Paper 3

Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=hbem20

Journal of Broadcasting & Electronic Media

ISSN: 0883-8151 (Print) 1550-6878 (Online) Journal homepage: https://www.tandfonline.com/loi/hbem20

Gender Differences in Selective Media Use for Mood Management and Mood Adjustment

Silvia Knobloch-Westerwick Ph.D.

To cite this article: Silvia Knobloch-Westerwick Ph.D. (2007) Gender Differences in Selective Media Use for Mood Management and Mood Adjustment, Journal of Broadcasting & Electronic Media, 51:1, 73-92, DOI: 10.1080/08838150701308069

To link to this article: https://doi.org/10.1080/08838150701308069

Published online: 05 Dec 2007.

Submit your article to this journal

Article views: 1328

View related articles

Citing articles: 16 View citing articles

Gender Differences in Selective Media Use for Mood Management and Mood

Adjustment

Silvia Knobloch-Westerwick

Mood management theory has found empirical support but was chal- lenged by gender-typed selections and exposure to negative content. These challenges are addressed with response style theory and the mood adjustment approach. A secondary data analysis and original ex- perimental data serve to test hypotheses. As expected, after a mood-impacting experience, men tend to distract themselves with ab- sorbing messages, whereas women tend to ruminate the experience and thus prefer messages with low absorption potential. When antici- pating a mood-impacting activity, men tend to distract themselves right before it by selecting absorbing content, whereas women focus on it and prefer less absorbing messages.

Media consumers’ moods play an important role in selections of media messages, as ample empirical evidence has shown. Based on mood management theory (Zillmann, 1988), many investigators have found affective states to influence which media content individuals attend to and which they avoid (for overviews, see Knobloch-Westerwick, 2006; Oliver, 2003; Zillmann, 2000). Experiments, quasi-experiments, field studies, and surveys have demonstrated in the United States and abroad that selections of electronic and broadcasting media such as TV entertain- ment genres, popular music, and Internet content in part result from the current feel- ing state of media consumers (e.g., Bryant & Zillmann, 1984; Knobloch, 2002; Knobloch & Zillmann, 2002; Zillmann, Hezel, & Medoff, 1980). Mood management considerations have been supported by amassed evidence but were also challenged by some observations. More specifically, both gender differences in hedonically moti- vated media selections and the exposure to upsetting, negative content cannot be ex- plained with mood management theory. This study aims to address these challenges by building (a) on response style theory (Nolen-Hoeksema, 1987, 1990), which con- ceptualizes gender differences in responses to dysphoric affects, and (b) on the mood adjustment approach (Knobloch, 2003), which offers explanations on why media us- ers might sometimes be drawn to upsetting content. Hypotheses derived from these

© 2007 Broadcast Education Association Journal of Broadcasting & Electronic Media 51(1), 2007, pp. 73–92

73

Journal of Broadcasting & Electronic Media/March 2007

Silvia Knobloch-Westerwick (Ph.D., Institute for Journalism & Communication Research, University of Music & Drama, Hanover, Germany) is a faculty member in the School of Communication, The Ohio State Univer- sity. Her research addresses media effects and message selection in entertainment media and news.

theoretical grounds will be tested with two data sets—a secondary data analysis of a mood management experiment and original data from a new mood adjustment study. The empirical investigation looks at selective music listening, although the same pat- terns are likely to apply to selections of other electronic and broadcasting media as well.

Mood Management Theory

Mood management theory (Zillmann, 1988) conceptualizes selections of media messages as motivated by affect optimization goals. Originally called the theory of af- fect-dependent stimulus arrangement (Zillmann & Bryant, 1985), its theoretical claims pertain to enhancement of both emotions and moods (see Zillmann, 2003, for the differentiation), although it became better known as mood management theory. This hedonistic objective is served by arousal regulation via media consumption to avoid boredom and stress, exposure to positively valenced content, and avoidance of messages that are associated with sources of negative affects. Thus, in states of stress, calming messages are preferred over stimulating messages to obtain agreeable arousal levels. On the other hand, bored individuals favor arousing messages accord- ing to the theory. Generally, messages with a tone that is more positive than the cur- rent affective state will be sought out, whereas any content with connections to ori- gins of disagreeable feelings will be avoided.

As Zillmann (2000) noted, gender differences have emerged repeatedly in mood management investigations. In these cases, men failed to comply with mood manage- ment predictions, whereas women selected messages in line with the theory (see Biswas, Riffe, & Zillmann, 1994; Masters, Ford, & Arend, 1983; Medoff, 1982). For in- stance, Anderson, Collins, Schmitt, and Jacobvitz (1996) found in a field study that men and women differed in their TV choices when under stress: Stressed women watched more game shows and variety programs, whereas stressed men preferred vi- olent action programs.

In light of psychological research on responses to one’s own affects, gender-split patterns of media-based mood regulation are not surprising at all. In fact, gender has been referred to as the most important interindividual characteristic when it comes to mood regulation (Thayer, Newman, & McClain, 1994). Although the genders do not differ in terms of emotional experiences (e.g., Johnson & Schulman, 1988), ways in which men and women cope with stress (Tamres, Janicki, & Helgeson, 2002) or try to change bad moods (Thayer et al., 1994) have clearly been shown to diverge.

Response Style Theory

As Nolen-Hoeksema (1987, 1990) postulated in her response style theory, the over- arching principle of these gender differences in affect regulation is that men tend to seek distraction to overcome a bad mood, whereas women tend to ruminate on bad

