Knock on Wood by Jeffrey S. Rosenthal

Author:Jeffrey S. Rosenthal
Language: eng
Format: epub
Publisher: HarperCollins Canada
Published: 2018-08-22T16:00:00+00:00

Chapter 17

Lucky Polls

One area of our modern lives where luck plays an important role is public opinion polls. Pollsters use random samples to study a population’s opinions about everything from product preferences to work habits to social attitudes. And most prominently, polls are used to predict the winners and losers of elections. If we are lucky, the polls will provide accurate predictions of the final election results. But if we are unlucky, the polls might be significantly off, causing confusion and embarrassment.

Sometimes, polls are spectacularly accurate in predicting election results. For example, just before the 2012 US presidential election, famed analyst Nate Silver sifted through the multitude of election polls and managed to not only correctly predict that Barack Obama would defeat John McCain, but also correctly predict which candidate would win in every single one of the 50 US states1—a very impressive triumph indeed.

Other times, polls are less successful. One example is the 2016 “Brexit” referendum about whether or not the United Kingdom should leave the European Union. This referendum was very heavily polled, with most pollsters confident that the Remain side would win2 by between two and eight percentage points. But when the referendum came, the Leave side won instead, by nearly four percentage points (51.89 percent to 48.11 percent). Many citizens were angry and frustrated that the polls had gotten it so wrong.

Actually, the Brits were already used to bad polling. In their 2015 general election, most polls showed a dead heat between the Conservative and Labour parties. Then the election came, and the Conservatives won by over six points (36.8 percent to 30.4 percent), a very different outcome. This discrepancy was so great that a formal inquiry was called. The subsequent report noted that “On average the final estimates of the polling companies put the Conservatives on 34 percent and Labour on 34 percent. . . . Yet in the event the Conservatives won 38 percent of the vote in Great Britain, Labour 31 percent. . . . In historical terms, the 2015 polls were some of the most inaccurate since election polling first began in the UK in 1945.”3 Others were less diplomatic; the BBC reported, “Following the outcome of the 2015 general election, a mixture of anger and contempt was showered on the pollsters who had spent six weeks suggesting a different result.”4 The Guardian called the error “notorious,”5 and forecasters admitted that “no company consistently showed anything approaching the seven point Conservative lead that happened in reality” and “no one had a good pre-election forecast.”6

So, which is it? Are polls helpful indications of a nation’s mood? Or are they misleading indicators that cause confusion and chaos? How do polls work, anyway?

Random Samples

One challenge for pollsters is basic randomness. If you flip a coin a bunch of times, you will probably not get exactly half heads, but you’ll get close to half heads. How close? Well, there are no guarantees. But if you flip a coin some number of times, the percentage of heads

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.