the polls.
Even if we interpreted the margin of error properly and lined our population up so that the only people we sampled were definitely going to the election, there’s still a problem – the elections are still months away. The Super Bowl will come and go before the first vote is cast. Anything can happen in that time – maybe everyone below 3% drops out and they all rally around a particular candidate, or some scandal breaks that forces a frontrunner out of the race.
We’re always happy to see our clients doing well! But it’s important to be especially careful about what we take away from polls like these and not chalk these poll results up to be more than they actually are.