Portrayals of Sexual Violence in Popular Hindi Films, 1997–99

Portrayals of Sexual Violence in Popular Hindi Films, 1997–99

Srividya Ramasubramanian1,2 and Mary Beth Oliver1

An exploratory content analysis was conducted to examine portrayals of sexual violence in

popular Hindi films. Nine films were randomly selected from box office hits (1997–99). The

findings suggest that moderate sexual violence is depicted as fun, enjoyable, and a normal

expression of romantic love. Victims were more likely to be women rather than men, and sexual

violence committed by heroes was a common portrayal, particularly moderate violence such as

harassment of women with whom the heroes ultimately became romantically involved. Severe

sexual violence was portrayed as criminal and serious, whereas moderate sexual violence was

treated as fun and romantic. Results are discussed in terms of script theory and social learning

perspectives.

KEY WORDS: Hindi films; content analysis; India; sexual violence; eve-teasing.

The incidence of sexual violence against women is greater in societies that have male-dominated ideologies and a history of violence, as is the case in India  (Burt, 1980; Check & Malamuth, 1985; Linz &  Malamuth, 1993). The number of registered cases of  sexual crimes against women in India increased from  b67,072 in 1989 to 84,000 in 1993 (“Crimes Against,” 1996). In 1995 alone, more than 25,000 cases of molestation  and 12,000 cases of rape were reported in the capital city of New Delhi (West, 1996). It is estimated  that well over 80% of sexual crimes go unreported (“Atrocities Against,” 2002). For example, only 7,643 of the estimated 50,000 instances of violence against women were reported to the police even in Kerala, a South Indian state with the highest women’s literacy  rate (“Atrocities Against,” 2002).  One specific form of sexual harassment called  “eve-teasing” is prevalent, especially in urban India.3

The term eve-teasing is used to refer to sexual harass-1Pennsylvania State University, University Park, Pennsylvania.  2To whom correspondence should be addressed at College of Communications,

Pennsylvania State University, 115 Carnegie Building, University Park, Pennsylvania 16802; e-mail: srivi@psu.edu. 3Although it is entirely possible that the word eve-teasing might

refer to the biblical story of Adam and Eve, we prefer to spell  eve-teasing with a small “e” rather than a capital “E” because this is the spelling used by Indian journalists.placement of women in public places such as the streets, public transportation, parks, beaches, and cinemahalls. This type of public harassment by a lone man or  gangs of men includes verbal assaults such as making passes or unwelcome sexual jokes; nonverbal assaults  such as showing obscene gestures, winking, whistling,and staring; and physical assaults such as pinching,  fondling, and rubbing against women in public places  (“Eve-teasing,” 1999; Stevens, 1984). In addition, inseveral instances eve-teasing has been followed by  more violent assaults such as rape and murder. In  trying to construct the profile of an eve-teaser, it is interesting to note that about 32% of eve-teasers are  college students (“Films,” 1998). The severity of these incidents coupled with their  high prevalence resulted in the legal declaration of eve-teasing as a punishable offense by the state government  of Tamil Nadu in 1999, where it was announced  that offenders would be penalized with up to  1 year of imprisonment or a fine of Rs. 10,000 or both  (“Ordinance,” 1999). Despite the seriousness of these  incidents, research suggests that they are frighteningly  commonplace. For example, a recent survey revealed  that approximately 90% of college women in New Delhi have experienced sexual harassment in some  shape or form (“Films,” 1998). Yet, it is estimated that  only about 1 in 10,000 eve-teasing occurrences are  reported to the police (“Atrocities Against,” 2002).  327 0360-0025/03/0400-0327/0 °C 2003 Plenum Publishing CorporationP1: ZBU Sex Roles [sers] PP779-sers-461827 March 11, 2003 13:51 Style file version June 3rd, 2002  328 Ramasubramanian and Oliver  The primary reasons why women abstain from reporting  incidents of sexual violence are the unwieldy  medico-legalprocess, concerns about continued violence,

and fear of stigmatization (Prasad, 1999). Mass Media and Sexuality in India.

The variables that give rise to sexual violence in India are undoubtedly numerous and complex. However,  for feminist media scholars, the idea that popular cinema plays a significant role in shaping notions  about gender roles and gender identities within  the Indian context is of special interest and concern  (Bagchi, 1996; Ram, 2002). Cinema has been a dominant  medium in India because of the sheer size and  reach of its indigenous film industry. The Indian film industry produces about 800 feature films annually— the highest in the world (National Film Development  Corporation, n.d.). Not only does India produce  the largest numbers of films in the world, but also a sizeable amount of film consumption is common  among almost all age groups, socioeconomic backgrounds, and geographical locations within India  (Derne, 1995). It is estimated that every week approximately 90–100 million Indian viewers go to the cinema halls to watch films (Nair, Barman, & Chattopadhyay, 1999). Many cinemagoers ritualistically make as many as 20–30 visits to the cinema hall in a month and repeatedly view a favorite film several dozens of times (Derne, 1999; Khare, 1985). Moreover, Indians films are popular not just in India but also amongst theIndian diaspora in countries such as the United States,United Kingdom, Canada, Fiji, Dubai, and Singapore (Bist, 2002). Indian-made films constitute the majority 

of the films watched by Indians; only about 5% of Indians watch non-Indian (mostly Hollywood) films

