Thinking Slow About Thinking Fast – Part IV – A New Perspective on the Framing Effect

Our tendency to choose options that appear less valuable than alternative options (such as choosing to stick with our original choice in the Monty Hall Problem) is often cited as evidence for our irrationality. However, the view that we are irrational derives too from inconsistency in our preferences.

Nobel Prize winner, Daniel Kahneman, and his colleague, Amos Tversky, described such inconsistency with the following example of the Framing Effect:

In response to the news that a disease outbreak is expected to kill 600 people, 2 programs are proposed:

  • Program 1 allows 200 people to be saved
  • Program 2 provides a 33.3% chance that 600 people will be saved and a 66.67% chance that 0 people will be saved

If asked to choose the superior program, which would you choose? Most people say Program 1, which ensures the safety of 1/3 of those in danger.

Imagine that instead of Programs 1 and 2, these programs are offered:

  • Program A results in the death of 400 people
  • Program B provides a 33.3% chance that no one will die and a 66.67% chance that 600 people will die

Which option do you prefer? A or B? People tend to choose Program B.

If we examine the programs, it is clear that Programs 1 and A are equivalent, and Programs 2 and B are equivalent, but people reliably choose different options depending on how the options are framed. That our preferences in these scenarios are inconsistent makes us, by several definitions, irrational.

But which program is actually superior? 1/A or 2/B? The question is perhaps philosophical, as there is no way to determine which choice will save the most lives/result in the least deaths. As economists point out, the options are equivalent from a utility perspective. Just as in Program 1, the expected average outcome in Program 2 is 200 lives saved (33.3% x 600 people). Similarly, 400 deaths is the expected average outcome of both Program A and Program B (66.67% x 600 deaths). This analysis leads to 2 essential questions:

(1) If all options are equivalent, why don’t we choose Programs 1/A  as often as we choose Programs 2/B?

(2) If all options are equivalent, does our vulnerability to frames carry any significant implications?

Let’s start with Question 1.

When there is no difference in the value of options, we expect choice to occur randomly. Accordingly, we should expect people to choose Program 1/A half the time and Program 2/B half the time. Many scholars point to psychological factors to explain why our behavior deviates from this prediction. Specifically, they describe us as “risk averse” with respect to gains and “risk seeking” with respect to losses. In this context, “risk” means having an uncertain outcome. In the positive frame above (“lives saved”), psychologists say we are risk averse because we choose the option where the outcome is certain (save 200 people). On the other hand, when the frame is negative (“deaths”), we appear risk seeking because we choose the uncertain outcome.

I think that the pattern of our choices illustrated through the Framing Effect is actually quite practical, considering the computational complexity of risk and our use of rules, or heuristics, to facilitate decision making. Heuristics such as “guarantee gains” and “avoid guaranteed losses” would reliably result in asymmetric preferences that have led researchers to conclude that we feel losses more deeply than we feel gains (more on this when we discuss the Endowment Effect).

In the context of the Framing Effect, considering whether to save some lives or take a gamble that involves a chance that no one is saved, a “guarantee gains” approach makes Program 1 the obvious choice. When the frame is switched so that the focus is on loss of lives rather than the saving of lives, an “avoid guaranteed losses” heuristic makes Program B (the risky option) appear superior. Conscious application of such heuristics is even plausible in the Situation Room (“You have to save people” ; “You can’t just let people die.”)

Question 2.

Does our vulnerability to frames even matter if outcomes are roughly equivalent? Rather than exposing a weakness in our decision making abilities, I think that the Framing Effect demonstrates our adaptive tendency to save time and energy when expending those resources would be fruitless. If outcomes are roughly equivalent, then allowing frames to focus us on an aspect of choice for the sake of efficient decision making should not be detrimental. Indeed, when the value of outcomes differs significantly, frames tend to have less of an impact on our choices.

Image via Sandro Donda / Shutterstock.

Nisha Cooch, PhD

Nisha Cooch, PhD, is a Senior Contributor to Brain Blogger. She holds a PhD in neuroscience and is an expert in the neural basis of decision making. You can follow her on Twitter @NishaKCooch.
See All Posts By The Author