74 Journal of Broadcasting & Electronic Media/March 2007

moods. Ample evidence has corroborated this postulation (Butler & Nolen- Hoeksema, 1994; Nolen-Hoeksema, Morrow, & Fredrickson, 1993; Nolen- Hoeksema, Parker, & Larson, 1994; Thayer et al., 1994). Rumination “is defined as thoughts and behaviors that focus the individual’s attention on the negative mood, the causes and consequences of this mood, and self-evaluations related to the mood” (Rusting & Nolen-Hoeksema, 1998, p. 790). Rumination, though, does not appear to be an effective mood-repair strategy because it leads to prolongation of the negative feeling state in both men and women, as experiments and field studies have shown (Ingram, 1990; Lyubomirsky & Nolen-Hoeksema, 1993, 1994; Morrow & Nolen- Hoeksema, 1990; Nolen-Hoeksema & Morrow, 1991, 1993; Nolen-Hoeksema et al., 1993; Nolen-Hoeksema et al., 1994). These studies also showed that distraction, on the other hand, is an effective mood-enhancement strategy for both men and women in negative moods. “Distraction involves focusing attention away from the mood and its causes onto pleasant or neutral stimuli that are engaging enough to prevent the mind from wandering back to the source of negative affect” (Rusting & Nolen- Hoeksema, 1998, p. 790). The fact that women are more likely to ruminate when in dysphoric states could explain why they are twice as likely as men to be depressed (Ingram, Cruet, Johnson, & Wisnicki, 1988; Nolen-Hoeksema, Larson, & Grayson, 1999; Nolen-Hoeksema & Morrow, 1991; Nolen-Hoeksema et al., 1994; Wood, Saltzberg, Neale, Stone, & Rachmiel, 1990). However, anger appears to be the excep- tion to the rule that women ruminate and men distract themselves from negative af- fects. As Rusting and Nolen-Hoeksema (1998) showed, women tend to avoid feelings of anger with the aid of distraction, instead of focusing on the anger. These differences can be explained with gender-specific emotion socialization and emotion-related gender stereotypes—expressing anger is socially acceptable for men but not for women, whereas the display of sadness, fear, and most other emotions is more ac- ceptable for women (see Brody & Hall, 1993; Fischer, 2004, for reviews).

Mood Management Research in Light of Response Style Theory

In light of the response style theory, it is surprising that mood management investi- gations have not encountered gender differences throughout in selective exposure to media stimuli. Although the key assumptions of mood management, as already cited, do not explicitly refer to rumination and distraction, Zillmann (1988) certainly en- dorsed these approaches as mood-impacting strategies while explaining the absorp- tion potential as a stimulus characteristic with relevance for mood management: “Per- sons seeking to terminate their moods would do well to expose themselves to highly absorbing messages; persons who seek to maintain their states, in contrast, should minimize distractions and consume minimally absorbing fare—or better yet, nothing at all” (p. 331).

Knobloch-Westerwick/GENDER AND MOOD ADJUSTMENT 75

It is furthermore surprising that, in the case of gender differences, men instead of women tended to choose content that was likely to sustain their negative affects. Ac- cording to response style theory, women should be avoiding distracting stimuli, whereas men should seek them out. The exception, again, should be anger, where these patterns should be reversed. Yet only some mood management investigations manipulated affects for hedonic differentiation, whereas others were interested in arousal management and thus induced states of boredom versus stress (e.g., Bryant & Zillmann, 1984). As response style theory does not lend itself to making predictions on arousal regulation, it is not puzzling from this perspective that mood management research did not reveal gender differences in selective exposure to calming and stimu- lating messages for bored versus stressed media users. For the studies that did exam- ine effects of moods with different hedonic tones, the absorption potential (Zillmann, 1988) of media selections was not always clearly differentiated because the hedonic valence of the material was more of interest.

Zillmann et al. (1980), for example, placed participants in bad, neutral, and good moods by ostensibly testing their social skills and providing predetermined feedback that served as mood induction. Then, in a purportedly second study, respondents were free to sample from sitcom, action drama, or game show programs. The absorp- tion potential of these choices may have been similar or at least ambiguous, thus it is plausible that no gender differences were reported. Likewise, choices of innocuous and hostile comedy in a study by Medoff (1982) may have provided similar absorp- tion levels, but Medoff found that especially frustrated men favored hostile comedy, whereas provoked men abstained from media consumption altogether, probably to ruminate their anger. On the other hand, frustrated and provoked women apparently tried to dissipate their anger by watching positively valenced comedy. It seems that, in line with Rusting and Nolen-Hoeksema’s (1998) findings reported earlier, women dis- tracted themselves from anger by using material without “semantic affinity” (Zillmann, 1988, p. 332) to the source of their anger, whereas men ruminated their frustration and anger while watching material with high semantic affinity or even nothing at all.

A study by Biswas et al. (1994) used the same mood induction procedure as Zillmann et al. (1980) and had respondents then select from good and bad news, for which equal ratings for “interesting” had been established in a pretest. Thus the choices were probably equally distracting. Biswas et al. (1994) found that women in bad moods preferred good news, whereas men in the same mood state failed to do so. It is possible that the participants in this study felt angry and provoked by the experi- menter’s mood induction, which might explain why men could have been motivated to ruminate negative, angry moods for a possible later retaliation (Knobloch- Westerwick & Alter, 2006; O’Neal & Taylor, 1989), whereas women aimed to down- play their anger (in line with findings from Rusting & Nolen-Hoeksema, 1998, re- ported earlier). However, in this situation they could do so only by selecting messages with negative or positive valence. Choosing more or less absorbing messages was not an option because, as mentioned, the reports had equal levels of “interesting” ratings.

76 Journal of Broadcasting & Electronic Media/March 2007

A mood management study with World Wide Web exposure (Knobloch, 2002) used a computerized form of the same mood induction as Zillmann et al. (1980) and Biswas et al. (1994), who had employed an experimenter. Thus anger toward a pro- voking individual is very unlikely in this situation. Yet the Web pages provided for se- lection had also been pretested to ensure equal ratings of “interesting” while differen- tiating the hedonic tone. In light of response style theory, again, it stands to reason that no gender differences were found because provided media selections offered about the same absorption potential.

However, the lack of gender differences in the results from Knobloch and Zillmann (2002) are puzzling when considering the research design from the perspective of re- sponse style theory. Again, the same mood induction from Zillmann et al. (1980) was applied, although also in a computerized format where anger responses seem un- likely. In a purportedly different study, participants were free to choose from a set of pop songs. The songs, offered via a computerized jukebox, had been pretested for lev- els of joyfulness and energy of musical expression—yet evaluations of these proper- ties were correlated in the pretest of the top-charts musical selections. However, the four songs with high pretest scores in energy and joyfulness certainly had a higher ab- sorption potential than the four more slow-paced songs with low scores on these di- mensions. Participants who had been placed in a bad mood spent more time on ab- sorbing music than those in a mediocre mood, who, in turn, also dedicated more time to uplifting music than the respondents in the good-mood condition. However, no gender differences or interactions with gender emerged, although women in a nega- tive mood should dedicate less time on the distracting, absorbing music compared to men in the same feeling state. Given women’s tendency to ruminate more than men, according to response style theory, an interaction between gender and mood state should be expected but was not found.

