UtahPolicy.com recently released a new poll declaring a several-point shift in public opinion regarding same-sex marriage in just a few months. Unfortunately, it’s also misleading and inaccurate.
The margin of error (MOE) of both polls is a cause for concern. While 2.5% is considered very good, the UtahPolicy.com polls only manage a MOE of 4.9%. This is a function of using a very small sample size, just 400 total respondents. Of course, using a larger polling sample will cost a lot more money, but getting away with a relatively large swing is possible because few people are going to look closely (much less understand what it means).
But the bigger problem here is that both polls use two entirely different groups. The August poll uses the more politically active group of likely voters. The recent poll uses just registered voters. Those groups are very different and drawing comparisons between the two is very much an apples and oranges type discussion.
If this were just some blogger (heh) hawking those results, I’d just shrug and say they don’t know any better. But Bryan Schott and UtahPolicy.com are “real” journalists. I’m sure they know better. In fact, I’d go so far as to say that the lines were intentionally blurred to make for a good click-bait headline. The cynic in me says that personal politics also probably got involved to create a “follow the trend” argument using polling.
I hope UtahPolicy.com will retract their headline since it is inaccurate and misleading and maybe dedicate some column-inches (or whatever the digital equivalent is) to explaining how to properly interpret and compare poll results.