English (pdf)
Article in xml format
Article references
Automatic translation
Permalink
Watching movies is one of the most common forms of fun. For example, in 2019, the movie industry earned about $100 billion, with 268 million people going to the cinema at least once in the United States and Canada (Motion Picture Association of America, 2019). With the development and popularization of streaming platforms (e.g., Netflix, Amazon Prime, HBO GO), access to hundreds of titles of various genres has been facilitated and expanded to various devices (e.g., TVs, notebooks, tablets, smartphones). For instance, Netflix and Amazon Prime have combined approximately 250 million subscribers worldwide (Statista, 2020a, 2020b).
However, when watching a movie, people are not merely passive receivers of information (Rentfrow, Goldberg, & Zilca, 2011). According to the paradigm of uses and gratification, people play an active role in selecting the media and its content (Katz, Blumler, & Gurevitch, 1974). Also, their individual differences mediate or moderate the effects of exposure to content conveyed in the media (Bartsch, Appel, & Storch, 2010; Krcmar, 2017). Thus, myriad psychological variables influence this complex selection process (González-Vázquez & Igartua, 2018). Furthermore, the preference for a movie genre is consistent or expresses specific personality traits (Bowes, Watts, Costello, Murphy, & Lilienfeld, 2018).
For instance, people who prefer horror movies tend to have higher levels of psychoticism (Weaver, 1991), psychopathic traits and negative affect, and lower levels of detachment (Blagov, Von Handorf, Pugh, & Walker, 2019). On the other hand, a preference for comedy and science fiction movies has been associated with higher levels of extraversion (Bowes et al., 2018; Weaver, Brosius, & Mundorf, 1993) and psychopathy (Bowes et al., 2018). War and Western movies are preferred by people with low levels of emotionality, and high levels of extraversion, narcissism, and psychopathy (Bowes et al., 2018). According to these authors, this last personality trait is also negatively associated with the preference for romantic movies.
Nevertheless, research in this field frequently performs exploratory factor analyses, resulting in non-interpretable factors composed of genres that do not present theoretical similarities (e.g., war, westerns, and sports; Bowes et al., 2018), in addition to the low construct validity – such studies do not present internal consistency coefficients for such factors. Therefore, inadequate instruments to operationalize the preference for movies can represent a barrier to progress in this field. Another problem is assessing the preference for movies through open questions, which has been done in the literature (Pimentel et al., 2014). This procedure may make it impossible to estimate the magnitude of such preferences, which would generate categorical variables for preference evaluation and reduce the possibilities of statistical analysis. Moreover, participants would be at the mercy of their memory and may forget to indicate some genres. Another commonly adopted practice is listing titles of different genres and asking participants to indicate their favorite (Weaver, 1991). However, familiarity with the title may vary among participants, which might be a bias (Wühr, Lange, & Schwarz, 2017).
The present research
Therefore, in the present study, we propose a new way of evaluating the preference for movies based on the Short Test of Music Preference – STOMP (Rentfrow & Gosling, 2003). We replaced music with movie genres, considering well-known types listed in previous studies (e.g., Bowes et al., 2018; Wühr et al., 2017). However, instead of adopting an exploratory approach that may result in meaningless factors, we analyzed the characteristics of 15 distinct genres, proposing a five-factor model based on their common characteristics (MOVIE; Melodrama, cOmic, Violent, Imaginative, and Exciting).