Taking a Process Perspective

A possible explanation why this interaction did not materialize can be drawn from a psychological study by Trask and Sigmon (1999). These authors argued that individ- uals in a bad mood are not likely to engage in just one strategy of mood regulation and that combinations should be accounted for. Earlier research had only studied effects of one particular strategy, represented in a distraction task or a rumination task. In Trask and Sigmon’s (1999) investigation, participants were placed in depressed moods and then asked to engage in two tasks. Depending on experimental group, participants performed two distraction tasks (Group A), a distraction task and then a rumination task (Group B), two rumination tasks (Group C), or a rumination task and then a distraction task (Group D). Results showed that Groups A and B reported lower levels of depression than after the mood induction in measurements after the first and the second task. Furthermore, Group D indicated a depressed mood after the first task but a better mood after the second task. Finally, Group C remained depressed

Knobloch-Westerwick/GENDER AND MOOD ADJUSTMENT 77

throughout both tasks. Thus immediate distraction leads to better mood, even if followed by rumination, whereas bad moods tend to sustain as long as the individual ruminates.

If Trask and Sigmon (1999) were correct in assuming that individuals do not only apply one mood regulation strategy, then accumulated selective exposure measures in mood management investigations might veil that participants engage in different behaviors across time. Possibly, men seek to distract themselves first to overcome a depressed state, whereas women might initially ruminate before finally seeking dis- traction. Such a pattern would comply with response style theory but can also explain why accumulated selective exposure patterns do not always reflect gender differ- ences in mood-regulation styles.

Mood Adjustment Approach

A study by Knobloch (2003) showed that different mood-regulating strategies via media use can indeed be pursued within a relatively brief period. This investigation examined the mood adjustment approach (Knobloch, 2003), which suggests that often media users do not only aim to optimize their mood, as original mood man- agement theory had postulated. Instead, media users frequently anticipate upcom- ing situations and activities that may call for specific moods. To regulate moods ac- cordingly, individuals may employ media stimuli. In the empirical study, participants performed an initial task and were then free to sample from pop music during an ostensible waiting period of 7 minutes. The findings indicated that partic- ipants allocated the first part of the listening period as mood management would predict, apparently pursuing mood optimization goals. However, as additional ac- tivities approached toward the end of the listening period, choices reflected mood adjustment purposes. Hence, coping with an anticipated task sets in right before this task.

The mood adjustment approach was further investigated with a specific interest for gender differences by Knobloch-Westerwick and Alter (2006). It was hypothe- sized that men and women would employ news choices differently while anticipat- ing a social encounter. This study investigated how provoked men and women ad- just their moods during media use, either knowing or not knowing that they get a chance to retaliate against the provoker. Participants received only negative feed- back about a social skills test from a supervisor. Half of the respondents were led to believe that they would get to evaluate the supervisor (retaliation opportunity) after they had examined a new online newsmagazine. The included news actually fea- tured both bad and good news, which had received equal ratings for “interesting” in a pretest and thus had about the same absorption potential. Gender differences in selective exposure to good and bad news emerged only for the respondents who anticipated the retaliation opportunity. As the end of the news-browsing period ap- proached, women who anticipated a retaliation opportunity looked at more good

78 Journal of Broadcasting & Electronic Media/March 2007

news, probably to dissipate their anger. At the same time, men who anticipated a retaliation opportunity decreased exposure to good news in favor of bad news, ap- parently to sustain their anger to get back at the provoker. Hence, like in the find- ings from Knobloch (2003), adjustment to an upcoming task or encounter occurred pretty much right before this activity and not throughout the whole media use pe- riod. Interestingly, women generally spent significantly more time reading the news, whereas men allocated more time to the overview page without actual news reading. This might also indicate that women tried to rid themselves of anger by distraction via news reading.

Furthermore, it is important to note that mood intensity showed no impact on selec- tive exposure patterns in this study (Knobloch-Westerwick & Alter, 2006), in contrast to the initial mood optimization patterns in the investigation by Knobloch (2003) de- scribed earlier. Most likely, the situations in the two studies differed in terms of rele- vance for self-perceptions. One employed a comparatively innocuous task of match- ing symbols (Knobloch, 2003)—the fact that participants engaged in some mood optimization behavior is probably due to low importance of such an assignment for self-perceptions. In contrast, there is apparently no room for mood optimization when the situation is highly relevant for self-perceptions, as it was in the study that involved an alleged test of social skills and related social behaviors (Knobloch-Westerwick & Alter, 2006).

Hypotheses

Taking these considerations and findings together, the following hypotheses can be proposed. Individuals’ media selections should reflect mood-enhancement strategies that depend on media users’ gender. It is expected that gender-specific mood regula- tion strategies will occur at different points in time, depending on whether the source of a mood has passed or is anticipated to become relevant. In the first case, men should engage in distraction and women in rumination right away. As mood manage- ment studies have often shown that women actually do employ media use to enhance their moods (e.g., Helregel & Weaver, 1989; Meadowcroft & Zillmann, 1987), it is likely that they also turn to distraction eventually. Yet when a stressor is anticipated, a reversed time pattern should occur, with men engaging in distraction and women in rumination as the stressing incident approaches. Finally, as mood enhancement is particularly important in negative affective states, these patterns could be more dis- tinct when individuals are experiencing negative feelings. On the other hand, de- pending on the relevance of the upcoming task for the self, as discussed earlier, the in- dividual might focus on mood adjustment and pursue no mood enhancement per se. Thus the investigation will differentiate affective states by hedonic valence—bad, neutral, and good—in comparative terms, as they are induced by an experimental treatment and defined by respondents’ assessments, to explore a research question about the importance of mood valence.

Knobloch-Westerwick/GENDER AND MOOD ADJUSTMENT 79

H1: After a mood-impacting experience, males initially prefer media content with high absorption potential and then decrease exposure to absorbing messages, whereas females initially prefer media content with low absorption potential and then in- crease exposure to absorbing messages.

H2: After a mood-impacting experience and while anticipating further related activi- ties, males prefer media content with high absorption potential as the activity ap- proaches, whereas females prefer media content with low absorption potential as the activity approaches.

RQ1: Are the patterns postulated in H1 and H2 more pronounced for mood states with negative valence?

Method

Overview

This investigation employs data from a mood management study presented by Knobloch and Zillmann (2002) to address H1 and RQ1 and furthermore analyzes new data from a mood adjustment investigation for H2 and RQ1. Hypotheses were further- more tested in a 3 × 2 × 2 design with repeated measures, with mood manipulation (bad, neutral, and good moods induced through feedback of different alleged perfor- mance levels), task anticipation (none vs. anticipated), and study (mood management study vs. mood adjustment study) as between factors and selective exposure to ab- sorbing music for different listening intervals as the within factor.

