Everything you ever wanted to know about polling

Aug 7, 2015, 10:36 AM Updated: Sep 8, 2015, 4:55 pm

A voter leaves a polling location in Detroit Tuesday, May 5, 2015. Michigan voters were deciding Tuesday whether to approve higher taxes to improve the state's deteriorating roads and bridges.Proposal 1, a constitutional amendment subject to a statewide vote, would raise Michigan's 6 percent sales tax to 7 percent to trigger $1.3 billion more a year for highways, roads, bridges and public transit. Lawmakers put the measure on the ballot in December after failing to directly increase taxes themselves or to divert money from elsewhere in the budget. (AP Photo/Paul Sancya)

(AP Photo/Paul Sancya)

Polls. You see them everywhere, and not just in politics.

Can you trust them? (Hint: It depends). How do you tell the good from the bad?

Think Tank co-host Mike O’Neil has spent the last 40 years learning and practicing the craft of survey research (the term pollsters prefer). It’s his day job. The Think Tank is what he does for fun.

This week’s Think Tank picks his brain. Mike will walk you through the history of polling, just under 100 years’ worth:

The Literary Digest: Thought to be the standard, at least until they picked Alf Landon by a landslide over FDR (oops!)
The 1948 “Dewey beats Truman” headline. How did that happen? And why was the University of Michigan was the only entity to get that election right? And why all polling since has used their core principles.
The switch to phones in the 1960s: How pollsters got around the problem of unlisted numbers (that one was easy) and how they are dealing with cell phones (that one is harder and more expensive).
Why you see more junk polls than good ones.
What to look for in a good poll. Who does the best polls?
What Nate Silver did starting in 2008 and 2012 that was revolutionary—and why the same methods may not work in 2016.

All that and more.

This week’s Think Tank. Sunday from 5-6 p.m. and again from 9-10 p.m.