effects of television violence on memory for …bbushman/bjepa98.pdfcannot remember the brand ......

Journal of Experimental Psychology: Applied Copyright 1998 by the American Psychological Association, Inc. 1998, Vol. 4, No. 4, 291-307 1076-898X/98/$3.00

Effects of Television Violence on Memory for Commercial Messages

Brad J. Bushman Iowa State University

Three experiments tested whether television violence impairs memory for commercials. Participants saw commercials embedded in violent or nonviolent film clips. The film clips were similar on other dimensions that might influence commercial memory (e.g., arousal and involvement). After viewing a film clip, participants completed several recall and recognition memory tests. In all experiments, participants who saw a violent film clip had poorer memory for commercials than did participants who saw a nonviolent film clip. In Experiment 3, participants also reported their mood after viewing a film clip but before completing the memory tests. Anger mediated between television violence and commercial memory. Television violence increased anger, and anger, in turn, impaired memory for commercials. These results suggest that sponsoring violent programs might not be a profitable venture for advertisers.

Since its introduction to American society, television has become an integral part of nearly every home. It is estimated that there are more television sets in the United States than there are toilets. Over 98% of American homes have at least one television set (American Psychological Association, 1993). In the United States, children spend more time watching television than they spend at school, and adults spend more time watching television than they spend on any other activity except sleeping and working (Huston et al., 1992). Much of what is shown on television is violent. The National Television Violence Study

I would like to thank Len Berkowitz, Bill Hoyt, Rick Gibbons, Gary Wells, Jim Weaver, and Dolf Zillmann for their helpful comments on an earlier draft of this article, Dan Russell for his assistance with the Linear Structural Relations analyses, and Ken Koehler for his assistance with the log-linear analyses. I also extend appreciation to Eric Seelau, Angela Stack, and Stacey Swaim for their help with the videotapes and to Joy Cart, Andrea Determan, Shawn Peters, David Laidig, Lori McGonigal, Sarah Peters, Angela Stack, and Dominic Starr for serving as experimenters.

Correspondence concerning this article should be addressed to Brad J. Bushman, Department of Psychol- ogy, Iowa State University, Ames, Iowa 50011-3180. Electronic mall may be sent to [email protected].

(1996, 1997) conducted a detailed content analysis of about 4,500 hr of programming on cable and broadcast television. The results showed that about 60% of the programs contained violence. By the time the average child graduates from elementary school, he or she will have seen more than 8,000 murders and more than 100,000 other assorted acts of violence on television (Huston et al., 1992).

Surveys indicate that most Americans believe that there is too much violence on television (e.g., Fischer, 1994; Zipperer, 1994). In one survey (TV Guide, 1992), for example, people were asked to say in their own words what made them angry about television programming. The most common complaint was "too much violence." If most people object to the amount of violence shown on television, then why is there so much violence portrayed on television? Television in- dustry leaders answer this question by claiming that their programs merely reflect the violence that already exists in society. For example, the television movie producer Jim Green said, "Mur- der happens. It happened a lot in Shakespeare" (Drevitch, 1994, p. 12). The amount of violence shown on television, however, far exceeds the amount of violence found in the real world. Even in reality-based police television programs, rio-

291

292 BUSHMAN

lence is grossly overemphasized (e.g., Oliver, 1994). A second possible answer to this question is that advertisers assume violent programs will attract larger audiences and will therefore gain more exposure for their products. However, the available scientific evidence does not support the idea that gratuitous displays of violence will draw more viewers. Television violence neither increases nor decreases the popularity of a program in the United States (e.g., Diener & DeFour, 1978; Diener & Woody, 1981; Sprafkin, Rubin- stein, & Stone, 1977). 1 Alternatively, advertisers might want to target certain audiences who have a preference for violent programming.

Regardless of the reason for the excessive amount of violence on television, one important question is whether violent content influences memory for commercial messages. A commercial may be interesting, enjoyable, or persuasive, but it may not be effective if the potential buyer cannot remember the brand advertised or the commercial message. Unfortunately, the question of whether television violence influences memory for commercials is difficult to answer because of a dearth of empirical evidence on the topic. The few studies that have been conducted on this topic have confounded violence with other variables that influence memory for commercials, such as arousal. Many violent programs are action-adventure type programs. Previous research has shown that arousing programs impair attention to, as well as processing and storage of, commercial messages (e.g., Mundorf, Zillmann, & Drew, 1991; Pavelchak, Antil, & Munch, 1988; Singh & Churchill, 1987). The present research is interested in determining whether violent programs influence memory for commercial messages when the arousal level of programs is held constant.

There are theoretical reasons for predicting that emotional responses other than arousal mediate the effect of television violence on memory for commercial messages. There is a large body of research evidence indicating that mood can affect memory (e.g., Bower, 1981; Burke, Heuer, & Reisberg, 1992; Christiansen, 1992; Mayer, McCormick, & Strong, 1995). Previous research has shown that violent programs put viewers in a bad mood and especially make them angry (e.g., Anderson, 1997; Bushman, 1995; Bushman & Geen, 1990; Hansen & Hansen, 1990). There are

at least two reasons why anger might interfere with memory for commercials. First, the mood state induced by a television program can prime or activate similarly toned material stored in memory. Berkowitz (1984) has suggested that "the aggressive ideas suggested by a violent movie can prime other semantically related thoughts, heightening the chances that viewers will have other aggressive ideas in this period" (p. 411). Berkowitz bases this priming hypothesis on the concept of spreading activation within a network in memory (e.g., Collins & Loftus, 1975): Thoughts send out radiating activation along associative pathways, thereby activating other related thoughts. In this way, ideas about aggression that are not identical to those observed in the media may be elicited by the latter. In addition, thoughts are linked, along the same sort of associative lines, not only to other thoughts but also to emotional reactions (Bower, 1981; Isen, Clark, Shalker, & Karp, 1978). Thus, the anger induced by a violent program could activate other hostility-related ideas that interfere with the elabo- ration of the information contained in the commercials, rendering the commercials less accessible.

Mood management theory provides a second reason why anger might interfere with memory for commercial messages (Isen, 1984; Parrott, 1993; Schwarz & Clore, 1983). Viewers in a program-induced negative mood should engage in processes to terminate or improve their negative mood, whereas viewers in a program- induced positive mood should engage in processes to prolong their positive mood. Because mood repair processes require more effort than do mood maintenance processes, viewers in a negative mood might focus more attention on themselves and less attention on their environment, including the commercials (Wood, Saltzberg, & Goldsamt, 1990). Thus, anger induced by a violent television program could lead viewers to focus more attention on repairing their mood than on the commercials, rendering the commercials less accessible. The present research does not, however, explicitly test these theories.

The first two experiments in this article tested

1Similar results have been reported in countries other than the United States, such as Finland (Mus- tonen, 1997; Mustonen & Pulkkinen, 1993) and Britain (Wober, 1988).

