#研究分享#【电影票房中的暴力研究】

#研究分享#【电影票房中的暴力研究】对电影中的暴力研究由来已久,引起了学者和公众的广泛关注。美国的研究人员研究了从1992年到2012年之间的2094部动作电影,通过反射(reflective)、抗拒(reactance)和市场三个模型,发现了电影中的暴力内容与广泛的美国文化和心理抗拒没有显著关系,而是与市场有强烈的关系。研究发现,PG-13级别的电影其中的暴力内容可以显著增长票房销量。从而得到了结论:观众(市场)的需求是暴力内容剧增的主要原因。

Violence at the Box Office

Considering Ratings, Ticket Sales, and Content of Movies

The negative effects of violent content in movies have recently been a hot topic among both researchers and the general public. Despite growing concern, violence in movies has persisted over time. Few studies have examined why this pattern continues. To fill this gap in the literature, we examine how Motion Picture Association of America (MPAA) movie rating descriptors predict ticket sales of 2,094 movies from 1992 to 2012. We test the validity of three theoretical models: (1) the reflective model, (2) the reactance model, and (3) the market model. We find that violent content is linked neither to violence in the broader U.S. culture (i.e., the reflective model) nor to a psychological reactance by adolescents (i.e., the reactance model). Rather, we find, especially among PG-13 (parents strongly cautioned) movies, that violent content leads to increased ticket sales, suggesting that market demand (i.e., audience preferences) is responsible for continued violent content. We discuss the implications of our findings.

Individuals may go to the movies for a variety of reasons, particularly to be entertained, enlightened, or to escape reality. Some have argued that Americans go to the movies because they are obsessed with violence and violent content (Savage, 2004). Inherent in these arguments is the idea that at the very least, individuals are indifferent to violence and at the most, they are obsessed with it. While people may find anecdotal evidence to support such claims, results of academic studies on violence in movies have been inconclusive and left unclear patterns for future researchers. One element that could be considered is who makes decisions about violent content in different genres of movies and how, if at all, these decisions might influence box office revenue. Some studies have focused on the ratings presented through the Motion Picture Association of America (MPAA; 2014). This governing body determines whether a movie should be rated G (general audience), PG (parental guidance suggested), PG-13 (parents strongly cautioned), R (restricted), or NC-17 (no one under 17 admitted). In addition, the MPAA provides descriptors to moviegoers that explain their rating. For instance, a movie may be rated R for “mature content.” Using the MPAA ratings, researchers have focused on how the MPAA selects its ratings, if these decisions have changed over time, and the impact of MPAA ratings on moviegoers.

These findings have two noteworthy shortcomings. First, many studies fail to take into account a movie’s popularity. Simply receiving a rating from the MPAA does not ensure that the movie will receive wide viewership. We believe the size of viewership should be an important criterion in selecting a sample of movies—which has not been done by most previous research (for an exception, see Leone & Barowski, 2011). Second, rarely have studies theoretically examined if violent content is a reflection of society’s love of violence or if there is some other theoretical reason for its appearance in movies.

In this article, then, we seek to overcome some of these shortcomings by selecting our sample in a more objective manner—using the popularity of a movie (e.g., ticket sales) to determine how the presence of violent content affects viewership. Specifically, we control for several important variables that may also predict popularity, including MPAA ratings, MPAA descriptors, and demographics about the movie (i.e., release month, runtime), and we also examine multiple categories of movie content (i.e., presence of language, sexual content, violent content, substance use, disturbing content, action, other material). While controlling for these variables, we use 2,094 movies to test three competing theoretical ideas: (1) violent content is higher because it reflects general societal trends (more violence in society), (2) violent content is higher because children are drawn to “adult content,” and (3) violent content is a product of the market. In other words, moviegoers want violent content in movies, and movie producers are finding ways to get violence into movies that will reach a wide audience and provide the most profit (such as PG-13 films). Our purpose is to determine which of these possibilities best explains the persistence of violence in popular movies over the past 20 years.

In 1968, the MPAA developed a movie rating system that was voluntarily adopted by the major Hollywood studios of the time. This ratings system would eventually become both the most well-known and the longest running rating system in the United States. Currently, the system rates movies as “G,” “PG,” “PG-13,” “R,” and “NC-17.” The MPAA says the following about the purpose of the rating system:

Movie ratings provide parents with advance information about the content of movies to help them determine what’s appropriate for their children. After all, parents know best their children’s individual sensitivities and sensibilities. Ratings are assigned by a board of parents who consider factors such as violence, sex, language and drug use and then assign a rating they believe the majority of American parents would give a movie. (Motion Picture Association of America, 2014)

Therefore, the rating system’s primary purpose is to help parents decide whether a movie’s content is appropriate for their children.

