Q: You make some excellent points about the value of opinion/issue polling. Would you say the same about the "horse race" polling (Candidate X is 17 points ahead of Candidate Y in State Z) that we get so much of in the news now? Or is it different? Does it have limitations? Pitfalls?

A: Oh, I think there is a very good justification for election polling.  In 1896, the Chicago Record, described as an “independent” paper, did a massive straw poll in Midwestern states (it mailed out over 800,000 “ballots” to a sample of registered voters, nearly 30% of which were returned).  It published results daily on its front page (including comments made on the ballots explaining “why” people were voting the way they were).  This was both news – there was no other such source of information – and a way of preparing for the eventual outcome.  In many emerging democracies, pre-election polling can actually help provide stability by managing expectations.  It also helps validate the election results (this requires competent polling, obviously).

Of course, this doesn’t work when there is no expected outcome and the election is very close--as was the case in 2000!  And while pre-election polls tell us much more than who is ahead, there always has been an inordinate amount of focus on the horse race – by politicians, the media, and the public itself.

Q: What was your most memorable moment during your career in polling?

A: I guess you remember the horrors more than the successes.  I used to say that the scariest eight words in the English language were “I’m really sorry, Dan [Rather]. We made a mistake.”  Although I wasn’t in charge of election projections at CBS News until after 2000, I was the person “assigned” to say those words that year!

I think that polling is most successful when it challenges conventional wisdom.  During the Clinton-Monica Lewinsky/impeachment year (1998), pundits expected the public to turn against Clinton; they never did. 

Q:  What lessons has working as a polling director taught you?

A: I have enormous regard for most of the journalists I worked with.  A decision made under a tight deadline may be a hard decision, but it isn’t necessarily a bad one.  As Herbert Hyman, a great social researcher, wrote describing his World War II work: “Under pressure and with strong motivation, the wartime staffs learned to be efficient and skillful, thoughtful without becoming obsessive, speedy but not sloppy…The deadlines spared us the pains of interminable conceptualizing and analysis…and there was an exhilarating sense of accomplishment. The need for speed not only improved the skills of the staff but also led to innovations in procedure.” 

I also have learned how complicated American opinion is.  I have little patience with those who assume everyone thinks like them and those who create simple explanations for the public’s mood and attitudes. 

Q: 1964 was an election year. So is 2016. What main differences do you see? Does polling contribute to those differences?

A: In 1964 there was an incumbent President running for re-election.  In June, it wasn’t quite clear who the GOP candidate would be.  Today is a little different. 

In 1964 polls were few and far between. Gallup, Harris, and Roper dominated the field.  Media polling as we know it today didn’t really begin until the 1970s.  The nomination processes are more open now—more primaries and much more coverage of caucuses.  So polling expands there too.

But the polling discipline is changing methods to confront declining response rates and changing technologies.  It’s a difficult time, I think.  While I’m still consulting and writing, I’m glad I don’t have to do it 24/7 anymore!