Although fans think a basketball player can have “hot hands,” psychologists disagree. We might just be displaying our tendency to select facts that support what we already believe–our confirmation bias.
Citing the “hot-hand phenomenon,” many basketball fans believe that players have a greater chance of shooting successfully following several previous points. Then, because they expect a hot hand, they are more likely to identify it. And yet, it just isn’t so. Studying more than a season of Philadelphia 76er games, researchers concluded that there was “no evidence for a positive correlation between the outcomes of successive shots.” Instead, the data echoed random sets of numbers.
Confirmation bias might also affect how we respond to the Rogoff-Reinhart controversy.
Very simply stated, the research of Harvard economists Kenneth Rogoff and Carmen Reinhart tells us that a 90% debt to GDP ratio is a tipping point. With more debt, GDP growth evaporates. Because their work indicates that austerity rather than stimulus can provide the path to economic health, the Rogoff-Reinhart studies have given academic validity to spending cutbacks.
Now, new research from the University of Massachusetts says the Rogoff-Reinhart work has quantitative inaccuracies. Finding a mistaken number and (what they say is) questionable data weighting and country selection, these economists challenge the underpinnings of the Rogoff-Reinhart conclusions.
Reading the exchange, I suspect we will see confirmation bias. Costly for any researcher to negate previous conclusions, I wonder whether each side will choose data to confirm a previously held point of view.
Sources and Resources: If you doubt the accuracy of a challenge to the “hot-hand” phenomenon, do read the original Tversky et al paper here. Then, this Wired article connects it to confirmation bias. (Please note that Wired questions the sources used by its author but I checked and all appear accurate.) As for the challenge to Rogoff and Reinhart, these two BusinessInsider articles, here and here, provide an ideal overview and the R/R response.