But who determines these ratings and content? According to the MPAA, an anonymous board of movie raters is appointed to make decisions about ratings and also about the content that merits particular ratings. This board is called CARA (Classification and Rating Administration) and uses a set of procedures in making ratings and determining content. The MPAA has outlined these procedures as follows:

A motion picture submitted for rating is viewed by designated members of the Rating Board, including at least one Senior Rater. After the Raters view the motion picture, each Rater submits a preliminary ballot to the designated Senior Rater, giving the Rater’s view of the appropriate rating of the motion picture. The Raters who viewed the motion picture then discuss the appropriate rating and reach agreement on a rating for the motion picture. After the Raters determine the rating of the motion picture, each Rater prepares a final ballot with his or her rating for the motion picture and details of the content of the motion picture that in his or her view require the rating. The Raters’ ballots are treated at all times as confidential and are not disclosed outside CARA. (Motion Picture Association of America, 2010, p. 5)

To determine specific content for movies above a “G” rating, CARA raters are given the following guidance:

Every motion picture assigned a rating of PG, PG-13, R, or NC-17 by the Rating Board also is assigned a “rating descriptor.” This rating descriptor helps guide parents on the type of content that resulted in the motion picture being assigned that rating; modifiers indicate the type and intensity of specific elements in the movie. The rating descriptor does not constitute an exhaustive list of the type of content in the motion picture but reflects only the type of content in the motion picture that is strong enough to merit the rating category assigned to the motion picture. (Motion Picture Association of America, 2010, p. 5)

For instance, a movie may be rated “R for violence” or “PG-13 for intense sequences of science-fiction violence and peril.” Such descriptors are important because studies have shown that parents prefer content descriptors over age descriptors to help them make the best decisions (Bushman & Cantor, 2003). What is clear then is that the content and rating are a subjective but agreed-upon decision by a board appointed by the MPAA that are often relied upon by parents in making decisions about what to allow their children to view.

Because of the unique nature of this subjective rating process and the various content descriptors associated with each rating, many articles consider aspects of the MPAA rating system. Some articles focus on the rating itself (e.g., PG-13), while others focus on the content descriptors behind each rating (e.g., intense sequences of science-fiction violence and peril). The purpose of most of these articles is to determine how accurate ratings are when empirically examining movie content, the clarity of content descriptors (specifically descriptors that pertain to violence and sexual content), and how ratings change over time. Generally speaking, research has found that parents welcome assistance in determining movie content and would prefer content-based information (e.g., violence) to age-based content (e.g., PG-13). Furthermore, many parents suggest that while they prefer content descriptors to age-based ratings, they find content descriptors provided by the MPAA to be confusing and too general (Thompson & Yokota, 2004).

Several research articles have focused on a specific rating and content. For example, Yokota and Thompson (2000) examined movies rated G. As stated above, according to the MPAA, G-rated movies should have no violent content and should be suitable for all audiences. Given this, Yokota and Thompson considered how much violence was actually in G-rated movies. They found that when examining 74 films over a 60-year time period, violence does exist in G-rated movies, typically in the form of bodily injury.

Other articles have focused on PG-13 movies, noting that sexual content and violence are the two primary criteria for a movie receiving a PG-13 rating or above (Leone & Osborn, 2004). Violence is a particular focus of these studies in part because it may influence aggression or violence in children (Savage, 2004). Sexual content is a focus because of fear of teens watching sex in movies cultivating sexual behavior (Gerbner, 1998; Leone & Osborn, 2004).

One particular subarea of movie rating discussions involves something called “ratings creep.” Put simply, ratings creep is the process by which adult content (such as graphic violence and sexuality) is increasingly allowed (i.e., “creeping”) into nonadult-rated movies (i.e., G, PG, and PG-13). In other words, researchers have argued that content in movies that would have received an R rating in the past are now given a PG-13 rating instead (Leone & Houle, 2006). The PG-13 rating is considered the most desirable for movie makers because it “excludes no one but suggests to children of all ages that the movie pushes the limits of unrestricted content” (Leone & Osborn, 2004, p. 89). Several articles have examined whether ratings creep exists.

Three articles written by Leone and colleagues (Leone & Barowski, 2011; Leone & Houle, 2006; Leone & Osborn, 2004) consider ratings creep. Leone and Osborn (2004) examined MPAA rating descriptions (six types: sex, theme, alcohol/drugs, violence, language, and an “other” category) from 2000 to 2002 for PG-13 movies. They found that there were higher amounts of violence in PG-13 movies, and although the sex category was higher, it showed only minimal ratings creep. Leone and Osborn concluded that the MPAA is more concerned about sexual content than violent content creeping into PG-13 movies. They argued this is in opposition to what many parents would prefer. In fact, a study conducted by Walsh and Gentile (2001) found that when asked about the ratings system, parents were more concerned about violence than sex in PG-13 rated movies.

A second article by Leone and Houle (2006) examined rating descriptions provided by the MPAA for the years 2000 to 2003. They examined the same rating categories (even types: sex, theme, alcohol/drugs, violence, language, and an “other” category) along with examining time modifiers (i.e., frequent) and intensity modifiers (i.e., low, mild) in one combined index. They found that the content areas of violence and sex were greater in PG-13 movies during this time frame. They also found no particular gains in sex or violence in R-rated movies. In other words, PG-13 movies became more violence/sex focused while R-rated movies remained the same. Hence, a ratings creep does exist. The authors concluded that ratings creep exists because PG-13 movies are more likely to make money through ticket sales and may be more popular. It is likely that PG-13 movies have a lot of profit potential. Interestingly, when examining an extra year of data, they noted that creep in 2003 was primarily in the area of sex instead of violence. They concluded that perhaps the MPAA was listening to parental concerns and had decreased the amount of violence creeping from R- to PG-13-rated movies.

