I spent the week in St. Petersburg, Florida, at a conference on analytics. The “real” title was “Moving the Needle.” But, between the predictive analytics of the conference and the incorrect predictive analytics of the election, by mid-week, I indeed felt the familiar pain of a needle stabbing in my eye.
I found myself awake on the morning of November 9 at 3:40 a.m. watching our new president-elect accepting having won the election, much to the chagrin of the media and the pundits who had been reporting the incorrect predictive analytics of the polls for weeks.
Donald Trump had been telling people correctly not to pay attention to the polls, but of course, it is how the media influences the election, pretending that they are simply reporting the data regarding the polls. And so, the media once again had egg on its collective faces when the pollsters failed to predict the large turn out among what were referred to as “uneducated” white males (only uneducated because they don’t hold traditional four-year college degrees; again an attempt by the media to insult Trump voters and thus influence the election.)
So, after getting only four rough hours of sleep I was in a conference session on “making predictive analytics actionable.” The presenter who introduced the topic said the real title should be “trying to make predictive analytics actionable.” I thought to myself, “how can we, at colleges, who are trying to take some of this information on what we know about students, for example students who display certain high risk characteristics for dropping out, or not completing, and predict which ones will and won’t be successful, and do it with any level of reasonable expertise, when the pollsters with all their money and statisticians and years of data and analytics, and knowledge, and computer software, can’t predict who is going to win an election because they can’t accurately predict voter-turnout of a certain group of voters.
It just made me want to pack up all my stuff and go home. It made me tired. I wanted to give up on even trying any kind of predictive analytics in my little job.
Watching the electoral college map on Tuesday night and early Wednesday morning was another great example of how predictive analytics work (or not). There were only a few states that were in “play” because voter analyst could predict, based on historical data, what precincts and states would vote for the Democratic or Republican candidates. In any given presidential election, some states will be the “swing” states and those electoral votes will decide the election. The pollsters got those predictions wrong too, and Trump won states such as Florida and Pennsylvania which he was not expected to win.
On a side note, I do want to mention that I’ve heard a lot of negativity about the electoral college lately. Some think we need to get rid of it, because a candidate can win the popular vote but not the electoral vote. Before you decide that, keep in mind that our country is not a democracy, but a democratic republic, which is a representative government. You need only take a look at the map above to see that it works. If we didn’t have this form of government, a few highly populated areas in our country would always choose our president. So, get informed and I think you’ll see that the electoral college is necessary and still works!
Anyway, the activities of this week have made my brain hurt, by stabbing a needle right in my eye with an overload of incorrect predictive analytics. It makes me not want to try. But, on a positive note, the weather in St. Pete was very nice in November! ⛱☀️
PS Isn’t that Barron Trump cute? He’s a 10-year-old version of his dad complete with the messy blond hair, especially when he’s sleepy at 3:40 a.m. I wish him the best growing up in the White House.