TELEVISION VIOLENCE AND MEMORY FOR COMMERCIALS 293

whether television violence impairs memory for commercial messages. The third experiment tested whether program-induced anger mediates the effect of television violence on memory for commercial messages.

Exper iment 1

Experiment 1 tested the effects of television violence on recall of brand names and commercial messages. Male and female college students were randomly assigned to watch either a violent or nonviolent film clip that contained two commercials. The violent and nonviolent film clips were preselected because they did not differ on measures of self-reported arousal (i.e., exciting, boring, and arousing ratings) or on measures of physiological arousal (i.e., systolic blood pressure, diastolic blood pressure, mean arterial pressure, and heart rate). Immediately after viewing the videotape, participants recalled brand names of products in the commercials and recalled details in the commercial messages. After completing the recall measures, participants reported the number of hours per week they spent watching various types of television programs, including violent programs. This measure was taken to ensure that violent content of films rather than habitual exposure to televised violence was affect- ing responses to commercials. Participants also reported whether they had seen the film from which the clip had been taken and whether they had seen the commercials imbedded in the film clip. Participants who saw the nonviolent video were expected to have greater recall of brands and commercial messages than were participants who saw the violent video. Both men and women were included in Experiment 1 to increase the generalizability of the results, but no sex differences were predicted for either dependent variable.

Method

Participants

Participants were 200 undergraduate students enrolled in introductory psychology classes (100 men and 100 women) who received extra course credit in exchange for their voluntary participation.

Videotapes

The videotapes for Experiment 1 were preselected to be differentially violent but equally exciting and arousing. The foUowing 2 videotapes were chosen from a pool of 52 tapes that were each about 15-min long. Each tape was recorded from a commercial or cable television channel.

Karate Kid III (violent). In a karate tournament, a young man defeats an arrogant opponent who repeatedly violates the rules by fighting dirty. Prior to the tournament, the opponent had insulted the young man, kicked the young man's girlfriend in the stomach, and destroyed the young man's property.

Gorillas in the Mist (nonviolent). A scientist observes gorillas in their natural habitat. At first, the gorillas do not trust the scientist. In time, however, the gorillas begin to trust the scientist and they allow her to interact with them. Al- though this film clip was intense and emotionally moving, it contained no violence.

A total of 2,080 undergraduate students enrolled in introductory psychology classes acted as judges. Each videotape was rated by 40 judges (20 men and 20 women). The judges filled out a set of 11-point rating scales indicating the degree to which the video was exciting, boring, arousing, and violent. Some other rating dimensions were added as fillers. The scales were anchored at one end by 0 (not at all) and at the other end by 10 (extremely). Exciting, boring, and arousing ratings were used to measure the arousal content of the tapes, and violence ratings were used to measure the violent content of the tapes. Judges also reported whether they had seen the movie from which the clip was taken; this measure was used as a covariate in all analyses. A cover sheet explained that the questionnaires were part of a media evaluation study being conducted jointly with the Depart- ment of Journalism and Mass Communication.

There were no differences between the Karate Kid III and Gorillas in the Mist videotapes on any of the self-reported arousal measures (see Table 1). As expected, the Karate Kid III videotape was judged to be more violent (M = 7.2, SE = 0.42) than the Gorillas in the Mist videotape (M = 2.2, SE = 0.39), F(1, 77) = 76.89, p < .05, d = 2.00.

In previous research using these two videotapes (Bushman, 1995), higher levels of anger

294 BUSHMAN

Table 1 Comparison o f the Violent and Nonviolent Videotapes Used in Experiments 1 and 2 on Self.Report and Physiological Measures o f Arousal

Violent Nonviolent videotape videotape

Measure M SE M SE

Self-reported arousal Exciting Boring Arousing

Physiological arousal Systolic blood pres-

sure Diastolic blood pres-

sure Mean arterial pressure Heart rate

7.1 0.30 6.9 0.36 1.9 0.36 1.7 0.36 5.8 0.37 5.0 0.43

119.8 1.39 119.5 1.75

69.7 0.90 67.2 0.97 80.3 0.89 81.2 0.82 81.2 0.82 80.3 0.83

Note. Means were adjusted for whether judges had seen the film from which the videotape was taken. Ratings were made using an 11-point Likert-type scale ranging from 0 (not at all) to 10 (extremely). The violent video was taken from Karate Kid III and the nonviolent video was taken from Gorillas in the Mist (see video descriptions in tex0. N = 40 (20 men and 20 women) for self-reported arousal measures. N = 296 (148 men and 148 women) for physiological arousal measures.

and aggressive behavior were found in participants who saw the Karate Kid III videotape than in participants who saw the Gorillas in the Mist videotape. The videotapes did not differ, however, in their effects on systolic blood pressure, diastolic blood pressure, mean arterial pressure, and heart rate (see Table 1). The cardiovascular measures are consistent with the self-report measures of arousal content (see Table 1).

Each videotape consisted of three segments (each about 5 rain long) that maintained a coher- ent story line. One commercial was inserted between the first and second film segments and the other commercial was inserted between the second and third film segments. The commercials were for Krazy Glue and Wisk laundry detergent. Both commercials advertised products with broad market appeal, were 30 s long, and were several years old.

The Krazy Glue commercial depicts a man hanging in the air from a construction beam, kicking his legs and trying to get down. The man is wearing a helmet that is fastened to the

construction beam with Krazy Glue. In the fore- ground is a table containing several fragile objects the man fixed using Krazy Glue. A woman is shown breaking the objects on the table, appar- ently because she is mad at the man for breaking a promise he had made to her. The commercial ends with the announcer stating, "Krazy Glue, for things that are broken, except promises."

The Wisk commercial shows people getting their clothes dirty in a variety of ways (e.g., children are shown jumping in mud puddles, chasing pigs in the dirt, and spilling food and beverages on their clothes). Meanwhile, the jingle "Tisk, Tisk, Tisk, Wisk, Wisk" plays in the background. A piece of white fabric, stained with the letters TSK, is shown coming clean after it is washed with Wisk detergent. The commercial ends with the announcer stating, "Advanced action Wisk dissolves adhesive bonds that lock dirt and stains onto your clothes. To clean a whole wash load of risks."

The order of the commercials was counterbal- anced so that half of the participants saw the Wisk commercial first and the Krazy Glue commercial second, whereas the other half saw the commercials in the reverse order. To enhance the realism of the videotape, a network identification message was inserted prior to each commercial.

Procedure

Participants were tested individually. They were told that the study was part of a National Consumer Research Project to evaluate clips of films that had recently been shown on television. Participants were randomly assigned to watch either the violent or nonviolent videotaped film clip. Immediately after the showing of the videotape, participants listed the names of brands advertised in the commercials and listed everything they could remember about the commercial messages. After completing the recall measures, participants reported the number of hours per week they spent watching various types of television programs, including violent programs. They also reported whether they had seen the film from which the clip had been taken and whether they had seen the commercials imbedded within the film clip. Finally, participants were questioned for suspicion and fully debriefed. No participant expressed suspicion about the study.


