The role of bleeps and warnings in viewers' perceptions of on-air cursing.

This study used a 2 x 2 x 2 design to examine the effects of warning labels, bleeping, and gender on viewers' perceptions and enjoyment of a docu-drama. We also examined the individual difference variable of verbal aggressiveness to test for possible interactions. Overall, the warning labels

increased enjoyment of the program containing profanity among college students. Bleeping had no effect on either program liking or perceptions of realism; however, bleeping decreased perceptions of the program's offensiveness, and increased viewers' perceptions of profanity frequently estimates. Lastly, verbally aggressive participants perceived the program as more realistic, and the language as less offensive

**********

The Telecommunications Act of 1996 mandated that broadcasters in the United States adopt program age and content ratings in order to help viewers make program viewing decisions. Since then, not only have program ratings and warning labels become a more familiar sight on television programs, but the very content that viewers are being warned about (e.g., profanity) appears to be occurring more frequently (Bauder, 2002). Research has examined the effect of age and content ratings and warning labels on children's program liking and perceptions of content, and found that in some circumstances, warnings and ratings have effects opposite those intended by the legislators (see, for example, Cantor & Harrison, 1996). Considerably less research has examined the effect on adults' perceptions of content (however, see Bushman, 1997). Furthermore, a majority of the research examines the effects of ratings and warnings on violent content (e.g., Cantor & Harrison, 1996; Cantor, Harrison, & Nathanson, 1998; Herman & Leyens, 1977) or educational content (Krcmar & Albada, 2000). Little, if any research has examined the effect of ratings and warnings on attitudes toward, and perceptions of, other potentially objectionable material such as cursing. It may be interesting to ask, therefore, if assigning a warning label affects how adult viewers interpret and recall cursing when it appears in television programs.

In addition to the use of program warnings, there are increasing numbers of cable channels available to viewers which may have served to loosen some norms regarding appropriate standards for programming. For example, in a recent airing of A Season on the Brink, over 100 instances of curse words (including those restricted by most network television; e.g., goddamn) were recorded. Another airing of the same program self-censored (or "bleeped") the profanity. Therefore, although program ratings have allowed viewers greater opportunity to filter their viewing, lowered norms of self-censorship have perhaps broadened the material available to them. These less restrictive norms, on one hand, and more frequent labeling of material, on the other, lead us to ask how labels and self-censoring affect the enjoyment and perceptions of program content by viewers. Therefore, the current study uses an experimental design to examine the effects of program labels and self-censoring on the perceptions and enjoyment of viewers. First, we will discuss the possible effects of warnings and self-censoring (bleeping) on viewers' enjoyment and perceptions of the content. Next, we will consider possible individual differences in viewers' responses to the manipulations.

The Effects of Program Ratings and Warning Labels

Although few studies on program warnings had been conducted prior to the 1990s (see, however, Austin, 1980; Herman & Leyens, 1977; Wurtzel & Surlin, 1978), the current use of program warnings and ratings by the networks, combined with the increase in parental concern about problematic media content (Cantor, Stutman & Duran, 1996), has created an environment in which research into this issue has become more prevalent. A majority of these studies is grounded in reactance theory (Brehm, 1981). Reactance theory states that people are motivationally aroused when they believe that their freedom or decision-making power is in some way threatened. To deal with this threat, people are motivated to regain their freedom, derogate the restrictive agent, or enhance the perceived value of the restricted item. Based on this premise, it seems likely that restrictive program ratings (e.g., TV-14 for an 11-yearold) and program warnings, which may be perceived by a viewer as limiting their decision-making power, to some extent, could generate reactance in the viewer. In other words, restrictive labels could enhance viewer enjoyment.

Consistent with reactance theory, several studies have found that restrictive labels and warning labels do enhance viewers' enjoyment or selection of those programs (Bushman, 1997), especially for boys (Cantor & Harrison, 1996) and older children (Cantor & Harrison, 1996). In general, the research on viewers' responses to labels seems to suggest that children experience reactance to restrictive labels, especially in cases where the rating restricts their viewing (e.g., PG-13 for a child of 11). Reactance occurs less for children when simple labels are used (e.g., "V"), most likely because no restriction is implied. Therefore, reactance seems to occur with older children; however, some studies have also found support for reactance among adults (for counter evidence, see Austin, 1980; Christenson, 1992; Wurtzel & Surlin, 1978).

