Of polls, expert opinion, second order questions, and single data points.

As we slide into the political season, you are going to see many polls, opinions, interviews and statistics. Let me warn you now.

Polls

Polls used to be taken to judge what people were thinking. Most today are taken to support an existing point of view or narrative. Trying to sway your thinking. Before you rely on poll results, you need to know, who funded the poll, who took the poll, how, when, and where was the poll taken. And if possible the questions asked. Good example. A laundry detergent maker, we will call them Ebb, pays a company to take a poll about Ebb laundry detergent. The company sends pollsters to grocery stores looking for Ebb laundry detergent already in a person’s shopping cart. So the pollster asks, why are you buying Ebb laundry detergent? Most likely answer, “I like it”. Poll results = 86% of the people we polled “Like” Ebb laundry detergent. Sway?

Expert opinion.

The question you have to ask is how did the “expert” come to give their opinion to the author. Did the author contact them directly? Did the author contact other “experts”. Did the author have “Reggie Lewis, syndrome” and kept contacting experts, till they found an opinion that supported their effort? How often in a “news” piece is only one opinion quoted?

2nd order question.

Too often no one asks the second question. Good example, question to the Minneapolis city council members. “Why do you want to disband the police?”

2nd order question, Why do most of you have private security firms protecting you, and paid for out of city dollars? Why is it ok for you to have city paid protection, and the residents not. If there is no hard 2nd order question, then the first question was a puff ball.

Single data points.

They mean nothing. Actually they mean less than nothing. Example.

“Deaths from opioid overdoses are down”. Means nothing if the next data point is “The number of doses of Narcan administered has tripled”.

Another example.

I had my first beer today at 7 a.m.

Right after I got home from working an all night shift.

Or, right after I got up. Same single data point, two different stories.

Bottom line, you need to think it through.

Newsletter