Specifically, romances (e.g., City of Angels), dramas (e.g., The Fault in Our Stars), and musicals (e.g., Phantom of the Opera) are usually movies that have plots centered on romantic relationships and conflicting situations around a romantic pair. Through the story, they pass situations that elicit strong sentimentality in the spectators, usually with happy endings. Because of such common characteristics, such genres would form the melodrama factor. Comedy (e.g., Dumb & Dumber) and animation (e.g., Angry Birds) movies operationalize the comic factor, marked by amusing stories that provoke laughter. Action (e.g., John Wick), war (e.g., Apocalypse Now), westerns (e.g., Django), and epics (e.g., Gladiator) movies present common aspects, such as explicit violence, with intense combat scenes, shootings, and bloody battles, leading them to operationalize the violent factor. Adventure (e.g., The Lord of the Rings), fantasy (e.g., Chronicles of Narnia), and science fiction (e.g., Twenty Thousand Leagues Under the Sea) movies usually involve imaginary, parallel and/or fantastic universes, with plots about challenging journeys, with magical beings and characters facing several challenging situations to achieve specific goals. We hypothesize that such genres will cluster in a dimension named imaginative. Finally, horror (e.g., Nightmare on Elm Street), thriller (e.g., The Suspects), and mystery (e.g., Se7en) movies often raise sensations of fear, anxiety, and excitability, forming a dimension named exciting.
In addition to testing the suitability of the MOVIE model, we relate it with the Big Five Personality Factors (i.e., openness, conscientiousness, extraversion, agreeableness, and neuroticism) and the Dark Tetrad (i.e., psychopathy, narcissism, Machiavellianism, and sadism). People with high scores in neuroticism watch drama movies to experience more intense negative emotions via dramatic scenes (Kallias, 2012). They also prefer romantic and musical genres because they are predictable, usually with happy endings and hopeful messages, helping them to regulate their emotions (Chausson, 2010). In addition, people with high scores in neuroticism can watch romantic movies motivated by nostalgia, remembering, and reviving past emotional experiences (Chamorro-Premuzic, Kallias, & Hsu, 2014). Thus, we hypothesized that higher levels of neuroticism predict a preference for melodrama movies (Hypothesis 1).
People with high levels of extraversion tend to appreciate humor more (Moran, Rain, Page-Gould, & Mar, 2014), having a greater tendency to laugh, which is a characteristic of this trait (Ruch & Deckers, 1993). This leads us to hypothesize that this personality trait will predict the preference for comic movies (Hypothesis 2). Sadism is a dark personality trait that describes a tendency to pleasure seeing other people suffering (Paulhus, 2014). People with a high level of such traits tend to present a fascination for weapons and violent video games (Gonzalez & Greitemeyer, 2018). Such characteristics might indicate that this personality profile will prefer movies with the violent factor (Hypothesis 3). We also hypothesize that openness will predict the imaginative factor (Hypothesis 4), as this personality trait describes those most prone to fantasy, being more inventive and creative (DeYoung, Quilty, Peterson, & Gray, 2014), and having a greater propensity to be absorbed into the fantastic universes created in such movies. Finally, we hypothesize that openness (Hypothesis 5), psychopathy (Hypothesis 6), and sadism (Hypothesis 7) will predict the exciting factor. Those with high levels of sadism and psychopath are characterized by low empathy. People with high empathy might dislike the genres that compose this factor, as these individuals tend to react negatively to the pain and ill-treatment of others (Tamborini, 1996). People with psychopathic traits are thrill seekers (Patrick, Fowles, & Krueger, 2009) and openness to experiences covers sensation seeking in the Big Five model (Chamorro-Premuzic et al., 2014). Sensation seekers like stimuli that elicit negative emotions (e.g., fear, anguish), because their intensity might help them reach their ideal levels of arousal (Hoffner & Levine, 2005).
Method
Participants and Procedure
Participants were 515 individuals, with age varying between 15 and 68 years old (M = 24.46; SD = 9.19), mostly women (70.5%), single (81.6%), with incomplete higher education (53%), and from middle class (42.7%). We collected the data using an online questionnaire, sharing the research link on social networks. To participate in the study, it was necessary to read and agree to the informed consent form. The participants were fully informed about the nature and aims of the study, as well as the anonymous and voluntary nature of their participation. Therefore, the resolutions 466/2012 and 510/2016 that guide research involving human participants in Brazil were taken into account.
Materials
Movie Preference Scale
Composed of 15 well-known genres, resulting in five factors: Melodrama (e.g., Romantic, Drama), cOmic (Comedy and Animation), Violent (e.g., Action, War), Imaginative (e.g., Fantasy, Sci-Fi), and Exciting (e.g., Horror, Suspense). Participants are instructed to indicate their preference using a seven-point scale (1 – Dislike Strongly; 7 – Like Strongly).