For example, Herman and Leyens (1977) examined adult viewing patterns based on 4 years of Belgian television movies. Their main finding was that films carrying violence or sex advisories had larger audiences than those that did not. Their results, then, may seem to support reactance theory. However, because the study was archival, not experimental, movies that received advisories were systematically different from movies that did not have advisories. This study may have demonstrated only that violent or sexual content with or without advisories attracts viewers.

More valid evidence comes from a series of studies by Bushman and colleagues (Bushman & Stack, 1996; Bushman, 1997). Arguing that warning labels may also cause reactance, especially in adult viewers, Bushman and Stack (1996) used made-for-television movies and systematically varied the warning labels. When asked how much they liked the movies, audiences showed a preference for movies with warnings. This result was especially true among those participants who were initially high in reactance. In a second study, Bushman (1997) compared the responses of participants who had read identical movie descriptions with either a content rating, a warning label, or no label. Overall, adult participants showed greater preference for the movie descriptions with a warning. Therefore, it appears that even among adult audiences, warning labels can affect enjoyment and selective exposure.

Therefore, it appears that while age ratings can affect children, warning labels may encourage reactance in adults. In accord with reactance theory, the age-based (e.g., TV-14) rating information can induce higher levels of enjoyment from the programming content among the children who are younger than the specified age group rating because it implies prohibition. With adult viewers, on the other hand, even the most stringent labels (e.g., TV-MA) do not restrict them. Therefore, for adults, it is the warning, not the rating, that implies restriction. By suggesting that the material may be too offensive for some viewers, warnings may imply that decisions should be made with caution, thereby limiting their freedom. Therefore, we predict:

H1: There will be a main effect for warning labels and self-censoring, such that those in the warning statement condition and the self-censored condition will like the program more than those in the no warning condition or uncensored condition.

In addition to examining the effect of the manipulations on program liking, we also investigated the role of warnings and self-censorship on perceived appropriateness of the program content. In the past, profanity was relatively uncommon on television. For example, there was less than one use of rough -- not even necessarily profane--language per prime-time hour on all the broadcast networks during the 1989-1990 TV season. By the 1999-2000 season, there were nearly five per hour, according to a study by the Parents Television Council (2002). Although audiences claim that this kind of realism is acceptable in some instances, such as adult programming in appropriate time slots and in documentaries (ASA, BBC, BSC, & ITC, 2000), it is unclear how audience members actually respond.

In a particularly vivid example of this debate, the weekend of March 9-10, 2002 saw the airing of two highly controversial programs. Several profanities were included in the footage on the CBS documentary about the World Trade Center attack, and ESPN's first-ever original movie, A Season on the Brink, contained numerous swear words in its docudrama about basketball coach Bobby Knight. Although the decisions to include the profanities were debated within the two networks, ultimately both decided to include them. Mark Shapiro, an ESPN senior vice president, stated that the decision to include the profanities was made in order to increase the realism of the program. He stated, "You couldn't produce a movie on Bobby Knight and use phrases like aw, shucks and golly gee. It just wouldn't be believable" (Bauder, 2002). Interestingly, ESPN aired an identical version with the curse words bleeped, indicating, at the very least, that the network recognized that some viewers might prefer the self-censored version.

Despite these controversies, little, if any research has examined the effect of cursing on viewers' perceptions of offensiveness in language and realism. Although the British Broadcasting Corporation (ASA et al., 2000) has tracked audiences' views of various expletives since 1998, no experiments have compared audience perceptions of programs with the expletives included compared to those with the expletives self-censored. In addition, the warnings themselves might affect how the expletives are viewed. Therefore, we ask:

RQI: Is there a main effect for the self-censoring of cursing, such that the uncensored version is perceived as more offensive by participants than the bleeped version?

RQ2: Is there a main effect for the self-censoring of cursing such that the uncensored version is perceived as more realistic by participants than the bleeped version?

