This unit is about representations of gender in Hollywood films. As part of this, you are reading an article about women action heroes. Therefore, for this discussion board I want you to closely ana

Findings showed that 58.6% of VFACs were portrayed in a submissive role to the male hero in the film, and 42% were romantically linked to him. The average VFAC was young, white, highly educated, and unmarried. VFACs engaged in masculine types of violence yet retained feminine stereotypes due to their submissive role and romantic involvement with a dominant male hero character.

The findings suggest continued gender stereotypes set within a violent framework of contemporary American cinema.

KeywordsFilm.

Violence.

Gender.

Stereotypes.

Content Analysis Introduction In 1979, Sigourney Weaver played Lt. Ripley in the movie Alien. Film theorists generally agree that the action character Lt. Ripley paved the way for a new type of female representation in American popular culture (Brown 1996,2005; Clover1992; Inness2004; Tasker1993). A sea change in filmic representations of female action charactersoccurred after the success ofAlienincluding theAlien sequels and such films series asTerminator,Lara Croft,and Kill Bill. It is now commonplace to see female action characters engage in hand-to-hand combat, wield swords, shoot machine guns, and employ high-tech weaponry to destroy people and property—behaviors once the exclusive domain of male action heroes. These tough female representations seem to have moved beyond traditional notions of femininity and have drawn attention from feminist theorists who have debated whether they are empowering images for real women (McCaughey and King 2001), represent the ability of women to draw upon their femininity as a source of power (Rowe-Karlyn2003), or are a kind of“post-woman”operating outside the boundaries of gender restrictions (Hills1999).

The present research adds to the ongoing academic debate by examining female characters in American action films to determine whether they are really moving beyond traditional gender roles and norms, or are re-articulating and re-presenting gender stereotypes in a new guise. There is substantial literature offering interpretive analyses of female action characters portrayed in cinema but relatively few studies that provide quantitative data. This research utilized an empirical approach through content analysis.

The content analysis focused on three major areas of research. First, it examined the gender stereotypes displayed by female action characters. Second, it created a profile or average type of female action character from the demographic data gathered. Third, it analyzed the quantity and type of violence committed by female action characters.

The findings of this research help to reveal the contradictory nature of female action characters appearing in contemporary mainstream American cinema. These characters are female, yet they engage in traditionally masculine forms of physical violence. They also appear in K. Gilpatric (*) Department of Social Sciences, Kaplan University, 6301 Kaplan University Lane, Fort Lauderdale, FL 33309, USA e-mail: [email protected] Sex Roles (2010) 62:734–746 DOI 10.1007/s11199-010-9757-7 action movies alongside male heroes engaging in violence.

The success of American action films at the domestic box office can be attributed to the targeted youth audience (MPAA2007). Media has long been recognized as an agent of socialization (Bandura1986; Blumer and Hauser1933; Gerbner1970; Lazarsfeld and Merton1948). Therefore, it is worth examining the representations of violent female action characters shown in popular action movies because of the potential to influence a young viewing audience and their ideas about gender and violence (Gerbner1998; Huesmann1999; Signorielli and Bacue1999).

In addition to the potential problems of violent media influencing the socialization process, a feminist critique can be applied. Feminist film critics have a long history of analyzing the representations of women in movies and comparing them to the inequities of real-world gender roles.

Rosen (1973) critiqued the one-dimensional characters that played stereotypical roles of good wife and mother. Haskell (1974) criticized the depiction of self-sacrificing female characters in romanticmelodramas. Mulvey’s(1975) psychoanalytic approach placed popular cinema inside a patriarchal society that imposed an imbalance between active/male and passive/female ways of looking at charac- ters in film. She asserted that female images serve as signs of visual pleasure for a“male gaze”, and that“the man’s role is the active one of advancing the story and making things happen”(p. 11).

By the 1990s, feminist film theorists argued against totalizing and essentialist readings of film. Instead they proposed an active viewer who engages in her own meaning-making processes as she uses what she sees on- screen to help construct her own identity (Dyer2002; McRobbie2004; Stacey2000). The very labels of‘male’ and‘female’were challenged as being essentialist by feminists who considered gender to be a social construction (Butler1990; Lorber1999). In addition, postmodern feminists have challenged the privileged representations of white, Western, hetero females in cinema and the resultant problems of audience identificatory practices (Butler1999; Hooks1996; Modleski1991; Tasker and Negra2007; Trinh 1991).

The theories of these feminist media critics seem disparate. However, they have in common an understanding that female identities are informed by social codes rooted in popular culture. Continued analyses of media representa- tions help to shed light on the ways in which our normative ideas about femininity are formed in American popular culture.

Gender Stereotypes in the Media The first focus of this research was on the gender stereo- types displayed by female action characters. Calvert et al.( 2001) and Dietz (1998) have found that female action characters rarely get to play the main hero character.

Magoulick’s(2006) analysis of television action heroines found that the heroine is usually shown in a romantic relationship with a male protagonist who retained power and control over her. King (2008) conducted a content analysis of 291“cop-action”films and found that female police characters, even though in a male-dominated profes- sion, usually maintained rookie status and were engaged in romantic relationships with fellow workers. King (2008) also found that women were twice as likely as men to take on lovers during the course of the film. Eschholz and Bufkin (2001) conducted a content analysis of 50 popular films and found that female characters were more likely to be romantic, domestic, sensitive, and manipulative, and male characters were more likely to be competitive, athletic, aggressive, and risk-taking. The researchers concluded that “on average, male characters were more masculine and female characters more feminine”(p. 324).

Consensus on gender stereotypes can be problematic, so this research relied on an established standard of gender traits employed in social science research (Eschholz and Bufkin2001; Lueptow et al.2001; Twenge1997). The gender traits for masculine stereotypes include:dominant, aggressive, competitive, independent, ambitious, self- confident, adventurous,anddecisive. Traits used for feminine stereotypes include:affectionate, submissive, emotional, sympathetic, talkative,andgentle.

In light of the literature above regarding gender stereo- types in the media, the first research question asked:

1. Do female action characters exhibit gender stereotypes?

Media Representations of Age, Marital Status, Race, and Occupation The second focus of this research was the demographics of the female action character. Previous media studies have shown a consistent pattern of female characters depicted as young and unmarried. Elasmer et al. (1999) studied prime time television programs in 1992–93 and found that over 60% of female characters were in their twenties and thirties and 12.6% were shown as currently married. Nearly a decade later, Glascock’s(2003) analysis of prime time programs in 2001 found 58.8% of female characters were ages 18 to 37 and 17.8% were shown as married. Similarly for film, Eschholz et al. (2002) found that 71% of the female characters in popular 1996 movies were in their twenties or thirties, and King (2008) found 80% were unmarried in action movies from 1968 to 2006.

Media studies have found increasing numbers of African American characters in prime time television. Mastro and Greenberg (2000) examined over two decades of programs Sex Roles (2010) 62:734–746735 and found that by 1997 African American representation on television reached parity with U.S. population statistics.

Glascock (2001) found African Americans represented 14% of characters on prime time television and was slightly higher than the U.S. population. Although there has been improvement in racial diversity shown in the media some groups have not fared as well. A Children Now (2004) report found that Hispanics, Asians, and Native Americans have been underrepresented in prime time television.