(Anjum, 2002).Apart from cinema halls, films also reach the Indian household through countdown shows on televisionthat feature film-based song-and-dance hit numbers (Nair et al., 1999). In addition, access to cabletelevision has also grown very rapidly in the last decade, with a penetration of over 50% of the urban Indian market as of 1997 (Nair et al., 1999). Futher more,over 85% of the cable television operators routinely screen two films a day through their own private  local channels to attract their customers (Nair et al.,1999).The importance of sexual portrayals in motion  pictures is particularly relevant to Indian audiences,not only because these portrayals are viewed in abundance,but also because issues of sexuality are rarely discussed in other contexts (Derne, 1999). According to Derne (1999), Indian film portrayals form a“privileged arena for construction of sexuality” for the common person, and serve as primary sources of information about how men and women are to

behave in sexual relationships (p. 548). A recent  study sponsored by UNICEF and Save the Children

Fund in the Indian subcontinent showed that the film medium is influential, especially with teenaged

boys, in teaching notions about masculinity, power,and violence in relationships with women (Poudyal,

2002). Similarly, researchers in the North American context have found that children and adolescents

use media narratives (especially teen magazines and prime-time television programs) as sexual scripts for learning about dominant norms concerning gender,love, and sexuality (Carpenter, 1998; Pardun, 2002;Ward, 1995; Wood, 2001; Wood, Senn, Desmarais, Park, & Verberg, 2002).Feminist scholars are particularly concerned that popular films in India too often portray women in  stereotypical roles of subordination—accepting sexual violence as a normal part of relationships with men (Dasgupta & Hegde, 1988; Gandhi & Shah, 1992).Further, they have pointed out that men’s abuse

of women is often glorified within Indian cinema(Derne, 1999). More specifically, critics have pointed

out that the repeated glamorization of eve-teasing infilms as a macho manifestation of a tough-acting, college student hero, who initially upsets the heroine but finally wins her attention, has fostered a climate supportive of such acts in real life (Birla, 2001; “Films,”1998; Ravindran, 2001). Although many critics have  voiced concerns, very few researchers have dealt with sexually violent portrayals in Indian films. In a rare study of its kind, Derne (1999) conducted a qualitative content analysis study of a selected few Hindi films in which violence and sexuality were often intertwined. Derne (1999) suggested that these films conveyed the notion that force and physical aggression were legitimate means of expressing romantic love. Therefore,sexual violence was not only “normal” but also “expected” in romantic relationships between heroes and heroines. Links Between Media and Sexual Violence Although little systematic research has explored

the causal influences of Hindi films on sexual violence in India specifically, there is research in other cultures,P1: ZBU  Sex Roles [sers] PP779-sers-461827 March 11, 2003 13:51 Style file version June 3rd, 2002 Sexual Violence in Hindi Films 329 particularly North America, that has explored the role of consumption of media portrayals of sexuality on viewers’ behaviors. In this regard, some researchers

have suggested that there is no causal relationship between access to sexually explicit material

and the incidence of sexual crimes (Kutchinsky, 1991), that effects are observed only for individuals who are predisposed to be aggressive (Zillmann & Sapolsky,1977), or that harmful effects are observed only for explicitly violent portrayals (Donnerstein, Linz, &  Penrod, 1987). However, results of meta-analytic research suggest that there is a relationship between  media consumption of sexually explicit materials (and  particularly violent materials) and a number of variables  related to sexual violence (Allen, D’Alessio, &   Brezgel, 1995; Allen, Emmers, Gebhardt, & Giery, 1995). These analyses reported that exposure to sexually explicit media (both violent and nonviolent) was associated with increased rape-myth acceptance and with increased subsequent aggression, especially  among angered participants (Allen, D’Alessio, et al.,1995; Allen, Emmers, et al., 1995). In addition, researchers have also reported that consumption of sexually explicit media (both violent and nonviolent) may lead to increased sexual callousness—the disregard or contempt for a woman’s right to deny  sexual access (Zillmann & Weaver, 1989). Similarly,  other researchers have argued that consumption of media portrayals of sexual violence may lead to target desensitization—the belief that certain individualsare appropriate, natural, and safe targets of violence who are deserving of aggression (Check &  Malamuth, 1985; Donnerstein & Berkowitz, 1981).  Behavioral effects of exposure to sexually explicit material can take the form of imitation of new behaviors  as well as lowered inhibitions to try out already  learned behaviors (Russell, 1988). Finally, other researchers have examined the idea of sexual objectification,and have reported that the viewing of pornography can lead some male viewers to interpret subsequent interactions with women in inappropriate  sexual or erotic terms (McKenzie-Mohr & Zanna, 1990).In summary, although most of the research on the effects of sexual portrayals suggests that some types of  explicit images, particularly those that contain aggression, can lead to harmful effects on viewers, there are

some inconsistencies in the literature that have made it difficult to determine causal effects. These differences have been attributed to the types of stimulus  materials employed, the types of populations studied  (e.g., rapists/noncriminals, hypermasculine men,whether or not the participants are under the influence  of alcohol, propensity to use force), the environment,and additional cultural factors (Harris & Scott,  2002). Despite these factors, however, meta-analytic research that synthesized the body of literature in this  area suggested that exposure to media portrayals of  sexually explicit material can have a variety of effects on viewers’ attitudes and behaviors, many of which  are causes for concern (Allen, D’Alessio, et al., 1995;  Allen, Emmers, et al., 1995; Harris & Scott, 2002).  The Present Research  Given the dearth of research on the effects of  sexual violence with respect to popular Indian films  and the need to investigate the effects of sexually explicit  media amongst diverse populations, there is a