In addition to examining the effects of warnings and bleeping on enjoyment and perceptions, we were also interested in how people processed material that, first, contains profanity and, second, contains a warning about that profanity. In the following section, we take an information-processing approach in order to explore how audiences respond to these variations in content.

Information Processing of Television Content

The information processing approach assumes that the mind attends to, uses, makes sense of, and organizes information systematically. The approach is "mechanistic but poses [problems] at the functional level" (Pashler, 1999, p. 6). In other words, researchers examine, in detail, the behaviors of individuals, varying the conditions systematically, in order to speculate about how information is processed. One aspect of information processing that has been studied extensively is that of attention. For example, it is now well accepted that various features of auditory stimuli (e.g., change in pitch, a novel tone) make attention more likely. In fact, these auditory variations are actually attended to automatically (Lang, Geiger, Strickwerda & Sumner, 1993; Underwood & Moray, 1971). Even when directed to ignore information heard over one channel, listeners tend to process the "ignored" information when it changes pitch (Treisman, 1964).

Therefore, it seems likely that any variations in pitch or tone might actually draw attention, even involuntarily, from the listener. How is this related to media presentations of profanity? It is common practice to audibly self-censor or "bleep" profanity. In fact, A Season on the Brink contained more than 100 instances of "bleeping" in its self-censored version. It may be, then, that bleeping curse words actually serves to draw attention to the cursing. Although the viewer may not hear the curse per se, they may attend even more to the language being used. Furthermore, in an experimental setting, prior warning also serves to increase the amount of attention paid to a stimulus (Lang et al., 1993). Therefore, it seems likely that a prior warning, such as a viewer advisory, may serve to increase the amount of attention paid to the profanity used. We predict:

H2: There will be a main effect for bleeping, such that those in the bleeped condition will perceive more frequent use of profanity than those in the non censored condition.

H3: There will be a main effect for warnings such that those in the warning condition will perceive more frequent use of profanity than those in the no warning condition.

Despite the fact that some processing is automatic (Cohen, Dunbar & McClelland, 1990), as argued earlier, it is also possible that some individual differences exist that are based on environmental or attitudinal differences. For example, a loud noise that may startle someone who typically spends a great deal of time in a quiet environment (a librarian, for example) might go ignored by someone accustomed to loud noise (a factory worker, for example). Similarly, profanity may typically be processed automatically and ignored by someone who hears it frequently, but not by someone who is unaccustomed to it. In the case of the latter example, we might expect that for those who use profanity themselves and are therefore largely inured to it, self-censoring curse words could draw even greater attention to the cursing itself. One individual difference variable that measures use of profanity is verbal aggressiveness (Infante & Rancer, 1996). Verbal aggressiveness may play a role in determining just who is affected by hearing profanity and who is not. In other words, processing may be automatic in some cases, but under certain conditions, or for certain people (i.e., those high in verbal aggression), the process may be brought to conscious awareness.

The personality trait of verbal aggressiveness has been defined as a stable trait that reflects a person's tendency to attack "a person's self-concept to deliver psychological pain" (Infante & Rancer, 1996, pp. 315-316). Furthermore, several of the items that measure verbal aggressiveness tap into an individuals' tendency to use profanity and insulting and strong language. Verbal aggressiveness is considered a highly destructive form of communication behavior (Infante, 1995) and is designed to threaten and to intentionally inflict harm. Research focusing on verbal aggressiveness has shown that it damages the receiver's self-esteem and serves as a catalyst to interspousal violence (e.g., Infante, Chandler, & Rudd 1989). Thus, verbal aggressiveness as a personality characteristic predicts a variety of behaviors from alcohol consumption (see Graham, Schmidt, & Gillis, 1996; Moss & Kirisci, 1995; Rohsenow & Bachorowski, 1984) to physical aggressiveness and substance use (Bukstein, 1996). It seems likely, therefore, that those who are verbally aggressive and who themselves use profane language, may be less likely to attend to and be offended by it. In fact, those who are high in verbal aggressiveness, as compared to their less verbally aggressive counterparts, may perceive fewer instances of profanity in the uncensored condition and find it both more appropriate and more realistic. Therefore:

H4: There will be an ordinal interaction between verbal aggressiveness and the self-censorship manipulation on viewer perceptions of how frequently profanity is used, such that those low in verbal aggression in the non censored condition will perceive more profanity, more offensiveness, and less realism than those high in verbal aggression.

