Cinematic Success Criteria and Their Predictors: The Art and Business of the Film Industry

Transcription

1 Cinematic Success Criteria and Their Predictors: The Art and Business of the Film Industry Dean Keith Simonton University of California, Davis ABSTRACT The author reviewed the empirical research on the factors underlying the success of feature-length narrative films. After specifying some methodological caveats, the review examined the three main criteria by which a film s success can be evaluated: critical evaluations (both early and post-theatrical run), financial performance (including first weekend and gross), and movie awards (including dramatic, visual, technical, and music categories). To what extent do these criteria represent distinct aesthetic and economic assessments? The review then turned to the various predictors of these success criteria. How is success connected with the film s production and distribution characteristics? To what extent do the predictors converge and diverge across alternative criteria? The article then closes with a discussion of some psychological issues raised by the reviewed findings Wiley Periodicals, Inc. The mainstream film industry faces a tough challenge every single year: The vast majority of the films released each year fail to make money during their theatrical run (De Vany & Walls, 1999; Vogel, 1998). In fact, most films lose millions of dollars. The only reason why the core industry survives at all is because a small percentage of films bring in substantial box office receipts, and an even smaller percentage of these attain the status of outright blockbusters with earnings in hundreds Psychology & Marketing, Vol. 26(5): (May 2009) Published online in Wiley InterScience ( Wiley Periodicals, Inc. DOI: /mar

2 of millions of dollars. 1 For example, the 2002 Sony film Spider-Man cost about $175 million to make but earned about $403 million in domestic gross. As the overall result, the industry is profitable on the average, but only on the average (Walls, 2005c). That is, the mean total box office exceeds the mean production costs (Prag & Casavant, 1994; Simonton, 2005a). At the same time, this highrisk system cannot prevent some pretty spectacular losses. For instance, the 2002 Disney film Treasure Planet cost $140 million to make but grossed only around $38 million. To be sure, additional income can be made in the postrelease period. Hence, some financial flops can recover some earnings when their DVD versions finally appear in video stores sometimes advertised as a special director s cut and many films even go directly to DVD to avoid the stigma of a disastrous theatrical performance. Even bad films can garner additional revenue in television showings in non-primetime program slots. Given the precarious nature of profit in the film industry, cinematic success becomes highly contingent on making the most optimal marketing decisions. What kind of film product is most likely to bring high returns? For example, will it enhance a film s prospects if it can get a star with top marquee value to sign on for the lead role? How should the completed product be best distributed? For instance, should the film appear in thousands of theaters on the first weekend, or should it debut in a few select theaters and then do the art-house movie circuit, building an audience slowly? And how should the film be promoted for best effect? Should the trailers, posters, and advertisements quote some well-known critics, emphasize that its director is an Oscar nominee, or stress that the screenplay is based on a bestselling novel? Given the high rate of financial failures, it would seem that the answers to these and related questions must be elusive if not totally unknown. The screenwriter William Goldman s famous remark that Nobody knows anything is frequently quoted as illustrating the degree of uncertainty in the business (e.g., Walls, 2005c, p. 177). This statement does not mean that there has not been a considerable amount of research devoted to the question of cinematic success. It s just that this research has not provided answers sufficiently consistent and precise to make the marketing of motion pictures a sure thing. To document this general conclusion, this article provides an overview of the relevant research findings. But before doing so, the reader should be first made aware of some technical issues. METHODOLOGICAL CAVEATS Researchers who publish on this topic hail from many distinct disciplines, including economics, marketing, advertising, communication, journalism, broadcasting, management, sociology, culture studies, statistics, mathematics, and even psychology. As a consequence, investigators often have very different substantive goals, apply distinct methods, and interpret their results using contrasting theoretical frameworks. Especially problematic are the methodological differences among the numerous inquiries. These differences can be allotted to the following four categories. 1 Of course, the governments in some countries provide substantial subsidies to filmmakers in order to maintain a domestic industry. Under such circumstances, the films do not have to be as profitable for the industry to survive. Yet those exceptions prove the rule. CINEMATIC SUCCESS 401