A third article by Leone and Barowski (2011) performed a longitudinal analysis of PG-13-rated films, rather than their rating descriptions, in order to determine changes in levels of adult content over time. The authors examined top-earning PG-13 rated films from 1988, 1997, and 2006 for depictions of violence, sex, nudity, substance abuse, and language. The researchers found more violence than other adult content across all of the films studied. The authors found significant increases in all types of violence (moderate, extreme, rough, and rough and persistent) over time, but did not find similar significant increases in the other types of graphic content. These findings indicate an MPAA double standard, suggesting that sexual content is treated as more threatening than violent content.

Thompson and Yokota’s (2004) work also provides guidance on how parents interpret content informed by rating descriptors. Using data from 1992 to 2003, they specifically examined rating descriptors and added into their data set information from another parent-based website regarding movies (Kids In Mind). They concluded that age-based ratings do little in the way of helping parents make decisions about movie watching. Instead, they recommended that parents use the content-based predictors to help make these decisions. Thompson and Yokota also noted that ratings creep exists in data from 1992 to 2003 with R-rated content creeping into PG-13-rated movies over time. Finally, this study found that the content descriptors were not specific and were perhaps not helpful to parents generally speaking. They concluded that other sites like Kids In Mind might provide a better way for parents to gain information about what their children should watch, one that is not based on age but based on accurate descriptions of movie content.

In sum, studies that have examined ratings creep have found that there is, indeed, a rating-creep effect with adult content “creeping” into nonadult-rated movies (e.g., G, PG, and PG-13). This creep seems to appear more for violence than for sexual content and is most concerning in the PG-13 category. In addition, studies have found that parents are unlikely to receive appropriate descriptors to make sense of violence and sexual content in PG-13 movies. Thus, while all of these studies provide useful information about MPAA ratings and the content descriptors from each rating decision (or lack of information), several important questions remained unanswered. For instance, do rating descriptors affect a movie’s popularity? If a PG-13 rating is so desirable, why not lessen the amount or graphicness of the violent content? In other words, why does violent content persist over time? Has ratings creep changed since 2006 (the studies noted above only focus on data up until 2006)? Furthermore, many of these studies provide anecdotal discussions of why PG-13 movies appear to be more violent and sexually based (e.g., viewers want it, it makes movie companies more money, etc.) but without considering movie popularity in addition to ratings and descriptors, it is difficult to understand the intention behind content in movies. In this article, we attempt to answer these questions by considering three competing theoretical models.

Our first theoretical model is based on reflection theory, which, put simply, states that culture reflects the society in which it is produced. This idea has been a part of sociological literature since 1954 (e.g., Albrecht, 1954) and has been systematically studied since as early as 1957 (e.g., Horton, 1957). Since its inception, it has been applied to music (e.g., Carey, 1969; Horton, 1957; Lomax, 1968), literature (Albrecht, 1954; Griswold, 1981), art (Kavolis, 1968, 1972; Wolfe, 1963), and movies (Boyd, 2004). While some researchers question that culture reflects the society in which it is produced, most studies have supported it in some capacity (Griswold, 1981)

For instance, Horton (1957) found that popular song lyrics reflected the different cycles of courtship faced by those growing up in the 1950s. Carey (1969) followed up Horton’s study by showing how the major cultural changes that took place in the 1960s also changed the way courtship was portrayed in popular American song lyrics. Even more important to the present study, Boyd (2004, p. 67) found that “popular dance films reflect cultural ideas concerning fundamental issues in life and art.”

In sum, our reflective model indicates that the content of media is simply a reflection of larger society. Therefore, violence in movies should mirror changes in actual violence in U.S. society.

Our second theoretical model is based on Brehm’s (1966) psychological theory of reactance. This theory states that when messages are perceived to limit or threaten personal freedom, an individual will be motivated to reestablish the threatened or lost freedom. This explanation has been used to explain teenage drinking (e.g., Engs & Hanson, 1989; Quick & Bates, 2010), smoking (e.g., Grandpre, Alvaro, Burgoon, Miller, & Hall, 2003; Miller, Burgoon, Grandpre, & Alvaro, 2006), and consumption patterns (e.g., Rummel, Howard, Swinton, & Seymour, 2000). Recently, Leone and Barowski (2011) applied reactance to children’s desire to view adult content in movies. The authors argue that most children are told adult content is off-limits, thereby motivating said children to seek out this forbidden content. This creates increased demand for PG-13 movies among children, because a PG-13 rating carries with it the

implication of more adult content than other unrestricted ratings like G or PG . . . Thus, from a child’s perspective the content assumes a “forbidden fruit” quality, and by engaging in the threatened behavior the child can restore their perceived loss of freedom. (Leone & Barowski, 2011, p. 56)

In sum, our reactance model indicates that children will be drawn to adult content, simply because they have been told it is off-limits. Therefore, the presence of adult content in PG-13 movies should result in greater ticket sales.

