Implicit Measure of Emotions: Distinguishing among Emotions of the Same Valence

Abstract

Investigators commonly assess emotions with explicit measures (e.g., the POMS; McNair, Lorr, & Doppleman, 1971) despite their shortcomings (e.g., transparency of research hypothesis, sensitivity to social desirability concerns). Although implicit measures of general positive and negative affect are available (Quirin, Kazen, & Kuhl, 2009), ones that discriminate among emotions of the same valence (e.g., sadness vs. anger) are lacking. This study is an initial step toward development of such a measure. We hypothesized that ratings of the emotional content of briefly presented visual stimuli would reflect current emotional state, in part via the feelings-as-information mechanisms (Schwarz, 2011), and thus index those states in an indirect manner. Participants heard a sadness- or relaxation-inducing story. Next, they saw 20 abstract paintings and rated whether each displayed anger, fear, happiness, sadness, or no emotion. We also administered the POMS. The two groups differed, as predicted, when inferring the levels of sadness and happiness portrayed in the paintings. Specifically, compared to the relaxed individuals, sad participants judged that paintings displayed less happiness and more sadness. Interestingly, the groups did not differ when inferring the levels of anger and fear portrayed. Thus the novel implicit measure thus was not only sensitive but also specific in detecting change in levels of sadness. The POMS did not uniquely detect the experimentally-induced shift in sadness.