Are the liberal good in Hollywood? Characteristics of political figures in popular films from...

Previous research (Prindle & Endersby, 1995) indicated that the leaders of Hollywood are far more liberal than the average American. The question that remained to be answered was to what extent this liberal bias leaks into the Hollywood product. In an effort to answer this question we examined

124 political characters in 47 popular political films spanning five decades. The findings suggested that the characters who were depicted as more liberal also tended to be depicted as more intelligent, friendly, and good. However the results also suggest that the liberal bias in Hollywood is not very robust. Hollywood movers and shakers appeared to have kept their liberal bias in check to some extent. It is speculated, (similarly to Prindle & Endersby), that Hollywood was more invested in producing entertainment that is politically generic and palatable to a general audience than in making movies that promote their liberal political leanings.

**********

* The term "power elite" was coined by C. Wright Mills in 1956 as a short-hand term referring to the select few in business and government who make key decisions for American policy. Lichter, Rothman, and Lichter (1986) updated the term to include political interest movements and the national media. Their research explored the impact of the power elite on American life, with particular emphasis on news journalists and their tendency toward political liberalism.

Those who control all forms of mass media can be considered members of the power elite. Newspaper, radio and television have traditionally been seen as the predominant forms of mass media, but as Macionis (1999) pointed out, videocassette recorders (VCRs) are becoming pervasive in American households. The dramatic increase of VCR use suggests that Americans are accessing films more than ever, thus placing movies in the realm of mass media, and filmmakers in the position of the power elite.

Casual speculation by radio talk-show hosts, politicians, and other public figures is that Hollywood has a strong liberal bias. For example, Cones (1997) asserted that Hollywood films reflect the values of a select few whom he identified as politically liberal, Jewish males of European heritage. Cones and Jaeger (n.d.) co-founded the Film Industry Reform Movement (F.I.R.M.), an organization designed to dethrone the power-base of liberal Hollywood.

While it is possible that Hollywood has a liberal bias, it is also possible that this is a misperception driven by the biases of those making the observations. Vallone, Ross, and Lepper (1985) demonstrated that people with a strong political opinion on an issue tend to perceive neutral media coverage on that issue as biased and against their side. They dubbed this the "hostile media phenomenon." Bias in perception was also highlighted by Rothman, Powers and Rothman (1993) who trained raters to evaluate films by objective criteria, and found no evidence for the then popular assertion that Hollywood filmmakers were portraying woman negatively (a "backlash" against the feminist movement). As Rothman et al. concluded, it is easy to become prey to bias if judgments about Hollywood's prejudice are made without a representative sample of films, a valid rating system, and trained raters.

Assertions that there is a liberal power elite in Hollywood has some empirical backing. Prindle and Endersby (1993) presented evidence that those with influence in Hollywood are indeed more liberal than the average person. They identified those in Hollywood who had the most influence on the ideological tone of the industry (including executives, actors, directors, producers, and writers) and through interviews determined that these individuals were far more likely than a control group to describe themselves as liberal and to support liberal views on a variety of issues. Powers, Rothman, and Rothman (1996) did a similar survey of Hollywood moviemakers, with similar results. Powers et al. also examined high-grossing films from 1946-1990, and found that, beginning in the mid-1960s, Hollywood's message has become more left-leaning and more critical of traditional American society. Specifically, they found a trend toward a more critical view of the military, police, wealthy people, businesspeople, and organized religion, and a more positive depiction of women and minorities.

While Powers et al. (1996) provided a wealth of information about the content of popular films, there are more questions to be answered. They found a trend toward more liberal content in films over time, but it should be noted that to our knowledge neither Powers et al. nor any other researcher has provided insight as to the original political anchor point from which films have drifted. That is, this trend may have moved film content from the right to the center, rather than from the center to the left. There is no point of reference from which to judge just how liberal film content has become over time. In addition, Powers, et al. demonstrated that films have moved toward more liberal depictions of social issues such as religion and the military, but they did not examine Hollywood's depiction of politics itself.

Films that are overtly political (i.e., those dealing with elections, presidencies, and so forth), provide a good gauge of how blatant Hollywood's liberal bias may be. It is one thing to paint wealthy people, or the police, in a negative light, it is quite another to depict Republicans negatively. Regardless of filmmakers' own political leanings, it cannot be overlooked that they are primarily in the business of making money, not furthering a partisan political agenda.