Methods

Participants

Participants were 208 undergraduate students (92 men, 116 women) at a large northeastern university. They received extra course credit in exchange for their voluntary participation. They ranged in age from 17 to 25 (M = 19.5, SD = 1.38).

Procedures

After informed consent was obtained, participants watched the ESPN program, A Season on the Brink, about basketball coach Bobby Knight. It was edited to be 30 minutes long (the original movie was 90 minutes) and contained 76 incidents of curse words being spoken (or bleeped). Participants were randomly assigned to watch one of four versions of the movie. The four versions were created by crossing two factors: Warning Label ("Due to strong adult language, viewer discretion is advised" versus no warning), and Censorship (swear words spoken versus bleeped out). While watching the movie, participants were asked to avoid interacting with others in order to eliminate possible effects of co-viewing; no interaction during viewing or questionnaire completion was observed.

Edited clips were used in the experiment instead of showing the entire movie in order to keep the experiment to a manageable length of time and to prevent fatigue. The clips were edited so that they would be identical in terms of scenes chosen. Scenes were chosen that both focused on the storyline about the relationship between Bobby Knight and the players and that contained many swear words. In other words, the edited 30-minute clip was a condensed version of the whole movie and retained the natural flow of the storyline. The manipulated material was based on a story of a well-known basketball coach. Therefore, there was a concern for the possible confounding effect of participants' familiarity and knowledge of Bobby Knight on perceived realism and perceived likeability. In other words, if a participant is already familiar with the real life story of how Knight conducted his coaching on and off the court and how and why he was fired from Indiana University, then familiarity and knowledge in the subject matter may affect how realistic the individual perceives the movie to be, regardless of the use of language, which is the focus of this study. Meanwhile, if a participant is already familiar with the real life story of Knight, which indicates his/her interests in college basketball, this could influence how likable s/he will find the movie. In order to control for this possible influence, several items measured the participants' familiarity with Bobby Knight's story and college basketball. These items were used as controlling variables in analyses, as explained below.

Moreover, the manipulated material was taken from a televised movie. Thus 29% of participants (60 out of total 208) had already seen the movie before they were exposed to the manipulation in the experiment. Two strategies were undertaken to deal with this issue. First, familiarity (i.e., having seen the movie previously) was used as a covariate in all analyses, as discussed. Second, additional analyses were done to consider this limitation. T-tests were conducted to investigate the difference between those with and without previous exposure, in terms of the variables of interest (i.e., perceived likeability, realism, language offendedness, perceived frequency of profanity, and verbal aggressiveness). Significant differences were found in language offendedness and verbal aggressiveness. Those with previous exposure rated the language less offensive (M = 2.56, SD = .94) than those without previous exposure (M = 2.97, SD = 1.01), t (202) = 2.70, p < .01. Also, those with previous exposure (M = 2.64, SD = .56) were more verbally aggressive than those without previous exposure (M = 2.32, SD = .53), t (202) = -3.79, p < .01.

Measures

Participants completed a questionnaire about the program after viewing it. The first 23 questions were examined using a principal component factor analysis with varimax rotation. The resulting four factors were re-examined through confirmatory factor analysis (CFA) to check for secondary loadings of items. The final factors came out as follows. The first factor (eigenvalue = 8.08 with 38.5% of variable explained; [alpha] = .92) contained 10 items and measured how much participants liked the movie (e.g., "I enjoyed watching the movie"). The second factor (eigenvalue = 1.52 with 7.2% of variable explained; [alpha] = .79) contained 3 items and measured how realistic participants thought the movie was (e.g., "The movie was realistic"). The third factor (eigenvalue = 1.99 with 9.5% of variable explained; [alpha] = .89) contained 3 items and measured how offensive participants thought the language was (e.g., "I found the language used in the movie offensive"). The fourth factor (eigenvalue = 3.20 with 15.2% of variable explained; [alpha] = .95) contained 4 items and measured how familiar participants were with the movie's story (e.g., "I am familiar with Bobby Knight's story in general"). The fourth factor was used as a covariate in all analyses. Questions were scored using a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree).

