Automatic and Controlled Components of Implicit Prejudice

University Of California-Davis, Davis CA

Investigators

Abstract

People may be unaware of their own important underlying attitudes and beliefs, particularly in the domains of stereotyping and prejudice. Nevertheless, these "implicit biases" significantly influence inter-group behavior in everyday social interactions. Indeed, the influence of these biases often is greater than that of explicitly stated inter-group attitudes and beliefs, affecting important outcomes for members of minority groups. As such, it is critical to understand the nature and operation of implicit bias, as well as the factors that may change them or reduce their influence. To achieve these goals, Dr. Jeffrey Sherman of the University of California - Davis uses mathematical modeling with a tool called the Quad model that he developed in his past research. Application of this tool has shown that implicit attitudes and beliefs sometimes may reflect the automatic activation of biased associations in memory or may reflect failures to regulate the influences of such associations on behavior. The Quad model can be used to estimate the independent roles of each of these factors in producing biased and unbiased behavior. There are four specific goals in this research. First, the Quad model will be applied to shed light on the contextual and individual factors that influence the extent of implicit bias. Implicit bias has been shown to vary significantly across social contexts and across individuals, often reflecting personal goals and motives. This research will model contextual and individual differences in implicit bias in order to better understand the factors that increase or decrease the automatic activation of biased associations and the ability to regulate the expression of those associations. The second goal is to apply a "treatment" approach to reducing implicit bias. In these studies, interventions that are designed specifically to influence either underlying associations or the ability to regulate those associations will be applied to individuals and contexts associated with enhanced activation of associations or diminished ability to self-regulate. The goal is to show how interventions can be tailored to address specific deficits in processing (over-activation of associations vs. failure of regulation) associated with increases in implicit bias. The third goal is to use the Quad model to better understand the relationships among different measures of stereotyping and prejudice, which frequently fail to correspond to one another. The present research proposes that dissociations among these different measures may reflect differences in the extents to which the measures reflect the automatic activation of associations versus the failure to regulate the associations. Application of the model will help to specify when and why different measures will and will not produce corresponding results. Finally, the fourth goal is to use the Quad model to better predict behavior in inter-group settings. Measures of stereotyping and prejudice are sometimes poor predictors of people's actual behavior towards members of minority groups. The Quad model can improve the ability of these measures to predict behavior by independently assessing the roles of automatic associations, the ability to regulate those associations, and interactions between these components. In sum, the purpose of this research is to improve the measurement of stereotyping and prejudice, increase understanding of the factors that increase or decrease such biases, and improve the ability of measures of bias to predict people's behavior. Implicit attitudes influence behavior in important domains of life including law enforcement, health, and employment. It is therefore critical to gain a better understanding of their nature and operation.