CBR Briefing #46

How a widespread fallacy causes bad decisions
People’s tendency to expect the opposite of what just happened could lead to costly mistakes.

A bar chart showing that judges were three-point-three percentage points less likely to grant asylum to an applicant when they granted asylum in their previous case; five-point-five points less likely when they granted asylum in their last two cases; three-point-seven points less when their last two cases were a denial then a grant; and zero-point-eight points less likely when their last two were a grant then a denial. A second bar chart shows a similar scenario with loan officers, who were twenty-two-point-eight points less likely to approve a loan when their last decision was an approval with flat incentives; four-point-eight points less likely when the last one was an approval with strong incentives; sixteen-point-five points less likely following two approvals; eight-point-three points less likely when their last two were a rejection then an approval; and nine-point-six points less likely when their last two were an approval then a rejection. A third bar chart shows another similar scenario with baseball umpires, who were one-point-five points less likely to call a strike after calling a strike on the previous pitch; two-point-one points less likely after calling two consecutive strikes; one-point-nine points less likely when their last two were a ball then a strike; and zero-point-seven points less likely when their last two were a strike then a ball. Note: All numbers are relative to what the decision-maker would have done following two consecutive negative decisions.

The likelihood of granting asylum fell even more after the granting of two previous requests in a row.

  • People who make repeated decisions as part of their jobs are likely to suffer from the “gambler’s fallacy”— the mistaken belief that they are unlikely to see a streak of good or bad cases in a row, even though this often occurs by chance. This belief leads decision-makers to mistakenly alternate their decisions, according to Daniel Chen of ETH Zurich and Chicago Booth’s Tobias J. Moskowitz and Kelly Shue.
  • The researchers find that asylum judges were up to 3.3 percentage points less likely to grant refugees political asylum if the previous case was approved than if it was denied (see chart). The likelihood of granting asylum fell even more after the granting of two previous requests in a row. The researchers examined decisions in 150,357 US refugee asylum cases from 1985 to 2013.
  • Baseball umpires were 1.5 percentage points less likely to call a pitch a strike if they called the previous pitch a strike. Analyzing 1.5 million called pitches from 2008 to 2012, the researchers find that the effect was stronger after the calling of two consecutive strikes.
  • The research suggests that the tendency to commit the gambler’s fallacy can diminish by changing incentives. Loan officers (in a field experiment in India) were much less likely to commit the gambler’s fallacy when they were given strong monetary incentives for accuracy.

Click here to download this briefing as a PDF.

More from Chicago Booth Review

Related Topics

More from Chicago Booth

Related Topics

Your Privacy
We want to demonstrate our commitment to your privacy. Please review Chicago Booth's privacy notice, which provides information explaining how and why we collect particular information when you visit our website.