Listening, Watching, and Reading: The Structure and Correlates of Entertainment Preferences

Abstract

People spend considerable amounts of time and money listening to music, watching TV and movies, and reading books and magazines, yet almost no attention in psychology has been devoted to understanding individual differences in preferences for such entertainment. The present research was designed to examine the structure and correlates of entertainment genre preferences. Analyses of the genre preferences of over 3,000 individuals revealed a remarkably clear factor structure. Using multiple samples, methods, and geographic regions, data converged to reveal five entertainment-preference dimensions: Communal, Aesthetic, Dark, Thrilling, and Cerebral. Preferences for these entertainment dimensions were uniquely related to demographics and personality traits. Results also indicated that personality accounted for significant proportions of variance in entertainment preferences over and above demographics. The results provide a foundation for developing and testing hypotheses about the psychology of entertainment preferences.

In an average week, the typical American spends approximately 38 hours watching television shows and movies, 8 hours reading books, magazines, and newspapers, and 18 hours listening to recorded music and radio (Motion Picture Association of America, 2007). Assuming the average person sleeps eight hours a night, people spend roughly 55% of their waking hours attending to entertainment media. Americans spend almost as much of their annual incomes on entertainment as they do on health care, and more money is spent on entertainment than on education, personal care, and charitable donations (United States Bureau of Labor, 2008). Of the money spent on entertainment, cable and satellite TV account for 39%, books, magazines, and newspapers for 23%, movie consumption for 19%, recorded music for 6%, and Internet, video games, mobile content, and satellite radio combined for 13% (Motion Picture Association of America, 2007).

Considering that entertainment media are nearly ubiquitous, it is astonishing that they have received so little attention in personality and social psychology. Indeed, of the nearly 15,000 articles published between 1932 and 2008 in the Journal of Personality and Social Psychology, Journal of Personality, and Personality and Social Psychology Bulletin, “television,” “movie,” “film,” “music,” “book,” “magazine,” or “media” were listed as subject headings in only 90 of them—a mere 0.6%. Many psychologists have argued that researchers need to broaden their research foci and pay more attention to ordinary aspects of people's daily lives so that we can develop an understanding of social behavior that is more ecologically sensitive (e.g., Funder, 2001; Rozin, 2001). Entertainment is undoubtedly important to people and permeates many aspects of social life, yet we know little about it. The present work was designed to explore the landscape of this undeveloped terrain with the aim of establishing a foundation on which to develop and test hypotheses about the psychology of entertainment preferences.

Effects of Media Exposure

A fair amount of the psychological research concerned with entertainment falls within the classic “media effects” paradigm (Adorno & Horkheimer, 1979), which is based on the assumption that certain media have a direct effect on individuals' attitudes, emotions, and behaviors. For instance, results from a recent meta-analysis indicated that exposure to TV and magazine advertisements that depict women as sex objects has a small but significant effect on body dissatisfaction (Grabe, Ward, & Hyde, 2008; but see Holmstrom, 2004 for an alternative position). Although data have been inconsistent, some studies have found that exposure to “violent” media (e.g., heavy metal or rap music, action movies, violent video games) increases aggressive thoughts and hostile feelings (e.g., Anderson, Berkowitz, Donnerstein, Huesmann, Johnson, et al., 2003; Wood, Wong, & Chachere, 1991). And, research suggests that regular exposure to educational TV programs in childhood is associated with higher achievement motivation and enhanced academic performance (Anderson, Huston, Schmitt, Linebarger, & Wright 2001; Comstock, 1995).

Although some researchers have concluded that the research supports a link between exposure to violent media and a variety of outcomes, others have questioned the validity of the media-effects paradigm, arguing that the research is politically charged and methodologically flawed (Ferguson, 2008; Gauntlett, 2005; Grimes, Anderson, & Bergen, 2008; Olson, 2004). Those critics have argued that the experimental conditions used to study media effects do not accurately reflect the complex conditions in which media are typically experienced in everyday life and that people are not passive recipients of entertainment but are active agents who seek out or avoid particular content. Thus, Gauntlett (2005) suggested that instead of just looking at how certain media affect people, researchers should first examine individual differences in entertainment preferences, identify the characteristics common to people with similar preferences, and then study how particular media affect those most likely to select them. Therefore, if we are to develop a full understanding of the role that entertainment plays in people's lives, we need to examine individual differences in preferences for a variety of entertainment media.

Individual Differences in Entertainment Preferences

Research on individual differences in preferences for entertainment is based on interactionist theories which assume that people prefer content that satisfies psychological needs. Thus, just as individuals seek out and create social and physical environments that reflect their personalities and self-views (e.g., Buss, 1987), it is postulated that people also seek out specific media content that reinforces their attitudes and dispositions. There are two general streams of research on this topic.

The first is concerned with the uses and gratifications of entertainment. For example, why do people listen to music? What are the motives underlying people's decisions to read a book? What do people expect to gain from watching TV? Research suggests that young people listen to music for purposes of identity development, emotion regulation, and relaxation (North, Hargreaves, & O'Neill, 2000). A few investigations of television viewing patterns suggests that young people watch TV for companionship whereas adults watch TV to acquire new information (Potts, Dedmon, & Halford, 1996), and that people high in neuroticism are more likely to watch television for entertainment and companionship than are people low in neuroticism (Weaver, 2003). Research also suggests that individuals seek out particular media content to manage their mood. For example, there is evidence that young people tend to seek out particular media to decrease negative affect (Dillman-Carpentier, Brown, Bertocci et al., 2008; Knobloch & Zillman, 2002). Moreover, people high in rumination are more likely to listen to mood-congruent music when feeling sad, whereas people low in rumination are more likely to listen to music high in positive affect when feeling sad (Chen, Zhou, & Bryant, 2007). Thus, it appears that entertainment can help satisfy some psychological needs and that people may differ in the particular media they use to satisfy those needs.

The second line of research on this topic investigates the links between personality traits and preferences within various media. Most of this research has examined preferences for music genres, but a few studies have examined preferences for book, TV, and movie genres. Research on music suggests that there are basic dimensions of music preferences that are related to personality traits. For instance, preferences for “reflective and complex” music (defined by classical, jazz, blues, and folk genres) are positively associated with openness to experience, verbal ability, and liberal political orientation, and negatively related to social dominance; preferences for “upbeat and conventional” music (pop, country, Christian, and film genres) are positively related to extraversion, agreeableness, conscientiousness, and political conservatism, and negatively related to verbal ability (e.g., Delsing, ter Bogt, Engels, & Meeus, 2008; Rentfrow & Gosling, 2003; Rentfrow & McDonald, in press). A few studies also suggest that more affluent and highly educated people prefer “highbrow” music (e.g., classical, opera, jazz) and a wider variety of musical genres compared to individuals from lower status groups, who tend to prefer “lowbrow” music (e.g., country, rock, rap; Katz-Gerro, 1999; Van Eijck, 2001). Two studies have examined the structure of book preferences and their connections with personality (Schutte & Malouff, 2004; Tirre & Dixit, 1995). Schutte and Malouff (2004) observed eight book-genre preference dimensions and Tirre and Dixit (1995) found 10 preference dimensions. Both studies found that openness and conscientiousness were positively related to preferences for “science” books and that openness was positively related to “culture” or “arts and crafts” books. Two studies have examined the factor structure of TV and movie preferences (Hirschman, 1985; North & Hargreaves, 2007), but the structures were very different and associations between the preference factors and personality traits were not examined.

