“Election polling is in near crisis, and we pollsters know. Two trends are driving the increasing unreliability of election and other polling in the United States: the growth of cellphones and the decline in people willing to answer surveys. Coupled, they have made high-quality research much more expensive to do, so there is less of it.”
“In short, polls and pollsters are going to be less reliable. We may not even know when we’re off base. What this means for 2016 is anybody’s guess.”
This is a quote from a recent NYT opinion article by Cliff Zukin entitled “What’s the Matter With Polling?” If you pay for access to the NYT website, the link is here (NYT Article). In short, the author points out that the polling industry is in crisis. It has become hugely more expensive, if not essentially infeasible, to do reliable polling anymore. Trends including the disappearance of land-lines and growth of the Internet have converged to undermine what little reliability polls used to have. The main takeaway is that polling “science” is really bad and is only getting worse and pollsters have no idea how to make it better.
The author focuses primarily on the growing financial and logistical challenges for the polling companies. Since pollsters must make a huge number of calls to obtain even a remotely valid sampling of reliable data, the cost of doing accurate polling has become extremely high – even prohibitively high.
However I prefer to focus on the problems that this situation creates for rational governance. Even good polls have undesirable consequences. Their mere existence creates an almost irresistible compulsion for politicians to pander to poll results, saying whatever the numbers tell them that likely voters want to hear. Even if the polls are accurate, this may not be the best way to govern – or even to campaign. But if polls have become woefully inaccurate to boot, and yet we continue to pander to them, then we have taken what was already a problematic approach to governance and made it far, far worse.
One specific problem I’d like to focus upon more deeply is the self-fulfilling prophesies that these polls create. If the polls tell us that, for example, Bernie Sanders cannot possibly win, then those polls influence huge numbers of people to respond by not voting for Bernie Sanders – creating a self-fulfilling feedback loop. And what if those admittedly unreliable polls were simply wrong? What if perhaps they were even disingenuously promoted as a stealth strategy by a big money Clinton campaign (theoretically) for exactly that purpose?
Maybe we’re better off without any polls. Good riddance, I say. Maybe we’re better off if politicians campaign and govern according to actual scientific data and humanistic ethical principles, not according to polling. Maybe we voters are far better off if we remain uninfluenced by polls as well.
Whoa there, you say. If you have read my book “Belief in Science and the Science of Belief” (found here) you know that a fundamental principle I champion is that decisions based on facts are inherently better than decisions based on beliefs. If that is the case, aren’t polling facts important information to consider in campaigns and in governance?
Yes but I’m also suggesting that polls aren’t the best facts to use and that they push all the air out of the space for actual, more important and more reliable facts to sufficiently drive political campaigns and decision-making.
Poll any group of alcoholics and the data will likely show that they want more alcohol. As antithetically “paternal” as this may sound to some, a government must provide what society needs, not necessarily what people want. Private corporations can be driven by market research to provide exactly what their customers want (when not unethical or illegal). But a government simply cannot make policy based primarily on polls if they are there to serve the common good.
Now if even accurate polling can create unhealthy pressures for governance, imagine the consequences if we continue to listen to polls that have now become even less reliable. Now we are making decisions that impact the lives of people and the life of the planet that are primarily based not merely on polling data but on really bad polling data.
I say again, good riddance to polls. My hope is that we turn this crisis into opportunity. This is our chance not to merely improve polling methodologies, but to start to weigh polling data far lower in our decisions and instead find ways to make policy decisions based more upon the best objective science and rational decision-making possible. I hope that our emphasis on “what people polled want” is permanently diminished and replaced by more indirect and sophisticated methods of data mining to understand what they actually need and what will best serve humanity and planet Earth in the long-term.