3 1. Sampling. Studies will often use rather disparate samples. In the first place, the sampled films will often span non-equivalent and even non-overlapping periods of film history. Because of historical trends in the structure of the film industry, studies can obtain contradictory results even when the investigations are conducted by the same researcher using the same methods (e.g., Litman, 1983; Litman & Ahn, 1998; Litman & Kohl, 1989; see also Smith & Smith, 1986). Furthermore, the samples will differ on whether they include non English-language films, documentaries, and animations cinematic products that may have dissimilar success determinants (Simonton, 2007b). After all, non English-language films often come from nations with government-subsidized film production; documentaries seldom possess a standard narrative with plots or characters; and animated films redefine the very nature of direction, acting, cinematography, and film editing. Moreover, the sampling criteria for the inclusion of a given film can vary immensely from study to study. Some possibilities include films that: (a) were drawn randomly from of all films released during a given period (e.g., Basuroy, Chatterjee, & Ravid, 2003), (b) earned a specified minimum amount in the box office (e.g., Chang & Ki, 2005), (c) were nominated for a prestigious award (e.g., Simonton, 2002), (d) featured at least one identifiable movie star (e.g., Wallace, Seigerman, & Holbrook, 1993), and (e) received critical acclaim or at least a threshold number of critical reviews (e.g., Eliashberg & Shugan 1997). Because the sampling criteria will have an impact on the range exhibited by certain variables, these decisions will affect the empirical outcomes. To illustrate, any investigation that confines its sample to just box office successes will, ironically, find it more difficult to identify the predictors of financial success because of the truncated variance in the dependent variable (e.g., Simonet, 1980). 2. Variables. The various published inquiries do not agree on the variables included in their investigations. For example, researchers vary in whether or not they include production budget, advertising expenditures, film distribution, and market competition. The omission of a particular variable obviously means that the predictive utility of that variable cannot be estimated. But even more critical is the fact that the effectiveness of individual predictors is almost always estimated in a multivariate framework. That is, the predictive power of a given variable is calculated after controlling for other variables in the equation. Hence, the omission of a crucial variable may permit the intrusion of spurious relationships. As an example, Prag and Casavant (1994) found that when marketing expenditures were taken into account, the impact of certain variables on box office earnings vanished. 3. Measurement. Even when several studies include the same variable in the prediction equation, that variable may be defined in totally incongruent ways. As a case in point, investigators have not reached consensus on how to determine whether or not a film features a movie star. Stars have been defined according to (a) whether they previously won a best-acting award (e.g., Delmestri, Montanari, & Usai, 2005), (b) their standing in surveys of frequent moviegoers (e.g., Canterbery & Marvasti, 2001), (c) the financial performance of their most recent films (e.g., Litman, 1983), (d) the total number of prior films in which they have performed (e.g., Chang & Ki, 2005), (e) subjective identification based on the researchers knowledge of films and movie stars (Prag & Casavant, 1994, p. 220), and (f) industry-based determinations of the most powerful, bankable, marquee value, or A and A list performers (e.g., Kindem, 1982). Yet there is absolutely no reason for believing that these rival indicators would correlate in the same manner with any criterion of cinematic success. Nor would 402 SIMONTON

4 they necessarily correlate highly with each other. Rendering measurement all the more complicated, no agreement has emerged on how to handle the presence of two or more stars who vary in status (e.g., A versus A list) or who occupy distinct roles (e.g., lead vs. supporting, hero vs. villain). In a word, measurement lacks consensus. 4. Analyses. Investigators do not always concur on the most suitable data analyses. Although most researchers rely heavily on ordinary least squares regression, some have argued that this may not be the most appropriate estimation procedure (e.g., Walls, 2005b). Consequently, some studies will introduce less commonplace procedures, such as Poisson regression (e.g., Cattani & Ferriani, 2008). These analytical choices are particularly significant when dealing with financial performance criteria, which tend to be highly skewed (Walls, 2005a). Not only will the variance substantially exceed the mean, but the variance may not even be finite (De Vany & Walls, 1999). Under these conditions, different estimation methods will yield contrary effect sizes for various predictors, and at times a predictor that is statistically significant according to one procedure might not be so by another (Walls, 2005b). Because of these complications, the review concentrates on what seem to represent the most robust empirical results across the divergent investigations. Additionally, the review will focus on archival research carried out on large samples of feature-length, English-language, live-action, narrative films the dominant cinematic products in the industry. By archival research I mean studies that apply quantitative measurement and statistical methods to data readily available in published reference works or film compendia in such electronic media as CD-ROMs and the Internet (e.g., the Internet Movie Database). Toward the end of the article, I will touch upon alternative methods that can be used to round out our understanding of cinematic success. SUCCESS CRITERIA The three main criteria of cinematic success will be defined first, and then the intercorrelations among these criteria will be presented. The Success Triad With few exceptions, almost every empirical inquiry in this area uses one or more of the following three success criteria: critical evaluations, financial performance, and movie awards. More properly speaking, these three are not single criteria but rather three sets of criteria. This criterion complexity will become evident as each is discussed in turn. Critical evaluations. Critics make a living passing judgments on the films released each year. Frequently these ratings take the form of stars, perhaps ranging from zero stars for a turkey or bomb to five stars for an unquestioned masterpiece. Such assignments can be easily converted into numerical scales (Simonton, 2002; Zickar & Slaughter, 1999). Significantly, these critical evaluations are concentrated at two distinct periods in the life of a cinematic product. The first batch consists of critical assessments published at the beginning of a film s theatrical run. Indeed, for many films the judgments will be based on a preview immediately prior to the official release. These informed evaluations CINEMATIC SUCCESS 403