In the mood management study, 116 U.S. respondents participated in a computer- ized research session that was ostensibly composed of two parts. During the first part, participants performed an emotion recognition test that actually served to place them in a bad, neutral, or good mood. In the second part, participants were free to choose from top-chart pop songs provided via computer to listen to whatever they liked dur- ing a 10-minute period. During the listening period, selective exposure to each song was unobtrusively logged, unbeknownst to participants. Finally, items on the current mood states and evaluation questions on the music were presented for closure before the debriefing.

The pop songs available for immediate listening had been categorized in a pretest into four songs that were high on the characteristics “energy” and “joyfulness” as mu- sic properties and four songs that were low on these dimensions. The two groups of songs offered clearly contrasting absorption levels, as they differed primarily on the energy dimension. An analysis of pretest ratings of “energy” yielded M = 1.4 for low and M = 7.2 for high on a scale from 0 to 10, F(1, 52) = 448.5, p < .001, and the dis- similarity for “joyfulness” was much smaller (M = 5.5 for low and M = 6.9 for high), F(1, 52) = 13.74, p = .001. Thus the contrast was 5.8 scale points for energy and only 1.4 for joyfulness.1 The two sets of songs also differed objectively in a comparison of beats per minute (M = 72, range = 63–86, for slow songs; M = 138, range = 116–168, for fast songs). Several studies have shown that fast, energetic music hinders concen- tration, whereas slow, soothing music helps listeners to focus (e.g., Borling, 1981;

80 Journal of Broadcasting & Electronic Media/March 2007

Smith & Morris, 1977; Wakshlag, Reitz, & Zillmann, 1982); furthermore, fast music is perceived to be more distracting than slow music (Mayfield & Moss, 1989). Hence, four songs featured high absorption potential and four offered only little absorption.

The procedure employed for the mood adjustment study with 79 German partici- pants was almost identical to the first one—an ostensible “emotion recognition test” served as mood induction before participants could sample from the same precategorized pop songs, featuring low versus high absorption potential. However, the procedure used here included an announcement before the music-listening pe- riod that tasks “similar to the one just performed” would follow. Hence, participants were led to form an expectation about upcoming activities. Furthermore, as the re- searcher no longer had access to the participant population employed in first experi- ment, this data collection was conducted with German students and with all instruc- tions in German. Earlier research (Knobloch, 2002) that used the same mood induction demonstrated that Germans’ media message selections were fully in line with mood management theory predictions; thus no major differences between sam- ples are to be expected. It is argued that the personal use of popular music, when bro- ken down into simple categories of songs with low versus high absorption potential, for mood-regulating purposes is unlikely to differ by country. Young people in Ger- many, similar to other countries (Winter, 1985), are typically exposed to a vast amount of popular songs originating in the United States (Deutsches Musikinformationszentrum, 2005) and are certainly used to listening to them on a daily basis.

Respondents

The sample of the mood management study consisted of 116 undergraduate stu- dents from a large southeastern university in the United States, recruited from intro- ductory communication classes for extra credit. Participants were evenly assigned to experimental conditions (ns = 39, 37, and 40) with even gender proportions in these groups (68% women in total, age M = 20.2, SD = 1.2). An additional 53 respondents drawn from the same population had served for a pretest of musical stimuli.

The sample of the mood adjustment study included 79 student respondents re- cruited in a central building on a large German university campus. They received a small financial compensation for participation. The average age was 21.2 years (SD = 2.2) and 54% were women. Participants were randomly assigned to experimental conditions (ns = 27, 26, and 26) while balancing for gender proportions.

Procedures

The procedures of the two studies were almost identical and are explained in the following. The instructions for the mood adjustment study were all German transla- tions from the first investigation. Any other difference is indicated in square brackets.

Knobloch-Westerwick/GENDER AND MOOD ADJUSTMENT 81

General Setting. The data collection was conducted in a computer laboratory with 17 [40] personal computers; up to 10 respondents were tested in one session. Respondents were informed about the basic research objective and procedure. Ini- tially, the computer screen only showed the text “Please wait for further instructions” along with a Start button. Participants initiated the procedure by clicking the button. All further instructions and presentation of stimuli were fully automated via software. Nonetheless, the experimenter was present to answer possible questions and to pre- vent interaction among respondents.

Mood Manipulation. The treatment was a computerized version of the procedure used by Zillmann et al. (1980) and Biswas et al. (1994), which is known to induce a range of different moods effectively. Participants were instructed that the task would be a test of their sensitivity to facial cues, which was said to be a crucial skill in suc- cessful human interaction. Twenty portrait photos with highly ambiguous facial ex- pressions were presented, each along with boxes for the six basic emotions of anger, disgust, fear, happiness, sadness, and surprise. A large bar graph continually updated the percentage of supposedly correct identifications of emotional expressions, fluctu- ating through zones from 0–25%, 50%, 75%, and 100%. Performance in the first quarter was labeled terrible, the second poor, the third good, and the uppermost quar- ter allegedly indicated excellent performance. Moreover, after each response for an image, the feedback Right or Wrong was displayed.

The performance feedback was randomly predetermined according to experimen- tal group. Correctness of responses was not obvious to participants, thus they were led to believe that they had performed poorly or excellently. Accordingly, they could be placed into a good mood by feedback on an allegedly favorable performance or into a bad mood by a discouraging feedback. This treatment was accomplished by making them believe that 85% (excellent), 55% (still in the good range), or only 15% (terrible) of their classifications were correct.

Anticipation Manipulation. The procedure of the mood adjustment study in- cluded an announcement before the music listening period that tasks “similar to the one just performed” would follow. Hence, participants were led to form an expecta- tion about upcoming activities. The music listening period was explained with indi- vidual ergonomic data that had been collected during the first test task and that would now undergo complex computations at the university’s computing center; a small blue bar indicated the ostensible progress of the “adjustment computation” during the music listening period. No further assignments were actually presented.2

Collection of Dependent Measures. After the mood manipulation, participants were instructed that they were free to select from music for immediate listening in the second and ostensibly unrelated session part. They were asked to put on headphones and to sample from pop songs for the next 10 [7] minutes. During the scheduled

82 Journal of Broadcasting & Electronic Media/March 2007

music listening period, the computer screens showed eight buttons, with song titles and names of performers, in randomized order. Once respondents clicked on a box, the song was played via the headphones. Unbeknownst to participants, their selective exposure to music was logged by the software. The eight songs available for listening had been taken from the top-charts listing and categorized in a pretest to be either low or high in energy and joyfulness, thus featured either low or high absorption potential (e.g., “I Knew I Loved You” by Savage Garden scored low and “Then the Morning Comes” by Smash Mouth scored high in energy and joyfulness). Further pretest details are reported by Knobloch and Zillmann (2002). The total length of all eight presented songs was 28.5 minutes. Thus the scheduled time allowed participants to listen to a third [fourth] of the stimuli.