need for research in the area of mediated sexual violence  in India. The literature on effects of filmed

sexual violence generally supports claims that Hindi films may be a contributory factor in sexual harassment. However, this would be true only if Hindi films  actually depicted the types of images that are thought  to play a role in influencing notions about sexuality in the Indian context. Hence, the first step in exploring  this issue is to examine the types of portrayals that  are commonly depicted in the films. At this point it is  unclear if popular films meant for mass consumption  would have any sexually violent material in the first  place. In particular, because popular Hindi films are  viewed by people of all age groups (rather than just by adults), one might expect that Hindi films would  be unlikely to show sexual images. On the other hand,  if Hindi films do provide an outlet for “discussions”  of sexual behaviors that serve to reinforce traditional  views of women, then one might expect that Hindi  films would be likely to show violence against women  as normal and perhaps even enjoyable.  Consequently, the purpose of this exploratory  study was to examine the manner in which popular  Hindi films portray sexual violence and the way in  which violence might be associated with gender and  romantic love. Specifically, we examined the following  research questions:  RQ1: What proportion of sexual scenes contains violence?  RQ2: Is there a relationship between gender and likelihood  of being the primary victim of sexual violence? RQ3: Is there a relationship between character role and likelihood of being the primary perpetrator of sexual violence?P1: ZBU  Sex Roles [sers] PP779-sers-461827 March 11, 2003 13:51 Style file version June 3rd, 2002  330 Ramasubramanian and Oliver RQ4: Is there a relationship between severity of sexual  violence and character role of the primary perpetrator of sexual violence?RQ5: Against what type of character roles are heroes most likely to perpetrate sexual violence? RQ6: Is there a relationship between scene type and  severity of sexual violence?

 METHOD

Sample  A sample of nine full-length feature films was randomly selected from a population of top-10 box office hits in the Hindi film industry released in the years 1997, 1998, and 1999.4 Within each of these  3 years, three films were randomly selected. Because  we were interested in mass entertainment, especially  films viewed by adolescents, films rated “U” (universal audience) and “UA” (public viewing with parental  guidance for children under age 12) were included  in the study but those rated “A” (films restricted to  adult audiences) were excluded (National Film Development Corporation, n.d.; see Table I for a list of  the films analyzed). Units of Analysis Two units of analysis were examined in this study:

characters and sexual scenes. A scene was defined as a division of the film that presents continuous action in one place such as a single situation or unit of dialogue in the film (e.g., love scene or fight scene). Because we  were interested in examining the nature of violence within the context of sexual interactions, only sexual  scenes (both violent and nonviolent) were coded. The  entire film was watched to locate the presence of sexual  scenes. One hundred and eight such scenes were included in this study. A sexual scene was defined as  one in which two or more characters were involved in activities such as having sex, kissing, petting, initiating or suggesting sexual contact, displaying nudity, engaging  in sexual talk, bathing in an erotic way, wearing provocative or revealing clothes, or shown as a sexual  object of gaze. This included actual depictions, suggestions  of, and preparation for sexual activities. No  4While we were making a random selection of three of the top-10 box office hits of 1997, the film Virasat was selected at first but it was  not accessible. Therefore, it was replaced with another randomly  selected film, Border.

Table I. List of Films Analyzed in This Study  Year Film title 1997  Pardes  Border  Hero No. 1

1998 Bandhan  Pyaar To Hona Hi Tha  Kuch Kuch Hota Hai  1999  Biwi No. 1 Hum Aapke Dil Mein Rehte Hain Sarfarosh  instances of homosexual relationships were portrayed  in any of the films selected. Therefore, only heterosexual  relationships were considered within the scope of this study. Also, because we were interested in examining  sexual interactions between individuals, two

or more characters had to be present in a scene for  it to be considered as a sexual scene. For example,

a woman undressing for a bath was not considered for the study, but if a man undressed a woman, it was  included within the study.  The second unit of analysis was the character. Seventy-seven characters were coded in this study. Only those characters who were shown speaking and  were present in a sexual scene were included in the  study. Characteristics of characters, such as gender  and type of character role, were coded. Characters  were observed for the entire film before coding their  characteristics.  Coding Scheme  A coding scheme was created for the variables

 of interest: presence of sexual violence, primary perpetrators/victims,  gender, character role, severity of  sexual violence, and fun/seriousness of scene. Presence of Sexual Violence Sexual scenes were of two types: mutually consenting  scenes and sexually violent scenes. Mutually consenting scenes were those in which the characters  involved showed interest in or expressed no objection  to engaging in the sexual behavior, and there  was no harm to any of the people involved. In contrast, a sexually violent scene was any sexual sceneP1: ZBU Sex Roles [sers] PP779-sers-461827 March 11, 2003 13:51 Style file version June 3rd, 2002  Sexual Violence in Hindi Films 331where there was actual depiction of, suggestion of, or  preparation for sexual violence. Sexual violence was  defined from the victim’s perspective as any forced sexual act that was inappropriate, offensive, and/or  harmful. Offensiveness to the victim was assessed using verbal and nonverbal expressions of disapproval,  anger, or disgust (e.g., saying no, crying, pushing away,  clenching fists). This included (but was not limited to)  acts such as rape, verbal comments, kissing, disrobing,  touching, staring, rubbing against, and obscene  gestures. Rape was defined as the actual depiction  of, suggestion of, or preparation for forced sexual intercourse.   Sexual violence was not just limited to rape but  also included sexual harassment, eve-teasing, and domestic  violence. Sexual harassment was defined as  inappropriate, offensive, and/or harmful sexual behavior  within the context of a workplace or academic  environment wherein a power differential existed between  the parties involved. For example, the sexual  harassment of a student by a professor or of a subordinate  by a boss was coded as sexual harassment. Eve-teasing was defined as sexual behavior displayed  in public places (especially between strangers or acquaintances  who are not committed to a relationship)  that was inappropriate, harmful, and/or offensive to  the victim. Domestic violence was defined as sexual  aggression (e.g., forced kissing, disrobing, pinching) between couples that were in an intimate, committed  sexual relationship (e.g., boyfriend/girlfriend, fi-anc´e/fianc´ee, husband/wife) where the victim was hurt and/or offended by the sexual act.