5 will appear in various media, such as newspapers, radio, television, and the Internet. Needless to say, these critics have taken it upon themselves to offer advice to prospective moviegoers on how to best spend their entertainment dollars in the upcoming weekend. The second batch of reviews does not come out until much later, namely, the evaluations that appear in popular movie, video, or DVD guides (e.g., Maltin, 2004; Walker, 2004). At the time of their publication, the rated films have almost invariably completed their theatrical run, and have recently been made available for purchase or rental for the purpose of home viewing. It should be stressed that this second wave of critical evaluations is not absolutely equivalent to the first wave. Part of the contrast can be ascribed to the influence of intervening events such as financial performance and movie awards as well as changes in the cinematic product (e.g., a director s cut or second thoughts about a box-office sleeper ). Four psychometric features of these ratings deserve special note. First, composite measures based on multiple critical ratings tend to display a roughly normal distribution (Simonton, 2004c). Bombs and masterpieces are relatively rare in comparison to the many motion pictures in the middle of the pack. Second, the critics in either wave of assessments demonstrate a strong consensus (Boor, 1990; Holbrook, 1999; Zickar & Slaughter, 1999). For instance, one 5-item composite based on movie-guide evaluations of 877 films had an internalconsistency reliability of 0.91 (Simonton, 2007c). Third, the critics judgments in the first wave correlate very highly with those in the second wave (Simonton, 2004c, 2005b). For example, one study found a correlation of 0.75 between earlier and later evaluations (Simonton, 2005b). And this test retest reliability is likely an underestimate insofar as the sample only included films that had been nominated for a major movie award (cf. Simonton, 2007b, which obtained 0.91 for a sample that included really bad films). Fourth, the opinions of film critics regarding the relative merit of various films correlate positively with the views of consumers (Boor, 1992; Holbrook, 1999; Wanderer, 1970). This correspondence is especially strong for those consumers who place a high value on film as a medium of artistic expression (Boor, 1992). In contrast, the agreement is weaker, but still positive, for those consumers who just see movies as a form of light entertainment (Holbrook, 1999). Hence, critical evaluations appear to be valid as well as reliable. Financial performance. This criterion encompasses a large number of separate criteria. To begin with, it can refer to gross box office earnings or at least receipts returned to the distributors (e.g., Sochay, 1994). Or it can be confined to the earnings in any given week, most commonly the film s first weekend gross (e.g., Basuroy, Chatterjee, & Ravid, 2003; Simonton, 2005b). Alternatively, the criterion might involve not earnings at all, but rather the total length of the theatrical run (Sochay, 1994). Although these alternative indicators correlate fairly highly with each other, they do not necessarily have identical predictors (e.g., Chang & Ki, 2005). For instance, a given film might have a long run by slowly progressing through the art houses or by being shown week after week in hundreds of multiplexes if not megaplexes. Because the United States film industry dominates the world market (Acheson & Maule, 1994), most investigations use domestic rather than world financial figures. Only occasionally will a study examine receipts outside North America (e.g., Hand, 2001) or include world-wide performance statistics (e.g., Litman & Ahn, 1998). This emphasis on the domestic market is important insofar as some 404 SIMONTON

6 film genres, such as comedies, do not transport well across linguistic and culture boundaries (see, e.g., Lee, 2006). An astute reader may have noticed that no mention has been made of a film s actual profit. This would seem to represent the most important criterion from a pure business sense. Yet very few empirical investigations actually examine this criterion (see, e.g., Litman & Ahn, 1998). The reason for this relative neglect is that profit cannot be determined without knowing a film s cost of production, or budget. Because the latter is considered proprietary information, it is seldom made available, and usually must be estimated by industry sources based on production information (cast, crew, shooting location, etc.). It is worth noting that a truly profitable film typically has to earn about double its production cost (Eliashberg, Hui, & Zhang, 2006). Movie awards. In 1928 the Motion Picture Academy of Arts and Sciences began bestowing what eventually became known as the Oscars. Since then the number of film awards has proliferated considerably in terms of both the range of award categories and the number and type of organizations granting them. Thus, Oscars alone are given in such major categories as best picture, directing, writing or screenplay (original and adapted), male and female lead acting, male and female supporting acting, film editing, cinematography, art direction, costume design, makeup, visual effects, sound editing, sound mixing, score, and song. And comparable awards are bestowed by the British Academy of Film and Television Arts, the Hollywood Foreign Press Association (the Golden Globes ), various critic organizations like the New York Film Critics Circle, and more specialized societies, like the Screen Actors Guild. The published research varies greatly on exactly how these honors are translated into empirical measures. Whereas some studies only look at awards, others include nominations. And while some investigations will include almost the full range of categories, others will focus on the main award categories, such as best picture, directing, and acting. Fortunately, a factor analysis of awards and nominations has shown that these various honors cluster into four orthogonal award dimensions: (1) the dramatic cluster of directing, writing, acting, and film editing, (2) the visual cluster of cinematography, art direction, costume design, and makeup, (3) the technical cluster of special visual effects, sound editing, and sound mixing, and (4) the music cluster of score and song (Simonton, 2004c). This result provides ample justification for treating certain awards separately from others, a point that we will amplify shortly. It should be pointed out that composite measures based on nominations and awards received from multiple organizations display an impressive magnitude of reliability. For instance, Simonton (2007b) obtained the following internalconsistency reliability coefficients (number of organizations bestowing that honor given in parentheses): picture 0.85 (8), directing 0.90 (8), writing 0.86 (8), lead male acting 0.89 (8), lead female acting 0.87 (8), supporting male acting 0.81 (8), supporting female acting 0.87 (8), film editing 0.79 (3), cinematography 0.88 (7), art direction 0.79 (3), costume design 0.58 (3), makeup 0.78 (2), visual effects 0.72 (2), sound effects editing 0.40 (2), sound mixing 0.75 (2), score 0.80 (8), and song 0.81 (4). Moreover, of the various awards, the Oscars provide the most reliable index of the overall consensus across organizations (Simonton, 2004b). 2 2 Although movie awards display a high degree of consensus, it is also true that the awards are probably subject to various contaminants, including contagion effects (Collins & Hand, 2006) and sympathy votes (Pardoe & Simonton, 2008). These errors are systematic rather than random. CINEMATIC SUCCESS 405