This past research provides some interesting clues to the nature of entertainment preferences. It appears that people seek out specific content because it can provide a sense of identity or companionship, that personality traits and demographic factors influence what people expect to derive from particular media, and that both personality and demographics are linked to preferences for certain media content. That work is useful and informative, but it is hard to know how robust the specific findings are because so few studies have actually examined the links between entertainment preferences and personality traits. Moreover, several of the studies relied on college-student samples, so it is unclear whether their results generalize to broader segments of society. Are the patterns of correlations between entertainment-media preferences and personality similar for young adults and older adults, for instance? Because most of the studies only examined preferences in one medium, we do not know whether preferences are driven by content or medium. That is, do people who enjoy sophisticated music (e.g., classical) also enjoy sophisticated movies (foreign), books (poetry), and TV shows (arts and entertainment), or are preferences in one medium independent of preferences in another?

The Present Research

The overarching goal of this research was to help situate personality theory and research in the flow of everyday life. Toward that end, we set out to explore the connections between personality traits and entertainment preferences. By examining such connections, this work will help gauge the impact and role that personality plays in everyday life. It is conceivable that the music people listen to, the movies they watch, and the books they choose to read have more to do with contextual variables and mood and less to do with personality. However, any observed relations between personality and preferences would suggest that personality plays a significant role in daily life and also shed light on how particular traits guide preferences and behaviors.

Another goal of the present research was to bridge theory and research on the psychology of entertainment preferences. Although previous research provides some insights into the types of rewards people derive from particular media, it is far from complete. If we are to develop a more comprehensive understanding of the reasons why entertainment is so important to people, how it affects people, and whether it affects certain people differently, we need to develop a firm foundation on which to develop and test hypotheses about the psychology of entertainment preferences.

As an initial step toward meeting those goals, this work set out to develop an empirically based classification of entertainment genres; the classification can serve as a framework within which future research can examine a broad array of individual differences in entertainment preferences. Specifically, our research aimed to examine (a) the latent structure of entertainment preferences, (b) the connections between the preference dimensions and other individual differences, and (c) the incremental validity of personality traits over and above demographic variables in the prediction of entertainment preferences.

Developing a Taxonomy of Entertainment Genres

Entertainment preferences can be measured at different levels of abstraction, ranging from a highly descriptive narrow subordinate level to a very broad superordinate level (Rentfrow & Gosling, 2003). For example, we could ask respondents to report their degree of liking for specific songs, books, films, and TV shows (e.g., respectively, Rehab, White Teeth, The Matrix, The West Wing), musicians, writers, directors, or actors (Amy Winehouse, Zadie Smith, the Wachowski Brothers, Martin Sheen), subgenres (blue-eyed soul, novel, cyberpunk, political drama), genres (soul, fiction, science fiction, drama), or general attributes (melancholic, satirical, exciting, intelligent).

Most previous research has focused on the subordinate level and used either specific examples, such as the titles of songs, films, or television programs (e.g., Rozin, Riklis, Margolis, 2004; Weaver, 1991), or genre labels (e.g., Han, 2003; Hirschman, 1985; Katz-Gerro, 1999; Kraaykamp & van Eijck, 2005; Rentfrow & Gosling, 2003). There is evidence that likeability ratings given to particular exemplars yield results similar to those given to their genre labels. For example, Rentfrow and McDonald (in press) found high convergence between preference ratings given to exemplary musical snippets and music-genre labels. Thus, it might seem as though exemplars and genres are equally useful. However, exemplars suffer from two important limitations. First, exemplars are specific, and as such they will be unfamiliar to many people. For instance, it is likely that more young people living in the U.S. will be familiar with The White Stripes (a Grammy Award-winning alternative rock band) than older people in the U.S. or most people living abroad. Second, the lifespan of exemplars can be short. A title exemplifying a specific genre (e.g., All in the Family for comedy television) may eventually become less representative of that genre and associated instead with a particular era (e.g., the 1970s).

Very few studies have examined preferences at the superordinate level of general attributes. One reason for this could be that such general attributes are highly ambiguous in the absence of exemplary stimuli. As a result, preference ratings for general attributes could be extremely subjective and unreliable.

In contrast to exemplary items and general attributes, genres are broad categories with which many people are familiar, have longer life spans than exemplars, and are more straightforward than attributes. Moreover, the set of genres for a given medium is considerably smaller than the set of potential exemplars, so an entertainment-preference measure based on genres should be more useful in diverse samples than one based on exemplars. Therefore, the genre level seems like the most sensible level of analysis for measuring entertainment preferences.

Overview of the studies

Using multiple methods, entertainment genre preferences were collected from three independent samples: a university-student sample, a community sample, and an Internet sample. The samples varied considerably in age, education, and ethnicity. These data enabled us to explore the structure of entertainment preferences and examine its generalizability across the samples. Goldberg's (1999) IPIP-AB5C measure of the Big-Five personality domains was administered to two of the three samples so that we could examine the associations between entertainment preferences and personality traits.

Methods

Participants

University sample

A total of 1,946 undergraduate psychology students at the University of Texas participated in the study. Of those who indicated, 1,270 (65%) were female and 672 (35%) were male; 65 (3%) were African American, 337 (17%) were Asian, 1,165 (60%) were Caucasian, 308 (16%) were Hispanic, and 64 (3%) were of other ethnicities. The average age of participants was 19 (SD = 2).

Community sample

In 1993, residents of the Eugene-Springfield community were recruited to participate in a series of mailed assessments over the next 5 to 10 years. Since then, a wide variety of measures of personality traits, social and political attitudes, and vocational and avocational interests have been administered, and informant ratings of personality have also been collected (for details, see Grucza & Goldberg, 2007).

A total of 736 participants returned completed surveys. Of those who indicated, 414 (56%) were female and 322 (44%) were male, 708 (98%) were Caucasian, 6 (1%) were Asian, and 6 (1%) were of other ethnicities. The average age was 60 (SD = 12). In terms of education, 9 (1%) did not complete high school, 108 (15%) completed high school or vocational school, 205 (28%) had some college education, 234 (32%) had a college degree and/or some post-college education, and 172 (24%) had a post-college degree.

Internet sample

A total of 545 individuals participated in the study. 369 (69%) were female and 164 (31%) were male (12 failed to indicate). The average age of the participants was 34 (SD = 10). In terms of education, 16 (3%) did not complete high school, 327 (61%) completed high school or vocational school, 150 (28%) had a college degree and/or some post-college education, and 40 (8%) had a post-college degree (and 12 failed to indicate).