Severity of Sexual Violence Sexually violent scenes were further categorized  as severe or moderate. Severe violence included  actual depictions of, suggested, attempted,  or preparation for rape or eroticized murder. Moderate  sexual violence included all other forms of sexual violence, sexual harassment, eve-teasing, and domestic violence—that did not involve rape or  murder.5

5It is important to point out that our use of the terms severe and moderate is not meant to imply that “moderate” acts are trivial.  Rather, both severe and moderate acts are understood to represent

sexual violence. However, our use of the term severe in this context refers to acts such as rape and murder that are extreme  forms of sexual violence that result in lasting physical harm to the

victim.  Primary Perpetrators/Victims in a  Sexually Violent Scene  All sexually violent scenes had at least one perpetrator and one victim. The perpetrator was the one  who initiated sexual aggression. A perpetrator was defined as someone who actually used, suggested the  use of, attempted to use, or made preparations for  using aggression in a sexual context. The victim was  defined as the character who expressed lack of consent to the sexual act and/or was harmed by the act. It is important to note that sexual violence was definedmore in terms of the harm caused to the victim ratherthan the intention of the perpetrator. This meant that even if the perpetrator did not intend to cause harm to the victim, it was considered as sexual violence if  the victim was harmed.  Character Role Every character was coded as playing one of five character roles: hero, heroine, villain, comedian, or supporting character. A hero was defined as a character who played the role of the main, leading, maleprotagonist of the film. The heroine was defined asthe main, leading, female protagonist in the film. Thevillain was anyone who was an antagonist (man orwoman). A comedian is a character who is similar toa “sidekick” in Hollywood films. A comedian was onewhose role in the narrative was to provide comic relief(man or woman). Supporting characters included anybody  who did not fall into the classification of hero,heroine, villain, or comedian. In a given film, more  than one person could play these roles. For example,there were some films with two heroes.  Fun/Seriousness of the SceneTo understand fully the context within which sexuality  was introduced into the plot of the films, itwas crucial to code for the type of scenes that depicted  sexuality. The sexual scenes were classified aseither fun scenes or serious scenes. Serious scenes  included drama, action, and mystery. Fun scenes includedcomedy, romance, and song–dance. Romance  was defined as scenes that showed sexually attracted,dating, engaged, or married couples interacting witheach other in a romantic fashion. Action was definedas scenes that showed fights, physical aggression, orviolence. Comedy was defined as scenes that depictedP1: ZBUSex Roles [sers] PP779-sers-461827 March 11, 2003 13:51 Style file version June 3rd, 2002332 Ramasubramanian and Oliver  jokes and humor. Song-dance scenes were definedas musical episodes accompanied by dances by charactersin the film. All other scenes were coded asdrama scenes. Typically drama scenes showed con- flict, were dialogue-oriented, and involved emotionssuch as anger or sadness. Coding ReliabilityAll coding and data reported here was conductedby the first author who was trained in coding proceduresand was familiar with the descriptive booklet.6 To examine reliability, a secondary coder fluent in thelanguage of the films but unaware of the specific researchhypotheses independently coded six of the ninefilms. Intercoder reliability was calculated by computingthe percent agreement for the five variables examinedin this study: presence of sexual violence (83%),character role (78%), severity (95%), gender of character(93%), and fun/serious (78%).7

RESULTS

Presence of Sexual ViolenceThe first research question concerned the prevalence

of sexual violence. An examination of the sexualscenes analyzed showed that slightly less than half

of the sexual scenes (40.7%, N = 44) contained violence(see Table II). The most common form of sexual

violence depicted was eve-teasing (57% of sexuallyviolent scenes, N = 25). Approximately 11% of the

sexually violent scenes contained severe sexual violencesuch as rape or eroticized murder (N = 5). It

should be noted here that one film (Border) showed6In this study, the primary coder was responsible for coding all ofthe data that were presented in this paper. The additional coder  was employed as a way of assessing the primary coder’s reliability.Although researchers often have coders resolve disputes andarrive at a mutual decision, there are numerous instances where  an additional coder is employed as a means of reliability computationor where multiple coders were employed, with their independentlycoded data collected after reliability checks had been  conducted (e.g., Fouts & Burggraf, 1999; Larson, 2001; Sharrer,2002; Schlenker, Caron, & Halteman, 1998).7There was disagreement among the coders concerning the unit ofanalysis for two of the sexual scenes. This disagreement is reflectedin the reliability indicators. Gender was coded as male/female.  There was some disagreement between coders regarding the genderof transvestites, bisexual persons, and transgendered individualsthat reduced the reliability of this measure.  Table II. Summary of ResultsSexual scenesWith violence 40.7%Without violence 59.3%Gender of primary victims of sexual violenceWomen 77.0%Men 23.0%Character role of primary victims of sexual violenceHeroines 95.0%Other roles 5.0%Primary perpetrators of sexual violenceHeroes 67.8%  Villains 32.2%Primary perpetrators of moderate sexual violenceHeroes 78.2%   Villains 21.7%Portrayal of moderate sexual violenceFun 69.2%Serious 30.7%only mutually consenting sexual scenes whereas HeroNo. 1 depicted 72.2% of the sexual scenes as violent.

