By Paul Zak

Elisabeth Hill and David Sally of University College London have recently completed a very interesting paper using the neuroeconomic method (real social interactions with payoffs) examining cooperation and fairness in adults and children with autistic spectrum disorder or the less devastating Asperger syndrome.  The authors specifically are examining the role of "mentalizing" or "theory of mind" (the ability to interpret another's intentions) in strategic interactions. This is a nicely designed study using normal controls and several unrelated control tasks to determine subjects’ basal theory of mind and cognitive abilities.  

Three games were used, the prisoner's dilemma (PD), the ultimatum game (UG) and the Dictator game (D) [I'll presume readers know these strategic social interactions, but if not, PD and UG admit equilibria with either cooperation or defection, and D measures altruism through gift giving.] Normal children had trouble finding the best strategy for all games, while normal adults quickly found good optimal strategies (and experimented on the parameters of them to further optimize).  Autistic adults were about as cooperative as normal adults, and autistic children were similar to normal children.  All children were more altruistic than adults.

One critique of this study is the (typical for psychologists) lack of use of monetary rewards to motivate attention to task.  This study "paid" for performance with stickers for children and chocolate for adults.  While these things are presumed desirable, their value across subjects varies (e.g. some adults don't care for chocolate).  Cash is king here and has much clearly interpretable effects.  A second critique is the very discursive writing of the authors that make it difficult to read (it is downloadable at www.ssrn.com).

Otherwise, this is a very nicely designed and executed study that tells us about social interactions in which the economically rational behavior requires little mentalizing ability.  More sophisticated strategic interactions clearly do require this (e.g. see McCabe et al, "A functional imaging study of cooperation in two-person reciprocal exchange" Proc. Nat. Acad. of Sci., 2001). 

Bottom line: many social and economic interactions do not require deep cognitive abilities, but are fairly quickly intuited using, e.g. market signals.  This is good; it is why economies chug along with little intervention needed as market participants can figure out easily what is best for them (and need not have someone or group tell them what they "should" do).  This study gives us an insight as to why this occurs.