Thus, we posit two competing hypotheses. With Prindle and Endersby (1993) and Powers et al. (1996) as precedent, one might safely hypothesize that popular Hollywood films that deal with politics will exhibit a clear liberal bias. These authors have demonstrated that filmmakers tend to be liberal, and the social messages in their films have become more liberal over time.

However, a case can be made for the competing hypothesis that Hollywood filmmakers will not interject a liberal political bias when "political" means party politics. That is, filmmakers may suppress the influence of their own personal political leanings from the movies they produce in the interest of making a product that will be profitable. If the prevailing political winds are conservative, for example, moviemakers may suppress their own liberal leanings and create a product that is more in line with the interests of what they perceive to be a relatively conservative public. Therefore, this hypothesis states that moviemakers are influenced by the stance of the current governmental administration, or by the political leaning of the population at large, rather than by their personal political leanings. This hypothesis was tested by examining the relations between political characters' political position and the other dimensions of interest (e.g., attractiveness, goodness, etc.) separately for characters from movies made during Republican and Democratic presidential administrations. Significant differences in correlations from the two types of administrations would lend support to this hypothesis.

In the present study popular films that directly deal with politics were examined. In order to gauge the extent to which Hollywood films exhibit a bias toward liberal or conservative ideas, depictions of the characters who represent the different sides of the political spectrum were examined. Thus the question asked was: are liberal characters depicted more positively than conservative characters? Specifically, are they depicted as smarter, more moral, more attractive, and so on? Do good things happen to liberal characters and bad things happen to conservatives? Additionally, the political climate at the time of the films release was also examined in order to test the contention that the climate influenced depictions of conservative versus liberal characters. Forty-seven popular political films spanning the past fifty-four years were examined. It was reasoned that the degree to which Hollywood depicts liberal characters in a favorable light would be manifest in positive correlations between a liberal versus conservative depiction and attractiveness, goodness, intelligence, friendliness, and romantic activity and a negative correlation between a liberal versus conservative depiction and aggressiveness.

METHOD

Selection of Target Films

Films were selected from The Ultimate Movie Thesaurus (Case, 1996). The year 1945 was chosen as the cutoff point because of difficulty locating videos of films made earlier. All topic headings that listed films pertaining to politicians were identified. The headings included were: Politicians, Politicians-Female, Politician's Wives, Political Activists, Political Campaigns, Political Drama, Political Satire/Comedy, and Political Thriller. A total of 207 different films were listed under these headings.

Next, two of the authors independently read the summaries of all 207 films listed under the topic headings selected. Each compiled a list of films that he or she judged to meet the following criteria:

1) contains at least one major character who is overtly political;

2) not a documentary;

3) released in 1945 or later;

4) not made for television;

5) includes U.S. political characters; and

6) given a rating of at least two out of six on The Ultimate Movie Thesaurus' rating of the film's box-office sales, with one being "A flop or totally ignored by moviegoers," two being "not a hit, but a fair amount of tickets sold," and six being "A blockbuster hit of the highest level." This ruled out obscure, low-budget films with very limited release, insuring that the films under consideration were films that moviegoers were interested in seeing.

Based on these criteria, one author generated a list of 61 films, while the other author generated a list of 65 films. Only films that appeared on both lists were selected, resulting in a total of 50 films. Because The Ultimate Movie Thesaurus only included films through 1995, the same individuals visited a large local video store and searched through movies released during 1996, 1997, and 1998 using the same criteria set out above, save for criterion number 6 which was not available. An additional five films were identified in this manner: Bulworth, Nixon, Wag the Dog, Primary Colors, and My Fellow Americans. We determined that eight films on our list (Beau James, Blind Ambition, Goodbye My Fancy, I Love Trouble [1947], Detective, Key to the City, The Seduction of Joe Tynan, and Mastergate) were currently out of print and thus unavailable for viewing. These were deleted from the list, resulting in a total of 47 films. (See appendix A for a complete listing of the films selected.)