Measures

Entertainment Preferences

As there were no established measures of entertainment preferences, we developed a genre-based entertainment preference measure. Drawing on the points raised earlier, two principles guided our decisions about which genres to study: We wanted lists of genres that were comprehensive in scope and lists composed of genres that are familiar to most people. To ensure comprehensiveness and familiarity we used a multi-step procedure. First, lists of genres were obtained from online stores that sell music, films, books, and television programs (e.g., Amazon.com, Barnesandnoble.com, iTunes, Tower.com), and from encyclopedic websites and books (e.g., IMDb.com, Infoplease.com, Videohound). Second, three judges independently generated lists of all the music, film, book, and TV genres that they could think of. The judges' lists reflected the genres with which people are likely to be most familiar and the reference materials were used as a more comprehensive source of potentially important but less familiar additional genres. Genres that appeared both on the judges' lists and the reference materials were selected as an initial set; genres that appeared in the reference materials but not on the judges' lists were added to the initial set only if all three judges were familiar with them; and those few genres that appeared on all three judges' lists but not in the reference materials were added to the initial set. The initial set of genres was then re-evaluated by all three judges and redundant genre labels were removed. This procedure resulted in a set of 108 genres (22 music, 34 book and magazine, 18 film, and 34 television) which comprise our Entertainment Preference Measure (EPM).

In all three samples, preferences for entertainment genres were assessed with the EPM. Respondents were asked to indicate their degree of preference for each of the 108 genres in the four media domains using a seven-point rating scale, with endpoints at 1 (Dislike strongly) and 7 (Like strongly).

Personality

Personality traits were assessed using Goldberg's (1999) IPIP-AB5C Inventory. The IPIP-AB5C measures each of the Big-Five personality domains (Extraversion, Agreeableness, Conscientiousness, Emotional Stability, and Intellect/Imagination) as well as 45 subordinate facets (nine facets for each domain). Specifically, each domain is measured with one subscale that is a direct marker of that broad factor and eight subscales that are blends of that factor and one of the other four Big-Five domains.

In the present studies, respondents indicated the extent to which each item was characteristic of themselves using a 5-point rating scale with endpoints at 1 (Very uncharacteristic of me) to 5 (Very characteristic of me). The reliabilities of the subscales ranged from .70 to .85 (M = .79) in the community sample and .66 to .89 (M = .82) in the Internet sample.

Intelligence

An estimate of the intellectual ability of the participants in the community sample was assessed by the Reasoning test (Factor B) from Cattell's 16 Personality Factor Questionnaire (16PF). This maximum-performance test consists of 15 verbal, numerical, and logical ability items. It has been shown to be internally reliable (coefficient alpha = .76; Goldberg, 1999) and to correlate highly with other measures of intelligence (Rieke & Conn, 1994). Scores on this test thus enabled us to examine the extent to which particular entertainment preferences are linked to intellectual ability.

Procedure

University sample

In the Fall of 2003, students registered for introductory psychology at the University of Texas at Austin were invited to participate in a mass testing session over the Internet in exchange for partial fulfillment of a course requirement. All who volunteered and provided consent were directed to a website where they could complete several psychology surveys, including the EPM. Due to time restrictions, participants did not complete a personality measure.

Community sample

In the Spring of 2003, a questionnaire packet that included the EPM was mailed out to members of the community sample. Participants received $25 for completing the survey. Between 1994 and 1996, participants completed the full 485-item IPIP-AB5C inventory.

Complete data for both the IPIP-AB5C and the EPM were available for 449 participants in the community sample.

Internet sample

In the Spring of 2007, advertisements were placed in several locations on the world wide web (e.g., Craigslist.com) inviting people to participate in an Internet-based study of personality, attitudes, and preferences. Approximately 1,600 individuals responded to the advertisement and provided their email addresses. They were then contacted and told that participation entailed completing several surveys, which included the EPM and IPIP-AB5C, on separate occasions. Those who agreed to participate were directed to a web-page where they could begin completing the first survey. After completing each survey, they were informed that they would receive an e-mail message within a few days with a hyperlink that would direct them to the next survey. Participants who completed every survey received a $25 gift certificate to Amazon.com.

Over a span of approximately two weeks, participants in the Internet sample completed the full version of the EPM and an abridged version of the IPIP-AB5C that included 356 items. The abridged IPIP-AB5C was intended to increase completion rates. Omitted from the abridged inventory were one facet of Extraversion (sociability), two facets of Agreeableness (sympathy, nurturance), three facets of Conscientiousness (rationality, perfectionism, orderliness), three facets of Emotional Stability (moderation, cool-headedness, toughness), and two facets of Intellect (imagination, depth).

Complete data for both the EPM and IPIP-AB5C were available for 503 participants in the Internet sample.

Results

Structural Analyses of the EPM

The primary aim of our investigation was to discover the structure underlying entertainment genre preferences. We sought broad dimensions that would not be restricted to one medium, so factor analyses were performed on the preference ratings of the 3,227 participants from the three samples combined, using all 108 music, film, book, and television genre labels as variables.

Multiple criteria were used to decide how many factors to retain: parallel analyses of Monte Carlo simulations, the scree test, replicability across factor-extraction methods, factor interpretability, and factor congruence across subsamples. Principal-components analysis (PCA) with varimax rotation yielded a large first factor that accounted for 15% of the variance, reflecting individual differences in general preferences for entertainment. Parallel analysis of random data suggested that the first 17 eigenvalues were greater than chance. Examination of the scree plot suggested an “elbow” at roughly 6 factors. Successive PCAs with varimax rotation were then performed for one-factor through seven-factor solutions. In the six- and seven-factor solutions, comparatively small, low-saturation factors composed of several items with large secondary loadings were added to the first five factors. Altogether these criteria suggested that we retain no more than five broad entertainment-preference factors.

We next examined the hierarchical structure of the one- through five-factor solutions using the procedure suggested by Goldberg (2006). First, a single factor was specified in a PCA, then two through five orthogonally rotated factors, in which factor scores for each solution were saved. Next, correlations between factor scores at adjacent levels were computed. The resulting hierarchical structure is displayed in Figure 1.

Figure 1.

Varimax-rotated principle components derived from 108 Music, Film, Book, and Television Genres.

There are several noteworthy findings that can be seen in that figure. The factors in the two-factor solution resemble the well-documented “Highbrow” and “Lowbrow” entertainment dimensions. The genres with high loadings on the Highbrow factor were mainly within the domains of music (e.g., classical, jazz, opera) and books (e.g., art, history, science). This factor remained virtually unchanged through the three- and four-factor solutions. The genres with high loadings on the Lowbrow factor were mainly within the domains of film (e.g., action, horror, suspense) and television (e.g., music television, comedy, reality television). This factor split in the three-factor solution into “Communal” and “Rebellious” subfactors. The Communal factor comprised mainly genres within the television domain (e.g., music television, reality television, made-for-TV movies). This factor remained consistent through the four- and five-factor solutions. The Rebellious factor comprised genres from the music (e.g., heavy metal, punk, rock) and film domains (e.g., action, science fiction, war). This factor split apart in the four-factor solution into “Dark” and “Thrilling” subfactors. The same music genres from the Rebellious factor load on the Dark factor (e.g., heavy metal, punk, rock), but the film genres changed (i.e., from action and science fiction to erotic, cult, and horror). The Thrilling factor comprised genres in the film (e.g., action, science fiction, war), book (e.g., action, adventure, espionage), and television domains (e.g., action, mystery, science fiction). Both the Dark and Thrilling factors remained the same down the hierarchy. In the five-factor solution, the Highbrow factor split into “Aesthetic” and “Cerebral” subfactors. The Aesthetic factor comprised genres from all of the entertainment media (e.g., classical music, foreign films, poetry) except television, and the Cerebral factor comprised mainly genres from the television (e.g., news, business and economy) and book domains (e.g., current events, health, computing).

