Structures of Mood and Personality

Until recently, investigations of the structures of mood and personality have proceeded largely independently of one another. I first summarize these investigations separately and then discuss their integration over the past decade.

Mood Structure

Working in the 1960s to understand the sources of life satisfaction or subjective well-being, Bradburn (1969) developed affect scales to measure emotional wellbeing, and he discovered that the positive emotions and negative emotions formed separate scales rather than opposite ends of a single bipolar scale. Through the 1970s and into the early 1980s, a great deal of evidence accrued to support the notion that two broad dimensions of affect emerged regardless of whether the data analyzed were self-reported mood scales, judged similarities or semantic differential ratings of mood terms, or even facial or vocal expressions of emotion. Watson and Tellegen (1985) reviewed this literature and presented a consensus structure of the two dominant mood dimensions, which is presented in figure 9-2.

Negative affect (NA) is a general dimension of subjective distress, encompassing a number of specific negative emotional states, including fear, sadness, anger, guilt, contempt, and disgust. The low end of the NA dimension is defined by such terms as calm and relaxed. Despite their conceptual distinctiveness and a modest degree of empirical differentiability, it is well established that these various negative mood states substantially co-occur both within and across individuals (Watson & Clark, 1992a, b).

Positive affect (PA), by contrast, reflects the co-occurrence among a wide variety of positive mood states, including joy, interest, attentiveness, excitement, enthusiasm, and pride. The low end of this dimension is marked by such terms as sleepy or drowsy. It is important to stress that these dimensions are descriptively bipolar, but affectively unipolar. That is, only the high end of each dimension reflects an engaged, emotional state. The low ends describe disengaged states that still may be called moods, but not emotions per se.

Despite the positive-negative terminology, which linguistically implies opposite ends of a single continuum, these mood dimensions are essentially orthogonal. A few examples can help to illustrate. Imagine that you are in a roller coaster, climbing that first big hill, rounding the top, looking and then plummeting down. Alternatively, imagine watching a horror movie or an action-sus-pense thriller. In either case, as the critical moment approaches, most people feel excited, scared, thrilled, and tense. In terms of the mood structure shown in figure 9-2, these feelings are at the high end of both the positive (excited and thrilled) and negative (scared and tense) affect dimensions. The positive af-fect—the excitement and thrills—keeps the experiences from being ones of stark terror, whereas the negative affect is what differentiates these experiences from pure comedy or from the feeling that one may get listening to Beethoven's Ode to Joy. Clearly, strong positive and negative affects can co-occur.

Coming at the independence of these dimensions from another angle, imagine waiting for a dental appointment. Most people feel at least a little nervous tension, that is, a little negative affect. However, this by no means implies the lack of positive affect, that is, feeling sleepy and drowsy. And afterward, or if the appointment were suddenly to be canceled, a likely reaction would be relief (a low negative affect state) rather than enthusiasm and joy (high positive affects). Thus, changes in negative affect occur independently of changes in positive affect, and vice versa.

In addition to research supporting the psychometric independence of these two mood dimensions, they have been shown to have qualitatively different correlates in both large-scale interindividual analyses and intensive longitudinal intramdividual studies. Reviews of this literature are available (e.g., Clark & Watson, 1991a), but I illustrate here some of the major relations using my own research and that of my collaborator, David Watson. Watson (1988) reported on 80 undergraduates who completed a mood rating form daily for 6–8 weeks, along with measures of physical health, social activity, exercise, and perceived

stress. In between-subjects analyses, NA was substantially related to perceived stress and health complaints, and the NA-stress relation also was obtained intraindividually. In a subsequent literature review of mostly between-subjects studies, Watson and Pennebaker (1989) found both of these relations to be very robust. By contrast, PA was correlated with social activity and frequency of exercise (Watson, 1988). In two later studies involving a total of 208 undergraduates, the PA-social activity relationship was replicated both intra- and inter-individually using a more extensive measure of social activity (Watson, Clark, Mclntyre, & Hamaker, 1992).

In a 3-month diary study of 18 Japanese undergraduates, Watson, Clark, and Tellegen (1984) demonstrated that NA and PA mood dimensions comparable to those found in the United States emerged in Japanese students as well. They then examined intraindividual relations between these dimensions and daily events reported by their Japanese subjects (Clark & Watson, 1988). Negative affect was related to interpersonal conflicts, irritants such as getting rained on while waiting for a bus or losing one's eyeglasses, health problems, and various reported concerns (e.g., about relationships or health). By contrast, PA was again correlated with social activities, particularly going to parties or going out to eat or drink, attending entertaining events (e.g., movies, concerts), vacation-related traveling, and engaging in physical activity. Positive affect was notably low on days that involved interpersonal conflict or that were spent working or at home alone. It is important to note that the direction of causality is indeterminate in these studies. For example, PA could have been elevated by entertaining events or, alternatively, being in a positive mood might increase the likelihood of deciding to attend such an event. What is noteworthy is the consistent pattern of correlates across studies both within the United States and elsewhere: NA and PA are systematically related to different types of variables.

In addition to their differential correlates, a second striking finding is that the PA dimension alone has a strong diurnal rhythm. In two separate week-long studies, a total of 196 undergraduates completed mood ratings every 3 waking hours (Clark, Watson, & Leeka, 1989). figure 9-3 shows the strong pattern of daily PA variation that emerged. Positive affect rose sharply from early morning upon awakening (average rising time was 9:30 A.M., ranging from approximately 9 A.M. for “morning” people to IOA.M. for “night” people) until noon, and then fell again after 9 P.M. Approximately 85% of all subjects had a substantial rise and fall at these times of the day, attesting to the robustness of the phenomenon. Notably, the shape and timing of this curve roughly parallel the diurnal rhythm of body temperature, and studies have found a relation between PA-related variables and body temperature variation (Thayer, 1989). These data clearly suggest a biological basis for PA mood variation.