7 The Criteria Intercorrelations The simplest outcome would be to find strong positive correlations among all of the success criteria defined above. A great film would launch its theatrical run with critical acclaim, do well in the first weekend box office, continue to bring in large audiences for many weeks thereafter, attain high gross earnings, receive multiple nominations and awards in the major categories of cinematic achievement, and be crowned with impressive movie-guide ratings upon release as a DVD or video. This congruence would greatly simplify our understanding of cinema from both psychological and marketing perspectives. Yet reality is not so friendly. To appreciate this unfortunate fact, let us look at the following three groups of correlations. Critical evaluations and financial performance. Most researchers have identified a positive relation between the critics judgments and box office success (e.g., Litman, 1983; Litman & Kohl, 1989; Sawhney & Eliasberg, 1996; Sochay, 1994). Nonetheless, some investigators have found no relationship (Delmestri, Montanari, & Usai, 2005; Reinstein & Snyder, 2005), still others have found a negative association (Hirschman & Pieros, 1985; Simonton, 2005a), or even a U-shaped curve (Wallace, Seigerman, & Holbrook, 1993). These contradictory results largely reflect the complexity of the phenomenon. For instance, the precise relation between the verdicts of the critics and box office returns varies across the successive weeks of the film s theatrical run (Basuroy, Chatterjee, & Ravid, 2003; Eliashberg & Shugan, 1997) and depends on whether the reviews are positive or negative in valence (Basuroy, Chatterjee, & Ravid, 2003). Of special interest is the finding that the connection between reviews and revenues is moderated by other variables, such as star power and genre (Desai & Basuroy, 2005). Critical evaluations and movie awards. The correlation between these two criteria of cinematic success is consistently positive (e.g., Cattani & Ferriani, 2008; Holbrook, 1999; Simonton, 2002, 2004b; Wallace, Seigerman, & Holbrook, 1993). For example, one study of 1322 films obtained a correlation of 0.49 between a composite measure of best picture awards and nominations based on seven professional organizations and a composite measure of critical evaluations based on five movie guides (Simonton, 2004b). Of course, because the later were published after the completion of the award season, the awards could have influenced the ratings. Nevertheless, best picture recognition still has a positive relation with critical evaluations published prior to the time that the awards were bestowed, albeit the correlation may be somewhat smaller (Simonton, 2005b; cf. Cattani & Ferriani, 2008). Critics are better at recognizing the validity of movie awards already granted than they are at anticipating which films will be so honored. It is important to recognize that other award categories may not necessarily correlate highly with critical acclaim (Simonton, 2002, 2004b). On the one hand, awards and nominations in the categories of directing and writing correlate about as highly with critics ratings as best picture recognition; on the other hand, the correspondences are noticeably smaller for the categories of acting, film editing, cinematography, and art direction, and smaller still for the remaining award categories. In fact, nominations and awards in the category of best song can actually exhibit negative correlations with critical evaluations (Simonton, 2002, 2004b, 2007b). In general, the critics opinions show the strongest association with the awards that represent the dramatic quality of a cinematic product. 406 SIMONTON

8 Movie awards and financial performance. It is commonly believed that awards in the major categories such as best picture, directing, and acting help enhance a film s box office returns. Most empirical research appears to bear this out, particularly when we focus on the Oscars, the most prestigious of all awards (e.g., Dodds & Holbrook, 1988; Litman & Ahn, 1998; Nelson et al., 2001; Sochay, 1994; Sommers, ; but see Hirschman & Pieros, 1985). A well-timed award may even save a film from becoming a box office flop. However, the causal relation between movie awards and financial performance is more subtle than first meets the eye. For example, Deuchert, Adjamah, and Pauly (2005) provided evidence that the financial impact of the best picture Oscar is concentrated in the nomination rather than in the award itself. That is, the actual victory adds little or no increment to earnings above what is already provided by the nomination (but see Ginsburgh, 2003; Nelson et al., 2001). In addition, the exact impact of the awards is not stable over time, but rather changes from decade to decade (Smith & Smith, 1986). A final complication is that movie awards in the major categories are more strongly associated with total gross earnings than with first weekend earnings (Simonton, 2005a). Part of this difference reflects the timing of theatrical release, and another part the type of film that is most likely to win the award (Simonton, 2005b, 2007c). This point will receive empirical documentation in the following section. CRITERIA PREDICTORS The three criteria of critical evaluation, financial performance, and movie awards not only display complex interrelationships, but also have been found to possess complex relationships with numerous predictor variables. The latter can be divided into two categories: production and distribution. Production The first set of predictors focus on the factors that go directly into the making of the film. The following three factors have received the most attention in empirical research: budget, screenplay, and personnel. Budget. First and foremost among these factors is the production budget, or what is often referred to as the negative cost (i.e., the cost of producing the negatives from which the positive prints are made for distribution). This represents the literal bottom-line factor that imposes constraints on all other aspects of filmmaking that contribute to cinematic success. Unfortunately, for reasons expressed earlier, it is also a variable that is not very accessible to researchers. Accordingly, empirical investigations incorporating this factor are relatively rare. Yet the studies that do include budget among predictors have discovered that it has an ambiguous relation with the three success criteria. For example, Simonton (2005a) found that production costs had (1a) a zero correlation with nominations and awards for best picture and in the dramatic cluster, (1b) positive correlations with nominations and awards in the visual, technical, and music clusters, (2) positive correlations with both first-weekend and total gross, and (3) negative correlations with critical evaluations registered both during and after the theatrical run (see also Simonton, 2005b). This pattern of relationships suggests that big-budget films CINEMATIC SUCCESS 407