However, most films depicted approximately 40% ofthe sexual scenes as sexually violent, suggesting thatalthough there is clearly variation in the percentageof sexually violent scenes portrayed, the majority ofthese films contained a substantial proportion of sexualscenes containing violence.

Gender and Primary Victim of a SexuallyViolent SceneThe second research question asked if there wasany relationship between gender and the primary victimin sexually violent scenes. A chi-square test of theprimary victims in sexual scenes revealed that womenwere more likely than men to be victims. Namely, ofall victims coded in sexually violent scenes, 77% werewomen and 23% were men, χ2(1, N = 43) = 12.30,p < .001, V∗ = 0.29 (see Table II). For example, inKuch Kuch Hota Hai, a typical college-based eveteasingis used as a means to enhance the sexual appealof the heroine Tina (played by Rani Mukherjee).When Tina enters the college campus wearinga very short mini-skirt and tight top, she is accostedby a gang of men in her college who stare at her legs,whistle, hoot, and make lewd remarks at her even  though she expresses her disgust at their behavior.However, it appears that the intention of the scene  is more to invite the audience to view the heroineas a sex object rather than to empathize with her  experience.P1: ZBUSex Roles [sers] PP779-sers-461827 March 11, 2003 13:51 Style file version June 3rd, 2002Sexual Violence in Hindi Films 333Character Role and Primary Perpetrator

The third research question focused on the relationshipbetween the character role (hero vs. villain)

and the primary perpetrator in sexually violentscenes. A chi-square analysis of heroes and villains

showed that heroes (67.8%) were more likely thanvillains (32.2%) to be the primary perpetrator in sexuallyviolent scenes. However, these differences onlyapproached statistical significance, χ2(1, N = 28) =3.57, p = .06, V∗ = 0.13 (see Table II). For example,films such as Biwi No. 1 and Hum Aapke Dil Mein

Rehte Hain show the hero eve-teasing women bysinging lewd songs, making sexual remarks, and touchingthe heroine in sexual ways despite knowing thatthe heroine does not like these acts. On the otherhand, in Pardes, the evil, villainous boyfriend tries toforce his fianc´ee to have sex with him and rips off partsof her clothes after taking her to a hotel room. However,such depictions are much fewer than instances where the hero is the perpetrator.Character Roles and Severity of Sexual Violence

The fourth research question examined the relationship between character of the primary perpetrator

and the severity of sexual violence portrayed. A chisquare analysis of character role and severity of sexual violence revealed that villains were more likelyto be featured as perpetrators of severe sexual violence whereas heroes were more likely to be feature das perpetrators of moderate sexual violence. Specifi-cally, primary perpetrators in severe scenes were moreoften villains (80.0%) than heroes (20.0%), whereas primary perpetrators in moderate scenes were more often heroes (78.2%) than villains (21.7%), χ2(1, N =28) = 6.39, p < .05, V∗ = 0.48 (see Table II). In filmssuch as Kuch Kuch Hota Hai and Hero No. 1 there areseveral instances in which the hero eve-teases attractiveyoung women to win their attention. These actsare treated in a very casual and trivial manner. However,later in the same film (Hero No. 1), the hero isshown enraged when the villains try to rape another young girl in the streets. The sexual violence in the latter scene is dramatized as something evil and wrong.Character Role and Primary Victim  The fifth research question focused on the typesof character roles against whom heroes were most  likely to perpetrate sexual violence. All sexually violent scenes where the hero was the primary perpetrator were selected for this analysis. A chi-square analysis of the primary victims in these scenes revealedthat heroines (95%) were much more likely to be the victim than were other characters (5%),

χ2(1, N = 19) = 15.21, p < 0.001, V∗ = 0.85 (seeTable II). For example, a typical scene showing a boy–

girl romantic confrontation using eve-teasing can beseen in Hero No. 1. The hero waylays the unsuspectingheroine (both are strangers to each other until thispoint in the story) at the airport, follows her around  to the train station, makes obscene passes at her, rubsagainst her body, and even sits on her lap in the train, even though all through the sequence, the heroineconstantly expresses her disapproval of these actions  by the stranger (hero). As the entire encounter isagainst the background of a catchy song-and-dance  sequence, the sexual harassment is presented to theaudience as light-hearted fun.  Severity of Sexual Violence andFun/Seriousness of SceneThe sixth research question asked if there was a  relationship between severity of sexual violence andthe type of scene. The findings suggest that severe sexualviolence is more likely to be portrayed as seriouswhereas moderate sexual violence is more likely tobe portrayed as fun. Specifically, a chi-square analysisrevealed that severe crimes were more often portrayedas serious (80.0%) than as fun (20.0%) whereasmoderate crimes were more often portrayed as fun(69.2%) than as serious (30.7%), χ2(1, N = 44) =4.64, p < .05, V∗ = .33 (see Table II). For instance, in Bandhan, the hero (played by Salman Khan) constantlyeve-teases the heroine (played by Rambha),  but the entire situation is couched in slapstick comedy,which distracts the viewer from the sexual harassmentper se. On the other hand, later in the samefilm when the villains stop the heroine in an isolatedfield and eve-tease her, the scene takes on more seriousproportions with dialogue and drama rather than song and dance.

