Testing the Circumplex Model of Emotions in a Consumer Setting

There is widespread belief that more positive emotional reactions to consumption situations will lead to positive business outcomes such as increased market share through the combination of increased loyalty, repeat purchase and strengthened reputation. However, most of the psychological work on emotions has not dealt directly with consumption experiences, but rather broader everyday experiences. In this study, psychological models of emotion were tested using magazine subscribers, specifically looking at their emotional responses to the magazine and the overall subscription package. The aim was to determine whether one of the major theories on emotional structure, the circumplex model, is relevant and consistent when applied specifically to a consumption experience. The results are positive, with the model being supported, and they provide insight into the structure and relations of different emotional responses (e.g., satisfaction, delight) consumers might have to a consumption experience. Introduction Emotion research is vast and covers a wide range of topics (Izard, 1972, 1977; Nowlis, 1965; Plutchik, 1980; Russell, 1978, 1980; Schlosberg, 1952; Thayer, 1967), but the theories of emotion, which are at the core of this paper, are limited to the work of a small number of academics (Plutchik, 1980; Russell, 1980) who were primarily concerned with the structure of emotions and the succinct definition of emotion space. The objective of this research was not only to understand the broader interactions of emotions, but also to provide a basic framework for identifying individual emotions within that framework. The use and application of these theories was particularly useful for accurately defining constructs like delight and satisfaction. Despite changing approaches to conceptualising the structure of emotions and the development of different theories based on these ideas, there is little consensus on what the structure of emotions should be. Although there are many, the two approaches that compete most predominantly in the consumer behaviour literature (Oliver, Rust and Varki, 1997) are the basic categories structure (Darwin, [1872] 1965; Izard, 1977; Plutchik, 1980) and the dimensional structure (Larsen and Diener, 1992; Russell, 1980; Thayer, 1989; Watson and Tellegen, 1985). The basic categories approach was one of the earliest to be developed and was applied by Darwin ([1872] 1965), Izard (1977) and Plutchik (1980). This theory allocates each emotion into a small set of mutually exclusive categories on some assessable basis. According to Plutchik, the eight primary emotions are joy, acceptance, surprise, fear, sorrow, disgust, anticipation and anger. To best understand the relationship of these emotions to all others, Plutchik arranges them in a circular pattern, which he likened to a colour wheel. The interpretation is that categories which are located next to one another are most similar and categories that are located on opposite sides of the wheel are most different. Despite the orientation, this depiction should not be confused with a circumplex model. Plutchik’s (1980) approach to defining the structure of emotions has, however, been subject to criticism from various authors (Ortony and Turner, 1990; Richins, 1997). The key criticism, that Plutchik’s (1980) model of emotion would not easily facilitate the measurement of nonprimary emotions, such as satisfaction or delight, meant it could not be adopted here. The second approach to defining emotion structure is the dimensional approach, which says all emotions can be identified as displaying association with one or more dimensions, such as intensity, pleasantness or degree of activation. Most often, the two broad dimensions of pleasantness and intensity (or arousal) have been identified using factor analysis and multidimensional scaling (MDS) techniques. Rather than emotion terms clustering at the axes, as is typical with these statistical methods, they tend to form a circular pattern around the dimensions hence being referred to as a circumplex. This approach to definition was pioneered and most widely used by Russell (1979, 1980), who now believes that the circumplex model is better suited to understanding core affect than prototypical emotional episodes (Russell and Feldman Barrett, 1999). Figure 1 provides an example of Russell’s (1980) circumplex structure, by placing eight emotion terms, indicative of each octant of the circumplex, around the two primary dimensions of pleasantness–unpleasantness (horizontal axis) and arousal–sleep (vertical axis). In his 1980 study, Russell used a list of “28 words or phrases that people use to describe their moods, feelings, temporary states, affect or emotions ... [as a] sample ... chosen to represent the domain of affect” (p. 1164). In that study, these 28 emotion terms were analysed using five different scaling techniques to illustrate that the structure of emotion was roughly circular (represented by a circumplex) and its’ properties held constant, regardless of which statistical technique was used. Figure 1: Circumplex structure of emotions Russell (1980, p. 1164) Misery Distress Excitement Depression Contentment Pleasure