9 often spend money on what moviegoers but not critics most appreciate lots of sights, sounds, and flash but little powerful drama. Other investigators have also found that production costs predict financial performance (e.g., Prag & Casavant, 1994). Ironically, these expenditures do not always ensure that the film will actually turn a profit. Although bigger budgets indicate larger receipts, the earnings may not sufficiently compensate for the increased investment (De Vany & Walls, 1999). In concrete terms, Wallace, Seigerman, and Holbrook (1993) found that, on the average, each additional dollar spent on production brought in only about 45 cents (cf. Litman & Ahn, 1998; Prag & Casavant, 1994). Part of the problem is that the money that goes to purchase outstanding contributions to the visual, technical, and music clusters may be money misspent. For instance, only the awards for best picture and in the dramatic and technical clusters are positively associated with box office gross, yet budget is only correlated with the technical cluster (Simonton, 2005a). In the long run, both the audiences and the critics want more than extravagant special effects, but the cash goes to big and loud explosions anyway. Screenplay. The production budget is largely devoted to the realization of a script, so it is not surprising that many researchers have examined the characteristics of the screenplay that contribute to cinematic success. The resulting investigations have tended to focus on the screenplay s most obvious features. The following seven are especially noteworthy: 1. Sequels. When a film becomes a blockbuster, a sequel becomes a very reasonable option. After all, the sequel is practically pre-marketed by its predecessor. Hence, sequels make up a sizable proportion of the mainstream movies released each year (nearly 10% according to Chang & Ki, 2005). Furthermore, from the standpoint of financial performance, sequels make good economic sense: They have higher odds of making money, especially in the early weeks of the theatrical run (e.g., De Vany & Walls, 1999; Litman & Kohl, 1989; Ravid, 1999; Simonton, 2005b). Yet by other success criteria, sequels are a bad bet. They are less likely to earn recognition for best picture or in the award categories making up the dramatic cluster (though they are prone to do well in the technical cluster), and they have a higher likelihood of getting panned by the critics (Simonton, 2005b; cf. Cattani & Ferriani, 2008). It is rare for a sequel to be as good as its predecessor. As if all that was not bad enough, sequels tend to cost substantially more to make than their predecessors (Simonton, 2005b). 2. Remakes. Films that are remakes of earlier films are something like sequels in the sense that much of their marketing is already in place. As a result, their cinematic success follows a very similar pattern: Remakes do slightly better in the box office, especially in the first weekend, but they are less inclined to receive nominations and awards in the dramatic cluster and they are more poorly received by film critics (Simonton, 2005b; cf. Chang & Ki, 2005). It is rare for a remake to outdo the original. 3. Adaptations. So many films are based on previous plays, novels, short stories, and other media that the Academy of Motion Picture Arts and Sciences has a separate award category for adapted screenplays. It turns out that films based on adaptations tend to receive higher critical praise and to do well in the award categories that make up the dramatic and visual clusters, but do not stand out at the box office, and may actually perform less well if the adaptation is based on a play (Simonton, 2005b; cf. Chang & Ki, 2005). Interestingly, films that receive 408 SIMONTON

10 nominations and awards in the visual cluster have a higher probability of being based on classic works, like the plays of Shakespeare or the novels of Jane Austen (Simonton, 2005b). These are often visually sumptuous costumers. 4. True stories. Some films are based on true stories, such as historical events or notorious episodes, and a subset of these films can be considered full-fledged biopics films that narrate the biography of some notable personality. Such films are often adaptations of historical or biographical nonfiction, but they need not be. Consequently, their correlations with success criteria do not follow the exact same pattern. In the broader case, screenplays based on real events tend to do less well with respect to first-weekend gross, but are more likely to earn critical acclaim and to receive nominations and awards for best picture and in the categories encompassed by the dramatic and visual clusters (Simonton, 2005a). Biopics exhibit a comparable set of correlations, with the exception that they are not quite as likely to receive best-picture honors and that they tend to open their first weekend on fewer screens (Simonton, 2005a). 5. Genre. Films are usually classified according to several genres, such as drama, comedy, romance, musical, action, thriller, adventure, fantasy, horror, scifi, mystery, crime, film-noir, western, war, family, and sports. These categories are not mutually exclusive (e.g., romantic comedy comedy romance). However, some genres are much more common than others. For instance, drama is by far the leading category, while film-noir is the rarest (e.g., Simonton, 2004c). Different genres also vary in their most likely magnitude of success, but that variation is contingent on the criterion of success. To illustrate, whereas the genre of drama scores high with the critics and tends to receive lots of nominations and awards for best picture and in the honors making up the dramatic cluster, dramas are less likely to receive recognition in the technical cluster, and they also tend to do less well financially, opening on fewer screens, earning less in the first weekend, and grossing less at the end of their theatrical run (Simonton, 2005b; see also Chang & Ki, 2005; Litman & Kohl, 1989; Prag & Casavant, 1994; Simonoff & Sparrow, 2000; cf. Sochay, 1994). Dramas may constitute serious art, but they are usually deficient as popular entertainment (see also Holbrook, 1999). 6. MPAA rating. Since 1968 the Motion Picture Association of America has rated films on a scale ranging from accessible to all audiences to restricted to audiences who are 18 and older. The current version of the scale runs as follows: G (general admission), PG (parental guidance advised), PG-13 (parents strongly cautioned because some material may be inappropriate for children under 13), R (restricted; children under 17 would not be admitted without an accompanying parent or adult guardian), NC-17 (no one 17 and under admitted), and Not Rated (most often used for foreign imports that will not appear in the theater chains). Although about half of films are rated either PG or PG-13, R-rated movies often constitute the single most common MPAA rating in film samples (Simonton, 2004c). Even so, R-rated films have a much higher probability of being financial flops with respect to both short- and long-term box office, whereas the most profitable films are to be found in the most accessible categories, especially PG-13 (De Vany & Walls, 2002; Medved, 1992; Ravid, 1999; Simonoff & Sparrow, 2000; Simonton, 2005b; cf. Litman & Ahn, 1998; Walls, 2005b). R-rated films are also less likely to receive nominations and awards in the visual, technical, and music clusters (Simonton, 2005b). So why make R-rated films at all? Evidently, like the genre of drama, they seem to be more artistic. They are prestige productions often the pet project of a movie star or producer who wants to rise above hoi polloi. Such films are more likely to CINEMATIC SUCCESS 409