DISCUSSION

The results of this study lend support to the idea that a substantial proportion of sexual scenes in popularHindi films depict sexual violence, even in thoseP1: ZBUSex Roles [sers] PP779-sers-461827 March 11, 2003 13:51 Style file version June 3rd, 2002334 Ramasubramanian and Oliver

films meant for viewing by audiences of all age groups or with parental guidance if under 12 years. However,it is not just the amount of sexual violence in the filmsthat is cause of concern but also the nature of theseportrayals. First, these films indicate a gender divide when it

comes to perpetrators and victims of sexual violence .Almost all films show female characters as victims ofsexual violence, whereas male characters are shownas perpetrators of these incidents. This seems to beconsistent with traditional gendered beliefs in India that women should be submissive and men should beaggressive in social relationships. This repeated pairing of women with violence is problematic because itmight reinforce existing beliefs that it is acceptable toaggress against women and that women should tolerateviolence from men.Another aspect of these portrayals that is a cause

for concern is that the perpetrators of sexual violencewere not just villains, but also heroes. Heroes

were somewhat more likely than villains to be the primary perpetrators in sexually violent scenes. It is a

cause for concern that heroes, who often representedthe essence of “ideal manhood” and male sexuality,were often perpetrators of sexual violence. This lendssome support to the idea that being aggressive is depictedas “being manly.”The idea that heroes would be shown engaging

in sexual violence is cause for concern, as social learningperspectives suggest that when likable, attractivecharacters such as heroes perpetrate sexual violenceon screen, they are more likely to be imitated by viewers.That is, research on social learning from mediaportrayals suggests that viewers are more likely to emulate behaviors that they see in the media whenthe modeled behavior is portrayed as rewarded (orat least not punished; see Bandura, 1994). This lineof reasoning suggests that film portrayals of women as victims of sexual aggression are particularly problematic because such behavior might be learned and imitated by the viewers. In addition, the viewer’s modeling of a media character’s behaviors is particularly likely to occur when the character is portrayed as attractive,

likable, and heroic. In terms of the presentresearch, this suggests that Indian male viewers may

be especially likely to emulate sexually violent behaviorperpetrated by heroes.

Heroes and villains differed in the types of sexual violence that they perpetrated. Heroes were more

likely to perpetrate moderate crimes such as eveteasing,sexual harassment, and domestic violence,

whereas villains were more likely to perpetrate severe Thus, moderate sexual violence seems not to be condemned and might even be rewarded. As we sawabove, heroes, by definition, seem to protect moral

good and to fight evil. Therefore, the association ofheroes with moderate sexual violence may run the

risk of sending a message to viewers that only severe crimes are bad and that moderate sexual violence isnot bad (and may be even perceived as good). Therefore,these findings suggest that only rape and eroticized murder might be considered crimes by the audiences,but that eve-teasing, sexual harassment, and domestic violence may be socially acceptable sexualbehaviors. Furthermore, moderate sexual violence is often depicted in the context of fun and happiness,whereas severe sexual crimes are depicted as serious and dramatic. This pairing of fun with moderate sexualviolence implies that such crimes are not bad but enjoyable for all involved.Moreover, the finding that heroes more often aggressed   against heroines than against any other charactersis consistent with the argument that aggression   is portrayed as a desirable attribute in Hindi films. Itshould be noted that in all the films in this study, the   hero and heroine were romantically involved. This suggests that it was appropriate, normal, and perhaps   even romantic for men to aggress against the women with whom they were romantically involved. From the perspective of sexual script theory, these portrayals may suggest to viewers (especially young adults, adolescents,and children) that these recurring themes ofviolence among romantically involved couples in the media represent acceptable ways of behaving in sexual relationships. The films analyzed in the current  study were not adult films but those rated U and UA.Thus, it is highly likely that these films’ audiences include younger age groups who are also likely to be learning social norms related to gender and sexuality. Moreover, as mentioned previously, eve-teasing statistics report that about one third of the perpetrators in real life are college-age youth (“Films,” 1998).This situation suggests that social learning and sexual script theories might be at work although clearly, experimental research needs to be conducted to determine the specific nature of the effects that these films may be having on their viewing audiences.Although the results of this study indicate that moderate sexual violence such as eve-teasing, domestic violence, and sexual harassment by men againstwomen is very often portrayed as appropriate, enjoyable, and romantic, there are several limitations thatdeserve attention and suggest directions for futureP1: ZBUSex Roles [sers] PP779-sers-461827 March 11, 2003 13:51 Style file version June 3rd, 2002  Sexual Violence in Hindi Films 335research. First, only top-10 box office hits were chosen   for the analysis. It can be argued that more (or fewer)types of films could have been included in defining   what is popular. However, the use of top-10 box office films allowed for the examination of films that   are clearly popular among a wide viewing audience both within India and amongst the Indian diaspora in   countries such as the United States, United Kingdom,Canada, Singapore, and Dubai. Second, we examined popular films in the Hindi language only and did not include regional Indian language films. Even though there are very large numbers of films produced each year in regional languages (especially Telugu and Tamil), their popularity and audience reach is smaller than that of Hindi films. Nevertheless ,future researchers could benefit by including these two-regional-language films to get a regional sample of sexually violent portrayals.Another limitation of this research is the small  sample of films analyzed. This limitation reflects thefact that Hindi films often average 3–4 hr in length,  which makes them time-consuming to analyze. However,the longer film length meant that the numbers  of scenes and characters analyzed in a given film were also proportionately higher. Nevertheless, future  studies would undoubtedly benefit from the inclusion of a larger sample of films.  An additional limitation of this study is that we coded only sexual scenes within the films. Although  the use of sexual scenes as a unit of analysis was appropriate for the questions examined in this study, it restricted the generalizability of our findings because it was not possible to say whether sexual scenes contained more or less violence than the nonsexual scenes. On the one hand, the potential effect of sexual violence may largely depend on the amount of violence in the film overall, with violence in the sexual scenes only arguably having different effects than violence running throughout all types of scenes. On the other hand, one might argue that any portrayals of sexual violence have the potential to affect viewers. Nevertheless, an examination of violence across all scenes, sexual and nonsexual, should be taken into  consideration by future researchers .Finally, a content analysis such as this one can  only describe the portrayals that exist on screen. Theethod is limited in its ability to predict attitudinal  and behavioral changes that could result from exposure. At best, the results can only be seen as  indicative of the likely effects on the audience. We cannot in any way claim that the increase in sex  crimes is due to the sexual violence portrayed in films. Such conclusions are best made using experimental methodologies. Despite these limitations, the results of this exploratory