Our third theoretical model is based on the work of Cohen and Young (1981), in which they put forth a market model of news media creation. This model argues that news worthiness is determined by what is of interest to the general public. In other words, news production responds to market demand. For example, if the public likes hearing about violence, then news stories about violent crime will dominate the news. Here, we adapt this model to a different medium—movies. Doing so is a logical next step in this literature, because, as Killingbeck (2001) pointed out, “The images and messages put forth by the news media are remarkably similar to those of the entertainment media” (p. 190).

In sum, our market model suggests that the media provide content according to consumer demand. Therefore, if a form of content is popular among moviegoers (e.g., violence), we would expect that form of content to be increasingly common in movies.

Based on our review of the literature, we have developed the following hypotheses:

  1. Hypothesis 1 (H1): If the reflective model is valid, we expect violent crime rates and the number of violent movies released to follow similar trajectories.
  2. Hypothesis 2 (H2): If the reactance model is valid, we expect all (or most) forms of adult content to be associated with greater ticket sales. In other words, we expect movies containing violence, sexual content, and adult language to increase ticket sales. This should be especially true for PG-13 movies.
  3. Hypothesis 3 (H3): If the market model is valid, we expect the effect of the presence of violent content on a movie’s popularity to mirror the change in the number of violent movies produced. In other words, if the number of violent movies increases, then violent content should make a movie sell more tickets.

We analyze 2,094 popular movies between the years 1992 and 2012. Our sample was created by gathering data on the 100 most successful movies (based on estimated ticket sales) from each year between 1992 and 2012. This resulted in data on 2,100 movies. However, five movies were removed from the data set because critics’ ratings (see description below) were unavailable, and one was removed because of its NC-17 rating (discussed below).

Data for our dependent variable, estimated ticket sales, were obtained from BoxOfficeMojo.com. BoxOfficeMojo is an online movie publication and box office reporting service that has become the most popular source of box office data worldwide, averaging over two million unique visitors per month (BoxOfficeMojo, 2014). Their website provides estimated ticket sales for all movies released since 1980 and has been used by other researchers examining movie popularity (e.g., Fu & Govindaraju, 2010; Leone & Barowski, 2011). Estimated ticket sales are calculated by dividing the gross amount of money a movie earns by the average ticket price for the year of release. For instance, in 2002, Spider-Man earned US$403,706,375. In the same year, the average cost of a movie ticket was US$5.81. This results in estimated ticket sales of 69,484,700 (403,706,375 / 5.81).

Independent Variables

Our first explanatory variable is the movie’s rating. As described above, movie ratings include G, PG, PG-13, R, and NC-17. Each rating was numbered consecutively starting with G = 1 and ending with NC-17 = 5. However, it should be noted that only one movie in our sample was rated NC-17 and was removed from the sample.1 Therefore, our movie rating variable varies from 1 to 4, and higher scores should be interpreted as the movie having more adult-oriented content.

Our second set of explanatory variables involves the rating descriptors previously mentioned. We include seven descriptor variables, each of which was coded as 1 if that particular content descriptor was included with the rating. For instance, if a language descriptor was assigned to a movie (i.e., “mild language”), we coded Presence of Language as 1. The seven descriptor variables are as follows: Presence of Language, Presence of Sexual Content, Presence of Violent Content, Presence of Substance Use, Presence of Disturbing Content, Presence of Action, and Presence of Other Material.

Each movie was coded based on all of the rating descriptors it received. Therefore, it is possible for a movie to be coded 0 on all seven variables or 1 on all seven variables. Rating descriptors were selected by reading through the list of terms assigned to each movie by the MPAA and determining if the word or phrase represented each category (i.e., a presence of violence, sexual content, disturbing content, substance abuse, language, and/or action). Any words or phrases that did not fit one of these six categories were put in the “other” content category.

Presence of Violence includes not only descriptors that specifically mention violence (i.e., “battle violence”) but also descriptors that are suggestive of violence (i.e., “gore” or “gunplay”). Presence of Sexual Content includes all descriptors mentioning sex, sexuality/sensuality, and nudity. Presence of Disturbing Content includes all descriptors mentioning horror, peril, and disturbing/intense scenes or images. Presence of Substance Abuse includes all descriptors referring to alcohol, drug, or tobacco use. All Presence of Language descriptors specifically mention language (i.e., “strong language”). All Presence of Action descriptors specifically mention action (i.e., “mild adventure action”). Any descriptor not meeting one of the previously mentioned six criteria was placed in the Presence of Other Material category. The most common descriptors in this category referenced “mature content” or “thematic” elements/material.