Similarly, Eschholz et al. (2002) stated that“Hispanics, Native Americans, and Asians were almost entirely missing from the 1996 popular films line-up”(p. 314).

Media studies have found that representations of women portrayed in professional occupations are increasing.

Signorielli and Bacue’s(1999) study of prime time tele- vision spanning three decades found that there has been a significant increase in the number of female characters cast as professionals. In a subsequent study, Signorielli and Kahlenberg (2001) found that“women were just as likely to be cast in professional (doctor, teacher, lawyer) and white- collar (secretarial, managerial, clerical) jobs as males” (p. 10). Children Now (2004) reported that 28% of females in TV programs were portrayed in high-status occupations such as executive/CEO, physician, attorney, judge, or other professional, and only 5% were depicted as homemakers.

Similarly, Steinke’s(2005) research found an increase in the number of female scientists and engineers depicted in movies from 1991 to 2001.

In light of the literature above regarding demographics of female representations in the media, the second research question asked:

2. What is the demographic profile of the female action character?

Violence and Gender in the Media The third focus of the research is on the quantity and type of violence committed by female action characters. Vio- lence is typically associated with masculinity. Eschholz and Bufkin (2001) compared the biological sex of movie characters to the specific gender traits listed above and found that the traits of masculinity were highly correlated with violent acts shown in the movies regardless of biological sex. The assertion of the study was that female characters were masculinized when they engaged in violence. Media studies indicate that there has been an increase in the number of female action characters over time. Signorielli and Bacue (1999) analyzed prime-time television spanning three decades and found an increase in the number of female characters within the action-adventure genre. Other studies confirm this trend in various action- oriented media including film, television, and video games(Dill and Thill2007; Jansz and Martis2007; King2008; Sapolsky et al.2003).

Several analyses of media have expressed concern over the victimization of women in various media (Crosby2005; Haskell1974; Linz and Donnerstein1994; Williams1991).

However, other media research has found that women are victimized less often than are men (Dietz1998; Eschholz and Bufkin2001; Smith et al.1998). This fact mirrors real- world statistics because male offenders most often victimize other males (U.S. Bureau of Justice1999). Sapolsky et al.

(2003) examined the top-10 action films from the 1990s and found that males were shown as victims significantly more often than females. Signorielli (2003) found a significant change in the ratio of female violence to female victimization in prime time television over the past three decades, stating“instead of 16 women being victimized for each woman who hurts or kills, the odds are even: women are equally likely to hurt or kill as be hurt or killed”(p. 51).

Signorielli also counted the number of characters that were killed in network television from 1969 to 1988 and found that 83% were male and 17% were female (in Linz and Donnerstein1994). This finding is also consistent with actual statistics because 64.4% of homicides in the U.S. are committed by male offenders, 7.1% by females, and 28.5% unknown (U.S. Department of Justice2007).

In this research real-world statistics were used to develop a gendered orientation of violence. A U.S. Bureau of Justice special report on women offenders (1999) compared female types of violence with that of males. The report found that females commit only 14% of all violent crimes.

Violent female offenders usually engage in simple assaults directed at other females (75%). Female offenders usually have had some relationship with their victim who was often an intimate, relative, or acquaintance (62%). Also, more than half of female offenders act alone (53%) and they rarely use weapons (15%). In contrast, male offenders account for 86% of all violent crimes. They are most often violent toward other males (70%) and act in groups with other men (51%). Men usually have no prior relationship with their victim (64%) and use weapons (28%) more frequently than women. The report also states that the “consequences of male violence are generally more serious in terms of weapon used, injury, and out-of-pocket losses to the victim”(U.S. Bureau of Justice1999, p. 3).

Not all violence shown in the movies is related to crime because both bad and good characters engage in violent acts. Eschholz and Bufkin (2001) found a consistent pattern of masculine violence in popular films stating that it was “significantly related to both offending and victimization in the movies. The‘good’characters, who successfully accomplish masculinity, and the‘bad’ones who challenge their dominators, repeatedly resort to violence and crime” (p. 670). 736Sex Roles (2010) 62:734–746 Violent means are also used to protect society and citizenry, as is the case with male-dominated occupations such as the military and police. In the past several decades women have increasingly entered male-dominated occupa- tions in the U.S. (Wootton1997). This change in American social reality has been reflected in the movies with the inclusion of female police and soldiers as action characters (King2008; Tasker1998).

Violence can also take on a feminine form. Powers (1991) suggests that violence has been linked with the archetype of protectress. Similarly, Tasker (1998)has argued that female action characters draw upon a heroic maternal motif to create stereotypes of mother and wife who risk all to save children and loved ones.

In light of the literature above regarding violence and gender in the media, the following research questions asked:

3. Has there has been an increase in the number of action films that contain female action characters?

4. Are female action characters engaging in more acts of violence over time?

5. Do female action characters tend to commit gender- oriented violence?

Method Units of Analysis Female action characters appeared in multiple scenes of violence within the films analyzed. Therefore, this research had two units of analysis: (1) the Violent Female Action Character (VFAC), and (2) the violent scene in which she appeared. For purposes of this study, a VFAC was a leading female character in the film who engaged in at least one act of physical violence. Physical violence was defined as physical force exerted (with or without weaponry) in an attempt to cause bodily injury, death, or damage to property.

Sample The films analyzed in this research were those ranked by Internet Movie Data Base (IMDB) as the highest grossing American action films. The sample consisted of the top-20 action films (as determined by IMDB) released over a 15-year period between 1991 and 2005, totaling 300 movies. The year 1991 was selected as a starting point because it attracted feminist attention with the release of Thelma and LouiseandTerminator 2: Judgment Day (Brown1996; Greenberg et al.1991). The action genre was selected as the appropriate target population because itis a male-dominated film genre that frequently contains violent female action characters (Sapolsky et al.2003; Tasker1993).

Blockbuster action films generate the highest box office revenue of any film genre (King2002;MPAA2007). By focusing on the most successful films in the most successful genre this research tapped into social codes, norms, and gender stereotypes ingrained in mainstream American popular culture.

Of the 300 films in the sampling frame only those containing a VFAC were included in the final sample for coding. To determine if a VFAC was present in the film various resources were used including IMDB’smovie synopses, cast overviews, and links to external movie reviews. Where there was any doubt the movie was viewed in its entirety to determine if there was at least one VFAC present. The final sample contained 112 films (37% of the sampling frame).

If an additional female action character was shown in the film committing violent acts, was central to the story, and was confirmed to be a main character in IMDBs cast overview, then that VFAC was also included. For example, when two female characters worked closely together, as in Thelma & Louise, or when they were arch enemies, as in Kill Bill, then the second VFAC was added. The number of VFACs was limited to three per movie. This limit ensured that the VFAC definition of a leading character was main- tained and that the coding process was not overly compli- cated. Of the 112 movies in the sample, 58% contained one VFAC, 40.2% contained two VFACs, and two movies (1.8%) contained three. The sample contained 157 different VFACs depicted in 786 scenes of physical violence.

Codebook and Variables The codebook had two sections corresponding to each unit of analysis. The first section was used to code the gender stereotypes and demographics of the VFAC. The second section was used to code each violent scene in which the various VFACs appeared.

