Working Paper
Racial bias and the validity of the Implicit Association Test

Implicit associations and biases are carried without awareness of conscious direction. In this paper, I develop a model to study giving behaviours under conditions of implicit bias. I test this model by implementing a novel laboratory experiment—a Dictator Game with sorting to study both these giving behaviours, as well as a subject’s willingness to be exposed to a giving environment.

In doing so, I adapt the Implicit Association Test (IAT), commonplace in other social sciences, for use in economics experiments. I then compare IAT score to dictator giving and sorting as a necessary test of its validity. I find that the presence of sorting environments identify a reluctance to share and negatively predict giving. However, despite the IAT’s ever-growing popularity, it fails to predict even simple economic behaviours such as dictator giving. These results are indicative that implicit bias fails to overcome selfish interests and thus the IAT lacks external validity.