study seem to suggest that Indian films tend to present moderate forms of sexual violence to its audience as normal, fun, and heroic. The effect that such sexually violent portrayals have on viewers is an area of study that is deserving of research attention. Overall,our data support the criticism that eve-teasing in Indian films is not generally portrayed as a crime that ought to be punished, but rather as an act of romantic  love aesthetically woven into the narrative as fun and enjoyable.

ACKNOWLEDGMENTS

The authors thank Srinithya for her assistance in data collection. They also thank Joan C. Chrisler

and the anonymous reviewers for their detailed and valuable comments.

REFERENCES

Allen, M., D’Alessio, D., & Brezgel, K. (1995). A meta-analysis

summarizing the effects of pornography: II. Aggression after

exposure. Human Communication Research, 22, 258–283.

Allen, M., Emmers, T., Gebhardt, L., & Giery, M. A. (1995). Exposure

to pornography and acceptance of rape myths. Journal of

Communication, 45(1), 5–26.

Anjum, Z. (2002, February/March). Hollywood calling, Bollywood

falling. Aaj Magazine. Retrieved August 13, 2002, from

http://www.aajmag.com/hollybollywood.html

Atrocities against women on the rise in state. (2002, April 11). The

Hindu. Retrieved August 12, 2002, from the Lexis-Nexis Academic

Universe database, http://web.lexis-nexis.com/universe

Bagchi, A. (1996, December 18). Women in Indian cinema. Retrieved

August 12, 2002, from http://www.cs.jhu.edu/∼bagchi/

women.html

Bandura, A. (1994). Social cognitive theory of mass communication.

In J. Bryant & D. Zillmann (Eds.),Media effects: Advances

in theory and research (pp. 61–90). Hillsdale, NJ: Erlbaum.

Birla, P. (2001, March 31). Helping kids get street-smart

on sex education. Retrieved August 18, 2002, from

the Health Education Library for People Web site:

http://www.healthlibrary.com/news/25 31 march/31 sex.htm

Bist, R. (2002, October 12). Bollywood takes on the world. Retrieved

December 6, 2002, from Asia Times (Online) Web site:

http://www.atimes.com/atimes/South Asia/DJ12Df01.html

Burt, M. R. (1980). Cultural myths and supports for rape. Journal

of Personality and Social Psychology, 38, 217–230.

Carpenter, L. (1998). From girls into women: Scripts for sexuality

and romance in Seventeen magazine, 1974–1994. Journal of Sex

Research, 35, 158–168.

Check, J. V. P., & Malamuth, N. M. (1985). An empirical assessment

of some feminist hypotheses about rape. International Journal

of Women’s Studies, 8, 414–423.

Crimes against women rise in India. (1996, August 27). Xinhua

News Agency. Retrieved February 11, 2002, from theP1: ZBU

Sex Roles [sers] PP779-sers-461827 March 11, 2003 13:51 Style file version June 3rd, 2002

336 Ramasubramanian and Oliver

Lexis-Nexis Academic Universe database, http://web.lexisnexis.com/universe

Dasgupta, S. D., & Hegde, R. S. (1988). The eternal receptacle: A

study of mistreatment of women in Hindi films. In R. Ghadially

(Ed.), Women in Indian society: A reader (pp. 209–216). New

Delhi, India Sage.

Derne, S. (1995). Culture in action: Family life, emotion and male

dominance in Banaras, India. Albany: State University of New

York Press.

Derne, S. (1999). Making sex violent: Love as force in recent Hindi

films. Violence Against Women, 5, 548–575.

Donnerstein, E., & Berkowitz, L. (1981). Victims’ reactions in aggressive

erotic films as a factor in violence against women.

Journal of Personality and Social Psychology, 41, 710–724.

Donnerstein, E., Linz, D., & Penrod, S. (1987). The question of

pornography: Research findings and policy implications. New

York: Free Press.