11 receive nominations and awards in the dramatic cluster and to earn higher praise from film critics (Simonton, 2005b; cf. Holbrook, 1999). They ensure fame, if not fortune. 7. Run time. It thus far seems that the criteria and their predictors tend to form two antithetical groups: film as art versus film as business. What determines success by the criteria of critical acclaim and awards will not affect success by the criteria of financial returns and predictors that are shared may have opposite effects. However, variables can correlate in the same manner with all three sets of criteria. A striking example is how long the film takes to unfold. Run time is positively correlated with critical evaluations, financial performance, and with awards and nominations for best picture, and in all four clusters, namely, the dramatic, visual, technical, and musical (Holbrook, 1999; Simonton, 2005b; Wallace, Seigerman, & Holbrook, 1993). In fact, motion pictures that win Golden Raspberry (Razzie) awards for worst picture, directing, screenplay, and acting are more likely to be shorter than those that win Oscars in the same categories (Simonton, 2007c). Apparently it takes a certain amount of time to realize fully all of the creative potential available in the medium. This complete realization cannot happen in an 80-minute film. Clearly, research has a long way to go before it completely deciphers the screenplay attributes that influence cinematic success. For example, nothing has been said about nature of the plot or characters that make or break a film as judged by the criteria triad (but see Beckwith, 2007; Eliashberg, Hui, & Zhang, 2006). Personnel. Ultimately, the quality of a cinematic product must depend on the artists and experts who collaborate on its production. The empirical literature has concentrated almost exclusively on whether movie stars ensure success by the criteria of financial performance (Albert, 1998; De Vany, 2004; Kindem, 1982). As noted earlier, the concept of star is defined in many different ways in the research, and so it should come as no surprise that the results are not consistent. Some studies indicate that the presence of stars has a positive impact on box office success (e.g., Canterbery & Marvasti, 2001; Litman & Kohl, 1989; Sochay, 1994; Wallace, Seigerman, & Holbrook, 1993; Sawhney & Eliasberg, 1996; Simonoff & Sparrow, 2000), other studies find that stars have no effect (e.g., Delmestri, Montanari, & Usai, 2005; De Vany & Walls, 1999; Litman, 1983; Litman & Ahn, 1998; Prag & Casavant, 1994; Smith & Smith, 1986), and a few even find a negative effect (Chang & Ki, 2005; Simonet, 1980). What is known for sure is that star power is a very unreliable predictor of financial performance. Besides the fact that only a small proportion of so-called stars has any demonstrable effect (De Vany & Walls, 1999), each star s impact is highly unstable over time (Wallace, Seigerman, & Holbrook 1993; see also Lehman, 1941). An A-list celebrity who is a big box-office draw in one year might become box-office poison in the next. Making matters worse, even when stars manage to attract crowds to the movie theaters, their fees may augment the production budget to such a degree that the increased receipts do not recoup the costs (Sedgwick & Pokorny, 1999). In simple terms, big marquee names do not always earn their keep. Their effect is most beneficial when they consent to earn less to perform in medium-budget films that could not otherwise afford their services (Sedgwick & Pokorny, 1999). Empirical research on the impact of star actors and actresses on the other two criteria of cinematic success have been few and far between. And the findings are not always consistent either. Holbrook (1999) found no relation between 410 SIMONTON