Control Variables

In addition to rating descriptors, we control for a number of other variables that might affect a movie’s popularity. For instance, we control for the season in which the movie was released. Four dummy variables were created: summer release (May, June, July, or August = 1), spring release (March or April = 1), fall release (September or October = 1), and winter release (November, December, January, or February = 1). Because movies released during summer months tend to sell the most tickets, summer release is our reference (omitted) group. Other control variables include a movie’s length in minutes (movie runtime) and critics’ ratings obtained from RottenTomatoes.com.2

In order to test the reflective model described above, we also include data on the U.S. homicide rates between 1991 and 2013.3 These data were obtained from the FBI’s Uniform Crime Reports (Uniform Crime Reports, 2010, 2013). Each yearly homicide rate is actually a 3-year average. For instance, the rate for 1992 is the average of the homicide rate for the years 1991, 1992, and 1993. We do this in order to avoid year-to-year fluctuations.

We analyze these data using ordinary least squares (OLS) regression. We test for multicollinearity by obtaining variance inflation factors (VIFs) for all of our models. No VIFs at or above the generally accepted level of 2.5 were observed.

The descriptive findings are presented in Table 1. The results show that, of the movies in our sample, nearly two thirds were assigned a language descriptor (66%) and just over half received sexual or violent content descriptors (54% and 51%, respectively). In addition, roughly three quarters of the movies were rated either PG-13 (40%) or R (36%).

 

Table

Table 1. Descriptive Statistics of Dependent and Independent Variables for 2,094 Popular Movies, 1992-2012.

 

Table 1. Descriptive Statistics of Dependent and Independent Variables for 2,094 Popular Movies, 1992-2012.

Note. G = general audience; PG = parental guidance suggested; PG-13 = parents strongly cautioned; R = restricted; NC-17 = no one under 17 admitted.

In order to test the reflective model (H1), we include Figure 1, which shows changes in U.S. homicide rates and the average number of tickets sold for movies containing violence from 1992 to 2012. If the reflective model was accurate, we would expect homicide and tickets sold to have similar trajectories. As Figure 1 demonstrates, the trajectories have been very different. Homicide has been steadily declining for the past 20 years, while violent ticket sales have been generally increasing over the same time period. Based on these patterns, we do not believe the reflective model is the best explanation for the relationship between movie ratings and ticket sales.

figure

Figure 1. Trends in homicide and tickets sold for movies containing violence.

Next, we turn our attention to the reactance model (H2). Table 2 shows the results of OLS regression models predicting the estimated number of tickets sold with various content control variables (presence of sexual content, language, violent content, etc.). We present results for both aggregated and disaggregated (by movie rating) samples. Within our four models, language has a negative effect on a movie’s popularity in one model (the full sample), sexual content has no effect in any of the models, and violent content has a positive effect in two out of four models (the full sample and the PG-13 sample).

 

Table

Table 2. Ordinary Least Squares Estimates Predicting Estimated Number of Tickets Sold for 2,094 Popular Movies, 1992-2012.

 

Table 2. Ordinary Least Squares Estimates Predicting Estimated Number of Tickets Sold for 2,094 Popular Movies, 1992-2012.

Note. Standardized coefficients (Beta) are shown in parentheses. PG = parental guidance suggested; PG-13 = parents strongly cautioned; R = restricted.

aAll coefficients were divided by 100,000 to make the chart easier to read.

*p ≤ .05. **p ≤ .01. ***p ≤ .001.

Given these results, we believe that the reactance model (H2) receives very little support. The presence of violence and disturbing content were associated with greater ticket sales for PG-13 movies, but there was no effect for the presence of language or sexual content. The latter is troubling, because previous research has shown that parents are most concerned about their children viewing sexual content (e.g., Breznican & Avitia, 2007; Rideout, 2007). In other words, if parents tell their children they cannot view sexual content, but are not as concerned about violence, then reactance should make children more likely to seek out sexual—not violent—content.

This leaves us with only the market model (H3) remaining. As previously noted, the results in Table 2 show that violent content is associated with increased ticket sales, but not language or sexual content. This is consistent with previous research (e.g., Bushman & Cantor, 2003) demonstrating that violent content is more attractive to consumers than sexual content. Rather than rely only on previous research, we include Figure 2 to help clarify the applicability of the market model. The figure shows the number of movies with violent content by movie rating for the years 1992 to 2012. Three findings stand out: (1) the number of violent PG movies remained relatively unchanged during the time period, which supports the market model (violence in PG movies has no effect; therefore, it is kept constant or decreased); (2) the number of violent PG-13 movies increased during the time period, which supports the market model (violence in PG-13 movies increases ticket sales; therefore, it is increased); and (3) the number of violent R movies decreased during the time period, which follows the market model (violence in R movies has no effect; therefore, it is kept constant or decreased).

figure

Figure 2. Violent content by rating.

Note. PG = parental guidance suggested; PG-13 = parents strongly cautioned; R = restricted.

These findings seem to suggest that the market model best explains the persistence of violent content in movies (i.e., violent content is a product of consumer demand). However, how such violent content is distributed across ratings has been a point of discussion by other researchers who discuss “ratings creep”—more and more R-rated content is being allowed in PG-13 movies. Or, put another way, movies historically considered “R” are now receiving a PG-13 rating. If this is true, it is possible that the relationship between ticket sales and violence in PG-13 movies is due to this shift (i.e., violence in PG-13 movies was not popular until “ratings creep” began). Figure 3examines this possibility.

figure

Figure 3. Number of movies by rating.

