Framing

I just finished reading a new article on the neurology of framing effects that is published in the latest issue of the Journal of Economic Psychology "The framing effect and risky decisions: Examining cognitive functions with fMRI" (Feb. 2005, available on Sciencedirect) by Gonzalez et al. They investigated what areas are activated when subjects are asked to answer the "Asian Disease Problem" that Tversky and Kahneman discussed.  They found ,just like nearly everybody else, that the subjects were risk-averse as to gains and risk-loving as to losses.  They also found that those who chose the risky answer in the gain frame were more likely to have differentially activated BA 9, middle and BA 46 as well as BA 8 in the prefrontal and BA 40, BA7 and the region in between them in the parietal, as well as interestingly bilateral BA 19 activation in the occipital.  They didn't find as much difference between the activations of those who chose risky or certain in the loss frame (although those who did chose risky losses did have  a little higher activations in the right dorsolateral prefrontal cortex and a little bit more in the lateral intraparietal sulcus).  This is an interesting finding.

After describing the finding, the article then goes on to discuss a theory of how losses have emotional consequences different from those of gains, and risk is treated differently than certainty.   While this theory is indeed consistent with the evidence they present, this theory brought to mind a couple of thoughts.  First, what they did is show that subjects who chose a certain gain had a different brain activation from those who chose a risky gain.  This does not mean that certainty on the gain sides recruits different brain mechanisms, but rather those who make the choice of a certain outcome recruit different brain mechanisms.  All subjects were confronted with the same problem.  Therefore, these results can tell us something about those who make particular choices, but it does not tell us that certainty or uncertainty necessarily invokes certain brain mechanisms.  For that, you would have to expose subjects to different conditions where in one condition their results are certain and in another they have risky choices.  Experiments like this have been done, and have interesting results, but that is not this experiment.  This experiment might more accurately be said to be telling us that subjects who chose the certain result are not all engaging in the expected value/utility calcuations that economic thoery would suggest.  That is, unlike standard economic models where all individuals calculate all the expected values or expected utilities and then make a decision based on this, the subjects who chose certain gains largely just went with the certain result and didn't appear to be making these calculations (as the lower activation in the parietal might suggest).  This is consistent with the authors' theory of subjects optimizing a combination of the use of cogntive resources and affective reward, but it might also be consistent with other stories too.  I want to make it clear, it is good to have discussions like this in papers, so that we can have some kinds of context for the results, and it truth this was pretty good discussion.  It is just that other explanations are also quite plausible.

The other thought results from the discussion in the paper where the authors refer to particular emotions as occurring in the brain.  While their discussion seems intuitively plausible, there is no way to know, or even really test if these are the emotions occuring in the subjects.  Before we can do this, we need to have a much more rigorous definition of what we mean by the various emotions.  This I think should, and actually already is, a priority for affective neuroscience.

On the whole, this is a pretty interesting article, examining an interesting question, which I wish I had gotten to work on.

Comments

[p]

While this research is interesting, I'm not sure how the use of expensive technology changes the basic findings produced by psychologists decades ago. Why is it good science to duplicate work that others conducted more cheaply with expensive technology? I do not understand how economics is adding any value in the venture. Can you help me out

[Terrence Chorvat]

The use of scanners and other methods of investigation is telling something about the internal mechanisms used to make these decisions. If all we have is behavior, then we can make guesses about what caused it. By knowing what parts of the brain are active, and we know what types of decisions are made in those areas, we understand the causes of that behavior more fully. Furthermore, it helps us to believe that phenomena like loss aversion are not just due to particularities of the experimental environment, but losses are often actually processed differently than gains. I hope this helps.

[p]

Terrence,

fair point