Results

Order, Commercial, and Sex Effects

The presentation order of the two commercials did not influence either dependent variable. There were no differences between the two commercials on either dependent variable. There also were no sex main effects or interactions for either dependent variable. Thus, the data for the two orders, two commercials, and two sexes were combined for subsequent analyses.

Exposure Effects

Habitual exposure to televised violence and previous exposure to the film and commercials shown in the videotape were not related to either one of the dependent variables, either alone or interacting with film content. The exposure variables were therefore excluded from the analyses.

Reliability of Coding Recall Measures

Two independent raters who were unaware of the conditions and experimental hypotheses coded the free recall test scores. Because the intraclass correlation coefficient (Shrout & Fleiss, 1979) was quite large (ricc = .86), scores from the two raters were averaged for subsequent analyses.

Effects of Television Violence on Commercial Recall

Brand name recall. Participants who watched the nonviolent video recalled more brands than did participants who watched the violent videotape, F(1, 198) = 7.85, p < .05, d = 0.40 (see Table 2). 2

Commercial message recall. On the free recall test, participants who watched the nonviolent video recalled more details in the commercials than did participants who watched the violent videotape, F(1 ,198) = 10.98, p < .05, d = 0.47 (see Table 2).

Discussion

As expected, television violence impaired memory for commercials. Participants who watched the violent video recalled fewer brand

Table 2 Effects of Television Violence on Memory for Commercial Messages (Experiment 1)


Measure M SE M SE

Brand name recall 1.22 0.08 1.50 0.07 Commercial message

recall 6.69 0.38 8.61 0.44

Note. Brand name recall scores could range from 0 (no brands recalled) to 2 (both brands recalled). Commercial message recall scores were continuous (scores ranged from 0 to 22).

names and commercial message details than did participants who watched the nonviolent videotape. It is unlikely that these memory deficts were due to differential arousal levels produced by the films. The violent and nonviolent videotapes used in Experiment 1 were preselected because they did not differ on self-report or physiological measures of arousal (see Table 1).

Exper iment 2

Experiment 2 was conducted to replicate and extend the results of Experiment 1. Experiment 2 included an additional memory measure--visual brand recognition. A brand choice decision made in a store may need only a recognition level of learning because the choices may be reviewed at the time a decision is made (Singh & Rothschild, 1983). Even when a consumer makes a shopping list of products in the home, the final brand selection generally occurs in the store (Betmaan, 1979).

Experiment 2 also included some manipulation check items. It is well known that active rehearsal facilitates encoding and later retrieval of stimuli (Craik & Lockhart, 1972), including commercial messages. Previous research has shown that if viewers are caught up in an involving television program, they are more likely to rehearse the program than the commercial message (Bryant & Comisky, 1978). Commercial interruptions in high involvement programs might also distract and annoy viewers (Worchel, Hardy, & Hurley,

2Although brand recall scores only ranged from 0 to 2, log-linear analyses yielded the same pattern of results, X2(2, N = 200) = 8.04, p < .05.

296 BUSHMAN

1976). One cannot rule out involvement level as an alternative explanation of the results of Experi- ment 1 because the videotapes were not rated on an involvement dimension. In Expe~ent 2, participants rated how involving the film clip was and how distracting and annoying the commercials were.

In Experiment 2, male and female college students were randomly assigned to violent or nonviolent videotape conditions. Immediately after viewing the videotape, participants recalled brand names, recalled commercial message details, and recognized advertised brands from competitor brands. Participants reported how involving the film clip was and how distracting and annoying the commercials were. Scores on all three dependent variables (i.e., brand recognition, brand recall, and commercial message recall) were expected to be higher in the nonviolent videotape condition than in the violent videotape condition.

Method

Participants


Procedure

Participants were tested in small groups, but they worked independently on tasks. The cover story and videotapes were the same as those used in Experiment 1. Immediately after the showing of the videotape, participants were given an unexpected free recall test. Participants listed the names of brands advertised and listed everything they could remember about the commercial messages. Participants then completed a visual brand recognition test. Participants were told that the film clip contained two commercials, one for a brand of glue and the other for a brand of laundry detergent. Participants were shown slides of 6 glue brands (Krazy Glue and 5 distractor brands) and 6 laundry detergent brands (Wisk and 5 distractor brands) and were asked to write the brands advertised in the commercials. The order of the products and the order of brands within each product were randomized for each session. The slides were shown at the rate of one slide every 5 s.

Next, participants reported the number of hours per week they spent watching various types of television programs, including violent programs. Participants also reported whether they had seen the film from which the clip had been taken and whether they had seen the commercials imbedded within the film clip. Participants also completed the following manipulation check items: (a) "The film clip was involving," (b) "The laundry detergent commercial was distracting," (c) "The laundry detergent commercial was annoying," (d) "The glue commercial was distracting," and (e) "The glue commercial was annoying." The statements were rated using a 10-point scale ranging from 1 (strongly disagree) to 10 (strongly agree). The first statement was included to test whether the Karate Kid III and Gorillas in the Mist videotapes differed in their involvement level. The remaining four statements were included to test whether commercial interruptions were more distracting and annoying when they were embedded in the violent film clip than when they were embedded in the nonviolent film clip. Finally, participants were questioned for suspicion and fully debriefed. No participant expressed suspicion about the study.

Results

Order, Commercial, and Sex Effects

The presentation order of the two commercials did not influence any of the dependent variables or manipulation check items. There were no differences between the two commercials on any of the dependent variables or manipulation check items. There also were no sex main effects or interactions for any of the dependent variables or manipulation check items. Thus, the data for the two orders, two commercials, and two sexes were combined for subsequent analyses.

Exposure Effects

Habitual exposure to televised violence and previous exposure to the film and commercials shown in the videotape were not significantly related to any of the dependent variables or manipulation check items, either alone or interacting with film content. The exposure variables were therefore excluded from the analyses.


Film Clip Differences in Involvement Ratings

There was no difference between the violent and nonviolent videotapes in involvement ratings, F(1, 198) = 1.44, p > .05 (see Table 3). Thus, any effects of media violence on commercial recall cannot be attributed to differential levels of viewer involvement in the violent and nonviolent videotape conditions. Film clip involvement ratings also did not influence any of the dependent variables, either alone or interacting with film content. The film involvement variable was therefore excluded from the analyses.



Measure M SE M SE

Brand name recognition 1.66 0.06 1 .85 0.04 Brand name recall 1.21 0.07 1.56 0.05 Commercial message

recall 7.10 0.39 8.01 0.32

Note. Brand name recognition and recall scores could range from 0 to 2. Commercial message recall scores were continuous (scores ranged from 1 to 17).