Perceived frequency of profanity was measured by asking participants to estimate the number of times they heard either swear words or bleeps. The question reads, "Please estimate how many times you heard either curse words/swear words or 'bleeped out' words in the movie you just watched (The curse/swear words here include: luck, bitch, God, pussy, ass, asshole, prick, whore, shit, and Christ)." The mean estimation was 84.7 (the actual number of swear words or bleeps used in the movie was 76), and ranged from 0 to 500 (SD = 73.5). Interestingly, seven participants reported "0" in their estimation; they all were from the experimental conditions in which uncensored versions were played. In addition, two reported 500 and one reported 350; neither of these estimates fell within the normal distribution. Therefore, those who were extreme statistical outliers or whose answers seemed somewhat incongruous (e.g., 0, 350, or 500) were eliminated from the analyses in which frequency estimates were used.

Verbal aggression was measured by utilizing an established 10-item, 5-point scale (strongly disagree to strongly agree) (Infante & Rancer, 1995). Five of the items were reverse coded to maintain directional consistency. The overall reliability of Cronbach's [alpha] for the 10-item scale was .82.

Results

In order to examine the effects of warnings and self-censorship on the dependent variables of interest, the first hypotheses and research questions were tested by conducting a 2 x 2 x 2 ANCOVA with warning label, self-censorship, and gender as the primary factors. Familiarity with Bobby Knight was used as the covariate, and liking, offendedness, perceived realism, and perceptions of profanity frequency were used as the dependent variables.

Program Liking

As expected, participants liked the movie more when it contained a warning label than when it contained no label, Ms = 3.80 and 3.64 respectively, F(1,204) = 4.80, p < .05. Participants also liked the movie more if they were familiar with A Season on the Brink, F (1, 196) = 32.86, p < .01. Moreover, self-censoring did affect program liking in that those in the self-censoring condition liked the movie more, Ms = 3.76 and 3.64 respectively, F (1, 204) = 3.47, p = .06. Therefore, warning labels and self-censoring both affected program liking.

Offendedness

As expected, participants thought the movie was more offensive when the swear words were spoken than when they were bleeped out, Ms = 2.98 and 2.68 respectively, F (1, 196) = 7.24, p < .01. Women thought the movie was more offensive than did men, Ms = 3.18 and 2.40 respectively, F (1, 196) = 16.31, p < .01.

Program Realism

Contrary to the point raised by ESPN's senior vice president, the presence of profanity did not increase viewer perceptions of realism, F(1, 196) = 0.75, p = .39. There was a main effect for participant sex, F(1, 196) = 5.39, p < .05. However, the main effect was qualified by an interaction between self-censorship and participant sex, F (1, 196) = 7.59, p < .01. Women gave the program higher realism ratings when swear words were spoken rather than bleeped, Ms = 4.25 and 3.98 respectively, F(1, 115) = 7.77, p < .01. Men gave the program similar ratings regardless of whether swear words were spoken or bleeped, Ms = 4.01 and 4.14 respectively, F (1, 90) = 0.86, p = .36. Participants also thought the movie was more realistic if they were familiar with A Season on the Brink, F(1,206) = 7.25, p < .01. No other effects were significant.

Estimates of Profanity Usage

Perceived frequency of profanity tended to be higher when swear words were bleeped, (M = 88.45, SD = 58.17), than when they were spoken, (M = 73.54, SD = 43.73), F(1, 192) = 4.00, p < .05. Therefore, Hypothesis 2 is supported. However, no other effects approached significance. Therefore, Hypothesis 3 is unsupported and the warning label had no effect on perceived frequency of profanity.