Although the factors depicted in Figure 1 are clear and interpretable, some of the factors (e.g., Communal, Cerebral, Thrilling) might be the result of differences in gender and/or age. To determine the extent to which the factor structure was driven by sex and age, PCAs with varimax rotation of residualized genre items, in which sex and age were statistically removed, were performed for solutions including 1 through 5 factors. The factor structures derived from the residualized items were virtually identical to those derived from the original genre items (mean factor congruence = .90). Additionally, analyses of the correlations between the corresponding factor scores derived from the original and the residualized genre items revealed high convergence across the factors (mean r = .86). These results indicate that even though there are significant sex and age differences in preferences for practically every entertainment genre, the factors underlying entertainment preferences are generally invariant to gender and age effects.

As a final step in determining how many factors to retain, we examined the replicability of the factor solutions across the three subsamples. Specifically, PCAs with varimax rotation for 1 to 5 factor solutions were performed on the 108 genre items separately for the university, community, and Internet samples. Targeted rotation was used to assess congruence between the factor structures derived from the subsamples and those derived from the total sample. The congruence coefficients, which are displayed in Table 1, were generally high. As can be seen in the sixth column of Table 1, the one-factor solution displayed the lowest degree of overall congruence, with a mean coefficient of .86, whereas the three- and five-factor solutions displayed the highest congruence, with mean coefficients of .95. Although there is no evidence to suggest that there are more than five broad entertainment-preference factors, the comparisons between the original and residualized genre items and the cross-sample factor congruence coefficients provide evidence to consider both three and five factor solutions. However, because five factors provide more information about entertainment preferences than three factors, findings based on the five-factor model are reported henceforth.

Table 1.

Factor and Total Congruence Coefficients for One through Five Factor Solutions in Three Independent Samples.

	Factor
	I	II	III	IV	V	Mean
One Factor
US	.93					.93
CS	.66					.66
IS	.88					.88
Mean	.86					.86
Two Factors
US	.92	.81				.88
CS	.89	.92				.91
IS	.96	.94				.95
Mean	.93	.90				.92
Three Factors
US	.97	.99	.89			.97
CS	.86	.92	.95			.92
IS	.95	.95	.94			.95
Mean	.94	.97	.93			.95
Four Factors
US	.99	.98	.70	.93		.95
CS	.96	.95	.91	.96		.95
IS	.97	.94	.84	.91		.93
Mean	.98	.96	.83	.94		.94
Five Factors
US	.99	.99	.80	.94	.99	.97
CS	.96	.89	.90	.94	.80	.91
IS	.97	.92	.91	.97	.70	.93
Mean	.98	.96	.88	.95	.91	.95

Note. US = University Sample (N = 1,946), IS = Internet Sample (N = 545), CS = Community Sample (N = 736). Factor congruence was computed using targeted rotation with factor structures from the combined sample as the target matrix. Mean congruence coefficients were computed using Fisher's r-to-z transformation.

Characterizing the entertainment-preference dimensions

The orthogonal five-factor structure derived from the combined sample is provided in Table 2. The factors have a remarkably simple structure, with only nine items (8%) loading .40 or greater on more than one factor; moreover, the factors are quite broad, with each factor including genres from multiple entertainment-media domains. Interestingly, virtually all of the factor loadings on each factor are positive (except gospel, which loads -.40 on the Dark factor), suggesting that the factors are unipolar. All five factors have discernable cores and are clearly interpretable.

Table 2.

Five Varimax-Rotated Principle Components Derived from 108 Music, Film, Book, and Television Genres.

Genre	Medium	Communal	Aesthetic	Dark	Thrilling	Cerebral
Romance	Film	.71	.15	−.06	−.08	−.05
Romance	Book	.64	.05	.02	−.01	−.02
Entertainment	Book	.61	−.05	.32	.02	.03
Day time talk shows	TV	.59	−.22	.09	−.10	.22
Made for TV Movies	TV	.58	−.02	−.12	.15	.13
Movie coverage	TV	.55	.01	.21	.10	.14
Music Television	TV	.54	−.11	.35	−.03	.05
Family	Film	.54	.23	.00	.12	−.07
Soap Operas	TV	.53	−.11	.15	−.04	.04
Dramas	TV	.53	.05	−.09	.12	.08
Pop	Music	.53	−.09	.09	.02	.01
Reality Television	TV	.53	−.31	.22	−.04	.15
Children	TV	.51	.14	.09	.11	.07
Show tunes	Music	.50	.17	.05	.10	−.05
Comedies	TV	.47	−.10	.27	.16	−.03
Game shows	TV	.47	−.21	.19	.12	.14
Drama	Film	.46	.27	.06	.07	.00
Religious	Music	.42	.12	−.37	.12	.02
Medical shows	TV	.41	−.02	.02	.14	.31
Variety shows	TV	.41	.04	−.01	.24	.24
Real-life Mysteries	TV	.41	.00	.04	.30	.23
Soul/R&B	Music	.39	.08	.19	.06	.07
Comedy	Film	.39	.01	.21	.17	−.06
Country	Music	.36	−.09	−.28	.12	.07
Religion	Book	.35	.25	−.16	.04	.12
Cooking	Book	.34	.25	−.10	.02	.25
Animation	Film	.33	.19	.23	.30	−.10
Lawyer shows	TV	.30	−.01	.04	.22	.27
Classical	Music	−.12	.66	−.07	.05	.08
Arts & humanities	TV	.11	.63	.03	.00	.35
Art	Book	.11	.62	.17	−.09	.21
Opera	Music	.01	.60	−.06	.00	.10
Foreign	Film	−.07	.59	.30	−.05	.09
Blues	Music	−.04	.57	−.07	.20	.10
World	Music	−.07	.56	.11	−.06	.07
Jazz	Music	−.03	.54	−.03	.12	.07
Folk	Music	−.14	.54	−.28	.12	.12
Poetry	Book	.24	.53	.20	−.09	.08
Classics	Film	.23	.51	−.16	.20	.07
Independent	Film	.02	.51	.35	−.01	.09
Fiction & Literature	Book	.15	.49	.07	.18	−.08
Philosophy	Book	−.02	.49	.28	−.05	.32
Bluegrass	Music	−.06	.47	−.23	.19	.13
Architecture	Book	−.02	.45	.08	.02	.37
Musicals	Film	.43	.45	−.21	.02	−.02
Africana	Book	.07	.44	.11	.10	.22
Academic	Book	−.07	.44	.06	.04	.42
History	Book	−.12	.43	−.12	.21	.40
Science & Nature	Book	−.16	.42	.08	.25	.41
Photography	Book	.26	.38	.24	−.05	.19
Psychology	Book	.22	.38	.31	−.19	.23
Mind & Spirit	Book	.33	.36	.18	−.14	.26
New Age	Music	.15	.35	.22	.04	.07
Oldies	Music	.22	.33	−.28	.19	.01
Punk	Music	.23	.04	.64	.00	−.10
Horror	Film	.25	−.12	.61	.25	−.09
Heavy Metal	Music	.00	−.01	.58	.18	.00
Horror	Book	.22	.00	.57	.32	−.07
Horror	TV	.25	−.12	.56	.33	−.03
Rap & hip-hop	Music	.39	−.18	.54	−.09	−.06
Cult	Film	−.12	.30	.54	.12	.05
Rock	Music	.12	.09	.51	.16	−.07
Alternative	Music	.17	.15	.50	.02	−.11
Late night talk shows	TV	.29	−.11	.50	.08	.11
Erotic	Film	−.20	−.03	.50	.17	.20
Erotica	Book	−.16	−.01	.49	.17	.20
Funk	Music	.19	.21	.46	.07	−.02
Gospel	Music	.30	.25	−.40	.12	.05
Sketch Comedy	TV	.15	.12	.37	.26	.07
Animation	TV	.26	.13	.30	.30	−.01
Dance & Electronica	Music	.20	.17	.26	.02	.07
Action	Book	.06	.12	.10	.66	.13
Action adventure	TV	.22	−.13	.13	.66	.12
Thrillers & Espionage	Book	.16	.10	.17	.64	.04
Action	Film	.18	−.13	.24	.62	.00
Adventure	Book	.08	.23	.03	.61	.09
Spy Shows	TV	.19	.04	.05	.58	.20
Science Fiction	Film	−.11	.25	.36	.57	−.01
Sci-Fi & Fantasy	TV	−.11	.22	.36	.57	.06
Western	Film	−.07	.12	−.19	.54	.25
Sci-Fi & Fantasy	Book	−.11	.29	.39	.52	−.04
Western	TV	−.02	.08	−.29	.52	.31
Mystery	TV	.40	.09	.00	.51	.06
War	Film	−.02	−.07	.24	.51	.17
Suspense	Film	.34	.02	.21	.50	−.03
Mystery	Book	.34	.24	.01	.45	−.06
Cop shows	TV	.20	−.19	−.08	.41	.34
Business & Economy	TV	−.03	.11	.08	.13	.64
Business	Book	−.02	.09	.12	.13	.58
Health	TV	.36	.14	−.05	−.07	.56
News/Current Events	TV	.13	.10	−.13	.14	.53
News/Current Events	Book	.09	.20	−.04	.09	.52
Educational	TV	.06	.41	−.09	.09	.50
Reference	Book	−.05	.41	−.05	.06	.48
Medical	Book	.18	.21	.07	−.02	.47
Computers	Book	−.15	.04	.26	.29	.45
Documentary	Film	−.08	.42	−.04	.06	.44
Science	TV	−.20	.31	.16	.32	.44
Health	Book	.36	.21	−.03	−.10	.43
Home improvement	TV	.31	.09	−.13	.08	.43
Biographies	Book	.11	.41	−.13	.00	.42
Society & Culture	TV	.22	.38	.13	−.05	.40
Public Access	TV	.13	.26	−.05	.03	.40
Travel	Book	.19	.32	−.11	.06	.38
Nonfiction	Book	.06	.36	−.01	.09	.37
Home & Garden	Book	.28	.33	−.30	.00	.37
Recreation & Sports	TV	.14	−.22	.09	.28	.35
Sports	Book	.11	−.20	.17	.29	.33