Note. G = general audience; PG = parental guidance suggested; PG-13 = parents strongly cautioned; R = restricted; NC-17 = no one under 17 admitted.

The figure shows that there was a substantial shift in the number of PG-13 and R movies between 1999 and 2000 (a dramatic increase in PG-13 rated movies and a dramatic decrease in R-rated movies). This shift appears to make the results from Table 2 seem suspect—perhaps the models are being dominated by the 2000-2012 period. However, Table 3 shows that this is not the case. Model 1 replicates Model 3 from Table 2—violent content increases ticket sales in PG-13 movies. Model 2 shows the period before the substantial changes in PG-13/R ratings (1992-1999) and Model 3 shows the period following this change (2000-2012). In both models, the presence of violent content is associated with increased ticket sales for PG-13 movies. Therefore, we feel even more confident in our support of the market model, noting that the relationship between ticket sales and violent content was not created artificially by a change in the way movies were rated. We discuss the implications of our findings below.

 

Table

Table 3. Ordinary Least Squares Estimates Predicting Estimated Number of Tickets Sold for PG-13 Movies, 1992-2012.

 

Table 3. Ordinary Least Squares Estimates Predicting Estimated Number of Tickets Sold for PG-13 Movies, 1992-2012.

Note. Standardized coefficients (Beta) are shown in parentheses. PG-13 = parents strongly cautioned.

aAll coefficients were divided by 100,000 to make the chart easier to read.

*p ≤ .05. **p ≤ .01. ***p ≤ .001.

The purpose of this article was to examine the persistence of violence in movies over time using ticket sales. While it was clear that violent content does appear in movies and some have argued we see more violence today than ever before, rarely have researchers tested the driving forces behind violent content in movies. We tested three hypotheses. First, we wanted to see whether violence in real life mirrored violence in the movies (i.e., reflective model). After our analyses, however, we determined that this was not the case. It was not violence in society driving violent movie content. Next, we examined whether the popularity of violent movies was due to children reacting against prohibitions against their viewing adult content (i.e., reactance model). Our results indicate that only violence, not other types of adult content, results in increased ticket sales for movies (PG-13 movies in particular). Finally, we concluded the best explanation for the persistence of violent content in movies involves market/consumer demand (market model). We draw several implications from these findings.

First, if one were to adopt a cynical approach and assume that the MPAA is working in collusion with NATO (National Association of Theatre Owners) and movie makers, the following conclusion could be drawn: Movie makers and/or NATO realized that violent content was popular among movie going audiences and pressured the MPAA to change how it was rating movies. Because violent content usually received an R rating, the number of potential ticket buyers was being limited (to only those 17 and older). Because the MPAA estimates that 25% of moviegoers are under the age of 17, this was a substantial limitation. In order to pacify NATO and movie makers, the MPAA agreed to allow more violent content in PG-13 movies (it should be noted that the MPAA only exists because it is supported by NATO and movie makers).

There is, of course, a less cynical explanation. It is possible that movie makers realized that violent content was popular among movie going audiences and changed the violence they portrayed to fit with the MPAA’s PG-13 restrictions, for instance, removing blood from violent scenes. This practice can be seen in the recent blockbuster film The Hunger Games. The movie is about a group of children forced to kill each other in a gladiatorial style setting. Most people would probably assume that such a movie would receive an R rating, yet The Hunger Games received a PG-13 rating. The key is the quick and bloodless violence that occurs in the movie. Removing the blood and not lingering on the violence that occurs allowed the movie to escape an R rating (Sullivan, 2013).

Second, our findings have serious implications for parents. While many parents rely on MPAA ratings to help make decisions about what their children should see and watch (especially pre-teenage children), many allow their children to decide for themselves. Our results make it clear that movies containing violence are the most popular choice. This may be particularly troubling to many parents when combined with research findings demonstrating that the graphic nature of violent content in PG-13 movies is increasing over time (Leone & Barowski, 2011). While we do not examine graphicness here, it is an important step for future research. It may be that the violent content, along with graphic content, is changing in PG-13 movies. In other words, PG-13 ratings mean something different than they did when most of today’s parents were young children. Perhaps this knowledge needs to be more openly spread to parents and researchers alike so that they can make more informed decisions about movies for children and teenagers. This is also an important avenue of future research, especially projects that combine Kids In Mind and other kid-friendly websites into analyses of movies’ popularity.

Given the two points mentioned above, concerning the MPAA and parents, a logical policy implication is that the MPAA should be more open in their discussions of ratings systems and allow parents greater access to their decision-making process. The MPAA could create manuals and procedures that were accessible to parents that inform and educate parents on the meaning behind ratings. Furthermore, the MPAA could be clearer in their content descriptors and limit the amount of descriptors used so that parents would have more, rather than less, information. Finally, there are currently only restrictive “suggestions” for children under the age of 13 to not attend PG-13 movies. One potential policy suggestion is theaters (perhaps in conjunction with the MPAA) should actually enforce the PG-13 rating (this suggestion is discussed also by Leone & Barowski, 2011).