Questionnaire. On completion of the scheduled music listening period, partici- pants filled out a computerized questionnaire for closure, in which they indicated their current mood, either with Thayer’s (1986) Activation–Deactivation adjective checklist in mood management study or with the German words for happy, tense, ac- tive, depressed, calm, and tired on scales ranging from 0 to 10 in the mood adjust- ment study.

Data Preparations. Analyses of the mood management study presented by Knobloch and Zillmann (2002) had been based on accumulated exposure times. In this secondary analysis, patterns of selective exposure are examined based on 1-minute time intervals, as constructed from computer logs of 115 participants (67% women, average age 20.1). For each respondent, the exposure to the four songs that had been categorized as absorbing was accumulated for 1-minute intervals.

To compare gender differences in mood management and mood adjustment pro- cesses, the two data sets were merged for combined analyses. As the mood adjust- ment study had featured a 7-minute music listening period, compared to 10 minutes in Study 1, only seven 1-minute intervals were included. A dichotomous variable in- dicated for each case from which study (mood management vs. mood adjustment) the case derived.

As different mood items had been used in the two studies, only those four adjec- tives where equivalent translations existed in the two original data sets were em- ployed (tense, tired, active, and calm). Furthermore, different scales had been pre- sented in the two studies, thus mood item scores were standardized for each data set before merging them. A factor analysis (principal component, varimax rotation) was conducted with the four standardized variables. It yielded two factors, one (39% ex- plained variance) with active (–.85) and tired (.85), the other one (32% explained vari- ance) with calm (–.82) and tense (.78). The two item pairs were condensed as mean scores so that two new variables represented “vivacity” and “tension.”

Knobloch-Westerwick/GENDER AND MOOD ADJUSTMENT 83

Results

Preliminary Analyses

To test the effectiveness of the mood manipulation, analyses of variance (ANOVAs) were performed with the manipulation and the data collection as between factors and tension and vivacity as dependent measures. A main effect of the mood manipulation failed to reach significance for tension (p = .11) but the main effect on vivacity, F(2, 189) = 3.83, p = .02, attested to an effective treatment (M = –.33 for bad, –.16 for neu- tral, and .47 for good mood). No other effects were significant in these analyses.

Effects of Mood and Gender on Selective Exposure

To examine H1 and RQ1, a mixed-design ANOVA with mood induction and gender as between factors and repeated measures for exposure to absorbing music by sam- pling interval as a within factor was performed with the mood management sample only. The time intervals amounted to 1 minute each, thus 10 intervals were investi- gated. This analysis yielded a significant main effect of mood, F(2, 109) = 4.13, p = .02, with a significant linear trend (p = .01), a key finding in the original analysis by Knobloch and Zillmann (2002). Although this effect was independent of time, Figure 1 depicts selective music exposure across time to illustrate the stability of the expo- sure pattern. In the first minute, listening to absorbing music was equally likely for all groups, as all participants obviously probed what type of music served their needs best in this situation. After first exploration, however, all experimental groups showed fairly stable exposure patterns in line with mood management theory. The only other effect that approached significance in this secondary data analysis was an interaction between time and gender (p = .11) that was independent of the mood induction.

The same analysis was performed with only the first seven time intervals as re- peated measurements. This served to provide an examination consistent with the sec- ond study that featured 7 minutes of music listening in total. The analysis revealed the same mood effect, F(2, 109) = 2.88, p = .06, and a significant linear trend (p = .02) for mood. Moreover, an interaction between time and gender in line with H1 was signifi- cant, F(6, 654) = 2.07, p = .05, and is depicted in Figure 2. Both genders were in the beginning equally likely to sample from absorbing music, which again indicates a phase where participants explored the stimuli to find what served their needs in this situation. However, women lowered their exposure to this music up to a distinct in- crease in the fifth minute and then remained at this higher level. Men essentially showed an opposite pattern: an initial increase in exposure to absorbing music, a clear decrease in the fourth minute, and then low levels toward the end of the listen- ing period.

84 Journal of Broadcasting & Electronic Media/March 2007

Effects of Mood, Anticipation, and Gender on Selective Exposure

To address H2 and RQ1, a mixed-design ANOVA with mood induction and gender as between factors and repeated measures for exposure to absorbing music by minute intervals as a within factor was conducted with the mood adjustment sample. It yielded only a significant effect in line with H2—an interaction between gender and the within factor of time, F(6, 438) = 2.44, p = .03. The associated selective exposure processes for men and women, who anticipated more tasks after completing a veiled mood induction task, are depicted in Figure 2. The patterns are almost reversed. Both genders originally exposed themselves to about the same levels of absorbing music during the first 3 minutes, which indicates an initial probing of the available stimuli and their affective impact in that situation. Women, however, showed a marked de- crease in the fourth interval and generally listened to less absorbing music during the last intervals. Men, on the other hand, increased their exposure to absorbing music al- most throughout the entire listening period.

Knobloch-Westerwick/GENDER AND MOOD ADJUSTMENT 85

Figure 1 Selective Exposure to Absorbing Music After a Mood Induction

as a Function of Mood

Combined Analysis of Effects of Mood, Anticipation, and Gender on Selective Exposure

A mixed-design ANOVA with study, gender, and mood induction as between fac- tors and seven 1-minute time intervals of selective exposure to absorbing music as a repeated measure (within factor) was conducted for a simultaneous analysis address- ing H1, H2, and RQ1. It yielded a main effect of study, F(1, 183) = 23.2, p < .001, η2 = .113, as participants in the mood adjustment study generally spent more time with ab- sorbing music (M = 281 vs. 191). More important, an interaction among study, gen- der, and time as the within factor emerged, F(6, 1,098) = 3.11, p = .01, due to the op- posite exposure patterns of the genders in the two studies (see Figure 2 with estimated marginal means across time). No additional effects were significant.

86 Journal of Broadcasting & Electronic Media/March 2007

Figure 2 Selective Exposure to Absorbing Music After a Mood Induction—With or Without

Anticipation of Further Tasks—as a Function of Gender and Time

Discussion

It is almost conventional wisdom that media preferences of men and women differ with great consistency. This investigation examined processes of media content selec- tions as they depend on gender and mood states. It was expected that men seek dis- tracting media stimuli right after mood induction, whereas women were thought to ruminate over less absorbing stimuli first and then eventually turn to more distracting messages. It was furthermore hypothesized that media users, when anticipating a mood-inducing activity, would show these gender-typed selection behaviors right be- fore this activity. A research question explored alternative considerations for these patterns in moods with different valence.