To address research question one, two variables were created to measure gender stereotypes: Love Interest and Heroine Status. Love Interest indicated who the VFAC was romantically involved with and included the categories (1) male hero, (2) villain, (3) none, and (4) other. Heroine Status reflected how the VFAC was represented in relation to the male hero, if one was present. These categories included: (1) main heroine, (2) assists male hero, (3) assists and is protected by male hero, (4) is protected by male hero, (5) villainess, and (6) other. The VFAC was coded as the main heroine if she was the principal character in the story and portrayed as strong, capable, and in charge. VFAC status declined as she went from being a main heroine, to Sex Roles (2010) 62:734–746737 assisting the male hero, to being protected by him— indicating a more submissive role.

The first section of the codebook also contained demo- graphic variables for the VFAC. The variables included: Age Range, Marital Status, Race (based on U.S. Census catego- ries), Education Level, Occupation, and Realism. For the Occupation variable, coders recorded the VFAC occupation as portrayed in the film. The variable Realism was added during the informal codebook testing and therefore was not part of the literature review. It became evident that the codebook needed a variable to distinguish those VFACs that were portrayed as superheroes, extra-terrestrials, and vam- pires. The Realism variable consisted of the following values:

(1) real, (2) part real, and (3) unreal. Real and unreal characters were based on human qualities or lack thereof.

The middle category indicated characters that were human beings but also engaged in unrealistic physical feats. For example, the Bride (played by Uma Thurman) inKill Bill Vols. I and II, was a human being with unrealistic martial arts abilities that enabled her to leap and flip through the air while killing dozens of enemies using a special sword.

The second section of the codebook contained variables for the violent scene unit of analysis. The variables addressed the quantity and type of violence the VFACs engaged in. Quantity of violence was determined using variables for the coding of each scene: Start-Time, End- Time, and Total Seconds (i.e. number of seconds calculated by subtracting Start-Time from End-Time). The type of violence committed by a VFAC was determined using five variables: Target, Relationship, Weaponry, Level of De- struction, and Motive. These five variables were used to determine if the VFAC engaged in masculine forms of violence and were based in large part on the U.S. Bureau of Justice statistics outlined in the literature review.

The Target variable identified who or what was the target of violence for the VFAC. Target included the following categories: (1) male, (2) female, (3) property, (4) alien or other life form, and (5) other. If the VFAC engaged in violence with many people at once, then coders were instructed to identify the predominant gender present. For example, if the VFAC shot at a group of police and they appeared to be mostly men, then they were coded as male.

Aliens were not gender coded.

The Relationship variable identified the relationship of the target to the VFAC. Relationship included the following categories: (1) intimate, (2) acquaintance, (3) stranger, (4) alien or other life form, and (5) other. If the VFAC engaged in violence with multiple people, then the coders were instructed to identify and code the VFAC’s relation to the primary target of her violence.

The Weaponry variable was created as a scale and coders were instructed to identify the highest level of weaponry used by the VFAC in the scene. Weaponry included thefollowing categories: (1) martial arts or hand-to-hand combat; (2) knives, clubs, non-firearms; (3) guns or firearms; (4) military and high-tech weapons; and (5) other.

The Level of Destruction variable was also set up as a scale and included the following categories: (1) no damage, (2) property damage only, (3) individual injury, (4) individual death, and (5) multiple injuries/deaths. Coders did not assume that property damage created collateral injuries or deaths unless these were specifically shown in the scene or indicated in dialogue.

The Motive variable was created to identify if there were any gendered motives for violence depicted in the scenes.

Motive did not relate to the plot of the film (i.e. issues of jealousy or revenge) instead it was specific to the single scene being coded. If the VFAC engaged in violence to protect a child or loved one, then Motive was coded (1) feminine. If the VFAC engaged in violence to protect a stranger or society at large, then Motive was coded (2) masculine. If a gendered motive did not occur, then Motive was coded using the following categories: (3) self-defense, only if the VFAC did not engage in the“first strike”defined below; (4) escape, if the VFAC engaged in violence in order to flee from her captors/ attackers; (5) evil, if the VFAC was a villainess; or (6) other.

Coding Before formal coding began the codebook and instructions were tested twice for reliability. During these pilot tests two coders were trained using verbal and written instructions and asked to select several movies from the sample that spanned the study period. Each coder gave feedback that was used to refine codebook instructions, increase reliabil- ity, and identify potential problems in coding films that spanned the 15-year period.

The coding tests revealed that over time action sequen- ces appeared to be getting shorter and occurring more rapidly, which made it difficult to establish the start and end times. Codebook instructions were refined to pinpoint the exact start and end time of each violent scene in order to achieve an acceptable level of reliability within a two- second margin of error. Specifically, timing for the violent scene began when the VFAC made the“first strike”, which was defined for coding purposes as the first moment she made a physical movement that resulted in an act of physical violence. For example, first strike occurred when Lara Croft pulled the trigger on her gun or when Catwoman cracked her whip. Timing stopped when the VFAC stopped engaging in the action. Intercoder reliability reached a minimum level of 80 percent agreement for all variables in the codebook before formal coding began.

As researcher, I watched and coded all movies. To ensure intercoder reliability a subsample was established using five coders (including the two coders from testing). Each coder 738Sex Roles (2010) 62:734–746 was asked to select seven films from the sample—one from every other year beginning 1991, 1992, or 1993. The subsample was purposive to ensure that at least one movie from each year was included. Of the 35 films selected by the five coders, eight were selected by more than one coder. This resulted in a subsample of 27 different films (24% of the sample) that contained 46 different VFACs shown in 212 different violent scenes. The subsample was coded by two coders (researcher included), and eight of the films in the subsample were coded by three coders.

Two tests of reliability were used for each unit of analysis. Kappa coefficients compute chance between coders’choices of the data available to them, whereas alpha and pi compute chance according to the probability of choices available within the data set (Krippendorff2004a).

Cohen’s kappa and Krippendorf’s alpha were used to determine reliability for the VFAC unit of analysis. The alpha coefficient was suitable for the nominal variables and the smaller subsample of 46 VFACs (Krippendorff2004a).

Cohen’s Kappa and Scott’s pi were used to determine reliability for the violent scene unit of analysis. The pi coefficient was suitable for continuous and ordinal varia- bles and the larger subsample of 212 violent scenes.

Intercoder reliability was strongest for manifest content such as Age, Marital Status, Race, all of which had over .90 kappa and alpha coefficients. Heroine Status, although latent and reliant upon subjective interpretation, achieved .91 kappa and alpha. Three variables for the VFAC unit of analysis fell below .80 including Realism (.79 kappa and alpha), Education (.73 kappa and .74 alpha), and Love Interest (.72 kappa and .73 alpha). According to Krippendorf, a minimum acceptable level of reliability is .80, but those variables with coefficients between .667 and .80 may be used for“drawing tentative conclusions”(2004b,p.241).

All variables for the violent scene unit of analysis reached above .80 for both kappa and pi coefficients including: Target (.81 kappa and pi), Relationship (.85 kappa and pi), Weaponry (.94 kappa and pi), Level of Destruction (.85 kappa and pi), Motive (.84 kappa and pi), and Total Seconds (.81 kappa and pi).