Effects of Television Violence on Distracting and Annoying Ratings of Commercials

There were no differences between the violent and nonviolent videotapes in ratings of how distracting and annoying the commercials were, Fs(1, 198) = 1.39 and 1.98, respectively, ps > .05 (see Table 3). Thus, commercial interruptions did not distract and annoy viewers more in the violent videotape condition than in the nonviolent videotape conditions. Distracting and annoying ratings of commercials also did not influence any of the dependent variables, either alone or interacting with film content. The distracting and annoying variables were therefore excluded from the analyses.

Reliability of Coding Message Recall

Two independent raters who were unaware of conditions and experimental hypotheses coded

Table 3 Effects of Television Violence on Manipulation Check Items (Experiment 2)


Measure M SE M SE

Film was involving 7.1 0.21 6.7 0.24 Commercials were

annoying 5.1 0.22 5.5 0.22 Commercials were dis-

tracting 5.4 0.22 5.8 0.22

Note. Ratings were made using a 10-point scale ranging from 1 (strongly disagree) to 10 (strongly agree).

the free recall data. Because the intraclass correlation coefficient (Shrout & Fleiss, 1979) was large (ricc = .95), scores from the two raters were averaged for subsequent analyses.


Brand name recognition. Participants who watched the nonviolent video recognized more brands from the slides than did participants who watched the violent videotape, F(1, 198) = 6.48, p < .05, d = 0.36 (see Table 4). 3

Brand name recall. Participants who watched the nonviolent video recalled more brands than did participants who watched the violent videotape, F(1, 198) = 15.70,p < .05, d = 0.56 (see Table 4). 4

Commercial message recall. Although the effect was in the predicted direction, videotape content did not significantly influence recall of commercial message details, F(1, 198) = 3.17, p > .05, d = 0.25 (see Table 4).

Discussion

Experiment 2 replicated and extended the results from Experiment 1. Participants who watched the violent video recalled fewer brand

3Although brand recognition scores only ranged from 0 to 2, log-linear analyses yielded the same pattern of results. However, the effect was not quite significant, ×2(2, N = 200) = 5.76, p < .06.

4Although brand recall scores only ranged from 0 to 2, log-linear analyses yielded the same pattern of results, ×2(2, N = 200) = 12.63, p < .05.

298 BUSHMAN

names than did participants who had watched the nonviolent videotape. Participants who watched the violent video, in comparison to those who watched the nonviolent video, were less likely to recognize the advertised brands from the competitor brands. The effect was in the same direction for recall of commercial message details, although it was not quite significant. It is unlikely that these memory deficits were due to involvement level because the violent and nonviolent videotapes did not differ in terms of how involving they were judged to be. In addition, the amount of violence in the program did not influence perceptions of how distracting and annoying the commercials were.

Experiment 3

The primary purpose of Experiment 3 was to determine whether anger mediates the effect of television violence on memory for commercial messages. Arousal and positive affect also were tested as potential mediators. Testing both anger and positive affect allows one to determine whether anger impairs memory for commercials or whether positive affect improves memory for commercials. Arousal was also tested to determine whether generalized arousal or emotions specific to aggression had a stronger mediating influence on the relation between television violence and memory for commercials.

A second purpose of Experiment 3 was to replicate the results from Experiments 1 and 2, using a broader set of programs and commercials. Although the violent and nonviolent videotapes used in Experiments 1 and 2 appear to be similar on a number of important dimensions (e.g., self-reported arousal, cardiovascular arousal, and involvement level), only one exemplar of each type of videotape was used. Thus, it is possible that differences between the violent and nonviolent videotapes are due to features of the videotapes other than the amount of violence they contain. In Experiment 3, four exemplars of each type of videotape were used. Although no differences were found between the two commercials used in Experiments 1 and 2, a different commercial also was used in Experiment 3 to increase the generalizability of results.

In Experiment 3, male and female college students were randomly assigned to watch either a violent or nonviolent film clip that contained

two commercials. There were four violent film clips and four nonviolent film clips that were preselected to be similar on measures of self- reported arousal (i.e., exciting, boring, and arousing ratings). Immediately after viewing the videotape, participants reported their mood, recalled brand names, recalled commercial message details, and recognized advertised brands from competitor brands. They also rated the videotape on several dimensions.

Scores on all three measures of commercial memory (i.e., brand recognition, brand recall, and commercial message recall) were expected to be higher among people who saw a nonviolent video than among people who saw a violent video. Structural equation analysis was used to test whether anger mediates the effect of television violence on commercial memory. Television violence was expected to increase anger levels in viewers, and anger, in turn, was expected to decrease commercial memory. This prediction follows from theories of the relation between mood and memory. The anger induced by violent programming could render commercials less accessible because violent programs activate other aggressive thoughts that interfere with the elabo- ration of the information contained in the commercials or because viewers focus more attention on repairing their negative mood than on the commercials. Arousal and positive affect were not expected to mediate the effect of television violence on commercial memory.

Me~od

Participants


Videotapes

From the pool of 52 tapes used to select videotapes for Experiments 1 and 2, 8 videotapes were selected (4 violent and 4 nonviolent) for use in Experiment 3. 5 The videos were preselected to

5The eight videotapes were taken from the following films: Cobra (male-on-male violence), Die Hard


Table 5 Comparison of the Nonviolent and Violent Videotapes Used in Experiment 3 on Various Rating Dimensions

Selection procedure Experiment proper

Violent Nonviolent Violent Nonviolent videotapes videotapes videotapes videotapes

Measure M SE M SE M SE M SE

Arousal content Exciting 7.0 0.20 6.7 0.20 6.7 0.18 5.9 0.19 Boring 1.8 0.21 2.2 0.21 2.9 0.17 3.6 0,20 Arousing 4.6 0.29 4.8 0.25 4.3 0.20 4.3 0,20

Involving 5.8 0.18 6.1 0.19 Violent 8.9 0.13 0.6 0.13 8.9 0.12 1.6 0.12

Note. The means were adjusted for whether or not judges had seen the movie from which the segment was taken. The nonviolent videos were taken from Awakenings, Chariots of Fire, Field of Dreams, and Never Cry Wolf, and the violent videos were taken from Cobra, Die Hard, Single White Female, and The Hand That Rocks the Cradle. There were no sex differences in ratings for any videotape.

be equally exciting but differentially violent. As can be seen in Table 5, the violent and nonviolent videotapes did not differ on any measure of arousal content. As expected, the violent videotapes were judged to be more violent than the nonviolent videotapes, F(1, 317) = 2,124.49, p < .05, d = 5.16. There were no sex differences in videotape ratings.

The commercials embedded in the film clips were for Krazy Glue and Plax mouth rinse. Both commercials advertised products with broad market appeal, were 30-s long, and were several years old. The Krazy Glue commercial, also used in Experiments 1 and 2, was described earlier. The Plax commercial depicted a man brushing his teeth vigorously. The message stated that plaque was the number one dental problem and that even brushing well could not remove all plaque. The commercial ended by having the announcer state, "Rinsing with Plax before brushing removes 300% more plaque than brushing alone."