Finally, our study shows the importance of using multiple years of data and using more objective measures of movie popularity. Other studies have analyzed only 2 or 3 years of data, and their findings have had different results than those presented here. We would suggest that looking at the popularity of movies over time is important for researchers to make good estimates of what is happening with violent content and how it affects moviegoers and society. Similarly, we would argue that using a more objective measure of moviegoers such as box office sales is important. Doing so, in combination with an examination of MPAA ratings (subjectively made) and descriptors (subjectively made), may present the best approach to understanding this topic. Future researchers could also consider some things we did not consider here such as what proportion of moviegoers read rating descriptors before viewing a movie and whether the amount of violence in a movie is related to ticket sales.

Limitations

While the findings of this study provide a variety of new avenues for research for those interested in violent media content in movies, this study has a few limitations. First, our use of box office sales has many strengths, but we are not able to account for those who rent/stream movies after release in terms of movie-watching patterns. It is possible that those who rent movies have different preferences for movie ratings and content. Second, our selection of rating descriptors into categories (violence, substance abuse, etc.) was subjective. We have tried to minimize bias by creating an “other” category and making sure that descriptors that did not clearly fit into one of our predetermined categories were not put into a category haphazardly.

Finally, our data do not allow us to determine whether people looked at the descriptors before going to the movie. We also do not know whether they actually chose the movie because it contained violence. Nevertheless, we do know that PG-13 movies containing violence sell more tickets. It is possible that none of our findings have anything to do with violence, only with the movie’s rating. In other words, it is PG-13 movies that are popular, not violence. PG-13 movies were already very popular, but once “ratings creep” took place and more violence was allowed in those movies, it made it appear as though violence was the driving force. However, we believe this explanation is unlikely for two reasons: (1) Our findings demonstrated that violence was linked to increased ticket sales before the substantial PG-13/R shift around 2000 and (2) if the popularity of adult content was merely an indirect result of “ratings creep,” language and sexual content should have effects similar to violent content. But, as our results show, this was not the case.

With these limitations in mind, we believe that this article substantially adds to the literature on violence in the media by showing that it is not the reality of crime (reflective model) or children reacting to prohibitions against adult content (reactance model) driving violent content. Instead, it is the market model driving violent content in movies, especially within the PG-13 category. Movie makers have determined that Americans are interested in watching violence and have found innovative ways to get violent content into one of the highest paying ratings—PG-13. As long as research continues to demonstrate that there are negative consequences to consuming violent media, this is a topic that should be further examined by researchers and parents alike.

Declaration of Conflicting Interests The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The authors received no financial support for the research, authorship, and/or publication of this article.

1. All models were run with and without this movie, but the results were not substantively different.

2. This score ranges from 1 to 100 and is based on the percentage of professional critic reviews that are positive for a given film. We originally included an audience score as well but, surprisingly, the two variables were highly correlated. Therefore, we were only able to include one of the variables, and we chose to go with the critics reviews. All models were run using each of the two variables, but the results were not substantively different.

3. We also obtained data on overall violence, robbery, rape, and aggravated assault, but all followed a similar pattern to homicide. Because homicide data are the most reliable and are more likely to be reported, less likely to be ignored by the police and less subject to definitional variation (LaFree, O’Brien, & Baumer, 2006; Schwartz, 2006; Steffenesmeier & Haynie, 2000), we include only these.