To assess the Five Factors of personality, we used the Brazilian adaptation (Pimentel, Ferreira, Vargas, Maynart, & Mendonça, 2014) of the Ten-Item Personality Inventory (Gosling, Rentfrow, & Swann Jr., 2003). This measure consisted of ten items, equally distributed among the Big-5 factors of personality. Participants must indicate to what extent each pair of adjectives describe them, using a seven-point scale (1 – Strongly Disagree; 7 – Strongly agree). For instance: “Open to new experiences, complex” (Openness), “Dependable, self-disciplined ” (Conscientiousness), “Extraverted, enthusiastic” (Extraversion),“ Sympathetic, warm” (Agreeableness) and “Anxious, easily upset” (Neuroticism).
To assess the aversive personality traits, we used the Brazilian adaptation (Gouveia, Monteiro, Gouveia, Athayde, & Cavalcanti, 2016) of the Dark Triad Dirty Dozen (Jonason & Webster, 2010). We also added the four items with higher factorial loadings from the Assessment of Sadistic Personality (Plouffe, Saklofske, & Smith, 2017). Participants are asked to indicate their agreement (1 – Strongly Disagree; 5 – Strongly Agree) to items such as “I tend to manipulate others to get my way” (Machiavellianism), “I tend to be cynical” (Psychopathy), “I tend to seek prestige or status” (Narcissism) and “Watching people get into fights excites me” (Sadism).
Data Analysis
To assess data, we used the software R (R Core Development Team, 2015) and PASW. With R, a Confirmatory Factor Analysis was performed to test the fit of the proposed theoretical model for the Movie Preference Scale, using the lavaan package (Rosseel, 2012). Specifically, the Diagonally Weighted Least Squares (DWLS) estimator was used, considering the following recommended fit indices (in parentheses, values for an acceptable model; Hu & Bentler, 1999): χ²/df ratio (< 3.0), Comparative Fit Index (CFI > .90), Gamma Hat (> .90), Goodness of Fit Index (GFI > .90), Adjusted Goodness of Fit Index (AGFI > .90) Root Mean-Square Error of Approximation (RMSEA < .08), and Standardized Root Mean Square Residual (SRMR < .08). With PASW, we calculated Pearson’s correlations and hierarchical regressions to verify how much movie preference is related to and predicted by personality traits.
Results
Initially, we assessed the adequacy of the proposed five-factor model using the DWLS estimator. The model fit presented values slightly below the recommended: (χ²/df = 4.07, CFI = .89; GammaHat = .94; GFI = .99; AGFI = .99; RMSEA = .077 [IC90% .069 - .086]; SRMR = .075). Since movies are combinations of different genres (Choi, Ko, & Han, 2012), we analyzed the modification indices and correlated the pairs of errors of the action and adventure genres (M.I. = 44.77). After that, the updated model (Figure 1) presented acceptable fit indicators: (χ²/df = 3.54, CFI = .91; GammaHat = .95; GFI = .99; AGFI = .99; RMSEA = .070 [IC90% .062 - .079]; SRMR = .068). In addition, the internal consistency coefficients (Composite Reliability) were acceptable, except for Melodrama, slightly below the recommended (Bagozzi & Yi, 1988): Melodrama (CR = .59), cOmic (CR = .64), Violent (CR = .73), Imaginative (CR = .66), and Exciting (CR = .81).
After, we verified the correlations between preference for movies and personality traits. In general, the correlations were weak (Table 1), indicating, for instance, that extraversion is related with preference for comic movies (r = .11; p < .01), as well as agreeableness (r = .10; p < .05). On the other hand, openness (r = .07; p < .05) is related with preference for imaginative movies, as well as sadism (r = .12; p < .01). This trait (r = .16; p < .01), as well as psychopathy (r = .08; p < .05) is associated with preference for exciting movies. Sadism (r = .18; p < .01) and psychopathy (r = .20; p < .01) also is related with preference for violent movies, and negatively with melodramatic ones (r = -.14; p < .01 for both).