Note. Loadings ≥ |.40| are in bold typeface. N = 3,227

The first factor, which we labeled Communal, is composed of 14 television genres (e.g., daytime talk shows, made-for-TV movies), five film genres (romance, family), five music genres (pop, show tunes), and four genres from the book and magazine domain (romance, religion, cooking). On the surface, most of the genres with high loadings on this factor appear lighthearted, uncomplicated, and popular. Closer inspection revealed that nearly all the genres on this factor focus on people and relationships.

The second factor, labeled Aesthetic, is composed of 12 book genres (e.g., art, poetry), nine music genres (classical, blues), four film genres (foreign, classics), and one television genre (arts and humanities). The majority of genres with high loadings on this factor may be described as creative, abstract, cultured, dense, and demanding.

The third factor, labeled Dark, is composed of eight music genres (e.g., punk, heavy metal), four television genres (horror, late night television), three film genres (horror, cult), and two genres from the book and magazine domain (horror, erotic). The genres on this factor are generally characterized by intensity, edginess, and hedonism.

The fourth factor, labeled Thrilling, is composed of six television genres (e.g., action adventure, spy shows), five book genres (action, thrillers and espionage), and five film genres (action, science fiction). None of the music genres have their highest loadings on this factor. The themes common to this factor include action, adventure, suspense, and fantasy.

The fifth factor, labeled Cerebral, is composed of 11 genres from the book and magazine domain (e.g., business, news and current events), nine television genres (business and economy, health), and one film genre (documentary). None of the music genres have their highest loadings on this factor. This is clearly an information-oriented factor that can be characterized by factual information about persons, places, or things.

Analyses of Demographic and Personality Correlates

Having identified a robust factor structure, our next aim was to examine the correlates of these five entertainment-preference factors. Demographic information was available for all three samples, so we examined the correlations between the five entertainment factors and age, gender, ethnicity, and level of education in each sample separately. We also examined the correlations between each of the entertainment-preference factors and the IPIP-AB5C domain and facet scores for the community and Internet samples. For all the correlational analyses, the entertainment-preference factors were based on the orthogonally rotated factor scores.

Demographic correlates

As can be seen in Table 3, the patterns of correlations between the entertainment factors and the demographic variables were fairly consistent across the three samples. Overall, gender was the demographic variable most strongly related to these entertainment factors (mean absolute r = .27), followed by intelligence, education, age, and ethnicity (mean absolute r = .16, .15, .13, .04, respectively). The first three columns in Table 3 indicate that females and individuals with low levels of education tend to have high scores on the Communal entertainment-preference factor. Preferences for communal entertainment are also negatively related to abstract reasoning ability. Columns 4-6 in Table 3 show that individuals with high levels of education, abstract reasoning ability, and females tend to have high scores on the Aesthetic factor. The correlations in columns 7-9 indicate that young people, males, and Hispanics tend to have high scores on the Dark entertainment factor. Preferences for dark entertainment are also positively related to education and reasoning ability. As can be seen in columns 10-12, preferences for the Thrilling factor tend to be higher among men than women, and among individuals with less education. Finally, the patterns of correlations in columns 13-15 suggest that older people and males tend to prefer the Cerebral entertainment-preference genres.

Table 3.

Demographic Correlates of Five Entertainment-Preference Factors in Three Independent Samples.

	Communal			Aesthetic			Dark			Thrilling			Cerebral
	US	IS	CS	US	IS	CS	US	IS	CS	US	IS	CS	US	IS	CS
Age	−.13	.05	.03	.05	.08	.08	−.04	−.38	−.39	.02	.10	−.11	.16	.11	.14
Sex
Female	.59	.41	.53	.16	.05	.22	−.31	−.16	−.22	−.36	−.12	−.36	−.19	−.12	−.16
Ethnicity
African American	.11	--	--	−.02	--	--	−.10	--	--	.02	--	--	−.01	--	--
Asian	−.06	--	.05	−.02	--	.00	.07	--	.03	−.01	--	.01	.04	--	.02
Caucasian	−.04	--	−.08	−.02	--	−.01	−.11	--	−.03	.01	--	−.04	−.05	--	−.02
Hispanic	.08	--	.02	.04	--	.01	.10	--	.05	−.01	--	.01	.02	--	.03
Other	−.04	--	−.01	.06	--	.03	.05	--	−.01	−.01	--	−.04	.03	--	−.02
Education	--	−.13	−.29	--	.20	.30	--	.00	.12	--	−.10	−.17	--	.13	.07
Intelligencea	--	--	−.27	--	--	.29	--	--	.16	--	--	−.03	--	--	−.03

Note. US = University Sample (N = 1,939), IS = Internet Sample (N = 503), CS = Community Sample (N = 725). -- = missing data.