In order to answer Hypothesis 4, which predicted an ordinal interaction between verbal aggressiveness and the self-censorship manipulation on 1) viewer perception of profanity frequency, 2) perceived language offensiveness, and 3) perceived realism, a series of hierarchical multiple regression analyses was conducted with three different dependent variables. First, the "verbal aggression" and "presence of censorship" variables were entered in the first block to filter out the main effects from the interaction, and the newly created interaction term was entered in the second block. The perceived frequency of profanity was used as a dependent variable in the first hierarchical multiple regression analysis. The first block accounted for less than 3% of total variance and approached significance, F (2, 194) = 2 .07; p = .06, however, there were no significant interaction effects found in the second block. Specifically, the presence of self-censoring had a significant positive effect on the perceived frequency of profanity ([beta] = .16, p < .05) as reported in research question

1. However, there were no interactions between censoring and verbal aggressiveness.

The second hierarchical multiple regression analysis was conducted with offendedness as the dependent variable. The first block accounted for 16.5% of total variance and was significant, F (2, 202) = 19.96; p < .01 ; however, there were no significant interaction effects in the second block. Only the main effects indicated that those lower in verbal aggression ([beta] = -.38, p < .01) and those who heard profanity ([beta] = -.19, p < .01) were more offended.

The third hierarchical multiple regression analysis was conducted with perceived realism as dependent variable. Again, there was no significant interaction between verbal aggression and presence of censoring on perceived realism. Thus, Hypothesis 4 was not supported. In sum, there were no interaction effects between verbal aggressiveness and censoring on perceived frequency of profanity, offendedness, and perceived realism. Instead, verbal aggressiveness had a main effect with those higher in verbal aggression less offended by the profanity. Those who were more verbally aggressive also found the movie to be more realistic.

Discussion

Summary of Findings

Several results of this study have practical implications for decisions made by broadcasters about on-air cursing. First, the presence of a warning statement did have a significant effect. Those in the warning statement condition found the movie more enjoyable. Second, those who heard the curse words spoken were more offended by the movie while they enjoyed the movie less. Third, the presence of profanities did not increase viewers' perception of realism. In fact, bleeping served to increase their estimates of profanity, an effect likely due to the attention called to the words.

However, program features were not the only variables to affect viewer enjoyment and perceptions. Specifically, those high in verbal aggressiveness perceived the programming content as more realistic than others. Verbal aggression was also found to be a negative predictor of offendedness; those with higher levels of verbal aggression perceived the language used in the movie as less offensive.

Theoretical and Practical Implications

It seems that bleeping was a win-win situation for ESPN. Although bleeping increased viewers' perceptions of the amount of cursing, it did not have a negative effect on their perceptions. Furthermore, bleeping served to increase the enjoyment of the movie. Contrary to the point raised by ESPN's senior vice president, cursing did not make the program more realistic to viewers in general. Therefore, when making program decisions, it seems that bleeping is an acceptable approach to dealing with on-air cursing. Bleeping doesn't seem to decrease enjoyment but it does seem to make the movie less offensive.

Therefore, self-censorship may appeal to some viewers--for example, those with young children. It does not, however, seem to hinder the enjoyment of viewers in general. Although it is interesting that those high in verbal aggression found the movie to be more realistic and less offensive, it is worth noting that this variable did not interact with the main manipulations. Simply put, those high in verbal aggression found the movie to be more realistic and less offensive, whether the cursing was bleeped out or not.

From a theoretical perspective, we argue that the findings are consistent with reactance theory (Brehm, 1966). Warning labels can serve to increase program liking by causing viewers to feel that their decision-making is being restricted in some way. Furthermore, self-censoring, or bleeping as it is commonly called, appears to affect information processing. Consistent with a classic information processing perspective (Treisman, 1964), bleeping calls attention to language that might otherwise be processed somewhat automatically (Underwood & Moray, 1971).

This study also contributes to theory regarding verbal aggression. Theory and research on the construct finds that verbal aggression is associated with a person's tendency to deliver psychological pain (Infante & Rancer, 1996). Verbally aggressive people are also more likely to engage in verbal and physical aggression towards others (e.g., a spouse) (Infante et al., 1989). Perhaps due in part to this communicative style, this study found that those who are more verbally aggressive find profanity less offensive and enjoy it more. Furthermore, they found the program more realistic. Why would this be the case? Recall that there was no interaction between the manipulations and verbal aggression on viewer perceptions. Whether profanity was included or bleeped out, more verbally aggressive people enjoyed and found realistic the content of the movie. Therefore, it is not the profanity per se, that verbally aggressive individuals find convincing. Rather, it seems that conflict itself, which is a main theme in the movie, could be consistent with their own experiences. These parallels may serve to increase identification with the movie and the resulting perceptions of realism.