Rating Procedure

All of the authors served as raters, thus there were a total of four. Ratings were carried out following procedures developed in previous research conducted by the first author (McIntosh, Smith, Bazzini, & Mills, 1999; Smith, McIntosh, & Bazzini, 1999). Before beginning data collection for the present study, raters practiced on a film that was not part of the sample. During the training session raters discussed the rating procedures at length until all raters felt competent to proceed.

At least two trained raters independently rated each film. Characters were rated on the following dimensions: 1) conservative/liberal, 2) outcome, 3) attractiveness, 4) aggressiveness, 5) friendliness, 6) goodness, 7) intelligence, 8) romantic activity. Definitions and criteria for assessment follow.

Conservative/liberal. Characters' political position was rated on a scale from 0 (very conservative) to 10 (very liberal). A rating of 5 indicated the character was a centrist, thus 4 denoted slightly conservative, 6 slightly liberal, and so on. This was in effect the "gatekeeping" rating. Only characters who could be rated on their political position were included in analyses. In the rare cases (fewer than 15) where some raters rated a character and others did not, final inclusion or exclusion of that character was determined through discussion after the film ended.

In order to be rated, a character had to explicitly reveal his or her political leanings during the course of the film. A character's political position could not be inferred by association. For example, working as an aide for a staunch Republican was not enough to rate a character as conservative. To be rated, a character's political opinions or political affiliation had to be made explicit in the film.

For the purposes of this study the following characteristics were judged appropriate evidence of political conservatism: 1) invested in tradition, 2) favors status quo over change, 3) believes there is wisdom in the past, 4) advocates for small government, 5) advocates self-responsibility, 6) favors military solutions, 7) pro-life, 8) favors capital punishment, 9) capitalistic, 10) belongs to the Republican party. By contrast, the following were judged evidence of a liberal political position: 1) not overly invested in tradition, 2) favors change, 3) emphasizes the future, 4) advocates a large role for government, 5) advocates group responsibility, 6) advocates diplomacy over force, 7) pro-choice, 8) opposed to capital punishment, favors rehabilitation of convicts, 9) socialistic, 10) belongs to the Democratic party.

Outcome and attractiveness. At film's end, each character's outcome was rated on a scale of 0 (extremely negative; e.g., death) to 10 (extremely positive; e.g., struck it rich and lived happily ever after). A rating of 5 indicated that he or she was no better or worse off at the end of the film than at the beginning.

Ratings of physical attractiveness were based on an 11-point scale ranging from 0 (extremely unattractive) to 10 (extremely attractive).

Aggressiveness. Aggressiveness was defined as physical violence inflicted on other people and was rated on a scale ranging from 0 (not at all aggressive) to 10 (extremely aggressive). Raters considered both the nature of the aggression (e.g., hitting someone with one's fists; shooting someone with a gun) and the frequency of aggressive acts when making their ratings. Therefore ratings take both frequency and intensity into account.

Friendliness, goodness and intelligence ratings. Raters assessed each character's friendliness on a scale ranging from 0 (extremely unfriendly) to 10 (extremely friendly). Raters were cautioned to distinguish between friendliness and goodness--a character can be personable yet immoral.

Alternatively, raters assessed the character's goodness as moral virtue on a scale ranging from 0 (extremely immoral) to 10 (extremely moral, saintly).

Finally, the character's intelligence was assessed on a scale ranging from 0 (extremely unintelligent, a moron) to 0 (extremely intelligent, a genius).

Romantic activity. A character's real or implied romantic and/or sexual involvement was rated on a scale ranging from 0 (totally inactive) to 10 (extremely active). Due to changing norms and laws, recent films portray this more explicitly. In an attempt to balance this inequity, raters attempted to estimate how much romantic/sexual activity a character was portrayed to be involved in, rather than consider only the amount of time the characters actually engaged in romantic and/or sexual activity.

RESULTS

Reliability of ratings

Reliability was assessed by randomly selecting 30 of the 124 characters rated. For each character the ratings provided by two raters (randomly selected when more than two individuals rated the character) were entered as variables and calculated coefficient alphas. For the majority of dimensions, reliabilities across raters were well above acceptable levels. Individual Coefficient Alphas (indicating reliability) were as follows: conservative/ liberal (.83), outcome (.86), attractiveness (.81), aggressiveness (.66), friendliness (.41) (1), goodness (.85), intelligence (.73), and romantic activity (.92). Because these interrater reliabilities were considered acceptable, mean ratings on each dimension for all 124 characters included were computed by averaging across all raters.