A secondary data analysis of a mood management experiment found the mood effect on selective media exposure to be consistent across time. Yet exposure patterns differed by gender nonetheless. After an alleged social skills test, women avoided distractions such as fast-paced music initially by favoring less absorbing stimuli. Women obviously preferred to ponder the results before they later turned to distractions. Men, on the other hand, preferred more distraction initially, possibly to better their mood, and then low- ered their exposure to absorbing music. This gender difference supported H1 and falls in line with what Nolen-Hoeksema (e.g., 1987) and Thayer et al. (1994) observed in terms of mood regulation patterns of the genders. However, regarding RQ1, the interaction be- tween gender and time was not mediated by mood valence.

In the mood adjustment study, where participants anticipated further mood-related activities after the media consumption period, valence of induced mood was irrele- vant for exposure to mood-bettering music. From a mood management perspective, this is surprising and contradicts the theory, yet the mood adjustment approach can explain this with the upcoming tasks. Instead, gender influenced selection patterns across time. Women lowered their exposure to absorbing music and, during the last 4 minutes before upcoming tasks, chose less absorbing music than in the beginning. Men essentially showed an opposite exposure pattern, as they increased listening to absorbing music almost throughout the entire listening period. In short, H2 was sup- ported. However, exploration of RQ1 yielded no mood valence impacts on gen- der-typed selections.

Finally, the combined data sets allowed examination of all assumptions subjected to empirical testing. The genders showed basically opposite exposure patterns for ab- sorbing music that were reversed for the two experimental situations. Men who had just taken a test favored absorbing music in the beginning and then lowered exposure to it abruptly, unlike men who anticipated more assignments, as the latter only slowly increased listening to this music type. Women on the other hand, when free to opti- mize moods after taking a test, first decreased exposure to absorbing music to fairly low levels until an abrupt increase occurred, followed by stable listening to this mu- sic. The other women who still anticipated more assignments initially favored absorb- ing music, before they reduced their exposure to it and kept it on lower levels.

Knobloch-Westerwick/GENDER AND MOOD ADJUSTMENT 87

An interaction between situation type—open versus anticipated assignment—and gender, as well as time and mood valence (as pondered in RQ1), would be expected in light of mood management theory. Yet the processes turned out to be simpler, as mood valence was not important. Instead, combined effects of gender, situation, and time emerged independent of hedonic valence of mood. Thus the same gender-typed selection patterns for absorbing music across time were observed for all participants, although they had been placed in bad, neutral, and good moods.

This observation converges with results from a study (Knobloch-Westerwick & Al- ter, 2006) about mood adjustment described earlier. Intensity of negative mood had been manipulated through level of provocation but did not result in any differences either. Only gender and anticipation emerged as relevant for selective media expo- sure across time. Accordingly, both the study by Knobloch-Westerwick and Alter (2006) and the findings reported here underscore that gender-specific strategies for af- fect regulation and anticipated tasks can be more important for mood adjustment through selective media use than motivations to overcome initial moods. However, this probably only applies when upcoming activities are relevant for self-perceptions. When the upcoming task appears to be merely a mundane assignment without high relevance for the self, the individual is probably more inclined to make the most of it and build in some mood optimization behavior, as pondered in the considerations re- garding RQ1.

The broader applicability of gender-typed mood regulation patterns indicates that men and women might generally differ in the ways in which they deal with affective states, as reflected in media choices. Women generally tend to employ more coping strategies than men but, in relative terms, men use distraction more when coping with stressors and women use more rumination (Tamres et al., 2002). The anticipation of a social skills test is very likely to function as stressor, thus the music listening was a pos- sible way to seek distraction. As men probably tried to distract themselves from the upcoming task more than women did, it is plausible that they favored comparatively absorbing music more than their female counterparts. It is furthermore possible that they thought that distraction would help them to attain a relaxed attitude, which could be functional for optimal performance in subsequent tests. Women, on the other hand, have a tendency to ruminate on thoughts related to a stressor, which could explain why they preferred subtle music with low levels of absorption poten- tial. These less distracting stimuli should be more suitable when trying to gather one’s thoughts while using rumination.

Two explanations can help us understand why men and women may differ funda- mentally in their strategies for affect regulation, also through selective exposure to en- tertainment fare, as indicated in the findings reported here. Research on emotion so- cialization of the genders and motivations to meet gender-typed social expectations can shed light on origins of this phenomenon.

Research shows that boys and girls are brought up differently regarding emotional feedback and expressions (Garside & Klimes-Dougan, 2002). Gender-specific social- ization of emotions is based on the common lay assumption that women experience

88 Journal of Broadcasting & Electronic Media/March 2007

and express most kinds of emotions, including sadness and fear, more often and more intensely than men; exceptions include anger and pride, which are perceived to be experienced and expressed more often by men (Birnbaum, Nosanchuk, & Croll, 1980; Plant, Hyde, Keltner, & Devine, 2000; Shields, 1984). If women are socialized in ways that encourage emotionality, in comparison to men, then it stands to reason that they generally respond to an affective experience or anticipation with rumination as an affect-intensifying strategy. Emotionality in males, on the other hand, is discour- aged in childhood, which may be why men generally distract themselves from affec- tive incidents to reduce affective responses. Yet for anger, as a “typically male” emo- tion that is already encouraged more in boys than in girls, men might do the opposite and sustain this affect via rumination (Knobloch-Westerwick & Alter, 2006; O’Neal & Taylor, 1989). These patterns, in turn, are then reflected in media choices that can fa- cilitate rumination or distraction, respectively.

It is also possible that media users prefer to expose themselves to media content that instigates emotions where the display thereof, depending on gender, is socially expected and tends to be rewarded by others. Empirical evidence supports this no- tion. Zillmann, Weaver, Mundorf, and Aust (1986) found that male participants’ en- joyment of a horror movie was higher when in the company of frightened female coviewers, whereas females enjoyed the film more when their male coviewers ex- pressed mastery. Harris, Hoekstra, Scott, Sanborn, and Dodds (2004) explored selec- tions of romantic movies during a date, which revealed that mostly the women picked the movie and liked it more. With this perspective, even making oneself sad or afraid has functionality for women because it supports behaviors and expressions that are consistent with their gender’s stereotype. Likewise, an aggressive posture that sup- presses display of any other emotion should be more easily attained after watching action- and conflict-laden media content. Possibly, men prime themselves by aid of violence-centered entertainment so that they adhere better to the prototype of the “manly man.” After all, men and women, as well as boys and girls, have a tacit under- standing of what kind of emotionality—or lack thereof for men—is expected from them and that converging with these expectations is usually socially rewarding. It is plausible, then, that men and women employ media use to display gender-typed af- fects as expected.