Table 1 Zero-order correlations and standardized regression coefficients for the Movie Preferences
| r (β) | |||||
|---|---|---|---|---|---|
| M | O | V | I | E | |
| Big Five | (R2 = .024) | (R2 = .029) | (R2 = .009) | (R2 = .018) | (R2 = .045) |
| OP | .01 (.00) | -.04 (-.07) | .04 (.03) | .07* (.07) | .10* (.14**) |
| CO | .02 (.01) | .02 (.00) | .00 (.00) | -.07* (-.07) | -.06 (-.04) |
| EX | .08* (.06) | .11** (.12**) | .05 (.06) | .03 (.03) | -.09* (-.09*) |
| AG | .08* (.05) | .09* (.07) | .02 (.05) | .06 (.12*) | -.08* (.02) |
| NE | .08* (.09*) | -.07* (-.05) | -.08 (-.06) | .00 (.01) | .12** (.12**) |
| Dark Tetrad | (R2 = .045) | (R2 = .035) | (R2 = .043) | (R2 = .039) | (R2 = .090) |
| MA | -.04 (.03) | -.07 (-.05) | .03 (-.07) | .05 (-.03) | .11** (.00) |
| PS | -.14** (-.09) | -.05 (.03) | .08* (.04) | .07* (.06) | .20** (.16**) |
| NA | .03 (.06) | .00 (.04) | .03 (.00) | .02 (.00) | .00 (-.09) |
| SA | -.14** (-.10) | -.09* (-.06) | .16** (.20**) | .12** (.13*) | .18** (.11*) |
Note: M = Melodrama; O = cOmic; V = Violent; I = Imaginative; E = Exciting; OP = Openness; CO = Conscientiousness; EX = Extraversion; AG = Agreeableness; NE = Neuroticism; MA = Machiavellianis; PS = Psychopathy; NA = Narcissism; SA = Sadism.
**p < .01; *p < .05 (Uni-tailed).
Finally, to assess the role of personality traits in predicting movie preferences, we performed five hierarchical regression analyses, having the five MOVIE factors as dependent variables. We included the Big Five personality traits in the first step of each analysis. In the second step, we included the Dark Tetrad to assess if these aversive traits explain additional variance in movie preference. The Big Five traits explained 2.4% (p = .028) of variance in the Melodrama factor and adding the Dark Tetrad traits accounted for significant variance above and beyond the Big Five (∆R² = .021, p = .026), but only neuroticism remain as a predictor (β = .09; p < .05). The Big Five explained 2.9% (p = .011) of variance in the Comic factor, but adding the Dark Tetrad did not result in more explained variance (∆R² = .006, p = .539), and extraversion was the only predictor of this factor (β = .12; p = .01). The Big Five traits explained 0.9% (p = .453) of variance in the Violence factor and adding the Dark Tetrad traits accounted for significant variance above and beyond the Big Five (∆R² = .034, p = .002), in the final step sadism was the only predictor of this preference (β = .20, p < .01). The Big Five traits explained 1.8% (p = .108) of variance in the Imaginative factor and adding the Dark Tetrad traits accounted for significant variance above and beyond the Big Five (∆R² = .022, p = .024), in the final step, agreeableness (β = .12, p < .05) and sadism (β = .13, p < .05) predicted this preference. Finally, the Big Five traits explained 4.5% (p < .001) of variance in the Exciting factor and adding the Dark Tetrad traits accounted for significant variance above and beyond the Big Five (∆R² = .045, p = < .001), in the final step, extraversion (β = -.09, p < .05), neuroticism (β = .12, p < .01), openness (β = .14, p < .01), psychopathy (β = .16, p < .01) and sadism (β = .11, p < .05) predicted this preference.