Coders wrote in occupation, if known. Coders also included a brief explanation of what happened in each violent scene and commented on the general character of the VFAC and her violent actions. This information was helpful in understanding the circumstances of VFAC actions as well as clarifying and supporting the overall research findings.

Results Research question one dealt with VFAC gender stereotypes.

Two variables, Heroine Status and Love Interest, were used to identify gender stereotypes present in VFACs. Frequencydistributions for the two variables are shown in Table1.Of the 157 VFACs coded, only 15.3% were depicted as the main heroine, and 58.6% were depicted in a submissive role to the male hero. Submissive included the categories of assists male hero (28.0%), assists and protected by male hero (5.1%), and protected by male hero (25.5%). Seventy percent of VFACs were portrayed in a romantic relation- ship. Of those VFACs shown in relationships, over 60% were involved with the male hero.

Crosstabs for the two variables contained several cells with a frequency less than five. In order to run chi square analysis the categories of assists male hero and assisted and protected by male hero were combined. In addition, the categories villain, villainess, and other were excluded. The results for the recoded data are shown in Table2. These results serve to highlight the relationship between the VFAC and the male hero. Chi square found a significant relationship between the recoded variables. The results showed that VFACs were most likely to be submissive in terms of being protected by the male hero when they were romantically linked to him. Further, those VFACs that had no romantic involvement were most likely to be main heroines and least likely to be protected by the male.

In answer to question one, the results showed that the majority of VFACs maintained gender stereotypes with respect to feminine traits of submission and affection. On average, most VFACs were depicted in some type of romantic relationship and often romantically linked with the male hero of the story. More than half of all VFACs were shown in a submissive role to the male hero of the story.

Further, the more submissive a VFAC was to a male hero, i.e. protected by rather than assisting, the more likely she was to be romantically involved with him, thus linking the feminine traits of submission and affection. The results also showed that main heroines were less likely to exhibit these feminine traits than were other types of VFACs. Table 1Frequency distribution for heroine status and love interest.

Heroine Statusn% Main Heroine 24 15.3 Assists Male Hero 44 28.0 Protected by Male Hero 40 25.5 Assists & Protected by Male Hero 8 5.1 Villainess 28 17.8 Other 13 8.3 Love Interest Male Hero 66 42.0 Male Villain 19 12.1 None 47 29.9 Other 25 15.9 Sex Roles (2010) 62:734–746739 Research question two focused on VFAC demographics.

Results showed that over 90% of VFACs were portrayed as young women, most in their twenties (55%) and thirties (38%). The percentage for VFACs in this age range was higher than the U.S. female population estimated to be 27.3% (U.S. Census2006). The VFAC’s age makes sense when compared to the youth audience targeted by film- makers because 55% of all movie-goers are between the ages of 12 and 39 (MPAA2007). The data also showed that 81% of VFACs were not married. This statistic was also high compared to the estimated 60% of U.S. females between the ages of 20 and 34 who are not married (U.S.

Census2006).

Results found that 74.5% of VFACs were white, 9.5% were African American, 9.5% Hispanic, 5.1% Asian, and 1.4% other. African American and Hispanic VFACs were underrepresented when compared to their estimated U.S.

populations of 13.5% and 15.1% respectively (U.S. Census 2008). It should also be noted that Halle Berry played one- quarter of the VFAC roles coded as African American.

The data revealed that over 60% of VFACs were employed and most tended to have high status jobs. These findings suggest the VFACs’occupations were more in line with real-world statistics because nearly 60% of all women over the age of 16 are employed. Women account for 51% of all U.S. workers in the high-paying management, professional, and related occupations (U.S. Department of Labor2008). The education level of VFACs was difficult to code because it was often not mentioned in the story.

However, 11% of VFACs had Ph.D. status inferred by their title or occupation. This far exceeds the U.S. female population who hold doctoral degrees, which is .8% (U.S.

Census2007). It was also estimated from occupational status that 24% of VFACs held bachelor’s degrees, which is higher than the U.S. female population of 17.8% (U.S.

Census2007). Over a third of the VFACs were portrayed in male-dominated careers. These consisted of 14.5% of VFACs who were scientists, engineers, or computer programmers, and 23.4% who were in the military, law enforcement, or security services. However, only one VFAC with a Ph.D. was a main heroine, the rest eitherassisted or were protected by a male hero. Only 4% were depicted as housewives.

Special attention was given to the Realism variable because during the coding process it appeared that VFACs were becoming more unrealistic over time. This assumption was confirmed using chi square analysis with post hoc testing. The results are shown in Table3. The percentage of real VFACs decreased over the time periods, while the percentage of part-real and unreal VFACs increased. Post hoc tests showed that the presence of real VFACs had a standardized residual of 2.0 in 1991–1995 versus−2.4 in 2001–2005 indicating a significant decline in realistic VFACs. Part-real VFACs had a standardized residual of 2.5 in 2001–2005 indicating that they had a significant impact on the change in realism. This finding was explained by looking at the data. A number of part-real VFACs were shown in several heroine action films and sequels during 2001–2005, includingKill Bill(2002, 2003), Lara Croft(2001, 2003), andResident Evil(2002, 2004).

In answer to research question two, the demographic profile of the VFAC showed that she was typically young, white, unmarried, highly educated, and often depicted in a male-dominated or high-status career. Additionally, VFACs became increasingly unrealistic over the 15-year period.

The final focus of the research was on the quantity and type of violence committed by the VFAC. Research question three asked if there had been an increase in the number of female action characters over time. Frequency distributions in Table4show that there appeared to be an increase in the number of VFACs. However, analysis of variance (ANOVA) revealed that there was no significant change,F(1, 13) = 3.82,p <.10.

Research question four asked if the quantity of VFAC violence had increased over time. The coding process produced the continuous variable Total Seconds, i.e. the total number of seconds of violence recorded for each scene. ANOVA tests for the variables Total Seconds and Year showed a significant change over the 15-year period, F(14, 771) = 2.30,p<.01). Both variables were then tested Table 3VFAC realism: 1991–2005.

RealismTime Periods Average 1991–1995 1996–2000 2001–2005 n%n%n%% Real 36 85.7 35 71.4 25 37.9 61.1 Part Real 2 4.8 4 8.2 18 27.3 15.3 Unreal 4 9.5 10 20.4 23 34.8 23.6 χ2= 29.0*,p≤.001,df=4,n= 157 *0 cells have an expected count of less than 5. The minimum expected count is 6.42 Table 2Percent heroine status within love interest: male hero only.

Love InterestHeroine Status Average Main HeroineAssists Male HeroProtected by Male Hero n=13n=45n=35 Male Hero 46.2% 60.0% 82.9% 66.7% None 53.8% 40.0% 17.1% 33.3% χ2= 7.49,p≤.05,df=2,n=93 740Sex Roles (2010) 62:734–746 using regression analysis which showed there was no significant linear relationship,β=−.048,t=−1.34,p< .20.