12 critical acclaim and the presence of movie stars. However, Zickar and Slaughter (1999) found that the critical evaluation of films was positively correlated with the number of Oscars the lead actors had received in previous films. Simonton (2004a) showed that although concurrent best acting awards and nominations predicted awards and nominations for best picture, the impact was far stronger for male than for female performances. Amazingly, an outstanding acting performance by a man in a supporting role has more predictive power with respect to best picture honors than an outstanding performance by a woman in a lead role! Finally, some investigators have tried to fathom the factors that predict exceptional acting performances that purport to contribute to cinematic success (e.g., Bennett & Bennett, 1998; Lehman, 1941; Pardoe & Simonton, 2008). Naturally, although the actors and actresses represent quite literally the most visible filmmaking personnel, they are far from being the sole contributors to the film product. This fact becomes quite evident in the seemingly endless list of credits that closes almost every motion picture. Even so, because not all collaborators are of equal importance to the final product, research tends to focus on the members of core crew who operate behind the camera, namely, the producer, director, writer, editor, cinematographer, production designer, and composer (e.g., Cattani & Ferriani, 2008). More accurately, investigations have concentrated on the producer, director, writer, and composer, the others receiving little or no attention (cf. Domino, 1974). Regrettably, even this minimal work is more piecemeal than coordinated, as the following representative findings indicate: 1. A producer s box-office success in prior films predicts better-than-average box-office for his or her current film, but the producer s prior film credits predict worse-than-average box-office (Simonet, 1980). This is a stark example of how the operational definition of a variable has critical consequences on the empirical outcomes. Although producers who make blockbusters (or flops) are more likely to make future blockbusters (or flops), those producers with the most total experience in making films tend to be the least likely to make blockbusters. The two producer assets operate at cross purposes. 2. Directors displayed the exact same outcomes as the producers with respect to the financial performance of their films: Prior box-office success had a positive effect but prior experience directing films had a negative effect (Simonet, 1980). This may help explain why the net impact of star directors on financial performance is often null (e.g., Litman, 1983; but see Litman & Kohl, 1989). Although top-notch directors have a positive impact on a film s best picture honors (Simonton, 2002, 2004c) and critical acclaim (Holbrook, 1999), the magnitude of that influence is likely a curvilinear inverted-u function of the director s credits or career age (Lehman, 1941; Zickar & Slaughter, 1999). Thus, directors manifest the same age achievement relation found in other forms of creativity (Simonton, 1997). 3. Empirical studies have shown that the quality of the screenplay, as judged by nominations and awards, rivals directing as a key predictor of both best picture honors and critical acclaim, and script quality actually surpasses the predictive power of acting quality (Simonton, 2002, 2004c). It should be mentioned that it makes a difference who is involved in writing the script. In the first place, films are more likely to receive acclamations from critics if the director is credited with the script, even if only as co-writer; unhappily, writer directors also create films that do more poorly at the box office (Simonton, 2005b). This is consistent with the concept of the Auteur in film theory insofar as the writer director has more control over the final cinematic product (see also Cattani & Ferriani, 2008). CINEMATIC SUCCESS 411

13 4. Although films often contain quite memorable music, the composer s direct contribution to cinematic success is less clear (Simonton, 2007b). Not only are the nominations and awards that make up the music cluster (score and song) orthogonal to the nominations and awards received in the dramatic, visual, and technical clusters (Simonton, 2004c), but also music awards and nominations are only weakly related to best picture awards and nominations and sometimes negatively related to critical evaluations (Simonton, 2002, 2004b, 2004c). Worse yet, music quality seems to be uncorrelated with the main financial performance criteria (Simonton, 2005a). Nonetheless, these results must be interpreted in the context of two additional findings. First, it is essential to distinguish between score and song because the two have contrary effects, the former often positive and the latter often negative (Simonton, 2002, 2004b). Second, Simonton (2007a) showed that cinema composers displayed career trajectories very similar to those of classical composers, trajectories that run independent of the films for which their music is written. Consequently, their most creative work does not necessarily correspond to the quality of the films for which they write. Contrary to the emphasis of so much psychological research, investigators in this area have begun to recognize that group-level factors have consequences beyond the individualistic contributions of the actors and members of the core crew (e.g., Delmestri, Montanari, Usai, 2005; Faulkner & Anderson, 1987; Ferriani, Corrado, & Boschetti, 2005). An excellent illustration is Cattani and Ferriani s (2008) demonstration that the award recognition received by individual crew members was a curvilinear inverted-u function of where the filmmaking team stood with respect to a core-versus-periphery dimension. That is, individuals attained higher levels of cinematic achievement when they belonged to teams that occupied an intermediate position in the social network that defines the film industry. Not surprisingly, the optimal crew composition also depends on the specific criterion of cinematic success. Delmestri, Montanari, and Usai (2005) showed that financial performance was enhanced when directors had strong working relationships with the producers and distributors, whereas critical acclaim and movie awards were associated with the directors having weak ties with writers, actors, and cinematographers. Distribution A film s distribution during its theatrical release has a major connection with its cinematic success. Indeed, what happens during this phase may make or break a film s theatrical run. Investigators have devoted the most attention to just five variables: Season of Release. Judged by financial performance, the big money-making movies are most likely to appear in the summer months (e.g., Chang & Ki, 2005; Litman & Ahn, 1998; Litman & Kohl, 1989; Sochay, 1994). Secondary periods with respect to box office returns are Christmas and, to a lesser and more inconsistent extent, Easter (Chang & Ki, 2005; Litman, 1983; Sochay, 1994), with family films having special prominence in the Christmas season (Litman & Ahn, 1998). In contrast, the first few months of each year are most likely to see the biggest flops. Yet if we judge the seasonal distribution by more aesthetic or artistic criteria, the picture changes dramatically. Films that receive nominations and awards in the major categories are most likely to be released during the 412 SIMONTON