When we controlled for participant’s biological sex (1 – Female; 2 – Male) in the regression analyses, we found that women show a preference for Melodrama (β = -.32, p < .01), and cOmic (β = - .10, p < .05). On the other hand, men prefer Violent (β = .27, p < .01), Imaginative (β = .17, p < .01), and Exciting (β = .15, p < .01) movies. Furthermore, after the insertion of sex, neuroticism ceased to predict preference for Melodrama, and sadism ceased to predict preference for Imaginative and Exciting. Finally, extraversion also ceased to predict Exciting.
General discussion
Watching movies is one of the most common forms of entertainment, which generates billions in profits for the movie industry (Motion Picture Association of America, 2019) and millions of subscribers to streaming services (Stastita, 2020a, 2020b). Nevertheless, studies in this area focus more on the effects of media on behavior than understanding which psychological variables predict the preference for a specific genre, even considering that the viewer has an active role in the content selection process (Katz et al., 1974; Rentfrow et al., 2011). One of the limitations is that there is no theoretically driven structure of preference for movies that contemplate the typical characteristics of the different genres. Previously, researchers have conducted exploratory factor analysis considering only the statistical aspects, which resulted in questionable and theoretically unrelated factors (Blagov et al., 2019; Bowes et al., 2018). In addition, these studies propose different factorial solutions, which makes direct comparisons and meta-analytic studies impossible.
In the present research, we considered lists of well-known genres used in previous studies (Bowes et al., 2018; Wühr et al., 2017), analyzing the common aspects of each type and hypothesizing a five-factor structure (Melodrama, cOmic, Violent, Imaginative, Exciting). The MOVIE structure presented a good model fit (Hu & Bentler, 1999), being the first structure of movie preference to be assessed at a confirmatory level. In addition, its factors showed acceptable internal consistency values for research purposes (Bagozzi & Yi, 1988), demonstrating that the instrument has acceptable psychometric parameters for future studies in the area.
Regarding the role of personality traits in predicting movie preference, we confirmed six of our seven hypotheses. For instance, neuroticism was the only trait that predicted a preference for melodramatic movies, supporting hypothesis 1. Such movies have intensely dramatic scenes, and people high in neuroticism might watch them to experience negative emotions more intensely (Kallias, 2012). Furthermore, such movies tend to have predictable endings, leaving hopeful messages, which can help neurotics with their emotional regulation (Chausson, 2010), in addition to commonly portraying romantic couples, which might lead them to recall affective experiences from the past (Chamorro-Premuzic et al., 2014). However, it should highlighted that this trait fails to predict the preference for melodramatic movies after controlling for participants’ sex, which may indicate some interaction between such variables.
Extraversion was the only predictor of the preference for comic movies, confirming hypothesis 2. One of the characteristics of this trait is the greater tendency to laugh (Ruch & Deckers, 1993). Previous research found that people high in extraversion appreciate humor more (Moran et al., 2014), one of the main characteristics of comic movies, which tend to present hilarious stories.
Confirming hypothesis 3, sadism was the only predictor of the preference for violent movies. These tend to present plots marked by explicit violence, with scenes of shootings, battles and bloody fights, perfect ingredients for people with high levels of this trait. The unique aspect that characterizes sadism is feeling pleasure in seeing or making other people suffer (Monteiro, Medeiros et al., 2020; Paulhus, 2014). This personality profile has a fascination for weapons (recurrent objects in such movies) and prefers other media with similar content, such as violent video games (Gonzalez & Greitmeyer, 2018).
Regarding imaginative movies, we did not confirm hypothesis 4, that the openness trait would predict its preference. In this case, the preference for such movies was predicted by agreeableness and sadism, but this dark trait fails to predict when controlling for participants’ sex. The protagonists of imaginative movies are frequently at risk, facing challenging situations and physically and emotionally suffering before reaching their goals. Such developments might attract spectators with high levels of sadism. Whereas for agreeableness, in such movies, there are recurrent help behaviors and the establishment of strong bonds of friendship, loyalty, and cooperation between the main characters, an essential aspect for the protagonists to achieve their goals. Therefore, such aspects can arouse the preference of more agreeable people for such movies.