^aAssessed by the Reasoning factor of the 16PF (N = 599). Correlations ≥ |.10| are significant at p < .05 and are in bold typeface.

Personality correlates

The correlations presented in Table 4 reveal a fascinating pattern of links between entertainment preferences and personality traits. As can be seen in columns one and two of Table 4, in both samples the Communal factor was positively related to tenderness, warmth, understanding, morality, empathy, pleasantness, cooperation (all facets of Agreeableness), reflection (Intellect), friendliness (Extraversion), and dutifulness (Conscientiousness), and negatively related to provocativeness (Extraversion), imperturbability (Emotional Stability), and creativity (Intellect). Thus, it appears as though the psychological characteristics most central to individuals who prefer the Communal entertainment factor are rather similar to the defining characteristics of that factor: pleasant, empathic, lighthearted, unadventurous, uncomplicated, and relationship-oriented.

Table 4.

Personality Correlates of Five Entertainment-Preference Factors in Two Independent Samples.

	Communal		Aesthetic		Dark		Thrilling		Cerebral
	IS	CS	IS	CS	IS	CS	IS	CS	IS	CS
Extraversion	.15	.04	.05	.09	.06	.16	.05	−.09	.18	.07
Gregariousness	.12	.09	.01	.08	.02	.11	.00	−.13	.13	.10
Friendliness	.25	.25	−.01	.06	−.06	−.02	.01	−.07	.08	.03
Provocativeness	−.13	−.22	.07	−.02	.20	.26	.02	−.04	.18	.09
Assertiveness	.14	−.06	.06	−.03	.00	.05	.08	−.05	.28	.08
Self-disclosure	.17	.09	.08	.16	.25	.21	.01	−.04	.06	−.05
Poise	.14	.04	.06	.16	−.05	.02	.08	−.02	.14	.10
Talkativeness	.04	−.03	−.03	−.01	.06	.15	.01	−.09	.07	.00
Leadership	.09	−.10	.10	.13	.01	.09	.08	−.15	.17	.07
Sociability	--	.19	--	−.17	--	.07	--	.12	--	.10
Agreeableness	.34	.41	.13	.23	−.14	−.20	.03	−.06	.03	−.04
Warmth	.36	.34	.15	.28	−.05	−.04	.03	−.11	.11	.05
Cooperation	.14	.22	.09	.16	−.23	−.28	−.05	−.12	−.10	−.12
Understanding	.28	.33	.17	.31	−.10	−.10	.03	−.07	.03	.03
Morality	.26	.27	.05	−.02	−.21	−.35	.08	.07	−.04	−.03
Sympathy	--	.40	--	.32	--	−.12	--	−.08	--	−.06
Pleasantness	.21	.27	.04	.12	−.12	−.08	.09	.02	.03	−.01
Tenderness	.49	.49	.07	.15	−.05	−.13	−.04	−.08	.04	−.10
Empathy	.26	.25	.20	.31	−.02	−.04	.02	−.06	.09	.01
Nurturance	--	.49	--	.03	--	−.21	--	.03	--	−.02
Conscientiousness	.14	.01	.09	−.09	−.17	−.16	.05	−.01	.16	.09
Efficiency	.12	.01	.03	−.09	−.13	−.10	.07	.01	.19	.04
Cautiousness	−.12	−.07	.00	−.15	−.22	−.17	−.05	−.05	.05	.13
Dutifulness	.28	.22	.04	−.07	−.20	−.20	.03	.04	.06	.06
Rationality	--	−.10	--	−.34	--	−.11	--	.18	--	.12
Conscientious	.10	.05	.06	−.13	−.17	−.14	.07	.02	.09	.04
Purposefulness	.12	−.01	.10	−.05	−.13	−.10	.10	.04	.16	.07
Perfectionism	--	.05	--	−.27	--	−.12	--	.06	--	.17
Organization	.17	−.06	.19	.06	−.03	−.07	.02	−.04	.19	.12
Orderliness	--	.21	--	−.24	--	−.25	--	.08	--	.08
Emotional Stability	−.11	−.12	.04	.05	−.01	.04	.05	.05	.12	.08
Happiness	.07	−.07	.04	.05	−.08	.03	.07	−.02	.14	.07
Impulse control	−.03	−.09	.02	−.01	−.09	−.07	.04	.07	.02	.03
Calmness	.01	.09	.10	.15	−.01	−.05	.00	−.04	.08	.01
Imperturbability	−.33	−.31	−.07	−.08	.08	.08	.04	.05	.08	.14
Moderation	--	−.05	--	.03	--	−.12	--	−.07	--	.06
Cool-headedness	--	.06	--	−.22	--	−.13	--	.08	--	.10
Stability	−.09	−.07	.07	.07	.01	.03	.04	.03	.14	.03
Toughness	--	−.21	--	.13	--	.13	--	.05	--	.03
Tranquility	.07	−.06	.00	.03	.03	.06	.08	.08	.10	.02
Intellect	.07	−.24	.38	.38	.13	.20	.09	−.17	.19	.15
Ingenuity	.07	−.24	.29	.19	.13	.18	.10	−.12	.18	.17
Introspection	−.05	−.24	.24	.25	.14	.08	−.03	−.08	.08	.01
Reflection	.36	.27	.36	.54	.07	.03	.05	−.16	.11	.07
Creativity	−.14	−.31	.36	.36	.18	.27	.05	−.13	.17	.17
Competence	.13	−.21	.19	.14	.06	.12	.12	−.03	.18	.08
Imagination	--	−.04	--	.58	--	.23	--	−.16	--	.09
Quickness	.06	−.24	.28	.33	.09	.19	.12	−.09	.16	.09
Depth	--	−.08	--	.17	--	.09	--	−.08	--	.02
Intellect	−.02	−.21	.37	.44	.08	.19	.03	−.25	.15	.13
Column vector correlations	.71		.86		.75		−.03		.47

Note. CS = Community Sample (N = 449), IS = Internet Sample (N = 503). -- = missing data. Column-vector correlations were computed using Fisher's r-to-z transformation. Correlations ≥ |.10| are significant at p < .05 and are in bold typeface.

The personality characteristics most consistently related to preferences for the Aesthetic factor come from the Intellect and Agreeableness domains. This factor was positively related to reflection, intellect, creativity, quickness, introspection, ingenuity, competence (all facets of Intellect), empathy, understanding, and warmth (facets of Agreeableness), as well as leadership (Extraversion) and calmness (Emotional Stability). In other words, individuals who enjoy the Aesthetic entertainment factor, which may be regarded as abstract, dense, and demanding, tend to be creative, calm, introspective, and in touch with their emotions.

