The research from the University of Iowa is based on previous studies indicating that people are particularly likely to stick to their original viewpoint when they’ve had to write their beliefs down– a phenomenon known as the ‘explanation effect’, which also affects future actions.
In the study, Tom Gruca, a professor of marketing at the Tippie College of Business, tried to find evidence of something called ‘confirmation bias’ – the tendency to give preference to existing information or beliefs, rather than considering alternative possibilities. He says equity analysts working on financial markets are particularly prone to this bias, with those who issue written forecasts being especially vulnerable to falling into the trap, despite having access to new data to influence them.
The study took place over a 10-year period, between 1998 and 2008, and focused on the Iowa Electronic Markets, an online futures market at the college, where payoffs are contingent upon real life events. In that time frame, Gruca had the students analyze the movie market. The students had to predict four-week opening box office totals for 18 movies, while also buying and selling contracts with real money.
It was discovered that, despite the initial box office figures giving a good indication of potential success, the student traders ignored the stocks and stuck to their original predictions. As a result, nobody was buying or selling – confirmation bias prevented the prices from becoming unstable.
In order to test for the presence of the bias, Gruca had the student traders explain why they had made the predictions they did prior to the beginning of trading. According to his results, it was this process of explaining – or the ‘explanation effect’ – that solidified the students’ beliefs and prevented their trading behavior from changing.
To keep from making erroneous conclusions, Gruca used a ‘control group’ in his research – in this case a separate group of people trading in markets, who weren’t asked to explain the logic behind their forecasts. With the explanation effect missing, it was found the respondents adjusted their opinions according to the new information presented much more readily.
Consequently, the stock prices changed much more fluidly than with the first group.
According to Gruca, “This study shows that when all traders in a market have the same bias — in this case, confirmation bias — market prices are not efficient and do not reflect all of the information available.
“However, if some traders are not biased, then market prices efficiently reflect new, relevant information,” Gruca writes.
The study, co-authored by Michael Cipriano, associate professor of accounting at the University of South Carolina Upstate, and former Tippie MBA graduate, was published in The Journal of Prediction Markets.