Pollsters recently suffered another well-deserved humiliation. Republican candidates for Senate and House did much better than polls predicted and some states predicted to support Joe Biden didn't. But if we only criticize polls for their inaccuracy we are missing far more fundamental problems with them.
Inaccurate polling is nothing new. We all remember 2016. But problems go back much further. The infamous 1936 Literary Digest poll predicted a landslide victory for Republican candidate Alf Landon. The actual landslide was for Democrat Franklin D. Roosevelt, who won all but two states. Wagsters updated the old aphorism "As goes Maine, so goes the nation" to "As goes Maine, so goes Vermont."
Pollsters subsequently improved their accuracy, but overconfidence in them still led a Chicago newspaper to print a "Dewey Beats Truman" headline in 1948.
Pollsters always vow to become more accurate. But the basic problem is not with polling inaccuracy but with polling itself. Increased accuracy could make polls even more troublesome.
What does polling contribute to our public discourse and democratic decision-making? What benefit is there to knowing an election's outcome before people have even voted? What's the hurry?
Aside from costing money that could be devoted to more constructive projects, polls actually inflict serious damage on our public life.
Polling has had an especially harmful impact on news reports. In recent decades newspapers, electronic media and pundits have devoted far too much of their limited space or time to the latest polls and speculations about their meaning. This is time or space that could have been devoted to factual reports about candidates and discussion of the costs and benefits of possible public policies.
"News" like this distracts people from the serious business afoot when we have elections. When the reported numbers heavily favor one side in a race, they may make it seem that the outcome is inevitable no matter how voters decide to act. When the numbers are very close, it turns the election into a horse race rather than a process for evaluating top leaders.
Polls undermine democratic control of the government in a second way. Candidates are always tempted to tell us what we want to hear rather than what they actually think. Polling has made it easier for them to lie to us systematically rather than just haphazardly. But the less we know about a candidate's actual views, the harder it is for us to maintain effective control over government, since after they are elected candidates will operate on the basis of their own actual opinions.
Registration complications prevented me from voting in 1964. Although I was then a Republican, I would have supported Democrat Lyndon B. Johnson. Johnson claimed that Barry Goldwater would escalate the Vietnam War, suggesting that Johnson himself opposed it. Re-elected, he promptly escalated and I am glad not to have supporting him on my conscience.
It would be wonderful if we could just ban public opinion polling. But the free speech and press guarantees in the First Amendment stand in the way.
Still, we could at least limit the damage. The more inaccurate polls are, the less seriously reporters, voters and candidates will take them. Totally inaccurate polls would be ideal!
To make it more difficult for pollsters to find the cross sample of voters needed to produce accurate results, we should refuse to answer pollsters' questions and encourage our friends to do likewise.
We should also criticize news media that talk about polling results and, when possible, switch our attention to newspapers and broadcasters that refrain from reporting on polls or at least minimize it.
We should do everything possible to create conditions in which candidates talk about and the media reports, things that encourage citizens to think seriously about public policy and to vote intelligently.
Paul F. deLespinasse is Professor Emeritus of Political Science and Computer Science at Adrian College. He received his Ph.D. from Johns Hopkins University in 1966 and has been a National Merit Scholar, an NDEA Fellow, a Woodrow Wilson Fellow and a Fellow in Law and Political Science at the Harvard Law School. His college textbook, "Thinking About Politics: American Government in Associational Perspective," was published in 1981 and his most recent book is "Beyond Capitalism: A Classless Society With (Mostly) Free Markets." His columns have appeared in newspapers in Michigan oregon and a number of other states. Read Prof. Paul F. deLespinasse's Reports — More Here.