Cumulative Analyses

In order to examine the hypotheses of interest, Pearson's product-moment correlations were computed between mean ratings of characters' political position and a variety of trait adjectives (attractiveness, goodness, aggressiveness, intelligence, friendliness), a life-situation measure (romantic activity), and the perceived outcome of the character at film's end. The first column of Table 1 displays these correlations. As indicated in the table, a liberal bias was found only with the traits of goodness, intelligence, and friendliness.

In many of the films rated, raters were forced to base their judgments on very little information, at times as little as one casual statement made by a character. After rating each film, raters discussed the amount of information their ratings were based on, and reached a consensus on whether the film should be classified as having little political discussion, or sufficient political discussion. The second column in the table displays the correlations computed using only a subset of characters (n = 105), from films considered by the raters to have a sufficient amount of political discussion. In general, all correlations became stronger compared to the first analysis, but none significantly so. The primary conclusion reached in the first analysis remained the same: the more liberal a character's political leanings, the more good, intelligent, and friendly that character is perceived to be.

To examine whether these findings apply only to those characters in main (i.e., starring) roles, the correlations in the two right-most columns pertain only to those characters that were judged to be main characters and secondary, respectively. We found no significant differences between correlations based on whether a character is a main or secondary character.

Character Portrayals as a Function of Current Administration and Public Political Ideology

A secondary question addressed by the current research was whether filmmakers are influenced by the prevailing political winds of the time. This question was explored in two ways.

First, the same correlations displayed in Table 1 were computed separately for characters in movies made during Republican administrations (as defined by the party in the president's office) and Democratic administrations. These results are displayed in Table 2. There were no significant differences in the correlations based on current administration, with one exception. Although characters' outcome at film's end was not significantly correlated with a character's political leanings on an absolute level, the data indicate that more liberal characters were associated with a more positive outcome in films made during Democratic administrations, r(80) = .248, and a more negative outcome in films made during Republican administrations, r(40) = -.254, z = 2.62, p < .05.

Second, whether the public's political leanings would affect the relationship between characters' political ideology and how they are portrayed was also examined. Party identification data from the American National Election Study was used to accomplish this. These data were first collected in 1952, so films released prior to 1952 were not included in these analyses.

In each year listed, a larger percentage of respondents described themselves as Democrats than Republicans. The difference between the percentage that described themselves Democrats and the percentage that described themselves as Republicans was used as our measure of political ideology. Because American National Election polls were conducted every other year, non-poll years were calculated using the mean of the preceding year and the subsequent year. Multiple regression was used to test for an interaction between political ideology and character ideology to predict character portrayal, entering each of the seven characteristics in turn. None of the interactions were significant.

Character Portrayals Across Time

As discussed earlier, Powers et al. (1996) found that Hollywood's message became more left-leaning and more critical of traditional American society beginning in the mid-1960s. To seek the same pattern in portrayals of political characters, two analyses were conducted.

First, the same correlations reported in table 1 were computed separately for films released before 1965 (n = 68), and those released from 1965-on (n = 56). There were no significant differences in the correlations based on film era. Second, multiple regression was used to test for an interaction between the year each film was released and character ideology to predict character portrayal, entering each of the seven characteristics in turn. None of the interactions were significant.

DISCUSSION

The questions posed by Prindle and Endersby (1993) were: (1) to what extent does the liberal ideology of Hollywood leaders leak into the content of their product; and (2) do Hollywood movers and shakers sell a liberal bias, or do they filter out their bias so as to produce a product that won't offend the general public? This study also answers these questions, but with respect to movies that have political content. Liberal characters are not cast as more wealthy, attractive, or romantically active, nor are they portrayed as less aggressive. However our findings suggest that Hollywood does tend to portray liberal characters as good, intelligent, and friendly. This positive spin on characters who can be judged as liberal may indeed constitute a bias on the part of Hollywood leaders. However, there appears to be more to this story.

