Another thorny issue researchers need to consider before making measurement decisions concerns whether emotions are discrete and/or dimensional constructs. The issue concerns whether emotions can best be conceptualized as two or three broad dimensions (e.g., Larsen & Diener, 1992) or as two (Meehl, 1975; Mower, 1960) or four (Gray, 1982) or five (Oatley & Johnson-Laird, 1987) or six (Frijda, 1993) or eight (Plutchik, 1980; Tomkins, 1984) or nine (Ekman, 1992; Izard, 1977) separate and distinct categories of experience. This issue goes back more than 100 years (Darwin, 1872/1965; Spencer, 1890; Wundt, 1897) and continues to be debated (cf. Diener, 1999). The most consensually agreed-upon dimensional view of emotion is represented by the circumplex model (Russell, 1980; Watson & Tellegen, 1985), which posits that emotions conform to a circular or radex arrangement with the coordinates of the circular space representing combinations of valence and arousal. The measurement implications of the circumplex model, including the need for multimethod measurement of emotion dimensions, are discussed in Larsen and Diener (1992) as well as Larsen and Fredrickson (1999).
Proponents of the discrete emotion viewpoint argue that the dimensional view often blurs meaningful distinctions between emotions (e.g., fear and anger and disgust are all similarly high-arousal unpleasant emotions). The main measurement issue implicit within this exchange of ideas concerns specificity: Measures that fit the discrete emotion view can be transformed to a dimensional arrange ment post hoc, but the converse is not always possible. For this reason, researchers should consider a priori whether distinctions between specific negative or between specific positive emotions are likely to be part of the theoretical or empirical agenda. If there is the slightest possibility for this need to discriminate basic emotions, then researchers should pursue emotion specificity rather than dimensional assessment.
Was this article helpful?