Procedure

Participants were tested in small groups, but they worked independently on tasks. The cover

(male-on-male violence), Single White Female (female- on-female violence), The Hand That Rocks the Cradle (female-on-female violence), Awakenings (nonviolent), Chariots of Fire (nonviolent), Field of Dreams (nonviolent), and Never Cry Wolf(nonviolent).

story was the same one used in Experiments 1 and 2. Immediately after viewing the videotape, participants completed a mood form that measured anger and positive affect. The anger measure consisted of 15 adjectives (e.g., angry, annoyed, and furious) from the Hostility subscale of the revised Multiple Affect Adjective Check- list (Zuckerman & Lubin, 1985). The positive affect measure consisted of 10 adjectives (e.g., alert, determined, and enthusiastic) from the Positive Affect subscale of the Positive and Negative Affect Schedule (Watson, Clarke, & Tellegen, 1988). Watson and his colleagues de- fined positive affect as a state of "high energy, full concentration, and pleasurable engagement" (p. 1063). All adjectives were rated along a 5-point Likert-type scale, where 1 = very slightly or not at all, 2 = a little, 3 = moderately, 4 = quite a bit, and 5 = extremely. Participants were told to "Indicate to what extent you feel this way right now, that is, at the present moment." The alpha coefficients for the measures of anger and positive affect were .88 and .89, respectively.

After completing the mood form, participants rated how exciting, boring, arousing, involving, and violent the film was. Several other rating dimensions were added as fillers. The ratings were made along a 10-point Likert-type scale ranging from 1 (not at all) to 10 (extremely). Participants then listed the brand names of products shown in the commercials and listed everything they could remember about the commercial

300 BUSHMAN

messages. Next, participants completed a visual recognition test. Participants were told that the film clip contained two commercials, one for a brand of glue and the other for a brand of mouth rinse. Participants were shown slides of 6 glue brands (Krazy Glue and 5 distractor brands) and 6 mouth rinses (Plax and 5 distractor brands) and were asked to write the brands of the products shown in the commercials. The order of the products and the order of brands within each product were randomized for each session. The slides were shown at the rate of one slide every 5 s.

Next, participants reported the number of hours per week they spent watching various types of television programs, including violent programs. Participants also reported whether they had seen the film from which the clip had been taken and whether they had seen the commercials imbedded within the film clip. Finally, participants were questioned for suspicion and fully debriefed. No participant expressed suspicion about the study.

Results

Order, Commercial, Videotape, and Sex Effects

The presentation order of the two commercials did not influence any of the dependent variables. There were no differences between the two commercials on any of the dependent variables. There were no differences between the four violent videotapes or between the four nonviolent videotapes on any of the dependent variables. There also were no sex main effects or interactions for any of the dependent variables. Thus, the data for the two orders, two commercials, four violent videotapes, four nonviolent videotapes, and two sexes were combined for subsequent analyses.

Exposure Effects

Habitual exposure to televised violence and previous exposure to the film and commercials shown in the videotape were not significantly related to any of the commercial memory measures, either alone or interacting with film content. These exposure variables were therefore excluded from the analyses.

Film Clip Differences in Arousing, Exciting, Boring, and Involving Ratings

An analysis of covariance was used to compare arousing, exciting, boring, involving, and violent ratings for the violent and nonviolent videotapes. Whether participants had seen the movie from which the clip was taken was used as a covariate. There was no difference between the violent and nonviolent videotapes in arousing and involving ratings, Fs(1, 317) = 0.00 and 0.79, respectively, ps > .05 (see Table 5). In contrast with the selection procedure findings, the violent videotapes were judged to be more exciting and less boring than the nonviolent videotapes, F(1, 317) = 9.61, p < .05, d = 0.35 and F(1,317) = 6.74,p < .05, d = 0.29 (see Table 5). Of course, the violent videotapes were judged to be more violent than the nonviolent videotapes, F(1, 317) = 1,738.27, p < .05, d = 4.68 (see Table 5). However, none of the videotape ratings influenced any of the dependent variables, either alone or interacting with film content. The videotape ratings were therefore excluded from subsequent analyses.

Reliability of Coding Message Recall

Two independent raters who were unaware of conditions and experimental hypotheses coded the free recall data. Because the intraclass correlation coefficient (Shrout & Fleiss, 1979) was fairly high ( r i cc= .80), scores from the two raters were averaged for subsequent analyses.


Brand name recognition. Participants who watched a nonviolent video recognized more brands from the slides than did participants who watched a violent videotape, F(1,318) = 5.80, p < .05, d = 0.27 (see Table 6). 6

Brand name recall. Participants who watched a nonviolent video recalled more brand names than did participants who watched a violent

6 Although brand recognition scores only ranged from 0 to 2, log-linear analyses yielded the same pattern of results, X2(2, N = 320) = 6.19, p < .05.




Measure M SE M SE

Brand name recognition 1.65 0.04 1.79 0.05 Brand name recall 1.14 0.06 1 .31 0.05 Commercial message

recall 5.50 0.23 6.28 0.23

Note. Brand name recognition and recall scores could range from 0 to 2. Commercial message recall scores were continuous (scores ranged from 1 to 18).

videotape, F(1, 318) = 4.50, p < .05, d = 0.24 (see Table 6). 7

Commercial message recall. Participants who watched a nonviolent video recalled more commercial details than did participants who watched a violent videotape, F(1 ,318) = 6.01, p < .05, d = 0.27.

Anger as a Mediator in the Effect of Television Violence on Memory for Commercial Messages

Linear Structural Relations (LISREL) VIII (Jtreskog & Stirbom, 1993) was used to test whether anger mediates the effect of television violence on memory for commercial messages. LISREL has at least two advantages over causal modeling procedures based on ordinary least- squares regression. First, it permits the removal of the effects of random measurement error that tend to attenuate path coefficients. Second, it provides a chi-square test of the overall goodness of fit of the model or the extent to which the hypothesized model is able to account for relations among the measured variables.

In specifying the model, an indicator variable was used to represent television violence (violent film clip = 1 and nonviolent film clip = 0). The anger variable consisted of the sum of participants' scores for the 15 hostile adjectives on the mood form. In order to reduce the number of parameters being estimated, coefficient alpha (.88) was subtracted from 1.00 to estimate the error variance in the anger variable. Commercial memory was treated as a latent variable, operationalized using three measured variables: (a) brand

name recognition, (b) brand name recall, and (c) commercial message recall. The variance--eovari- ance matrix used in the LISREL analysis is given in Table 7.

