Before something can be measured it must be defined. Is there an adequate definition of emotion? Although emotion researchers do not fully agree on a single, consensual definition (cf. Ekman & Davidson, 1994), researchers have long debated its facets and components. A consideration of these components provides a working definition of emotion useful in planning measurement strategies. Moreover, a working definition is appropriate because it may be refined and extended over time as new findings about the nature of emotion emerge.
A useful working definition can be gleaned from a systems perspective on emotions (e.g., Bradley & Lang, 2000), which holds that emotion is a multi-faceted construct inferred from multiple indicators. The multiple facets of emotions are manifest in multiple response systems. According to Lang's bioinformational theory (1979), the systems of primary importance are the behavioral system, the language system, and the physiological system. Each of these systems, in turn, consists of multiple subsystems. For example, the language system contains evaluative self-reports of emotion as well as vocal and paravocal features of speech associated with emotions. The physiological system contains subsystems such as the central nervous system (brain), the peripheral nervous system (sympathetic, parasympathetic), and the hormonal system. Each of these subsystems can be thought of as a channel that carries potential information about the emotional state of the person.
To make measurement matters more complicated, the multiple channels themselves are not tightly connected to each other, thereby not offering well-calibrated or interchangeable indicators. To paraphrase Venables's (1984) statement about
Preparation of this manuscript was supported by grant R01-MH63732 from the National Institute of Mental Health.
arousal, the multiple response systems involved in the manifestation of emotion are themselves loosely and imperfectly coupled, plus the response systems complexly interact with, and mutually influence, each other. For example, facial expressions appear to amplify subjective feelings (e.g., Larsen, Kasima-tis, & Frey, 1992). Emotion generates loosely organized and temporally cascading changes across a wide array of psychological and physiological domains, including subjective experience, facial actions, central and peripheral nervous system activation, cognitive appraisals, information processing changes, and behavioral action tendencies.
In the optimal case, assessing emotions would involve measurement across many of these multiple components simultaneously. Streams of data obtained from multiple response systems may converge on the underlying construct of emotion and increase the relative confidence that we can place in the composite indicator. However, given the loosely coupled and complexly interacting nature of emotional response systems, researchers should not expect perfect or even substantial convergence across multiple indicators.
Examples of response discordance between different measures of the "same" emotion are not difficult to find. For example, Nesse et al. (1985) used nine different measures of distress obtained when subjects were in the presence of a phobic object and found only modest convergence across methods. Bradley and Lang (2000) concluded that covariation between measures of different response systems, supposedly indicating the same latent construct, "seldom accounts for more than 10-15% of the variance" (p. 244). Even when using multiple, synchronized measures, the underlying psychological construct of "emotion" still remains some inferential distance from (i.e., is only probabilistically related to) the composite of observable indicators.
The situation just described implies that the use of multimethod assessment, although not an absolute solution, is nevertheless acutely necessary in the emotion domain. Researchers should be particularly on guard against the false virtue of reduc-tionistic interpretations or "operational definitions" of emotion in monomethod terms. Emotion is not a self-report of emotion, nor is it a P300 response to stimulus, or a potentiated startle reflex, nor is it an appraisal of some event. Emotion must be conceived in a postpositivistic fashion, as a theoretical construct to be inferred from multiple observables. As a scientific term, emotion carries surplus meaning beyond any single measure and thus is best represented by a multimethod composite. Although this is generally good advice for any theoretical construct, it is especially appropriate for emotion because of the multiple channels involved in the creation and manifestation of emotion.
Was this article helpful?
This guide will help millions of people understand this condition so that they can take control of their lives and make informed decisions. The ebook covers information on a vast number of different types of neuropathy. In addition, it will be a useful resource for their families, caregivers, and health care providers.