The non-linear trend is understandable when looking at the frequencies displayed in Table4. The total seconds of violence had peaks and valleys. Total seconds of violence in years 2003 (20.4%) and 2004 (14.4%) accounted for over one-third of the violence shown during the 15-year period. These higher numbers were attributable to a few films:Kill Bill Vols. I and II(2003, 2004);Lara Croft: The Cradle of Life(2003);Terminator 3(2003);Underworld (2003);Catwoman(2004); andResident Evil: Apocalypse (2004). Several years had small amounts of total seconds of violence. Referring to the data for the VFAC unit of analysis revealed that four main heroines appeared in years 2003 and 2004 when total seconds of violence peaked.

Conversely, no main heroines appeared in any of the films during the years 1998, 1999, and 2000 when total seconds of violence were low. This suggests that the appearance of main heroines in the films impacts the quantity of violence.

In answer to question four, the results showed a significant change in the total seconds of violence shown in each scene over the 15-year period, but there was no linear relationship.

Therefore, an increase in the quantity of VFAC violence over the study period could not be established. Instead, there were a few years that contained movies with more seconds of violence, especially years 2003 and 2004, and several years that contained fewer seconds of violence.

Research question five asked if VFACs committed gender-oriented violence. In order to easily interpret and display the findings in tables, the violent scenes weregrouped into three time periods and treated as ordinal variables. Chi square analyses were used to analyze the data for the five variables: Target, Relationship, Weaponry, Level of Destruction, and Motive.

The data for the Target variable are shown in Table5.

Target changed significantly over the time periods. VFACs were shown in violent scenes most often fighting against males. Fighting against males declined in periods 1995– 2000 and 2001–2005, while violence against aliens in- creased. Post hoc tests showed a standard residual of−3.2 for aliens in 1991–1995 and 2.2 in 2001–2005 indicating that the absence or presence of aliens had a significant impact on the VFAC target of violence. A standard residual of−2.3 for females in 1996–2000 also had a significant impact and called attention to the lack of female-on-female violence during that period. The category of other included such targets as dinosaurs and animals. On average, VFAC’s Table 4Frequency distribution for films, VFACs, and violent scenes by year.

Year Films VFACs Violent Scenes Total Seconds of Violence Mean Scene Length (sec.) n%n%n%n%n 1991 6 5.4 9 5.7 35 4.5 1,444 7.6 41.3 1992 4 3.6 5 3.2 19 2.4 575 3.0 31.9 1993 5 4.5 5 3.2 27 3.4 867 4.6 32.1 1994 7 6.3 9 5.7 34 4.3 894 4.7 26.3 1995 10 8.9 14 8.9 50 6.4 1,039 5.5 20.8 1996 5 4.5 6 3.8 36 4.6 984 5.2 27.3 1997 10 8.9 14 8.9 49 6.2 1,225 6.5 25.0 1998 5 4.5 7 4.5 31 3.9 622 3.3 20.1 1999 10 8.9 15 9.6 64 8.1 591 3.1 9.2 2000 6 5.4 7 4.5 27 3.4 419 2.2 15.5 2001 6 5.4 8 5.1 49 6.2 1,330 7.0 27.1 2002 6 5.4 10 6.4 53 6.7 608 3.2 14.3 2003 14 12.5 19 12.1 129 16.4 3,853 20.4 29.9 2004 9 8.0 16 10.2 114 14.5 2,725 14.4 23.9 2005 9 8.0 13 8.3 69 8.8 1,706 9.0 24.4 Total 112 100.0 157 100.0 786 100.0 18,882 100.0 24.2 Table 5Target of VFAC violence: 1991–2005.

TargetTime Period Average 1991–1995 1996–2000 2001–2005 n%n%n%% Male 114 71.3 133 66.2 223 54.7 61.1 Female 18 11.3 10 5.0 50 12.3 10.1 Property 13 8.1 17 8.5 29 7.1 7.7 Alien 15 9.4 41 20.4 106 26.0 21.1 χ2= 29.07,p≤.001,df=6,n= 769 Sex Roles (2010) 62:734–746741 target was a male or group of males in 61.1% of the violent scenes and female in 10.1% of the scenes. The findings here support a masculine orientation for VFAC violence because statistics show that real world violence occurs most often between offenders and victims of the same gender (U.S. Bureau of Justice1999).

The data for Relationship variable were analyzed using chi square. The results found that Relationship did not change significantly over the time periods,χ 2(8,n= 773) = 13.98,p<.10. On average, the target of VFAC violence shown in each scene was most often a stranger or group of strangers (36.7%,n= 284). Acquaintances were the targets in 26.5% (n= 205) of the scenes, followed by aliens 19.0% (n= 147), other 10.0% (n= 77), and intimates 7.8% (n= 60).

The findings here support a masculine orientation for VFAC violence because statistics show that real-world male violence is most often directed toward strangers and less often toward acquaintances or intimates (U.S. Bureau of Justice1999).

The data for the variables Weaponry and Level of Destruction are shown in Table6. Weaponry was treated as a scale ranging from martial arts and hand-to-hand combat to high-tech weaponry. Weaponry changed signif- icantly over the time periods. The use of martial arts was lower in periods 1991–1996 and 1996–2000 than in 2001– 2005. A reverse trend occurred for firearms. The increased use of military and high-tech weapons in 1996–2000 can be attributed to several VFACs who were depicted as futuristic military personnel. These characters included Lt. Ripley in Alien: Resurrection(1997), Private Flores and Navy Pilot Ibanez inStarship Troopers(1997), and Lt. Cmdr.

Devereaux inWing Commander(1999). Each VFAC used high levels of weaponry against alien attackers resulting inhigh levels of destruction. Table6illustrates that on average VFACs used some form of weaponry in over 59% of the violent scenes (excluding other weaponry such as cars, fire, and poison). U.S. crime statistics show that 28% of male offenders use weapons versus 15% for female offenders (U.S. Bureau of Justice1999). Therefore, the high use of weaponry by VFACs supports a masculine orientation for VFAC violence.

Data for the variable Level of Destruction are shown above in Table6. Level of Destruction was set up as a scale ranging from no damage to multiple injuries/deaths. Level of Destruction changed significantly over the time periods.

The highest category, multiple injuries/deaths, appeared in 14.6% of the violent scenes in period 1991–1995 and increased to 22.9% by 2000–2005. This finding suggests that VFACs were causing more death and damage as a consequence of their violence. On average, individual death occurred in 40.7% of the scenes and multiple injuries/ deaths occurred in 20.7% of the scenes.

Weaponry and Level of Destruction appear together in Ta b l e6because their relationship was statistically signifi- cant,χ 2(16,n= 786) = 191.51,p<.001). The use of martial arts by VFACs resulted most often in individual injury (68.3%) and resulted less often in individual death (11.0%) or multiple injuries/deaths (9.8%). Conversely, the use of military and high-tech weaponry resulted most often in multiple injuries/deaths (44.8%) versus individual injury (19.0%) or individual death (12.1%). Post hoc tests found that nearly all of the cells had a standard residual value of +/- 2.0 that validated the high level of correlation between the two variables. For example, the variable martial arts had a standard residual of 6.8 for individual injury versus−3.8 for multiple injuries/deaths. High-tech weaponry had a standard Time Period Average 1991–1995 1996–2000 2001–2005 n%n%n%% Weaponry Martial arts, hand-to-hand 61 37.2 59 28.5 126 30.4 31.3 Knives, clubs 28 17.1 23 11.1 94 22.7 18.4 Firearms 53 32.3 75 36.2 134 32.3 33.3 Military and High Tech 8 4.9 28 13.5 22 5.3 7.4 Other 14 8.5 22 10.6 39 9.4 9.5 Level of Destruction No damage 5 3.0 22 10.6 46 11.1 9.3 Property damage 21 12.8 16 7.7 39 9.4 9.7 Individual injury 75 45.7 89 43.0 156 37.6 40.7 Individual death 39 23.8 36 17.4 79 19.0 19.6 Multiple injuries and deaths 24 14.6 44 21.3 95 22.9 20.7 Table 6Weaponry used by VFAC and level of destruction inflicted: 1991-2005.