The hypothesized mediation model contained a path from television violence to anger and another path from anger to commercial memory (see Figure 1). The mediation model fit the data extremely well, X2(5, N = 320) = 10.60, p > .05, goodness of fit index (GFI) = .99, comparative fit index (CFI) = .98, root mean square error of approximation (RMSEA) = .059. As can be seen in Figure 1, television violence increased anger, z = 6.64, p < .05, and anger, in turn, impaired memory for commercial messages, z = -4 .71 , p < .05.

An alternative model that included a direct path from television violence to commercial memory was tested. Because the mediation model is nested within this alternative model, a chi- square difference test can be used to compare the two models. The difference test was nonsignifi- cant, ×z(1, N = 320) = 0.08, p > .05. Thus, television violence did not have a direct effect on commercial memory.

Other Mediators in the Effects of Television Violence on Memory for Commercial Messages

LISREL also was used to test whether arousal or positive affect mediate the effect of television violence on memory for commercial messages. Each analysis is described in turn below.

Arousal. In specifying this model, an indicator variable was used to represent television violence (violent film clip = 1 and nonviolent film clip = 0). Arousal was treated as latent variable, operationalized using three measured variables: (a) arousing film ratings, (b) exciting film ratings, and (c) boring film ratings. Commer- cial memory was treated as a latent variable, operationalized using three measured variables: (a) brand name recognition, (b) brand name recall, and (c) commercial message recall. The

7Although brand recall scores only ranged from 0 to 2, log-linear analyses yielded the same pattern of results. However, the effect was not significant, X2(2, N = 320) = 4.48, p > .05.

302 BUSHMAN

Table 7 Variance-Covariance Matrix for Linear Structural Relations Analyses

Measure 1 2 3 4 5 6 7 8 9

1. Anger 104.09 .11" .01 -.02 .04 -.27* -.26* 2. Arousing film rating 2.84 6.14 .41" -.31" .42* -.07 -.08 3. Exciting film rating 0.29 2.47 6.02 -.60" 1.48" .01 -.01 4. Boring film rating -0.55 - 1.89 -3.58 5.90 - .31" .06 .08 5. Positive affect 2.77 7.90 8.92 -5.59 56.34 .02 -.02 6. Brand name recognition - 1.46 -0.09 0.02 0.08 0.08 0.29 .54* 7. Brand name recall -1.93 -0.15 -0.02 0.13 -0.11 0.21 0.51 8. Commercial message recall -7.11 -0.16 0.15 0.16 -0.00 0.50 1.20 9. Television violence 1.83 0.05 0.28 -0.25 -0.59 -0.04 -0.04

--.24* .36* -.02 .04

.02 .23*

.02 -.20* --.00 --.16"

.32* --.13"

.58* --.12" 8.18 --.14"

-0.20 0.25

Note. Variances are on the diagonal and covariances descriptive purposes. *p < .05.

are below the diagonal. Correlations are reported above the diagonal for

variance-covariance matrix used in the LISREL analysis is given in Table 7.

The mediation model contained a path from television violence to arousal and another path from arousal to commercial memory (see Figure 2). The mediation model fit the data extremely well, ×2(13, N = 320) = 20.44, p > .05, GFI = .98, CFI = .98, RMSEA = .042. As can be seen in Figure 2, television violence increased arousal, z = 4.20, p < .05. Arousal did not, however, impair memory for commercial messages, z = -0.80, p > .05.

The alternative model that included a direct path from television violence to commercial memory was tested. This alternative model did not fit the data better than the mediation model, ×2(1, N = 320) = 3.48, p > .05. Once again, television violence did not have a direct effect on commercial memory.

Positive affect. In specifying this model, an indicator variable was used to represent television violence (violent film clip - 1 and nonviolent film clip = 0). The positive affect variable consisted of the sum of participants' scores for the 10 positive affect adjectives on the mood form. In order to reduce the number of parameters being estimated, coefficient alpha (.89) was subtracted from 1.00 to estimate the error variance in the positive affect variable. Commercial memory was treated as a latent variable, operationalized using three measured variables: (a) brand name recognition, (b) brand name recall, and (c) commercial message recall. The variance-covariance matrix used in the LISREL analysis is given in Table 7.

The mediation model contained a path from television violence to positive affect and another path from positive affect to commercial memory

Television I .3S* ,-.1 vi°lence I " l

g= .e~ if= .15 ~= .6o

Brand name I Brand name I Commercial reco0n,ion I I ,'.ca,, I I r.essa0ereoa,,

Anger I -.33* .~/" ~'~'1'2~'~.~*' ~

Figure 1. Anger as a mediator between television violence and memory of commercial messages. *p < .05.


Television violence

~'= .65 ~'= .15 ~'= .60

Brand name ] I Brand name Commercial recognition recall message recall

.26" ~ -.05 / " ,.,u,,,,,,=,,.,,=,, \

Arousing [ I Exciting film rating film rating

Boring [ film rating

T 1 T ¢ = .79 C = .24 ¢ = .52

Figure 2. Arousal as a mediator between television violence and memory of commercial messages. *p < .05.

(see Figure 3). The mediation model fit the data extremely well, X2(5, N = 320) = 7.69, p > .05, GFI = .99, CFI = .99, RMSEA = .041. As can be seen in Figure 3, television violence decreased positive affect, z = -3 .01 , p < .05. Positive affect did not, however, facilitate memory for commercial messages, z = 1.69, p > .05.

An alternative model that included a direct path from television violence to commercial memory was tested. This alternative model did not fit the data better than the mediation model,

X2(1, N = 320) = 1.82, p > .05, As was found in the previous models, television violence did not have a direct effect on commercial memory.

In summary, the LISREL analyses suggest that anger is a more important mediator of the effect of television violence on commercial memory than either arousal or positive affect. Although all of the models fit the data extremely well, only anger was significantly related to commercial memory. Anger explained 11% of the variation in commercial memory, positive affect explained

Television [ -.76* ,,., violence I "-

~' = .68

Brand name recognition

I Positive affect I •

(R 2 = .58) I

(= .08 (= .63

II-°-" I I recall message recall

Figure 3. Positive affect as a mediator between television violence and memory of commercial messages. *p < .05.

304 BUSHMAN

only 2% of the variation in commercial memory, and arousal explained none of the variation in commercial memory.

Discussion

As expected, television violence impaired memory for commercial messages. Participants who watched a violent video had lower scores on all three memory measures (i.e., brand name recognition, brand name recall, and commercial message recall) than did participants who watched a nonviolent video. These findings replicate those obtained in Experiments 1 and 2 using a broader set of film clips and commercial messages.

The primary advantage of Experiment 3 is that it provides some insight about why television violence impairs memory for commercial messages. It is unlikely that the observed memory deficits were due to differential arousal produced by the films. The films did not differ in terms of how arousing and involving they were, and none of the film ratings influenced memory for commercials. Further, arousal was not related to commercial memory (see Figure 2). It also is unlikely that the observed memory deficits were due to positive affect. Although television violence decreased positive affect, positive affect was not related to commercial memory (see Figure 3). The observed memory deficits were due to anger. Television violence increased anger, and anger, in turn, decreased memory for commercial messages (see Figure 1). Television violence had an indirect rather than a direct effect on commercial memory.

