After

I was a high school student in the late 1960s. Education was a lot different back then, though not necessarily the "Leave-it-to-Beaver" high school experiences depicted on television. Starting with my junior year, our English classes switched up a bit from the standard fare English coursework to mini courses. I don't know who came up with this idea, but it was a brilliant one - teaching students the necessities of high school level English through engaging content that seemed to reflect the time we were living in.

One of the most impactful mini courses I took was called “Propaganda and Prejudice”. For the record, this was a public high school in New Hampshire, and the usual English skills of reading, writing, and speaking were still taught, just through the lens of examining things like Thomas Paine or even advertising materials. I don't think my abilities in English suffered from the shift in focus.

The lessons of examination, consideration, and verification were not lost on me. In fact, if I had to point to something I learned in high school that has stayed with me for more than 50 years, the lessons learned in this mini course would be at the top. This week, after reading how polls show the current occupant of the White House is more popular than ever, I needed a moment of reflection on exactly what I was looking at.

My first inclination when reading about polls and how the data is analyzed is to wonder who actually participates. Quinnipiac University is a highly regarded poll after all. Who - and how many participants - are part of this survey?

From this quotation on the Quinnipiac Poll website, I got part of the answer:

The Quinnipiac University Poll uses what has long been considered the gold standard methodology in polling: random digit dialing using live interviewers, calling both landlines and cell phones. This methodology has been the key to our accuracy over our many years of polling.

When a pollsters’ reliance is on cellphones and landlines, I need to question the reliability of the data. A telephone poll, even one using what appears to be randomized information fitting a particular criteria, would seem to be skewed. Who answers random calls from unidentified callers? I’m a Boomer; I do not. Most people younger than me do not answer random unknown calls either.

Data collection and polling are complicated processes in the information age. With spoofed numbers, scam calls, unless I recognize the phone number or the person calling me, I’m not apt to pick up. So are large number of "potential voters" being left out of the data? Would those who are not being queried drive the data in another direction?

The questioning of information taught to me by Mr. Temple and Ms. Podulke way back in high school are, in my opinion, essentials. Who is answering the questions is just as important as what the answer might be. Full stop.

When the headline writers and news analysts tell me Mr. Trump is more popular now than on Election Day last November, I am skeptical. Dig a little deeper. Maybe the full picture is not being represented.

Examine. Consider. Verify. Or at least point out where data collection might be impacting the analysis.