Wednesday, April 27, 2011

Lucy's Lesson on The Fundamental Attribution Error

The fundamental attribution error is pervasive. On yesterday's walk, our black standard poodle, Lucy, simply would not heel. I held a dog biscuit in my left hand, usually a surefire cure for bad behavior, but she insisted on roaming to the right, where my wife was walking. I quickly came up with an explanation, "She's always liked you better." Margaret agreed, "Yes, I spend more time with her and I feed her."

I ran out of dog biscuits, so I asked Margaret if she had any. She gave me her bag of treats, tiny bits of liver biscotti, Lucy's favorite. Everything changed. Now Lucy paid no attention whatsoever to Margaret. She heeled perfectly, her eyes riveted to my left hand, waiting, salivating. Lucy's behavior had nothing to do with her love for Margaret, and everything to do with who held the liver biscotti.

Why do we so readily attribute behavior to enduring traits and neglect the power of the situation? I suspect it's because over the past few hundred thousand years, information about what individuals are like has been enormously useful. People, and dogs, have personalities that stay remarkably consistent for decades. Attending closely to what people are like, coming up with theories about their traits, and gossiping about them has been important. Using theories about persons to explain behavior comes naturally, so naturally that we routinely underestimate the power of the situation.


Brandon Hidaka said...

One of things I enjoy most about evolutionary biology, is that the theory allows one to ask "why?" E.g. why do some people secrete more amylase in their saliva than others? Or, why do children get Reye Syndrome?
I find your explanation convincing. Perhaps throughout most of our ancestral past, it was adaptive to give less credence to attribution of environmental causation because it was so inaccurate: understanding of physical, psychological, and medical phenomena largely relied upon religious/supernatural explanations. I'm afraid that the Fundamental Attribution Error may directly contribute to scientific misunderstanding.

Wesley Goodman-Levy said...

Or perhaps as a scientist you have been trained (or innately do so, hence becoming a scientist) to look for longitudinal data and avoid self-confirmation biases?
Perhaps another person, who is not so aware of these behavioral tendencies, might easily spot the bag of treats instead of considering the walking factor.

Furthermore, your analysis may have been correct in the absence of the treat bag. Without the bag of treats in the situation, who's to say Lucy wouldn't have been conditioned to expect walk's from Margaret, which would cause her to obey her more readily as doing so would be more likely to result in walks, the "treat bag" substitute in this case.