General Discussion

In previous research examining the effect of violent programming on commercial memory, program violence has been confounded with other variables that could influence memory, such as arousal. The present research attempted to disentangle these variables by using programs selected to be differentially violent but equally arousing. The present findings suggest that violent programming impairs memory for commercial messages even when the level of program arousal is held constant or is statistically controlled.

The negative effects of television violence on memory for commercial messages is at least partly due to the anger induced by the violent films. Information in commercials is retained to the extent that the viewers rehearse this information in their minds. The angry mood that was induced by the violent media might have acti- vated other hostility-related ideas that interfered with rehearsal of commercial messages. Partici- pants who saw a violent film clip might also have tried to repair their angry moods. During the time that advertisers hope viewers are absorbing their commercial messages, viewers may actually be focusing on themselves, trying to calm the anger brought on by what they have just seen on the screen.

It should be noted that the experiments in this article did not directly test the effect of television violence on purchasing behavior. However, to the extent that a positive purchasing decision is based on the reception of information in commercials, impaired information acquisition and retrieval can only decrease commercial effectiveness (Mun- dorf et al., 1991). It also should be noted that the experiments in this article did not test the effect of television violence on delayed memory for commercial messages.

The commercials used in the present research advertised products with broad market appeal and were relatively neutral in content. No attempt was made to manipulate whether the ads were consistent or inconsistent with the content of the program. One might predict that hostility based ads would be more memorable if placed in violent than in nonviolent programs. Current research in our laboratory is in the process of testing this hypothesis.

The participants in the present research were all college students, and college students do differ from other people in some respects (e.g., Sears, 1986). Therefore, an important question is whether these findings generalize to other groups of people. It is unlikely that memory processes differ between college students and people of most other ages who do not have brain damage, but college students may differ from other people on nonmemory processes. College students may be heavier (or lighter) consumers of violent programming than people of other ages. Habitual exposure to television violence did not, however, influence the results in any of the experiments


reported in this article. Because the effect of television violence on commercial memory in Experiment 3 was mediated by anger, another potential obstacle to generalizing the results of the present research would be a differential propensity on the part of college students (rela- tive to other groups) to react to violent material with anger. Although I know of no research documenting that college students are more (or less) likely than other people to feel anger when watching violent programming, the present findings suggest that if such differences in anger response were found, they would lead to corre- sponding group differences in memory for advertising content. In summary, group differences in the effects of television violence on commercial memory, if they exist, could be either smaller or larger than those obtained for college students in the present research. There is little reason to believe that the present findings are limited to college students. Television violence probably impairs memory for commercial messages in most viewers.

Consider the magnitude of the effect of television violence on memory for commercials. The average effect-size estimate for the three experiments reported in this article was 0.34, with a 95% confidence interval ranging from 0.19 to 0.49 (for a reference on combining correlated effect-size estimates, see Gleser & Olkin, 1994). It is interesting to note that television violence has a similar sized effect on aggressive behavior (Hearold, 1986; Paik & Comstock, 1994; Wood, Wong, & Chachere, 1991). Thus, the effects reported in this article are by no means trivial.

Several years ago, a spokesperson for the J. Walter Thompson Company stated, "The more we probe the issue, the more we are convinced that sponsorship of television violence is poten- tially bad business, as well as a social risk" ("Lousy Frames for Beautiful Pictures," 1977, p. 56). The data from the present experiments are consistent with this conclusion. There are at least three reasons why advertisers might want to think twice about sponsoring violent programs. First, most Americans are upset by the excessive amount of violence shown on television (e.g., Fischer, 1994; TV Guide, 1992; Zipperer, 1994). Second, television violence has a negative impact on society because it increases aggressive behavior in viewers (e.g., Hearold, 1986; Huston et al.,

1992; National Institute of Mental Health, 1982; Paik & Comstock, 1994; Surgeon General's Sci- entific Advisory Committee on Television and Social Behavior, 1972; Wood et al., 1991). Third, violent programs decrease viewers' memories of brand names and commercial messages. Thus, sponsoring violent programs might not be a profitable venture for advertisers.

References

American Psychological Association. (1993). Vio- lence and youth: Psychology's response. Washing- ton, DC: Author.

Anderson, C. A. (1997). Effects of violent movies and trait hostility on hostile feelings and aggressive thoughts. Aggressive Behavior, 23, 161-178.

Berkowitz, L. (1984). Some effects of thoughts on anti- and prosocial influences of media effects: A cognitive-neoassociation analysis. Psychological Bulletin, 95, 410-427.

Bettman, J. R. (1979). Memory factors in consumer choices: A review. Journal of Marketing, 43, 37-53.

Bower, G. (1981). Mood and memory. American Psychologist, 36, 129-148.

Bryant, J., & Comisky, P. W. (1978). The effect of positioning a message within differentially cogni- tively involving portions of a television segment on recall of the message. Human Communications Research, 5, 63-75.

Burke, A., Heuer, E, & Reisberg, D. (1992). Remem- bering emotional events. Memory and Cognition, 20, 277-290.

Bushman, B. J. (1995). Moderating role of trait aggressiveness in the effects of violent media on aggression. Journal of Personality and Social Psy- chology, 69, 950-960.

Bushman, B. J., & Geen, R. G. (1990). Role of cognitive-emotional mediators and individual differences in the effects of media violence on aggression. Journal of Personality and Social Psychology, 58, 156--163.

Christiansen, S.-A. (1992). Emotional stress and eye- witness memory: A critical review. Psychological Bulletin, 112, 284-309.

Collins, A., & Loftus, E. (1975). A spreading- activation theory of semantic memory. Psychologi- cal Review, 82, 407--428.

Craik, E I. M., & Lockhart, R. S. (1972). Levels of processing: A framework for understanding memory

306 BUSHMAN

research. Journal of Verbal Learning and Verbal Behavior, 11, 671--684.

Diener, E., & DeFour, D. (1978). Does television violence enhance program popularity? Journal of Personality and Social Psychology, 36, 333-341.

Diener, E., & Woody, L. W. (1981). Television violence, conflict, realism, and action: A study in viewer liking. Communication Research, 8, 281- 306.

Drevitch, G. (1994, February 11). Murder, she saw. Scholastic Update, 12-13.

Fischer, R. L. (1994, July). Is it possible to regulate television violence? USA Today, 72-75.

Gleser, L. J., & Olkin, I. (1994). Stochastically dependent effect sizes. In H. Cooper & L. V. Hedges (Eds.), The handbook of research synthesis (pp. 339-355). New York: Russell Sage Foundation.

