The Huge Problem with Polls: My Letter to Frank Newport

This letter was sent on Sept. 24 -- via an intermediary who knows him well -- to Frank Newport of Gallup, the pollster adviser to the bipartisan Commission on Presidential Debates. I've received no response. Ironically, Newport is author of Polling Matters: Why Leaders Must Listen to the Wisdom of the People. I think a close reading of the letter shows that Newport has hardly taken his own advice. 
-- Sam Husseini


Dear Frank Newport --

I believe I have found a significant blind spot in the exclusion criteria used by the CPD. When some suggested alternative criteria for inclusion in presidential debates, like if a majority wanted another candidate to be in the debates, the heads of the CPD rejected the effort. Then-CPD Director and former Republican Senator Alan Simpson said: “The issue is who do you want to be president. It’s not who do you want to do a dress rehearsal and see who can be the cutest at the debate.” Similarly, Paul Kirk, the then-co-chair of the CPD (now co-chairman emeritus) and former head of the Democratic National Committee, said: “It’s a matter of entertainment vs. the serious question of who would you prefer to be president of the United States.” (Citation in google books, "No Debate" by George Farah.) 

But none of the polls the CPD is relying on for its exclusion criteria actually ask the "serious question of who you would prefer to be president of the United State" -- nor do they ask "who do you want to be president."

They all ask some minor variation of "if the election were held today which of the following would you vote for". I hope that it's apparent to you that for many people who they "want" to be president among the choices given (Clinton, Trump, Johnson and Stein) is different than who they would vote for. Voting is a tactical choice based largely (especially in this election) on wanting to ensure the candidate you least like does not become president. Thus, millions intend to vote for Trump because they don't want Clinton and millions more will vote the opposite. But many of those people prefer or actually want Johnson or Stein. Those who "want" or "prefer" Johnson or Stein could even constitute a plurality and we'd never know it because the question that would gauge that is never asked. 

[This wording, "if the election were held today] -- which David Moore has described as starting as a "gimmick" may well ironically now be a serious impediment to understanding the affirmative preference of the public, since it has displaced other measurements of public opinion and preference in this critical regard. 

As the pollster adviser to the CPD, it's my view that it's incumbent upon you to ensure that the polls the CPD relies upon actually gauge the "serious" question the CPD officials publicly claim the CPD is concerned with: Who do you prefer/want to be president. 

I hope you will concur, but in either case, I would most welcome your thoughts on this important matter. As I've talked to pollsters since submitting a legal brief on this matter in June, it's become apparent that many pollsters are not free to ask the questions they want to ask, they are frequently at the mercy of the media outlets they work for. I hope that your intellectual honesty will compel you to address this potentially fatal blind spot immediately. (See "How Presidential 'Non-Opinion' Polls Drive Down Third Party Numbers and Facilitate Debate Exclusion")

Look forward to your positive and enlightened response.

best regards, 

Sam Husseini