Strength of affective reaction as a signal to think carefully

Abstract

Analytic processes reduce biases, but it is not known how or when these processes will be deployed. Based on an affective signal hypothesis, relatively strong affective reactions were expected to result in increased analytic processing and reduced bias in judgement. The valence and strength of affective reactions were manipulated through varying outcomes in a game or evaluative conditioning of a stimulus. Relatively strong positive or negative affective reactions resulted in less desirability bias. Bias reduction only occurred if participants had time to deploy analytic processes and indicators of the degree of analytic processing (in the form of attentional control) predicted less bias. Affective processes have long been acknowledged as a source of bias, but these findings suggest affective processes are also integral to bias reduction.

Keywords:

• Affective reactions
• Dual process
• Bias reduction
• Optimistic bias
• Desirability bias

Thanks to Daniel Lench for technical assistance and to the Emotion and Motivation lab for data collection and entry, particularly Regina Herpin and Alexandra Sweeney.

Thanks to Daniel Lench for technical assistance and to the Emotion and Motivation lab for data collection and entry, particularly Regina Herpin and Alexandra Sweeney.

Notes

¹ Confusion persists about the meaning of terms that describe emotional processes, despite recent attempts at consensus (Eich, Kihlstrom, Bower, Forgas, & Niedenthal, 2000; Lench, Flores, & Bench, 2011). Affective reactions are elicited immediately upon perception of stimuli and involve only appraisal of goodness/badness and are separable from general affect (feeling good/bad unrelated to stimuli/event), and emotions (reactions to specific events).

² Unrealistic optimism effects may partially result from factors other than desirability. These include a tendency to focus only on the self (Chambers, Windschitl, & Suls 2003), and the underlying frequency of events (Harris & Hahn, 2011). These criticisms are limited to studies that ask participants to compare themselves to an average peer and are not relevant to studies that use other paradigms.

³ No participants expressed suspicion. In fact, in the positive high-importance conditions several participants continued to ask about credits after debriefing. To alleviate any emotional consequences, debriefing included mood-repair. The outcomes are similar to those often given (noise blasts in studies of punishment, rewards in studies of choice). Ethically, we regard the deception as necessary to manipulate affective reactions and as less harmful than experiencing the outcomes.

⁴ Sample is larger relative to other studies because we anticipated more variability in the attention measure.