Hansen, C. H., & Hansen, R. D. (1990). The influence of sex and violence on the appeal of rock music videos. Communication Research, 17, 212-234.

Hearold, S. (1986). A synthesis of 1043 effects of television on social behavior. In G. Comstock (Ed.), Public communication and behavior (Vol 1, pp. 65-133). New York: Academic Press.

Huston, A. C., Donnerstein, E., Fairchild, H., Fesh- bach, N. D., Katz, P. A., Murray, J. P., Rubinstein, E. A., Wilcox, B. L., & Zuckerman, D. (1992). Big world, small screen: The role of television in American society. Lincoln: University of Nebraska Press.

Isen, A. M. (1984). Toward understanding the role of affect in cognition. In R. Wyer, Jr. & T. Srull (Eds.), Handbook of social cognition (pp. 179-236). Hillsdale, NJ: Erlbaum.

Isen, A. M., Clark, M., Shalker, T. E., & Karp, L. (1978). Affect, accessibility of material in memory and behavior: A cognitive loop? Journal of Person- ality and Social Psychology, 36, 1-12.

J6reskog, K. G., & S6rbom, D. (1993). Windows LISREL 8.12. Chicago: Scientific Software.

Lousy frames for beautiful pictures: Are they chang- ing? (1977, May 23). Advertising Age, 48, 56.

Mayer, J. D., McCormick, L. J., & Strong, S. E. (1995). Mood-congruent memory and natural moods: New evidence. Personality and Social Psy- chology Bulletin, 21, 736-746.

Mundorf, N., Zillmann, D., & Drew, D. (1991). Effects of disturbing televised events on the acquisition of information from subsequently presented commercials. Journal of Advertising, 20, 46-53.

Mustonen, A. (1997). Nature of screen violence and its

relation to program popularity. Aggressive Behav- ior, 23, 281-292.

Mustonen, A., & Pulkkinen, L. (1993). Aggression in television programs in Finland. Aggressive Behav- ior, 19, 175-183.

National Institute of Mental Health. (1982). Television and behavior: Ten years of scientific progress and implications for the eighties (Vol. 1), Summary Report. Washington, DC: U.S. Government Print- ing Office.

National Television Violence Study. (1996). National television violence study (Vol. 1). Thousand Oaks, CA: Sage.

National Television Violence Study. (1997). National television violence study (Vol. 2). Studio City, CA: Mediascope.

Oliver, M. B. (1994). Portrayals of crime, race, and aggression in "reality-based" police shows: A content analysis. Journal of Broadcasting and Elec- tronic Media, 38, 179-192.

Paik, H., & Comstock, G. (1994). The effects of television violence on antisocial behavior: A meta- analysis. Communication Research, 21, 516--546.

Parrott, G. W. (1993). Beyond hedonism: Motives for inhibiting good moods and for maintaining bad moods. In D. M. Wegner & J. W. Pennebaker (Eds.), Handbook of mental control. Englewood Cliffs, NJ: Prentice Hall.

Pavelchak, M. A., Antil, J. H., & Munch, J. M. (1988). The Super Bowl: An investigation into the relation- ship among program context, emotional experience, and ad recall. Journal of Consumer Research, 15, 360-367.

Sears, D. O. (1986). College sophomores in the laboratory: Influences of a narrow data base on social psychotogy's view of human nature. Journal of Personality and Social Psychology, 51, 515-530.

Schwarz, N., & Clore, G. (1983). Mood, misattribu- tion, and judgments of well-being: Informative and directive functions of affective states. Journal of Personality and Social Psychology, 45, 513-523.

Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86, 420--428.

Singh, S. N., & Churchill, G. A. (1987). Arousal and advertising effectiveness. Journal of Advertising, 16, 4-10, 40.

Singh, S. N., & Rothschild, M. L. (1983). Recognition as a measure of learning from television commercials. Journal of Marketing Research, 20, 235-248.

Sprafkin, J. N., Rubinstein, E. A., & Stone, A. (1977).


The content analysis of four television diets (Occa- sional Paper 77-3). Stony Brook, NY: Brookdale International Institute.

Surgeon General's Scientific Advisory Committee on Television and Social Behavior. (1972). Television and growing up: The impact of televised violence. Washington, DC: U.S. Government Printing Office.

TV Guide. (1992, October 10-16). TV Guide poll: Would you give up TV for a million bucks? TV Guide, 40, 10-13, 15, 17.

Watson, D., Clarke, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063-1070.

Wober, M. (1988). The extent to which viewers watch violence-containing programs. Current Psychol- ogy: Research and Reviews, 7, 43-57.

Wood, J. V., Saltzberg, J. A., & Goldsamt, L. A. (1990). Does affect induce self-focused attention? Journal of Personality and Social Psychology, 58, 899-908.

Wood, W., Wong, E Y., & Chachere, J. G. (1991). Effects of media violence on viewers' aggression in unconstrained social interaction. Psychological Bul- letin, 109, 371-383.

Worchel, S., Hardy, T. W., & Hurley, R. (1976). The effects of commercial interruption of violent and nonviolent films on viewers' subsequent aggression. Journal of Experimental Social Psychology, 12, 220--232.

Zipperer, J. (1994, February 7). Violence foes take aim: Advertisers and affiliates caught in the cross- fire. Christianity Today, 40--42.

Zuckerman, M., & Lubin, B. (1985). Manual for the MAACL-R: The Multiple Affective Adjective Check- list Revised. San Diego, CA: Educational and Industrial Testing Service.

Received September 15, 1997 Revision received April 17, 1998

Accepted April 28, 1998 •

New Editors Appointed, 2000-2005

The Publications and Communications Board of the American Psychological Associa- tion announces the appointment of three new editors for 6-year terms beginning in 2000.

As of January 1, 1999, manuscripts should be directed as follows:

For Experimental and Clinical Psychopharmacology, submit manuscripts to Warren K. Bickel, PhD, Department of Psychiatry, University of Vermont, 38 Fletcher Place, Burlington, VT 05401-1419.

For the Journal of Counseling Psychology, submit manuscripts to Jo-Ida C. Hansen, PhD, Department of Psychology, University of Minnesota, 75 East River Road, Minneapolis, MN 55455-0344.

For the Journal of Experimental Psychology: Human Perception and Perfor- mance, submit manuscripts to David A. Rosenbaum, PhD, Department of Psy- chology, Pennsylvania State University, 642 Moore Building, University Park, PA 16802-3104.

Manuscript submission patterns make the precise date of completion of the 1999 volumes uncertain. Current editors, Charles R. Schuster, PhD; Clara E. Hill, PhD; and Thomas H. Carr, PhD, respectively, will receive and consider manuscripts through December 31, 1998. Should 1999 volumes be completed before that date, manuscripts will be redirected to the new editors for consideration in 2000 volumes.

effects of television violence on memory for …bbushman/bjepa98.pdfcannot remember the brand ......

Documents