Finally, confirming hypotheses 5, 6 and 7, results showed that the preference for exciting movies is predicted by openness, psychopathy and sadism. Exciting movies are also predicted by extraversion and neuroticism, but these turn nonsignificant after controlling for sex. These movies elicit fear and anxiety and bring physiological excitement. In such movies, scenes of pain and suffering are recurrent, which generates a more sadistic interest in taking pleasure from the suffering of others and having low empathy (Paulhus, 2014; Tamborini, 1996). People with higher levels of psychopathy are thrill seekers (Patrick et al., 2009). Sensation seekers like stimuli that elicit negative emotions, helping them to reach their ideal levels of arousal (Hoffner & Levine, 2005), an explanation also valid for the openness trait, as it covers the sensation seeking in the Big Five model (Chamorro-Premuzic et al., 2014). Individuals with high neuroticism experience negative emotions (e.g., fear, anguish) more intensely, have stronger reactions to aversive stimuli (Weibel, Wissmath, & Stricker, 2011) and are more immersed in media with harmful content (Weibel, Wissmath, & Mast, 2010). Finally, more extraverts might prefer movies with funny stories that elicit laughter (Moran et al., 2014; Ruch & Deckers, 1993). Consequently, individuals that present lower scores on this trait (more introverts) might present a preference for movies that explore themes that involves pain, suffering, fear, and anguish.
Even though the effects have low magnitudes, it is worth mentioning that the process of choosing a media or content is multidetermined, influencing contextual and mood variables (Rentfrow et al., 2011). However, these authors point out that any observed relation indicates the role that personality traits can play in the real world, influencing preferences and behaviors, indicating that the viewer has an active role in choosing the content (Katz et al., 1974) that may reflect, to some extent, specific personality traits (Blagov et al., 2019; Bowes et al., 2018).
Implications and Future Studies
From a clinical point of view, movies can be used in a psychotherapeutic context (Wedding & Niemiec, 2003). For example, it is possible to psychoeducate the patient about their problem through movies (Oliva, Vianna, & Lotufo Neto, 2010), which elicit emotions and can lead the patient to come into contact with stifled emotions (Hesley & Hesley, 1998). Within this context, preference for movies can be an essential variable to understanding a patient’s emotions better, increasing the chance of effectiveness of treatment if the chosen stimulus gravitates around the patient’s preferences.
In the psychology of music, musical preference is associated with the perception of similarity between the personality profile of the fan and the artist (Greenberg, Matz, Schwartz, & Fricke, 2020). Paralleling the movie preference, which may express specific psychological characteristics, an individual may prefer a specific genre or factor because he perceives similarities between his personality and some character in the movie. In the clinical context, this can be useful when the patients relate to a movie character and their conflicts and problems. Thus, psychologists can help patients reflect on how they would act and face such problems, reducing distorted thoughts and increasing the behavioral repertoire since this type of media can influence behavior (Oliva et al., 2010).
Despite our study’s robust evidence and potential applicability, we should highlight some limitations. It is worth mentioning the non-probabilistic nature of the sample used, making it impossible to generalize the results, with a prevalence of university students. Furthermore, because of their length, the instruments used to assess personality have some limitations, not covering the full range of characteristics of the evaluated traits (Maples, Lamkin, & Miller, 2014) and considering multidimensional traits as unidimensional (Monteiro, Nogueira, Reis, Monteiro, & Nascimento, 2022).
As future possibilities, it is important to assess to what extent continued exposure to such movie types can affect people’s mental health. For instance, in the psychology of music, research has shown that a preference for intense music is related to internalizing problems (Monteiro, Coelho, Vilar, Pimentel, & Andrade, 2020). However, due to the lack of a solid model, little is known about the effects that different movie types could have on people’s well-being, being an important topic to be explored in future possibilities. Another point is to test the MOVIE model with a larger and more heterogeneous sample from Brazil and other countries, gathering evidence of factorial invariance, as well as testing other psychometric parameters (e.g., test-retest stability), and checking whether the measure predicts behavioral tasks (e.g., choosing a movie).