Chi square:Weaponryχ2= 28.20,p≤.001,df =8,n=786; Level of Destructionχ2= 19.44, p≤.05,df=8,n=786 742Sex Roles (2010) 62:734–746 residual of -2.6 for individual injury versus 4.0 for multiple injuries/deaths.

These findings support a masculine orientation for violence because statistics show that the consequence of male violence is more serious in terms of weapons used and injuries inflicted (U.S. Bureau of Justice1999).

Chi square analysis found that the variable Motive changed significantly over the time periods,χ 2(10,n= 786) = 51.48,p<.001. Feminine motive for violence shown in the scenes decreased from 22.0% in 1991–1996 to 8.2% in 2001–2005. This meant there were fewer scenes of VFACs using violence to protect a child or loved one.

Masculine motive increased from 17.1% in 1991–1995 to 28.0% in 1996–2000, and then decreased to 16.1% in 2001–2005. This meant there were more scenes of VFACs using violence to protect citizenry or society during 1996– 2000. This time period contained several VFACs who were depicted as futuristic military personnel and as such took on the role of protecting society.

Self-defense as a motive rose steadily in each period, while escape and evil declined. Self-defense was coded when VFACs responded to attack and did not initiate the first strike of violence in the scene. On average, masculine motive was 19.5%, feminine motive was 11.2%, self-defense was 24.2%, escape was 12.1%, evil was 18.1%, and other was 15.0%. VFACs were shown using masculine motives less often than shown using self-defense. Feminine motives were shown least often. Therefore, the results do not support a gender orientation for motive of VFAC violence.

In answer to research question five, the results showed that VFACs exhibited gender-oriented violence for four of the five variables tested. Target, Relationship, Weaponry, and Level of Destruction were all found to be masculine.

Weaponry and Level of Destruction showed significant change over the time periods, and the relationship between the two variables was significant.

In summary, the results of the research found that VFACs maintained feminine stereotypes of submission and affection, especially in relation to male heroes present in the films. VFACs on average were young, white, highly educated, and unmarried. There was a significant change in VFAC realism over time. VFACs engaged in masculine types of violence because they most often fought against males and against strangers, and they often used weapons and caused high levels of destruction. However, there was no significant increase in the total seconds of VFAC violence over the 15-year study period.

Discussion The VFAC is a recent addition to contemporary American cinema and has the potential to redefine female heroines, forbetter or worse. From a superficial or visceral viewpoint it appears that the emergence of the VFAC is an indicator of gender equality. However, by looking more closely this research found that the VFAC was most often portrayed in a submissive role and was romantically involved with a dominant male hero character, thus reifying gender stereotypes.

Instead of accepting female action characters as empower- ing role models, it is useful to take a critical view and understand that VFACs are a market-driven commodity.

Mainstream movies are created to capture the largest audience possible (Sklar1994). It is no accident that the most succes- sful films adhere to gender stereotypes and strive to be non- offensive in order to appeal to a mainstream audience.

Within the film industry, the action genre commands the highest box office revenue. Of the top 100 domestic box office hits of all time, 59 are action blockbusters with a main male hero, and each has grossed over $185 million dollars (IMDB2008). Hollywood has attempted to break the action genre barrier for females with movies such asG.

I. Jane(1997),Lara Croft(2001, 2003),Kill Bill(2003, 2004), andAlienseries (1979, 1986, 1992, 1997, 2001).

However, these films have generated far less box office revenue than blockbuster male-hero action movies. Accord- ing to IMDB (2008), only four action-heroine movies have earned over $100 million at the domestic box office:Lara Croft: Tomb Raider(2001), bothCharlie’s Angelsmovies (2000, 2003), andMiss Congeniality(2000). The sample used in this research is evidence of the fact that action movies with a main heroine are relatively rare. Of the 300 movies included in the sampling frame, 37% (n= 112) con- tained a VFAC and only 7% (n= 22) contained a main heroine. Instead of breaking gender barriers and portraying empowering female roles, most VFACs were shown as sidekicks and helpmates to the more dominant male hero and were frequently involved in a romantic relationship with him. Over 40% of all VFACs were portrayed as girlfriends or wives to the male heroes in the movies. The findings suggest that VFACs seem to be inserted into the story to support and promote the actions of the male hero.

The VFAC often appeared as a damsel-in-distress providing the impetus for a male hero to overcome obstacles in order to save her. This was more likely to occur if the VFAC was also linked romantically to the male hero. The exceptions were the few VFACs depicted as main heroines. Main heroines were less likely to be romantically linked to the male hero, and therefore less likely to assist or be protected by him. This was to be expected because main heroines took on the role of the central hero figure and therefore were less likely to exhibit feminine stereotypes of submis- sion and affection.

A more troubling finding was the number of VFACs that died in the films. Coders took notes on what happened in each violent scene and VFAC deaths were counted. Nearly Sex Roles (2010) 62:734–746743 30% of all VFACs died by the end of the movie. Of the VFACs who died, four (8%) were main heroines. These included the two heroines ofThelma&Louise(1991), Captain Walden inCourage under Fire(1996), and Lt.

Ripley inAlien 3 (1992). Another 47% were evil and consequently killed as punishment for their bad acts, and 45% were categorized as submissive to the male hero. A critical review of these death scenes revealed disturbing imagery because VFACs died heart-wrenching deaths in the arms of male heroes. Some examples include: Dizzy, the jilted girlfriend inStarship Troopers(1997), who sacrifices her life to save Johnny’s; Nyssa, the good vampire inBlade II(2002), who lets Blade carry her into the morning light as she crumbles into dust; Trinity, the faithful girlfriend in Matrix Revolution(2003), who saves Neo and then dies in his arms; Elektra, the heroine inDaredevil(2003), who battles Bullseye then crawls to Matt as he listens to her heart stop beating with his ultra-sensitive hearing; and Valerious, the immortal Princess inVan Helsing(2004), who willingly lets Van Helsing impale her and end her tragically-cursed life. These examples of dying, self- sacrificing women illustrate the extreme end of submissive affection and feminine stereotypes shown too often in popular action movies.

The profile of the VFAC as being young, white, unmarried, and highly educated is aimed at the youth audience. Many young viewers may not match up with the VFAC profile, yet they identify with the normative social codes that are embedded in these films—social codes that reflect what is valued in American culture.

VFACs also appeared to become more unrealistic over time. VFACs included superheroines, extra-terrestrial beings, and vampires, all of which were aided by special effects and computer generated imagery. In addition, coders noticed that violent scenes seemed to occur more frequently and in short bursts, thus increasing the rapid pace of action.