The personality characteristics most common to the Dark entertainment-preference factor include low Conscientiousness, low Agreeableness, high Extraversion, and high Intellect. Specifically, this factor was negatively related to dutifulness, cautiousness, conscientiousness, purposefulness, efficiency (all facets of Conscientiousness), morality, cooperation, and understanding (Agreeableness), and positively related to creativity, ingenuity (Intellect facets), provocativeness, and self-disclosure (Extraversion). Thus, it appears that individuals with preferences for the Dark entertainment factor may generally see themselves as defiant, reckless, and immodest. Not surprisingly, those descriptors also reflect some of the key characteristics of the media genres on that factor.

The personality characteristics most common to the Cerebral entertainment-preference factor include high Intellect and Extraversion. Specifically, Cerebral was positively related to ingenuity, creativity, intellect (all facets of Intellect), and to poise and gregariousness (Extraversion). Scores on this factor were also positively related to organization (Conscientiousness facet) and negatively related to cooperation (Agreeableness). In other words, individuals who prefer the Cerebral factor tend to see themselves as enterprising, innovative, intellectual, self-assured, and detail oriented.

Finally, there was no consistent pattern of correlations between the personality facets and Thrilling preferences. Indeed, the patterns of correlations for the Internet and community samples were often in the opposite direction. For example, ingenuity, competence, and quickness (all facets of Intellect) were positively related to the Thrilling factor in the Internet sample, but negatively related in the community sample.

Generalizability of the personality correlates across samples

As one would expect for such a broad array of personality traits, the sizes of the correlations varied greatly across traits and entertainment dimensions. Therefore, to examine the generalizability of the correlations across the two samples, column-vector correlations were computed for each of the dimensions. Specifically, the absolute values of the correlation coefficients in Table 4 were transformed using Fisher's r-to-z formula and then correlations between the two columns of transformed correlations were computed. As shown in the bottom row of Table 4, the pattern of correlations was similar for four of the entertainment factors (Communal, Aesthetic, Dark, and Cerebral; mean = .62). However, the column-vector correlation for the Thrilling factor was very low (−.03), suggesting that the personality characteristics associated with that factor may not generalize to other samples.

Generalizability of the personality correlates across raters

We also examined whether self-reported entertainment preferences were related to informant reports of personality. Informant reports of participants' personalities and self-reported entertainment preferences were available for 598 participants in the community sample. Self-report data from participants and informant report data from knowledgeable informants were available for John and Srivastava's (1999) Big Five Inventory and Saucier's (1994) Big Five Mini-markers. As can be seen in Table 5, the patterns of correlations for the entertainment-preference dimensions were very similar across the self- and informant-reports, which were very similar to the results using the AB5C (see Table 4). These results provide cross-method convergence and further validation for the correlations between entertainment preferences and personality traits.

Table 5.

Correlations between Five Entertainment-Preference Factors and Self- and Peer-Ratings of Personality.

	Communal		Aesthetic		Dark		Thrilling		Cerebral
	Self	Peer	Self	Peer	Self	Peer	Self	Peer	Self	Peer
Big Five Inventory
Extraversion	.15	.16	.04	.06	.00	−.06	−.08	−.10	.09	.03
Agreeableness	.24	.15	.08	.05	−.18	−.09	−.10	−.10	.02	.04
Conscientiousness	.08	.07	−.09	−.07	−.13	−.10	.01	.01	.05	.01
Emotional Stability	−.06	−.11	.04	−.02	−.04	−.03	.02	.07	.06	.08
Openness	−.14	−.13	.50	.46	.19	.16	−.16	−.22	.11	−.01
Big Five Mini-markers
Extraversion	.11	.09	.06	.08	−.02	−.04	−.06	−.08	.08	.03
Agreeableness	.29	.20	.17	.15	−.12	−.08	−.16	−.15	.00	.08
Conscientiousness	.06	.08	−.09	−.05	−.13	−.11	.01	−.02	.05	.01
Emotional Stability	−.07	−.10	.17	.06	−.11	−.03	−.11	.02	.00	.09
Intellect	−.19	−.17	.38	.43	.16	.18	−.18	−.21	.12	.00

Note. Self- and peer-ratings of personality were obtained from members of the Community Sample (N = 598). Correlations > |.10| are significant at p < .05 and are in bold typeface.

Multiple correlations

We next examined the incremental validity of the personality traits as predictors of entertainment preferences. To do so, in the Internet and community samples each of the orthogonally rotated entertainment-preference factor scores were regressed onto three demographic indices (age, gender, and level of education) at Step 1, followed by the five IPIP-AB5C scale scores representing each of the Big Five domains at Step 2. As can be seen in Table 6, at Step 1, the demographic indices predicted the five entertainment factors with multiple correlations ranging from .22 (Cerebral in the Internet sample) to .56 (Communal in the community sample). At Step 2, the inclusion of the five personality factors to the regression equation raised the multiple correlations substantially, now ranging from .27 (Thrilling in the Internet sample) to .64 (Communal in the community sample). The F ratios for the Δ Rs indicated that the addition of personality traits to the demographic indices led to significant increases in the proportion of variance accounted for in every entertainment-preference factor except for Thrilling in the Internet sample. That is, although age, gender, and social class each significantly affect entertainment preferences, personality plays a significant role.

Table 6.

Multiple Correlations of Five Entertainment-Preference Factors with Demographic and Personality Variables.

	Communal		Aesthetic		Dark		Thrilling		Cerebral
	IS	CS	IS	CS	IS	CS	IS	CS	IS	CS
Step 1: Demographicsa	.42	.56	.23	.42	.39	.42	.25	.45	.22	.24
Step 2: Personalityb	.55	.64	.49	.60	.51	.50	.27	.49	.33	.30
Δ R F-ratio	15.47	15.04	20.31	25.13	12.14	8.78	0.91	4.31	5.67	2.81

Note. CS = Community Sample (N = 449), IS = Internet Sample (N = 503).

^aDemographic variables included age, sex, and education.

^bPersonality variables included aggregate scores of each Big Five dimension. F statistics in bold typeface are significant at p < .05.

Discussion

The first aim of this work was to explore the structure of entertainment preferences. The results obtained from multiple samples, methods, and geographic regions converged to suggest that entertainment preferences can be conceptualized in terms of five independent factors. Each factor comprises genres that are similar in content and came from different media domains. These findings indicate that entertainment preferences are more a function of substance than style. Thus, individuals prefer genres that share similar content irrespective of the medium through which it is conveyed.

This research also aimed to provide an initial examination of the relations between these basic factors and the basic dimensions of personality. The results indicated that the preference factors were related moderately to age and ethnicity, and related strongly to sex and level of education. The preference factors were also uniquely related to a wide variety of personality traits in each of the Big-Five domains. And, in two independent samples the patterns of correlations between the preference dimensions and personality traits were highly similar for four of the five factors. Furthermore, the results from multiple regression analyses indicated that the Big-Five personality domains account for significant proportions of unique variance in the preference factors even when demographics were held constant. Thus, entertainment preferences are not determined exclusively by age, gender, or education, but also by psychological dispositions. Overall, the findings provide a solid foundation on which to develop and test hypotheses about the causes and consequences of entertainment preferences.