Limitations and Future Research

Despite the findings of this study, several limitations exist, namely the contaminated participants who were either already familiar with Bobby Knight's story, or who had seen the televised movie before the experiment. The issue of the contaminated participants was dealt with by using two approaches. First, familiarity with Bobby Knight's story was used as a covariate. Second, t-tests were conducted to determine if there was any significant difference between those who were exposed to the manipulated material prior to the experiment and those who were not.

As presented in the method section, the results of the t-tests suggest that those who are verbally aggressive and those who are less offended by profanity are more likely to seek out material such as A Season on the Brink. Second, there was no difference between those two groups in terms of other dependent variables (i.e., liking, realism or perceived frequency of profanity), which were, in the present study, affected by the manipulation. Therefore, it can be concluded that although the participants with previous exposure were included in the study, their inclusion was not likely to alter the results to a great extent. Nevertheless, it is important to keep this limitation in mind when considering the results.

The second limitation, as is the case with most experimental research, is that it is possible that neither the viewing situation (a laboratory) nor the college sample can be generalized beyond this study. In terms of the college sample, it is possible that because younger people use profanity more than older individuals, the results would not be similar if a more representative sample were used. Perhaps older people would prefer the self-censored version even more than the college students, it is also possible that they would notice the profanity more than the students who may generally be more inured to it. In any case, it may be interesting to explore how bleeping and warnings are perceived by a more general audience. In terms of the laboratory situation, although efforts were made to minimize any contamination due to co-viewing, which would limit this study's external validity, it is possible that effects in other situations (the living room, or a bar, for example) may be markedly different. Future research should be done to replicate these results in various settings.

Lastly, it might be an interesting to investigate how less offensive language (e.g. "jerk" rather than "asshole") affects perceptions of a movie in terms of, for example, perceived realism or offendedness. Because the profanity used in this study did not increase perceptions of realism, it is important to know what specific combination of language use, warnings, and bleeping might serve to attract and hold some audiences, without deterring others. As always, empirical research can aid in understanding how audiences respond to and interpret program content.

Appendix

Items used to measure variables (5-point scale, ranging from 1 = strongly disagree to 5 strongly agree)

1. I am familiar with Bobby Knight's story in general.

2. I enjoyed watching the movie.

3. I was interested in watching the movie.

4. The movie was believable.

5. I am familiar with the movie shown.

6. I found the language used in the movie offensive.

7. I was absorbed by the movie.

8. The movie was stimulating.

9. I am familiar with Bobby Knight's coaching career

10. I liked watching the movie.

11. The movie was exciting.

12. The language used in the movie bothered me.

13. The movie was involving.

14. I am familiar with the story portrayed in the movie.

15. The movie was interesting.

16. I want to follow up the story of the movie.

17. The movie was realistic.

18. It was an entertaining movie.

19. The movie was true-to-life.

20. I found the language used in the movie vulgar.

21. The movie was arousing.

22. I think movies with curse words should be edited to beep out the offensive words.

23. I think movies with curse words should be allowed instead of being bleeped out.

References

Advertising Standards Authority, British Broadcasting Corporation, Broadcasting Standards Commission, and Independent Television Commission. (2000, December). Delete expletives? Retrieved May 17, 2004, from http://www.asa.org.uk/research/documents/ upl_10.pdf

Austin, B. A. (1980). The influence of MPAA's film-rating system on motion picture attendance: A pilot study. The Journal of Psychology, 106, 91-99.

Bauder, D. (2002, March 11). A red-letter day for TV swearing. The Associated Press. Retrieved May 17, 2002 from The Enquirer Web site: http://www.enquirer.com/editions/2002/03/ 11/tem_a_red-letter_day.html

Brehm, S. S. (1981). Oppositional behavior in children: A reactance theory approach. In S. S.S. Brehm, M. Kassin, & F.K. Gibbons (Ed.), Developmental social psychology: Theory and research (pp. 117-124). New York: Oxford Press.

Bukstein, O. G. (1996). Aggression, violence, and substance abuse in adolescents. Child and adolescent psychiatric clinics of North America, 5(1), 93-109.

