Like 2016, political pollsters got it wrong again this time. After 2016, pollsters were left to figure out the how and why of such a miss. Polling is based on data and data is a science. But finding the correct scientific formula can be baffling.
Once a formula is discovered however, when followed, the results will be the same everytime. For example, we all know the correct formula for water is H2O. But, what if we vary that slightly by combining two Oxygen molecules with two Hydrogen molecules, rather than just one? We can expect very different results. H2O2 is not extra-hydrogenated water, it is Hydrogen Peroxide. Not the same at all.
Some may think it is as simple as targeting only those who are likely to vote this time, because they voted last time. But that’s not enough. What if instead of using the 2016 formula, the pollsters changed it up? To predict who will win an election, it’s important to poll as many segments of society as possible. This means, several variations of the data should be combined, in order to produce the most accurate results.
Some givens for each formula are things like, age, gender and location. Duh, right? But what happens when education, profession, marital status and religion are added in? When polls are conducted on college campuses, the results will likely vary significantly from results obtained from a segment of farmers or auto workers in the Midwest, from retirees, or even from single parents? Are the same issues important to these groups? Likely not.
So, how can this be done better in 2024? Well, the post mortem is still being conducted. (Heck, votes are still being counted). But, I think we all agree that predictions were off, calls reversed and results different than what we had been told. The good news is, data is a constant and ever increasing. Let’s hope the pollsters throw away the old formula and mix things up a bit next time. Until then, let’s work together to make these next four years a success, even if the polls weren’t.