We tweeted this on Tuesday morning: "Lesson from the #IowaCaucus: Political polling is hard. Despite advances in technology, accurate sampling remains the challenge." We were referring to the Republican race, where the ideas of actual results vs. expectations have became media spin topics in the wake of pollsters missing the boat.
Early in my career, I was involved in some political polling. It's hard. It's harder than most other types of opinion surveying. The biggest problem lies within the keystone tenet of inferential statistics (i.e. trying to infer from a sample what a population thinks); in an election, we don't know what the population will be (actual voter turnout and the mix of views among those that vote), so accurately trying to sample that population is virtually impossible. The best that can be done is to look at a range of polls that model likely voters in different ways. Advances in technology allow us to more easily analyze a variety of models and scenarios. One of my favorite sites that does this is FiveThirtyEight which looks at a wide range of topics, including politics, through a data lens.
So the title of this post teases about how what happened in Iowa relates to your business- what's the story with that? Simply this- look at the online reviews of your property(ies). Are all the reviewers a representative sample of the complete population of your guest base? Will a potential new guest properly infer what their experience might be like based on those reviews? That's the point of our services; ask all guests to give feedback, thus providing a more accurate sampling of that population both for your internal use and to push more reviews to sites where potential new customers are researching new places to go.
The blog of Database Sciences and its GuestInsight & ListenKeenly services.