Albrecht M. (1954). The relationship of literature and society. American Journal of Sociology, 59, 425-436. Google Scholar CrossRef
BoxOfficeMojo. (2014). Yearly box office. Retrieved from http://www.boxofficemojo.com/yearly/chart/?yr=2014Google Scholar
Boyd J. (2004). Dance, culture, and popular film: Considering representations in Save the Last Dance. Feminist Media Studies, 4, 67-83. Google Scholar CrossRef
Brehm J. W. (1966). A theory of psychological reactance. Oxford, UK: Academic Press. Google Scholar
Breznican A., Avitia L. E. (2007, April 10). R rating under review: Hollywood tries to set its standards straight. USA Today, p. 1D. Google Scholar
Bushman B. J., Cantor J. (2003). Media ratings for violence and sex: Implications for policymakers and parents. American Psychologist, 58, 130-141. Google Scholar CrossRef, Medline
Carey J. T. (1969). Changing courtship patterns in the popular song. American Journal of Sociology, 69, 720-731. Google Scholar CrossRef
Cohen S., Young J. (1981). The manufacture of news: Social problems, deviance and the mass media. Thousand Oaks, CA: Sage. Google Scholar
Engs R., Hanson D. J. (1989). Reactance theory: A test with collegiate drinking. Psychological Reports, 64, 1083-1086. Google Scholar Link
Fu W. W., Govindaraju B. (2010). Explaining global box-office tastes in Hollywood films: Homogenization of national audiences’ movie selections. Communication Research, 37, 215-238. Google Scholar Link
Gerbner G. (1998). Cultivation analysis: An overview. Mass Communication and Society, 1, 175-194. Google Scholar CrossRef
Grandpre J. R., Alvaro E., Burgoon M., Miller C., Hall J. R. (2003). Adolescent reactance and anti-smoking campaigns: A theoretical approach. Health Communication, 14, 349-366. Google Scholar CrossRef
Griswold W. (1981). American character and the American novel: An expansion of reflection theory in the sociology of literature. American Journal of Sociology, 86, 740-765. Google Scholar CrossRef
Horton D. (1957). The dialogue of courtship in popular songs. American Journal of Sociology, 62, 569-578. Google Scholar CrossRef
Kavolis V. (1968). Artistic expression: A sociological analysis. Ithaca, NY: Cornell University Press. Google Scholar
Kavolis V. (1972). History on art’s side: Social dynamics in artistic efflorescences. Ithaca, NY: Cornell University Press. Google Scholar
Killingbeck D. (2001). The role of television in the construction of school violence as a “moral panic.” Journal of Criminal Justice and Popular Culture, 8, 186-202. Google Scholar
LaFree G., O’Brien R. O., Baumer E. (Eds.). (2006). Is the gap between Black and White arrest rates narrowing? New York: New York Press. Google Scholar
Leone R., Barowski L. (2011). MPAA ratings creep. Journal of Children and Media, 5, 53-68. Google ScholarCrossRef
Leone R., Houle N. (2006). 21st century ratings creep: PG-13 and R. Communication Research Reports, 23, 53-61. Google Scholar CrossRef
Leone R., Osborn L. (2004). Hollywood’s triumph and parents’ loss: An examination of the PG-13 rating. Popular Communication, 2, 85-101. Google Scholar CrossRef
Lomax G. (1968). Folk song style and culture. Piscataway, NJ: Transaction Publishers. Google Scholar
Miller C., Burgoon M., Grandpre J. R., Alvaro E. (2006). Identifying principal risk factors for the initiation of adolescent smoking behaviors: The significance of psychological reactance. Health Communication, 19, 241-252. Google Scholar CrossRef, Medline
Motion Picture Association of America. (2010). Classification and rating rules. Sherman Oaks, CA: National Association of Theatre Owners. Google Scholar
Motion Picture Association of America (MPAA). (2014). Understanding the film ratings. Retrieved from http://www.mpaa.org/film-ratings/ Google Scholar
Quick B. L., Bates B. R. (2010). The use of gain or loss frame messages and efficacy appeals to dissuade excessive alcohol consumption among college students: A test of psychological reactance theory. Journal of Health Communication, 15, 603-628. Google Scholar CrossRef, Medline
Rideout V. (2007). Parents, children, and media: A Kaiser Family Foundation Survey. Menlo Park, CA: The Henry J. Kaiser Family Foundation. Google Scholar
Rummel A., Howard J., Swinton J. M., Seymour D. B. (2000). You can’t have that! A study of reactance effects and children’s consumer behavior. Journal of Marketing Theory and Practice, 8, 38-45. Google Scholar CrossRef
Savage J. (2004). Does viewing violent media really cause criminal violence? A methodological review. Aggression and Violent Behavior, 10, 99-128. Google Scholar CrossRef
Schwartz J. (2006). Family structure as a source of female and male homicide in the United States. Homicide Studies, 10, 253-278. Google Scholar Link
Steffenesmeier D. J., Haynie D. (2000). Gender, structural disadvantage, and urban crime: Do macrosocial variables also explain female offending rates? Criminology, 38, 403-438. Google Scholar CrossRef
Sullivan K. (2013, March 15). How did “catching fire” avoid an R rating? MTV News. Retrieved from http://www.mtv.com/news/1717894/catching-fire-hunger-games-r-rating-guidelines/ Google Scholar
Thompson K. M., Yokota F. (2004). Violence, sex, and profanity in films: Correlation of movie ratings with content. Medscape General Medicine, 6, 3. Google Scholar Medline
Uniform Crime Reports. (2010). Table 1: Crime in the United States by volume and rate per 100,000 inhabitants, 1991-2010. Retrieved from http://www.fbi.gov/about-us/cjis/ucr/crime-in-the-u.s/2010/crime-in-the-u.s.-2010/tables/10tbl01.xls Google Scholar
Uniform Crime Reports. (2013). Table 1: Crime in the United States by volume and rate per 100,000 inhabitants, 1994-2013. Retrieved from http://www.fbi.gov/about-us/cjis/ucr/crime-in-the-u.s/2013/crime-in-the-u.s.-2013/tables/1tabledatadecoverviewpdf/table_1_crime_in_the_united_states_by_volume_and_rate_per_100000_inhabitants_1994-2013.xls Google Scholar
Walsh D. A., Gentile D. A. (2001). A validity test of movie, television, and videogame ratings. Pediatrics, 107, 1302-1308. Google Scholar CrossRef, Medline
Wolfe A. W. (1963). Social structural bases of art. Current Anthropology, 10, 3-44. Google Scholar CrossRef
Yokota F., Thompson K. M. (2000). Violence in G-rated animated films. Journal of the American Medical Association, 283, 2716-2721. Google Scholar CrossRef, Medline

 

链接:http://journals.sagepub.com/doi/full/10.1177/0093650215614363

作者:Raymond E. Barranco, Nicole E. Rader, Anna Smith

来源:communication research


Comments are closed.



无觅相关文章插件