Understanding CHOICE


Understanding how human beings make decisions is critical in fields like cybersecurity, public health, elections and governance, and economics. How often do people make rational choices, weighing all the options? How often do they use mental shortcuts, short-circuiting good choices?

Do our choices vary because we make mistakes, or are our minds split between different priorities? The answers to these questions matter whether we’re asking why people don’t choose healthy lifestyles, why they don’t come out to vote in elections or why they end-run security measures. The field suffers in particular from a large number of theoretical models that have not been sufficiently tested. 


Michel Regenwetter and colleagues at the University of Illinois at Urbana-Champaign believe that understanding decision making will mean statistically testing the different theories against real people making decisions—and against each other. With help from PSC staff, and utilizing PSC’s Blacklight system, the group was able to test tens of thousands of theories against the data at once—thousands more than the largest previous analyses. With new colleagues in the neurosciences, Regenwetter plans to test whether people who tend to make rational decisions have different patterns of brain activity than those who make decisions by shortcut. 

“There are basically two reasons we need supercomputing. One is that there are a lot of theories out there that we need to test, and most don’t actually explain variability of behavior. The other is that the statistical methods we use are very complex, and so computationally expensive. We’re scaling up the speed at which the research is being done by two to three orders of magnitude.” 

—Michel Regenwetter, University of Illinois at Urbana-Champaign 

figure3 cobaltBlue

figure4 cobaltBlue

In one possible explanation for decision making (left), the decider has a fixed preference between three choices (L, M and S) but makes mistakes. Each shaded box represents a different preference between the three, its size determined by the decider’s rate of error. In another explanation (right), the decider’s preference isn’t so set, and can change, so that the possible choices smear out to include a much larger possible set of choices. Data from people making decisions in experiments can help researchers compare the two explanations’ ability to predict those decisions.