The approach of this study can be improved in future research by drawing par- ticipants from only one population, which was not possible here. Ideally, a repli- cation or extension of this investigation should compile and pretest new sets of songs or of other electronic-media stimuli that fit into what is popular and up-to-date at the time. The pretest could also collect explicit information on how distracting or absorbing listeners perceive the various songs to be. Furthermore, to go beyond mere exploration of the research question in this study, experimental manipulation of both mood valence and kind of upcoming activity, with low or high task relevance for self-perceptions, would offer insight into circumstances that foster or reduce patterns of mood optimization and mood adjustment in selec- tive media exposure.

Knobloch-Westerwick/GENDER AND MOOD ADJUSTMENT 89

Additional research in this area is certainly needed. The outlined gender-typed af- fect regulation strategies, either due to emotion socialization or motivations to meet social expectations, are likely to affect any entertainment choices from electronic and broadcasting media. Women are more likely to watch tragedies, daily soaps, and ro- mance, as they express greater appreciation for these genres than men. Men, on the other hand, tend to like horror, sports, and action movies more than women do (Col- lins-Standley, Gan, Yu, & Zillmann, 1996; Herzog, 1944; Oliver, 1993, 2000). Women also display more intense reactions to tragedies and fright-evoking movies but also express greater enjoyment, whereas men favor aggression-laden entertain- ment such as action movies, sports, horror, and heavy-metal music (Hansen & Hansen, 2000; Oliver, 2000). Future research should pursue origins of these gender differences in selective media use, as the observation of these long-standing and sta- ble gaps still awaits full explanation.

Notes

1When categorizing musical stimuli by absorption potential, a small differentiation of the va- lence aspect is almost unavoidable because these two dimensions are correlated in perceptions of music stimuli (r = .27 in this pretest).

2Another investigation employed the same cover story and reports all details of it (see Knobloch, 2003).

References

Anderson, D. R., Collins, P. A., Schmitt, K. L., & Jacobvitz, R. S. (1996). Stressful life events and television viewing. Communication Research, 23, 243–260.

Birnbaum, D., Nosanchuk, T., & Croll, W. (1980). Children’s stereotypes about sex differ- ences in emotionality. Sex Roles, 6, 435–443.

Biswas, R., Riffe, D., & Zillmann, D. (1994). Mood influence on the appeal of bad news. Jour- nalism Quarterly, 71, 689–696.

Borling, J. E. (1981). The effects of sedative music on alpha rhythms and focused attention in high-creative and low-creative subjects. Journal of Music Therapy, 28, 101–108.

Brody, L. R., & Hall, J. A. (1993). Gender and emotion. In M. Lewis & J. M. Haviland (Eds.), Handbook of emotions (pp. 447– 460). New York: Guilford.

Bryant, J., & Zillmann, D. (1984). Using television to alleviate boredom and stress: Selective exposure as a function of induced excitational states. Journal of Broadcasting, 28, 1–20.

Butler, L. D., & Nolen-Hoeksema, S. (1994). Gender differences in response to depressed mood in a college sample. Sex Roles, 30, 331–346.

Collins-Standley, T., Gan, S., Yu, H. J., & Zillmann, D. (1996). Choice of romantic, violent, and scary fairy-tale books by preschool girls and boys. Child Study Journal, 26, 279–301.

Deutsches Musikinformationszentrum. (2005, April). Statistiken zur Musikwirtschaft [Statis- tics on the music business]. Retrieved September 20, 2005, from http://www.miz.org/ suche_1019.html

Fischer, A. (Ed.). (2004). Gender and emotion: Social psychological perspectives. Cam- bridge, UK: Cambridge University Press.

Garside, R. B., & Klimes-Dougan, B. (2002). Socialization of discrete negative emotions: Gender differences and links with psychological distress. Sex Roles, 47, 115–128.

90 Journal of Broadcasting & Electronic Media/March 2007

Hansen, C. H., & Hansen, R. D. (2000). Music and music videos. In D. Zillmann & P. Vorderer (Eds.), Media entertainment: The psychology of its appeal (pp. 175–196). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.

Harris, R. J., Hoekstra, S. J., Scott, C. L., Sanborn, F. W., & Dodds, L. A. (2004). Autobiograph- ical memories for seeing romantic movies on a date: Romance is not just for women. Media Psy- chology, 6, 257–284.

Helregel, B. K., & Weaver, J. B. (1989). Mood-management during pregnancy through selec- tive exposure to television. Journal of Broadcasting & Electronic Media, 33, 15–33.

Herzog, H. (1944). What do we really know about daytime serial listeners? In P. F. Lazarsfeld & F. N. Stanson (Eds.), Radio research, 1942–1943 (pp. 3–33). New York: Duell, Sloan & Pearce.

Ingram, R. E. (1990). Self-focused attention in clinical disorders: Review and a conceptual model. Psychological Bulletin, 107, 156–176.

Ingram, R. E., Cruet, D., Johnson, B. R., & Wisnicki, K. S. (1988). Self-focused attention, gen- der, gender role, and vulnerability to negative affect. Journal of Personality & Social Psychology, 55, 967–978.

Johnson, J. F., & Schulman, G. A. (1988). More alike than meets the eye: Perceived gender differences in the subjective experience and its display. Sex Roles, 19, 67–79.

Knobloch, S. (2002). “Unterhaltungsslalom” bei der WWW-Nutzung: Ein Feldexperiment [”Entertainment Zigzag” in www use: A Field Experiment]. Publizistik, 47, 309–318.

Knobloch, S. (2003). Mood adjustment via mass communication. Journal of Communica- tion, 53(2), 233–250.

Knobloch, S., & Zillmann, D. (2002). Mood management via the digital jukebox. Journal of Communication, 52(2), 351–366.

Knobloch-Westerwick, S. (2006). Mood management: Theory, evidence, and advancements. In J. Bryant & P. Vorderer (Eds.), The psychology of entertainment (pp. 239–254). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.

Knobloch-Westerwick, S., & Alter, S. (2006). Mood adjustment to social situations through mass media use: How men ruminate and women dissipate angry moods. Human Communica- tion Research, 32, 58–73.

