The study of emotion in politics has been active, especially as it relates to the personality of political leaders and as an explanation for how people evaluate significant features around them.
Researchers have been divided into two groups—those who study leaders and those who study publics. The research programs have also been divided between those who use emotion to explain reliance on early experience that dominates contemporary judgment and those who use emotion to explain why people respond to the immediate contemporary circumstances around them. More recently, theory and research have attempted to reconcile these two seemingly contradictory roles by integrating them. Emotion’s role in politics is pervasive both because emotion enables past experience to be encoded with its evaluative history and because emotion enables contemporary circumstances to be quickly evaluated. More recently still, theoretical models and supporting evidence suggest that there are multiple channels of emotional evaluations.
1) The power of affective evaluations in predicting, for example, the vote (Kelley 1983) naturally led to considerable research on people’s emotional reactions to a wide range of political stimuli. Describing the characteristic emotional responses of political leaders has been one area of research on affect (Marcus 1988). The influence of attractiveness on candidate evaluation was the focus of Rosenberg’s research (Rosenberg et al 1986). Klein (1991) has found that overall feelings about presidential candidates display a negativity bias, such that a voter weights negative personality traits more heavily than positive ones to arrive at an overall feeling toward a candidate. Masters and Sullivan have shown that politicians characteristically provoke two independent emotional responses, one “hedonic” and one “agnostic” (McHugo et al 1985; Sullivan & Masters 1988; Masters & Sullivan 1989a, b, 1993;, Masters 1991;, Masters et al 1991).
2) This view is confirmed by work on the effects of anger (Lerner et al 1998). The inducement of anger yielded simpler cognitive processing, less attention to available information, and greater reliance on heuristics. However, these effects were reversed if subjects were told they would be held responsible for their views (i.e. if they anticipated having to justify their reactions). Thus, the intrusion of emotion, in and of itself, is not necessarily detrimental to the quality of decision making. Work on emotion and stereotypical thinking (Bodenhausen 1993, Bodenhausen et al 1994a, b) also suggests a more complex set of relationships, with different emotional states having different effects. Isbell & Wyer (2000) found that the effect of mood manipulation on subjects judging political candidates was counterassimilated when subjects were motivated and assimilated when they were not, indicating that the motivational status is highly relevant to the role of emotion in decision making. But although there are many studies on the role of emotion in politics, much more work will be necessary to achieve a sound theoretical and empirical understanding.
Recently, Green and Salovey (personal communication) have clarified that their original work (Green et al 1993) was meant only to claim that happy and sad mood terms reside on the same dimension and not to claim that only one bipolar dimension suffices to describe emotional response.
The measurement problem is not primarily a measurement theory problem but a theoretical underspecification problem. Moreover, the controversy has largely been confined to measures of self-report, ignoring the studies that use electromyographic measures of facial muscles (Cacioppo et al 1986) or techniques that allow mapping of brain activity (Tomarken et al 1990;, Tomarken et al 1992;, Wheeler et al 1993;, Davidson & Tomarken 1994;, Davidson 1993, 1995;, Robinson 1995;, Sutton & Davidson 1997). Still other measurement approaches use the startle reflex as a method of ascertaining emotional response (Bradley et al 1990;, Lang et al 1990, 1993;, McNeil et al 1993;, Lang 1994, 1995;, Ito et al 1998a).
All of these studies conclude that at least two dimensions are required to explain the variance in emotional response. Hence, a substantial array of results from a variety of methodologies points to the need for a two-channel model of emotion, a point on which Green and Salovey now agree with Russell, Cacioppo, Watson, Tellegen, and Marcus. Such controversy as remains turns on whether the description of emotion is better described by valence and arousal dimensions or by enthusiasm and anxiety dimensions (somewhat mislabeled as “positive” and “negative” psychology). Determining which of these two alternatives is more useful requires research that weighs the evidence for the substantive claims of the competing accounts.
As noted, valence accounts fail to adequately account for emotional response, but a considerable literature continues to thrive, in part driven by continued reliance on “feeling thermometers” and the like in various communal data-gathering programs. It is hoped that multiple-channel theories of emotion will guide future research.
Multiple-channel theories of emotion presume that affective reactions derive from multiple evaluative processes resulting in multiple affective dimensions. Although work in the 1950s seemed to establish that evaluation was global and formed a single dimension (Osgood et al 1957), in fact this conclusion derived from the reliance on paired oppositions, the semantic differential, which imposes a single-valence structure on the data. Once data gathering enables subjects to disaggregate their emotional responses, then it becomes clear that salient stimuli often evoke multiple, simultaneous emotional reactions (Lang et al 1993, Lang 1994, Ito et al 1998b).
Multiple-channel theories, such as those of psychologist John Cacioppo (Cacioppo & Bernston 1994;, Cacioppo et al 1997, 1999b), neuroscientists C Robert Cloninger (1986), Jeffrey Gray (1987a, b, 1990;, Gray & McNaughton 1996), Jaak Panksepp (1989, 1998), or political scientist George E Marcus (1991;, Marcus et al 1995, 2000), each argue that more than one evaluative process is ongoing and subserved by emotional processes at any given time.
There is evidence of multiple-channel responses to groups (see Hass et al 1992, Nelson 1998). More generally, if there are multiple channels of evaluation, then there should be evidence of multiple motivational consequences. Evidence supporting multiple-channel theories of evaluation can be shown by differential effects of each channel on learning and political judgment (Marcus & MacKuen 1993, Brader 1999). Psychology has been primarily focused on differential responsiveness to each channel, as in Cacioppo’s bivariate model (Cacioppo & Berntson 1994, Cacioppo et al 1997), whereas political science is likely to be more interested in the differential effects of each channel (Marcus et al 2000). Evidence of the asymmetric effects of the two channels in politics has been reported (Marcus & MacKuen 1993).
Although a full understanding of emotion is not yet realized, there has been a general shift from presumption of disruption and distortion to a more functional and less normatively biased view. Such a shift in normative orientation was recommended more than 50 years ago in psychology (Leeper 1948). That recommendation is something else political science could borrow from psychology