Study: Agents like Snowden prone to irrational decision-making

By Karene Booker
Reprinted from Cornell Chronicle, July 11, 2013

U.S. intelligence agents – like the embattled Edward Snowden – are more prone to irrational inconsistencies in decision making than college students and postcollege adults, reports a study to be published in a forthcoming issue (as yet unscheduled) of the journal Psychological Science.

“With increasing age and experience, people are less likely to engage in literal, quantitative analysis and more likely to use simple qualitative meaning or gist when making decisions,” said Valerie Reyna, professor of human development in Cornell’s College of Human Ecology, and lead author of the study. “While the growth of experience-based intuition can enhance performance, it also has predictable pitfalls.”

For the study, 36 agents from a federal intelligence agency, 63 college students and 54 adults were presented with scenarios involving risk and asked to make choices – the options were systematically varied to omit information or emphasize gain or loss, while leaving the literal meaning the same.

For example: A dreaded disease is threatening a town of 600. Do you: Save 200 people for sure or choose the one-third probability that 600 will be saved and a two-thirds probability that none will be saved? Or: Do you pick the option where 400 will surely die, or instead a two-thirds probability that all 600 will die and a one-third probability that no one dies?  Both versions of the decision are equivalent—if 400 die, then 200 are saved.

While we would expect rational decision makers to treat such equivalent options the same, the results showed agents treated them differently based on superficial wording changes. Agents were more willing than college students to take risks with human lives when outcomes were framed as losses, and they were more confident in their decisions.

When lives are at stake, simple categorical distinctions like saving some or none become pivotal, Reyna said. According to her research, decision-making gravitates to the simplest bottom-line gist of options, which boils down, in the gain scenario, to saving some people versus either saving some or saving none. Decision makers choose the sure option because saving some lives is better than saving none. Conversely, in the loss scenario, the options boil down to some people die versus either some die or none die. Valuing none die over some die, decision makers choose the risky option, which offers the categorical possibility that none die.

“The irony is that being biased by context (gains vs. loss wording) is a hallmark of the most advanced thinking – the kind of intelligence that intelligence agents should have,” said Reyna “Our results shed light on the underlying mechanisms of decision making at work in intelligence agents and others who make life-and-death decisions.

“And framing questions, like some other laboratory gambling tasks, has been shown to predict real-world behavior,” she added.

The article, “Developmental Reversals in Risky Decision-Making: Intelligence Agents Show Larger Decision Biases than College Students,” which was co-authored by graduate students Christina Chick and Jonathan Corbin, and Andrew Hsia ’12, was funded by the National Institutes of Health.

Karene Booker is extension support specialist in the Department of Human Development.