The connections between personality and the entertainment-preference dimensions suggest that people seek out entertainment that reflect and reinforce aspects of their personalities. This interpretation is consistent with the view that people are not passive recipients of information, as the media effects paradigm implies. Researchers concerned with entertainment media, and in particular the associations between media exposure and behaviors, should consider media consumption as less of a passive process and more of an active one. Indeed, if we are to develop an understanding of the ways that particular media content affect individuals, it is necessary that we also consider whether people like the entertainment genre to begin with and what they derive from listening, watching, or reading it.

Limitations and Future Directions

One of the limitations of the present work is that all the data analyzed were based on self-reports. This is potentially problematic because people might report preferences for genres that they consider socially desirable. Thus, it is not clear whether the results would be similar if entertainment preferences were assessed using behaviorally revealed preferences (e.g., CD, book, and DVD collections, or purchasing habits at amazon.com or iTunes). However, a limitation with that method is that people do not listen to, watch, or read every CD, video or book that they buy, nor do they do not necessarily enjoy them (purchasing behavior is undoubtedly imperfectly correlated with enjoyment). Nevertheless, Rentfrow and Gosling (2003) found that the factor structure of music preferences assessed by individuals' online music collections was practically identical to the structure observed from self-reported preferences. So it seems reasonable to expect behaviorally revealed preferences for other media to yield results similar to those reported here.

Our preliminary research suggested that genres are the optimum level of analysis for measuring entertainment preferences and the current results indicate that there are broad dimensions of entertainment-genre preferences, which are linked to personality traits. However, genres are quite abstract categories and it is conceivable that preferences for entertainment subgenres could provide further insight into the nature of entertainment preferences. For example, the television genre ‘comedies’ includes a wide range of different forms of comedy—from slapstick and family humor to insult comedies and stand-up performances—and not everyone who enjoys comedy television will enjoy all forms of it. Thus, future research should explore the structure and correlates of preferences for entertainment subgenres.

The present work does not provide any direct information about the temporal stability of the entertainment preference dimensions, so we do not know the frequency or extent to which such preferences change over time. Research on music-genre preferences suggests that preferences are stable over several weeks and even up to three years (Delsing et al., 2008; Rentfrow & Gosling, 2003), so it seems reasonable to suppose that preferences for entertainment genres are at least as stable.

Another limitation with the present research is that the participants were predominantly Caucasian and middle class. Consequently, it is not clear whether the same preference dimensions and correlates would emerge in a more heterogeneous sample. Future research with more diverse samples would shed light on the generalizability of the current results.

The present work examined connections between entertainment preferences, demographic variables, and a wide range of personality traits, and therefore it would now be useful to also explore associations with other types of individual differences. Some of the preference factors, particularly Aesthetic and Cerebral, are characterized by genres that could be considered complex, demanding, and erudite. It seems possible that preferences for such genres should be related to intellectual ability, cognitive complexity, and need for cognition. Moreover, research suggests that political orientation is related to music preferences and personal possessions (Carney, Jost, Gosling, & Potter, 2008), such that politically conservative individuals tend to prefer conventional styles of music, own a smaller variety of books, and display more sports memorabilia in their rooms than do politically liberal individuals. It therefore seems likely that individual differences in entertainment preferences might be related to ideological variables like political orientation and religiosity.

In addition to studying other individual differences, it would also be useful to examine the links between mental states and entertainment preferences. Although individuals have stronger preferences for some dimensions than others in general, the five factors differ considerably in tenor and substance, so certain entertainment factors may be more preferable on some occasions than on others. It is likely that energy, mood, and goals could affect the particular types of genres that people seek out on a given occasion. For example, the genres comprising Communal entertainment tend to be light and uncomplicated, which might be particularly enjoyable when people are fatigued and wanting entertainment that does not require too much effort or attention. Thrilling entertainment might be especially appealing when individuals are bored or restless and in the mood for something exciting. Considering that the Cerebral entertainment factor is defined by non-fiction content, it is probably the case that people seek such genres to inform their decisions, like where to go on vacation, which stocks to invest in, how to repair a broken appliance, or how to cope with an illness. Given that films, music, and stories have been shown to effectively induce a range of emotions (e.g., Westermann, Spies, Stahl, & Hesse, 1996), it is also likely that individuals select certain entertainment dimensions to maintain or achieve a desired mood. Thus, even though relatively stable individual differences contribute to entertainment preferences, future research might benefit from considering a state-trait distinction in preferences. Such an approach would shed light on the effects of mental states on preferences as well as the effects of social contexts (e.g., alone vs. with friends, in the car vs. at home).

The results from our work have useful implications for research in the media effects tradition. The entertainment taxonomy provides researchers concerned with the effects of media on individuals' attitudes, emotions, and behavior a new framework with which to examine individual differences in preferences, which could inform our understanding of the possible moderators of such effects. For example, work concerned with the deleterious consequences of exposure to “violent” media could measure entertainment preferences prior to such exposure and examine whether people with preferences for Dark or Thrilling entertainment genres respond differently than do people who do not like such genres. Additionally, researchers interested in the effects of particular media on body satisfaction could investigate whether preferences for Communal content, which include romance books and movies, pop music, and soap operas—genres that tend to portray men and women in traditional roles—are associated with body image concerns. From an interactionist perspective, it is conceivable that the effects of being exposed to certain content might be different for people who seek out such genres than from those who do not.

There is substantial anecdotal and empirical evidence that individuals use entertainment in the service of self-expression. Consider, for instance, members of such social groups as Punks, Goths, Trekkies (aka Trekkers), or Dead Heads. Such groups are defined almost entirely by their preferences for particular types of music, film, books, or TV. Not only do members of such groups share similar entertainment preferences, but research suggests that members share similar psychological characteristics as well (Bešić & Kerr, 2009). There is also evidence that less radical individuals use entertainment preferences as badges to communicate information about themselves to others (North et al., 2000). For example, Rentfrow and Gosling (2006) found that music and movie preferences were popular topics of conversation among young people, and that people were able to form accurate impressions of each other on the basis of the their music preferences. It seems reasonable to suppose that observers might also be able to form impressions of others' personalities from preferences for movies, books, or TV shows. After all, most users of online social-networking websites (e.g., Facebook.com) share information about their preferences in each of these domains, and research suggests that people are able to form accurate impressions of each other from their online profiles (Vazire & Gosling, 2004). Research concerned with impressions based on entertainment preferences could also consider the extent to which each medium is important to observers. Perhaps observers are better at inferring valid information from domains that are personally important (e.g., movies) because they have knowledge and expertise that provides them with more insight into the genres in the domain and the characteristics of the people who enjoy them.

Considering that individuals spend so much time watching TV and movies, listening to music, and reading books, it is interesting to consider whether similarities in preferences for such media affect the quality of interpersonal relationships. Indeed, if two people, whether romantic partners or roommates, share similar entertainment preferences they should agree more often about which music to listen to and which movies and TV shows to watch than people with different preferences. In a recent study, Delsing, ter Bogt, Engels, and Meeus, (2009) found that close friends were more likely to share similar preferences for music than were less intimate friends. Although that study focused on music, the results nevertheless suggest that entertainment preferences could affect the quality of interpersonal relationships.