Lyubomirsky, S., & Nolen-Hoeksema, S. (1993). Self-perpetuating properties of dysphoric ru- mination. Journal of Personality & Social Psychology, 65, 339–349.

Lyubomirsky, S., & Nolen-Hoeksema, S. (1994). Effects of self-focused rumination on nega- tive thinking and interpersonal problem-solving. Journal of Personality & Social Psychology, 69, 176–190.

Masters, J. C., Ford, M. E., & Arend, R. A. (1983). Children’s strategies for controlling affective responses to aversive social experience. Motivation & Emotion, 7, 103–116.

Mayfield, C., & Moss, S. (1989). Effect of music tempo on task performance. Psychological Reports, 65, 1283–1290.

Meadowcroft, J. M., & Zillmann, D. (1987). Women’s comedy preferences during the men- strual cycle. Communication Research, 14, 204–218.

Medoff, N. J. (1982). Selective exposure to televised comedy programs. Journal of Applied Communication Research, 10, 117–132.

Morrow, J., & Nolen-Hoeksema, S. (1990). Effects of responses to depression on the remediation of depressive affect. Journal of Personality & Social Psychology, 58, 519–527.

Nolen-Hoeksema, S. (1987). Sex differences in unipolar depression: Evidence and theory. Psychological Bulletin, 101, 259–282.

Nolen-Hoeksema, S. (1990). Sex differences in depression. Stanford, CA: Stanford University Press.

Nolen-Hoeksema, S., Larson, J., & Grayson, C. (1999). Explaining the gender difference in depressive symptoms. Journal of Personality & Social Psychology, 77, 1061–1072.

Nolen-Hoeksema, S., & Morrow, J. (1991). A prospective study of depression and distress fol- lowing a natural disaster. Journal of Personality & Social Psychology, 61, 105–121.

Nolen-Hoeksema, S., & Morrow, J. (1993). Effects of rumination and distraction on naturally occurring depressed mood. Cognition & Emotion, 7, 561–570.

Knobloch-Westerwick/GENDER AND MOOD ADJUSTMENT 91

Nolen-Hoeksema, S., Morrow, J., & Fredrickson, B. L. (1993). Response styles and the dura- tion of episodes of depressed mood. Journal of Abnormal Psychology, 102, 20–28.

Nolen-Hoeksema, S., Parker, L., & Larson, J. (1994). Ruminative coping with depressed mood following loss. Journal of Personality & Social Psychology, 67, 92–104.

Oliver, M. B. (1993). Exploring the paradox of the enjoyment of sad films. Human Communi- cation Research, 19, 315–342.

Oliver, M. B. (2000). The respondent gender gap. In D. Zillmann & P. Vorderer (Eds.), Media entertainment: The psychology of its appeal (pp. 215–234). Mahwah, NJ: Lawrence Erlbaum As- sociates, Inc.

Oliver, M. B. (2003). Mood management and selective exposure. In J. Bryant, D. Roskos-Ewoldsen, & J. Cantor (Eds.), Communication and emotion (pp. 85–106). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.

O’Neal, E. C., & Taylor, S. L. (1989). Status of the provoker, opportunity to retaliate, and inter- est in video violence. Aggressive Behavior, 15, 171–180.

Plant, E. A., Hyde, J. S., Keltner, D., & Devine, P. G. (2000). The gender stereotyping of emo- tion. Psychology of Women Quarterly, 24, 81–92.

Rusting, C. L., & Nolen-Hoeksema, S. (1998). Regulating responses to anger: Effects of rumi- nation and distraction on angry mood. Journal of Personality & Social Psychology, 74, 790–803.

Shields, S. A. (1984). Distinguishing between emotion and nonemotion: Judgements about experience. Motivation & Emotion, 8, 355–369.

Smith, C. A., & Morris, L. W. (1977). Differential effects of stimulative and sedative music on anxiety, concentration, and performance. Psychological Reports, 41, 1047–1053.

Tamres, L. K., Janicki, D., & Helgeson, V. S. (2002). Sex differences in coping behavior: A meta-analytic review and an examination of relative coping. Personality & Social Psychology Review, 6, 2–30.

Thayer, R. E. (1986). Activation–Deactivation adjective check list: Current overview and structural analysis. Psychological Reports, 58, 607–614.

Thayer, R. E., Newman, J. R., & McClain, T. M. (1994). Self-regulation of mood: Strategies for changing a bad mood, raising energy, and reducing tension. Journal of Personality & Social Psy- chology, 67, 910–925.

Trask, P. C., & Sigmon, S. T. (1999). Ruminating and distracting: The effects of sequential tasks on depressed mood. Cognitive Therapy and Research, 23, 231–246.

Wakshlag, J. J., Reitz, R., & Zillmann, D. (1982). Selective exposure to and acquisition of in- formation from educational television programs as a function of appeal and tempo of back- ground music. Journal of Educational Psychology, 74, 666–677.

Winter, J. P. (1985). American music and Canadian youth. Communication Research, 12, 345–352.

Wood, J. V., Saltzberg, J. A., Neale, J. M., Stone, A. A., & Rachmiel, T. B. (1990). Self-focused attention, coping responses, and distressed mood in everyday life. Journal of Personality & So- cial Psychology, 58, 1027–1036.

Zillmann, D. (1988). Mood management through communication choices. American Behav- ioral Scientist, 31, 327–340.

Zillmann, D. (2000). Mood management in the context of selective exposure theory. In M. F. Roloff (Ed.), Communication yearbook 23 (pp. 103–123). Thousand Oaks, CA: Sage.

Zillmann, D. (2003). Theory of affective dynamics: Emotions and moods. In J. Bryant, D. Roskos-Ewoldsen, & J. Cantor (Eds.), Communication and emotion (pp. 533–567). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.

Zillmann, D., & Bryant, J. (1985). Affect, mood, and emotion as determinants of selective ex- posure. In D. Zillmann & J. Bryant (Eds.), Selective exposure to communication (pp. 157–190). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.

Zillmann, D., Hezel, R. T., & Medoff, N. J. (1980). The effect of affective states on selective exposure to televised entertainment fare. Journal of Applied Psychology, 10, 323–339.

Zillmann, D., Weaver, J. B., Mundorf, N., & Aust, C. F. (1986). Effects of an opposite-gender companion’s affect to horror on distress, delight, and attraction. Journal of Personality and Social Psychology, 51, 586–594.

92 Journal of Broadcasting & Electronic Media/March 2007