Bushman, B. J. (1997). Effect of warning labels on attraction to television violence in viewers of different ages. Paper presented at the Telecommunication and Policy Research Conference, Alexandria, VA.

Bushman, B.J., & Stack, A. D. (1996). Forbidden fruit versus tainted fruit: Effects of warning labels on attraction to television violence. Journal of Experimental Psychology: Applied, 2(3), 207-226.

Cantor, J., & Harrison, K. S. (1996). Ratings and advisories for television programming. In National television violence study (Vol. 1) (pp. 361-410). Thousand Oaks, CA: Sage.

Cantor, J., Harrison, K., & Nathanson, A. (1998). Ratings and advisories for television programming. In J. Federman (Ed.), National television violence study (Vol. 2) (pp. 267-322). Thousand Oaks, CA: Sage.

Cantor, J., Stutman, S., & Duran, V. (1996). What parents want in a rating system for television. Report of the National PTA, the Institute for Mental Health Initiatives, and the University of Wisconsin. Retrieved May 14, 2004, from the National Parent Teacher Association Web site: http://www.pta.org/programs/tvreport.htm

Christenson, P. (1992). The effect of parental advisory labels on adolescent music preferences. Journal of Communication, 42(3), 106-113.

Cohen, J., Dunbar, K, & McClelland, J. (1990). On the control of automatic processes: A parallel distributed processing account of the stroop effect. Psychology Review, 97(3), 332-361.

Graham, K., Schmidt, G., & Gillis, K. (1996). Circumstances when drinking leads to aggression: An overview of research findings. Contemporary Drug Problems, 23(3), 493-557.

Herman, G. & Leyens, J. P. (1977). Rating films on TV. Journal of Communication, 27(4), 48-53.

Infante, D. A. (1995). Teaching students to understand and control verbal aggression. Communication Education, 44, 51-63.

Infante, D. A., Chandler, T. A., & Rudd, J. E. (1989). Test of an argumentative skill deficiency model of interspousal violence. Communication Monographs, 56, 163-177.

Infante, D. A., & Rancer, A. S. (1996). Argumentativness and verbal aggressiveness: A review of recent theory and research. In B. Burleson (Ed.), Communication yearbook 19 (pp. 319-351). Thousand Oaks, CA: Sage.

Krcmar, M., & Albada, K.F. (2000). The effect of E/I ratings on children's interest in and recall of an educational program. Journal of Broadcasting & Electronic Media, 44, 674-689.

Lang, A., Geiger, S., Strickwerda, M., & Sumner, J. (1993). The effects of related and unrelated cuts on viewers' attention, processing capacity and memory. Communication Research, 20, 4-29.

Moss, H. B., & Kirisci, L. (1995). Aggressivity in adolescent alcohol abusers: Relationship with conduct disorder. Alcoholism: Clinical and Experimental Research, 19, 642-646.

Parents Television Council. (2002). Wired for raunch: A content analysis of basic cable's original prime-time series. Retrieved May 18, 2004, from http://www.parentstv.org/ptc/ publications/reports/cablestudy/main.asp

Pashler, H.E. (1999). The psychology of attention. Cambridge, MA: MIT Press.

Rohsenow, D. J., & Bachorowski, J. (1984). Effects of alcohol and expectancies on verbal aggression in men and women. Journal of Abnormal Psychology, 93, 418-432.

Triesman, A. (1964). The effect of irrelevant material on the efficiency of selective listening. American Journal of Psychology, 77, 533-546.

Underwood, G., & Murray, N. (1971). Shadowing and monitoring for selective attention. Quarterly Journal of Experimental Psychology, 23, 284-295.

Wurtzel, A., & Surlin, S. (1978). Viewer attitudes toward television advisory warnings. Journal of Broadcasting, 22(1), 19-31

Marina Krcmar (Ph.D., University of Wisconsin, Madison) is an Associate Professor in the Department of Communication Sciences at the University of Connecticut. Her research interests include effect of violent television on children and the impact of media policy on families.

Steve H. Sohn is a doctoral candidate in the Department of Communication Sciences at the University of Connecticut. His research interests include media effects on body image and media effects on consumers' advertising message process.