Eve-teasing—The menace refuses to die. (1999, June 12). The

Hindu. Retrieved August 12, 2002, from the Lexis-Nexis

Academic Universe database, http://web.lexis-nexis.com/

universe

Films, TV serials have contributed to increasing acts of eve-teasing.

(1998, September 22). The Hindu. Retrieved February 11,

2002, from the Lexis-Nexis Academic Universe database,

http://web.lexis-nexis.com/universe

Fouts, G., & Burggraf, K. (1999). Television situation comedies:

Female body images and verbal reinforcements. Sex Roles, 40,

473–481.

Gandhi, N., & Shah, N. (1992). The issues at stake: Theory and

practice in the contemporary women’s movement in India. New

Delhi, India: Kali for Women.

Harris, R. J., & Scott, C. L. (2002). Effects of sex in the media. In J.

Bryant & D. Zillmann (Eds.),Media effects: Advances in theory

and research (2nd ed., pp. 307–331). Mahwah, NJ: Erlbaum.

Khare, V. (1985). The Dinman Hindi film inquiry: A summary. In

B. Pfleiderer & L. Lutze (Eds.), The Hindi film: Agent and

re-agent of cultural change (pp. 139–148). New Delhi, India:

Manohar.

Kutchinsky, B. (1991). Pornography and rape: Theory and practice?

Evidence from crime data in four countries where pornography

is easily available. International Journal of Law and Psychiatry,

14, 47–64.

Larson, M. S. (2001). Interactions, activities and gender in children’s

television commercials: A content analysis. Journal of

Broadcasting and Electronic Media, 45, 41–56.

Linz, D., & Malamuth, N. M. (1993). Pornography. Newbury Park,

CA: Sage.

McKenzie-Mohr, D., & Zanna, M. P. (1990). Treating women as

sexual objects: Look to the (gender-schematic) male who has

viewed pornography. Personality and Social Psychology Bulletin,

16, 296–308.

Nair, N. K., Barman, A. K., & Chattopadhyay, U. (1999,

December 15). Study on copyright piracy in India. Retrieved

August 11, 2002, from the Government of India Department

of Education Web site: http://www.education.nic.in/htmlweb/

cr piracy study

National Film Development Corporation. (n.d.). Cinema history.

Retrieved February 11, 2002, from http://www.nfdcindia.com/

history.htm

Ordinance against eve-teasing issued. (1999, July 30). The Hindu.

Retrieved February 11, 2002, from the Lexis-Nexis Academic

Universe database, http://web.lexis-nexis.com/universe

Pardun, C. J. (2002). Romancing the script: Identifying the romantic

agenda in top-grossing movies. In J. D. Brown, J. R. Steele, &

K. Walsh-Childers (Eds.), Sexual teens, sexual media: Investigating

media’s influence on adolescent sexuality (pp. 211–225).

Mahwah, NJ: Erlbaum.

Poudyal, R. (2002, June 12). Boys on film: Challenging masculinities

in South Asia. Retrieved August 11, 2002, from

http://www.id21.org/society/6arp1.html

Prasad, S. (1999). Medicolegal response to violence against women

in India. Violence Against Women, 5, 478–506.

Ram, A. (2002). Framing the feminine: Diasporic readings of gender

in popular Indian cinema. Women’s Studies in Communication,

25(1), 25–52.

Ravindran, V. (2001, September 10). Victims of whims. The Hindu.

Retrieved August 12, 2002, from the Lexis-Nexis Academic

Universe database, http://web.lexis-nexis.com/universe

Russell, D. E. (1988). Pornography and rape: A causal model. Political

Psychology, 19, 41–73.

Scharrer, E. (2001). Tough guys: The portrayal of hypermasculinity

and aggression in televised police dramas. Journal of Broadcasting

and Electronic Media, 45, 615–634.

Schlenker, J. A., Caron, S. L., & Halteman, W. A. (1998). A feminist

analysis of Seventeen magazine: Content analysis from 1945 to

1995. Sex Roles, 38, 135–149.

Stevens, W. K. (1984, March 17). For women of India, a rite of

spring is sour. The New York Times, p. 24. Retrieved February

11, 2002, from the Lexis-Nexis Academic Universe database,

http://web.lexis-nexis.com/universe

Ward, M. (1995). Talking about sex: Common themes about sexuality

in the prime-time television programs children and adolescents

view most. Journal of Youth and Adolescence, 24, 595–

615.

West, J. (1996, August 11). Police ‘tease’ is slapped in jail: India

applies the law to curb gropers. Sunday Telegraph, p. 25. Retrieved

February 11, 2002, from the Lexis-Nexis Academic

Universe database, http://web.lexis-nexis.com/universe

Wood, J. (2001). The normalization of violence in heterosexual romantic

relationships: Women’s narratives of love and violence.

Journal of Social and Personal Relationships, 18, 239–261.

Wood, E., Senn, C. Y., Desmarias, S., Park, L., & Verberg, N. (2002).

Sources of information about dating and their perceived influence

on adolescents. Journal of Adolescent Research, 17, 401–

417.

Zillmann, D., & Sapolsky, B. S. (1977). What mediates the effects

of mild erotica on annoyance and hostile behavior in males?

Journal of Personality and Social Psychology, 35, 587–596.

Zillmann, D., & Weaver, J. (1989). Pornography and men’s sexual

callousness toward women. In D. Zillmann & J. Bryant

(Eds.), Pornography: Research advances and policy considerations

(pp. 95–126). Hillsdale, NJ: Erlbaum.