These changes can be explained by advances in digital technology that have made special effects and the editing process easier and less costly (Metz2006). As Pfeil (1995) suggests, the movie industry“offers us an altogether different economy of pleasure, in which a giddying blur of high-speed chase and the gratifyingly spectacular release of aggressive impulse occurs at regularly recurring intervals throughout the film”(p. 23). Coder comments reinforced this view with references to VFACs engaging in a constant stream of battles with bad guys and aliens, dodging bullets, jumping out of windows, and being pursued in high-speed chase scenes.

The rapid action and heart-pounding special effects in action movies translate into income for Hollywood—the more explosions and carnage the better. Blockbuster action films generate the highest box office revenue of any film genre and are the most internationally exported films (King2002;MPAA2005,2007). Due to the success and popularity of the action genre, which is disseminated to a broad audience both nationally and internationally, we need to be concerned with continued gender stereotypes set within a violent framework. In addition, citizens around the world are exposed to the violence in action media and we should be concerned with the violent way in which American culture is portrayed globally (Gerbner1994).

This research found no significant increase in the number of seconds of VFAC violence shown over the 15-year period. Instead, there were clusters of films that displayed certain types of VFAC violence. During 1996– 2000 several VFACs were depicted as futuristic military personnel who used high-tech weaponry to fight off alien attackers in outer space. Subsequently, in 2001–2005 several partially realistic VFACs engaged in martial art combat, and used knives, clubs, or other objects when fighting. VFAC violence seemed to move away from a distanced type of fighting that relied on powerful weaponry toward a more up-close combat style with exaggerated physical action. Further research to extend the study period is warranted to see if this was a continuing trend.

VFACs tended to be masculine in their orientation for violence, which was expected because VFACs by definition were violent. The target of VFAC violence was usually a male or group of males (61.1%) and often strangers (36.7%). VFACs used weapons most of the time (59.1%) and frequently caused a high level of destruction with multiple injuries and deaths (20.7%). These findings alone might indicate that the VFAC was a strong, empowered character. However, most often VFAC violence was guided by or served the interests of a dominant male hero, thus reinforcing gender stereotypes.

As Hall ( 1997) suggests, images and representations in media stem from gender norms, roles, and values deeply embedded in our culture. They are the means by which power and ideology are made to signify particular dis- courses, create maps of meaning, and“re-present”social reality to the viewer. In this way, the majority of the VFACS examined in this research both re-presented and re- articulated a social reality based on normative social codes and gender stereotypes. Those few VFACs who were main heroines were more outliers to, rather than representations of, social norms and stereotypes.

The debate continues as to whether the few action heroines that we are familiar with, such as Lt. Ripley, Sarah Connor, or Lara Croft, have broken down gender barriers in action films. This research provides evidence that the majority of female action characters shown in American cinema are not empowering images, they do not draw upon their femininity as a source of power, and they are not a kind of“post woman”operating outside the boundaries of gender restrictions. Instead, they operate inside socially 744Sex Roles (2010) 62:734–746 constructed gender norms, rely on the strength and guidance of a dominant male action character, and end up re-articulating gender stereotypes.

This research has limitations because it examined only female action characters and therefore is unable to make claims about changes occurring for male action characters.

Male action heroes dominate the action genre and their roles may be changing along with shifts in VFAC violence.

Further research in this area would provide a more complete understanding of VFACs in relation to their male counterparts. The rise in VFACs and the number of violent scenes during 2003 and 2004 is an indicator that there could be a trend of increasing amounts of violence that this particular research was unable to establish. The study should be continued to see if a trend in the quantity of violence occurred after 2005. Finally, the focus of this research was on successful action films in order to understand the gender stereotypes exhibited in mainstream American popular culture. However, a larger random sample might provide different results and comparative research could be warranted.

References Bandura, A. (1986).Social foundations of thought and action: A social cognitive theory. Englewood Cliffs: Prentice Hall.

Blumer, H., & Hauser, P. (1933).Movies, delinquency, and crime.

New York: MacMillan.

Brown, J. (1996). Gender and the action heroine: Hardbodies and the “Point of No Return”.Cinema Journal, 35(3), 52–71.

Brown, J. (2005). Gender, sexuality, and toughness: The bad girls of action film and comic books. In S. Inness (Ed.),Action chicks (pp. 47–74). New York: Palgrave MacMillan.

Butler, J. (1990).Gender trouble: Feminism and the subversion of identity. New York: Routledge.

Butler, J. (1999). Gender is burning. In S. Thornham (Ed.),Feminist film theory(pp. 336–349). New York: New York University Press.

Calvert, S. L., Kondla, T. A., Ertel, K. A., & Meisel, D. S. (2001).

Young adults’perception and memories of a televised woman hero.Sex Roles, 45,31–52.

Children Now. (2004).Prime time diversity report. Retrieved from http://www.childrennow.com.

Clover, C. J. (1992).Men, women, and chain saws: Gender in the modern horror film. Princeton: Princeton University Press.

Crosby, S. (2005). Female heroes snapped into sacrificial heroines. In S. Inness (Ed.),Action chicks(pp. 153–178). New York: Palgrave MacMillan.

Dietz, T. L. (1998). An examination of violence and gender role portrayals in video games: Implications for gender socialization and aggressive behavior.Sex Roles, 38, 425–442.

Dill, K. E., & Thill, K. P. (2007). Video game characters and the socialization of gender roles: Young people’s perceptions mirror sexist media depictions.Sex Roles, 57, 851–864.

Dyer, R. (2002).The matter of images: Essays on representation. New York: Routledge.

Elasmer, M., Hasegawa, K., & Brain, M. (1999). The portrayal of women in U.S. prime time television.Journal of Broadcasting & Electronic Media, 44,20–34.Eschholz, S., & Bufkin, J. (2001). Crime in the movies: Investigating the efficacy of measures of both sex and gender for predicting victimization and offending in film.Sociological Forum, 16, 655–676.

Eschholz, S., Bufkin, J., & Long, J. (2002). Symbolic reality bites:

Women and racial/ethnic minorities in modern film.Sociological Spectrum, 22, 299–334.

Gerbner, G. (1970). Cultural indicators: The case of violence in television drama.The Annuls of the American Academy of Political and Social Science, 388,69–81.

Gerbner, G. (1994). Making a killing.Psychology Today, 27(4), 18.

Gerbner, G. (1998). Cultivation analysis: An overview.Mass Commu- nication & Society, 1(3/4), 175–194.

Glascock, J. (2001). Gender, roles on prime-time network television:

Demographics and behaviors.Journal of Broadcasting & Electronic Media, 45, 656–669.

Glascock, J. (2003). Gender, race, and aggression in newer TV networks’primetime programming.Communication Quarterly, 51,90–100.

Greenberg, H. R., Clover, C. J., Johnson, A., Chumo, P. N., Henderson, B., Williams, L., et al. (1991). The many faces of “Thelma & Louise”.Film Quarterly, 4(2), 20–31.

Hall, S. (Ed.). (1997).Representation: Cultural representations and signifying practices. Thousand Oaks: Sage.

Haskell, M. (1974).From reverence to rape: The treatment of women in the movies. New York: Holt Rinehart and Winston.

