Dan Gardner is the New York Times best-selling author of Risk, Future Babble, Superforecasting (co-authored with Philip E. Tetlock), and How Big Things Get Done (co-authored with Bent Flyvbjerg). His books have been published in 26 countries and 20 languages. Prior to becoming an author, Gardner was an award-winning investigative journalist. More >

My Fearless Predictions For The Election

This is the last column I'll publish before the election and so I will lay out how I think the big day will unfold. I predict there will be an election. Many people will vote. The election will either produce major change, in line with recent polls, or little change, contrary to the polls. Either way, there will be surprises. But - and this is the part I'm most certain of - political analysts will instantly explain why things happened the way they did and what the results mean for the future. I can also say with considerable confidence that most of this insta-analysis and insta-forecasting will be worthless. In part, I base this prediction on the past. It's what happens every election. But I also expect to see this on Monday because it's human nature. People are explainers. If they see dots, they connect them. They have to. It's a compulsion. In a wide array of clever experiments, cognitive scientists have shown that even in circumstances where people absolutely do not know the explanation for an observed phenomenon, they patch together a story, no matter how rickety and easily disproven, and convince themselves it's true. The hardest thing for anyone to say is "I don't know." And when you're a pundit on television paid to deliver insta-analysis, it's close to impossible. The compulsion to explain has consequences. Most importantly, it can lead us to believe things that aren't supported by evidence and reason. But more germane to what will happen Monday, it enhances a serious psychological bias. And it promotes a dangerous delusion. The bias is "hindsight bias." The delusion is that the future is predictable. Imagine it's August, 2008. I ask you how likely it is that Barack Obama will win the presidential election in November. You tell me. I write your answer down. Now, it's April, 2011. I ask you to think back to August, 2008. Back then, I ask, how likely did you think it was that Barack Obama would win the presidential election? You tell me. I compare your answer now to your answer then. They're different. Your answer now is higher than your answer then. That's hindsight bias: After we know what actually happened, we remember ourselves thinking that outcome was more likely than we actually did. The strength of the bias varies. And it can be manipulated. In one study, people were asked to list reasons why a football team won. The team had a better quarterback, they may have said. Or the other team's star is injured. Whatever. The simple act of explaining why things turned out as they did caused hindsight bias to get stronger. Now think about those pundits on the TV shows. Know what every one of them is doing? They're enhancing their hindsight bias. That matters. The expert who thinks, a decade ago, that the Liberals will hold power for a generation or more will be much more cautious about predicting the political future today if he accurately remembers what he believed back then. But the expert who is convinced by hindsight bias that he suspected the Liberals might lose a string of elections and see their support steadily eroded - or even thought it likely that they would - will be much more inclined to think he can see the big stuff coming today. And that's a delusion. Know what pundits were actually saying about the Liberals a decade ago? They called it "Gritlock." The "friendly dictatorship." Historian Michael Bliss wrote that "Canada is more than ever... a one-party state." "If anything is certain in politics," wrote veteran journalist Val Sears, "it is that Paul Martin will be our next ruler and some kid still running the Liberal club in college after that." I could fill this newspaper with similar statements. It wasn't long ago. But I doubt anyone predicted a Liberal party fighting for its very existence in 2011. Maybe some hoped for it, or saw it in a dream. But a serious prediction? No. Twists of this sort happen remarkably often in politics. In 2004, it was widely agreed that George W. Bush had cemented Republican dominance for a generation. In 1994, Bill Clinton couldn't possibly be re-elected. In 1991, George H.W. Bush couldn't possibly not be. And how's this for a story to keep Liberals awake at night: In 1906, Britain's Liberal party, the colossus of the 19th century, was swept to power in the biggest landslide ever. Nine years later, they fractured. Nine years after that, they were a minor third party. They never recovered. Natural scientists have shown that complex systems - systems that are more than the sum of their parts - are inherently unpredictable. Most human systems are complex. People are also self-aware, and aware of others' self-awareness, so we change our minds when we perceive that others will change their minds. And people have hardwired psychological biases that routinely steer thoughts wrong. Add it all up and there is heaps of theory to explain why attempting to predict the political future beyond one's nose is a fool's errand. But we don't need theory to know that. Memory will do. Accurate memory, that is.