14 Christmas season and are least likely to see summer release (Simonton, 2007c). Conversely, films dishonored with Razzie recognition are most likely to appear in the summer (Simonton, 2007c). In the case of critical acclaim, I obtained a correlation of 0.12 (p 0.001, n 1711) between the number of stars a film receives in movie guides and the number of the month in which it was released (1 January to 12 December, using data collected for Simonton, 2005b).The later in the year a film is released, the higher is its critical evaluation. Number of Screens. Films commence their theatrical runs in two diametrically opposed ways. On the one hand are the films that start their first weekend on thousands of screens, and on the other hand are those that open on fewer than a dozen, perhaps even just one or two. Wide-release films often tend to earn the most in the first weekend of release (De Vany & Walls, 1999), and also are prone to earn greater box office by the end of their theatrical run (Litman & Ahn, 1998; Litman & Kohl, 1989; Simonton, 2005b). Even though such wide distribution on the first weekend is positively associated with awards and nominations in the technical and musical clusters (Simonton, 2005b), wide release has no correlation with awards and nominations for best picture or for honors in the dramatic and visual clusters, and, even worse, has a negative correlation with critical acclaim (Simonton, 2005b). In other words, the critics favor films that do the art-house circuit. Major Distributor. Many independent filmmakers display their products at major film festivals, such as Sundance, in the hope that they will sign a contract with a major distributor (Warner, Universal, Columbia, Paramount, Twentieth- Century Fox, etc.). However, distribution by one of the majors has a somewhat ambivalent relation with financial performance, sometimes having either no effect or even a negative effect, depending on the specific success criterion (e.g., Litman, 1983; Litman & Kohl, 1989). For instance, Chang and Ki (2005) found that the distributor only determined the length of run: Films released by one of the majors stayed in the theaters longer. Marketing Expenditures. Because this information is even less likely to be available than production budget, it is extremely rare to find advertising costs included in any empirical analysis. The most significant exception is a study by Prag and Casavant (1994) that was able to obtain prints and advertising figures for 195 films out of a sample of 652 films for which they had acquired data on production costs. For this subgroup, advertising was found to be positively associated with critical evaluations and movie awards. It merits observation that the primary predictors of marketing expenditures included production budget and the presence of movie stars. The former predictor suggests that the filmmakers try to recoup their investment by extensive advertising, while the latter predictor suggests that movie stars provide useful ammunition in such advertisements. Surprisingly, production budget and star power are not necessarily correlated (Basuroy, Chatterjee, & Ravid, 2003), and hence these two factors may make independent contributions to the amount spent on advertising. CINEMATIC SUCCESS 413

15 Market Competition. Probably the most critical distribution factor besides the preceding is studied the least, namely, the amount of competition that a film faces from other films during its theatrical release (Krider & Weinberg, 1998). A film that might otherwise do very well in the box office may find itself pushed aside by an abundance of competing films that opened on the same weekend or shortly before. In brief, timing may be everything. Indeed, if all films are equal in merit but differ in the number of films released at the same time, then we might expect that each film s market share would simply be the reciprocal of the number of films in head-to-head competition. In any case, empirical research does indicate that market competition does adversely impact a film s box office performance (e.g., Sochay, 1994) as well as net profit (Litman & Ahn, 1998). Of course, the deficit gets even worse when a mediocre film opens against a superior film. Hence, many inferior motion pictures come out in the early months of the year so as to not compete with the summer blockbusters or winter art products aiming at movie awards. Curiously, no empirical investigation to date has examined how market competition might affect other success criteria, be they critical evaluations or movie awards. It would be expected that competition would have a much bigger effect on the latter than the former. Nothing prevents a critic from assigning five stars to more than one film, but only one film can win the Oscar for best picture. Sadly, not a single empirical investigation has examined all of the foregoing distribution variables in a single multivariate equation. So it is impossible to say anything about their relative importance even with respect to the criterion of financial performance. SOME PSYCHOLOGICAL ISSUES Psychologists were directly responsible for only a very small minority of the investigations just reviewed (e.g., Boor, 1990; Lehman, 1941; Simonton, 2004c; Zickar & Slaughter, 1999). Accordingly, the work was seldom motivated by psychological theories, hypotheses, or conjectures. On the contrary, the literature is largely dominated by prevalent ideas in economics and marketing. Even so, the findings often raise issues that that have genuine psychological interest. Some of these issues can provide the basis for future research in the area of psychology and marketing. Consider the four examples: social influence, group problem solving, creative talent, and cinematic aesthetics. Social Influence At various points during production and distribution there intrude occasions in which individuals influence the ideas or behaviors of other individuals. Indeed, the very fate of a potential film often depends critically on the outcome of pitch meetings in which screenwriters try to persuade studio executives and producers to green light a proposed project (Elsbach & Kramer, 2003). Once a film is actually made, its financial performance is heavily contingent on various kinds of influence processes. Besides the impact of direct advertising already noted (Prag & Casavant, 1994), various influences operate throughout the theatrical release. Two have received a great deal of attention. 414 SIMONTON