Hills, E. (1999). From figurative males to action heroines: Further thoughts on active women in cinema.Screen, 40(1), 38–50.

Hooks, B. (1996).Reel to real: Race, sex, and class at the movies.

New York: Routledge.

Huesmann, L. R. (1999). The effects of childhood aggression and exposure to media violence on adult behaviors, attitudes and mood: Evidence from a 15-year cross-national longitudinal study.

Aggressive Behavior, 25,18–29.

Inness, S. (Ed.). (2004).Action chicks: New images of tough women in popular culture. New York: Palgrave MacMillan.

Internet Movie Data Base (“IMDB”). (2008). Retrieved fromhttp:// www.imdb.com/.

Jansz, J., & Martis, R. G. (2007). The Lara phenomenon: Powerful female characters in video games.Sex Roles, 56, 141–148.

King, G. (2002).New Hollywood cinema, an introduction. New York:

Columbia University Press.

King, N. (2008). Generic womanhood: Gendered depictions in cop action cinema.Gender & Society, 22, 238–260.

Krippendorff, K. (2004a). Reliability in content analysis: Some common misconceptions and recommendations.Human Commu- nications Research, 30,411–433. Krippendorff, K. (2004b).Content analysis: An introduction to its methodology. Thousand Oaks: Sage.

Lazarsfeld, P. F., & Merton, R. K. (1948). Mass communication, popular taste, and organized social action. In L. Bryson (Ed.), The communication of ideas(pp. 95–118). New York: Harper & Row.

Linz, D., & Donnerstein, E. (1994). Dialogue: Sex and violence in slasher films: a reinterpretation.Journal of Broadcasting & Electronic Media, 38, 243–246.

Lorber, J. (1999). Embattled terrain: Gender and sexuality. In M. M.

Ferree, J. Lorber, & B. B. Hess (Eds.),Revisioning gender(pp.

416–48). Thousand Oaks: Sage.

Lueptow, L. B., Garovich-Szabo, L., & Lueptow, M. B. (2001). Social change and the persistence of sex typing: 1974–1997.Social Forces, 80,1–35.

Magoulick, M. (2006). Frustrating female heroism: Mixed messages in Xena, Nikita, and Buffy.The Journal of Popular Culture, 39, 729–747. Sex Roles (2010) 62:734–746745 Mastro, D. E., & Greenberg, B. S. (2000). The portrayal of racial minorities on prime time television.Journal of Broadcasting & Electronic Media, 44, 690–703.

McCaughey, M., & King, N. (Eds.). (2001).Reel knockouts: Violent women in the movies. Austin: University of Texas Press.

McRobbie, A. (2004). Post-feminism and popular culture.Feminist Media Studies, 4(3), 235–264.

Metz, C. (2006, May). Hollywood reboots.PC Magazine,62–74.

Modleski, T. (1991).Feminism without women: Culture and criticism in a‘postfeminist age’. New York: Routledge.

Motion Picture Association (2005). U.S. entertainment industry: 2005 MPA market statistics. Retrieved fromhttp://www.mpaa.org/ InternationalTheatricalMarket.pdf.

Motion Picture Association (2007). U.S. movie attendance study.

Retrieved fromhttp://www.mpaa.org/MovieAttendanceStudy.pdf.

Mulvey, L. (1975). Visual pleasure and narrative cinema.Screen, 16 (3), 6–18.

Pfeil, F. (1995).White guys: Studies in postmodern domination and difference. New York: Verso.

Powers, M. A. (1991).The heroine in western literature. Jefferson:

McFarland & Company, Ltd.

Rosen, M. (1973).Popcorn Venus: Women, movies, and the American dream. New York: Coward, McCann and Geoghegan.

Rowe-Karlyn, K. (2003).Scream, popular culture, and feminism’s third wave:“I’m not my mother.”Genders, 38. Retrieved from http://www.genders.org/g38/g38_rowe_karlyn.html.

Sapolsky, B. S., Molitor, F., & Luque, S. (2003). Sex and violence in slasher films: Re-examining the assumptions.Journalism & Mass Communication Quarterly, 80(1), 28–38.

Signorielli, N. (2003). Prime-time violence 1993–2001: Has the picture really changed?Journal of Broadcasting & Electronic Media, 47,36–57.

Signorielli, N., & Bacue, A. (1999). Recognition and respect: A content analysis of prime-time television characters across three decades.Sex Roles, 41, 527–544.

Signorielli, N., & Kahlenberg, S. (2001). Television’s world of work in the nineties.Journal of Broadcasting and Electronic Media, 45,4–22.

Sklar, R. (1994).Movie-made America. New York: Vintage Books.

Smith, S. L., Wilson, B. J., Kunkel, D., Linz, D., Potter, J., Colvin, C.

M., et al. (1998).Violence in television programming overall:University of California, Santa Barbara study. In National television violence study: Vol. 3(pp. 5–220). Newbury Park:

Sage.

Stacey, J. (2000). Desperately seeking difference. In A. E. Kaplan (Ed.),Feminism and film(pp. 456–465). New York: Oxford University Press.

Steinke, J. (2005). Cultural representations of gender and science.

Science Communication, 27,27–63.

Tasker, Y. (1993).Spectacular bodies: Gender, genre, and the action cinema. London: Routledge.

Tasker, Y. (1998).Working girls: Gender and sexuality in popular cinema. New York: Routledge.

Tasker, Y., & Negra, D. (Eds.). (2007).Interrogating postfeminism:

Gender and the politics of popular culture. Durham: Duke University Press.

Trinh, T. M. (1991).When the moon waxes red: Representation, gender, and cultural politics. New York: Routledge.

Twenge, J. (1997). Changes in masculine and feminine traits over time: A meta-analysis.Sex Roles, 36, 305–325.

U.S. Bureau of Justice Statistics. (1999). Women offenders. Retrieved fromhttp://bjs.ojp.usdoj.gov/content/pub/pdf/wo.pdf.

U.S. Census Bureau. (2006). American community survey. Retrieved fromhttp://factfinder.census.gov/servlet/ACSSAFFPeople?_sub menuId=people_0&_sse=on.

U.S. Census Bureau. (2007). Educational attainment. Retrieved from http://www.census.gov/population/socdemo/education/cps2007/ Table1-01.xls.

U.S. Census Bureau. (2008). Minority links. Data set 10—resident population by race, Hispanic origin, and single years of age:

2008. Retrieved fromhttp://www.census.gov/compendia/statab/ cats/population.html.

U.S. Department of Justice. (2007). FBI crime report: Expanded homicide data table 3. Retrieved fromhttp://www.fbi.gov/ucr/ cius2007/offenses/expanded_information/data/shrtable_03.html.

U.S. Department of Labor. (2008). Women’s bureau: Quick stats of women workers, 2008. Retrieved fromhttp://www.dol.gov/wb/ stats/main.htm.

Williams, L. R. (1991). Film bodies: Gender, genre, and excess.Film Quarterly, 44(4), 2–13.

Wootton, B. (1997, April). Gender differences in occupational employment.Monthly Labor Review,23–33.

746Sex Roles (2010) 62:734–746 Copyright of Sex Roles is the property of Springer Science & Business Media B.V. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.