Although an overall liberal bias can be found in terms of character portrayal, the more striking finding was that most political films were ambiguous in terms of partisan politics. Very little partisan political discussion took place in most of these films, despite the fact that they are classified as political movies. Most political films present commentaries on the American political system, but pointedly avoid the riskier business of commenting on partisan political issues. Indeed, in four of the films (The Fountainhead, Kisses for My President, True Colors, Wag the Dog) not one character was judged to have espoused a political position that would indicate a conservative or liberal perspective. In 15 other films, raters indicated that they made their judgments of political position based on very sparse information, often one passing comment during the course of the entire film. Partisan politics was especially difficult to determine in star characters. We interpret this finding as one that strengthens the conclusion that Hollywood appears to be quite cautious in making primary characters politically extreme.

As Prindle and Endersby (1993) speculated, Hollywood leaders were careful to produce entertainment that would not offend the general audience. In fact, the findings here suggested that when liberals were in office Hollywood tends to produce political movies in which liberal characters fair well, and when conservatives were in office the political movies end with conservative characters enjoying a favorable outcome. Whether this tendency to mold movies to the current political climate stemmed from an unconscious sensitivity to the public or from an overt marketing strategy remains to be seen. It should also be noted that this pattern was not evident when public ideology (American National Election data) was used as the indicator of current political sentiment rather than current administration, so the trend is not very robust.

Of course, this study will not be the final word on Hollywood and politics. One direction future study might take would be to look at the impact of the political messages in Hollywood films. When a film does clearly exhibit a political bias, does it affect the social and political attitudes of viewers? Most importantly, do movies directly influence voting behavior? To our knowledge, a thorough investigation of the causal relations between Hollywood movies and voting behavior has not been done. In addition, it would be interesting to explore whether films from other cultures exhibit a political bias, either to the left or the right, and to what extent the political orientation of a nation's government predict this.

In conclusion, this research has shown that Hollywood leaders tend to keep their liberal leanings in check. Most political films presented characters that carefully avoided distinguishing themselves as liberal or conservative. When the ambiguity of political stance was brought slightly into focus more liberal characters were portrayed in a positive light in terms of goodness, intelligence, and friendliness.

APPENDIX A

Films Rated

Advise and Consent            1962
All the King's Men            1949
All the President's Men       1976
American President            1995
Blaze                         1989
Bob Roberts                   1992
Born Yesterday                1950
Born Yesterday                1993
Buccaneer                     1958
Bulworth                      1998
The Candidate                 1972
Clear and Present Danger      1994
Dave                          1993
The Dead Zone                 1983

A Face in the Crowd           1957
Fail Safe                     1964
Farmer's Daughter             1947
The First Monday in October   1981
Flamingo Road                 1949
A Foreign Affair              1948
The Fountainhead              1949
Jefferson in Paris            1995
JFK                           1991
Kisses for My President       1964
A Lion in the Streets         1953
The Last Hurrah               1958
Magnificent Yankee            1950
The Manchurian Candidate      1962
My Fellow Americans           1996
Nixon                         1995
The Parallax View             1974
Primary Colors                1998
Running Mates                 1992
Seven Days in May             1964
The Senator Was Indiscreet    1947
The Shaggy DA                 1976
Speechless                    1994
State of the Union            1948
Sunrise at Campobello         1960
Taxi Driver                   1976
Twilight's Last Gleaming      1977
True Colors                   1991
Wag the Dog                   1998
Wilson                        1944
The Young Philadelphians      1959

TABLE 1
Pearson-Product-Moment Correlations between Political Position
and Relevant Trait and Life-Style Dimensions

                                 Characters
                                  in Film
                       All      w/Sufficient
                    Character    Political        Main      Secondary
                      Rated      Discussion    Characters   Characters
Rated Dimension     (n = 124     (n = 105)      (n = 50)     (n = 32)

Attractiveness       .192         .200           .239         .109
Goodness             .599 **      .639 **        .474 *       .777 **
Aggressiveness      -.012         .001          -.187        -.022
Intelligence         .461 **      .466 **        .492 *       .403
Friendliness         .480 **      .504 **        .465 *       .485
Romantic activity    .280         .285           .227         .201
Outcome in film      .090         .095          -.037         .167

* Denotes p < .05

** Denotes p < .001

TABLE 2
Pearson Product-Moment Correlations Between Political Position
and Relevant Dimensions as a Function of Political Party in
Power at the Time the Film Was Made

                    Character From Films Made During:

                      Democratic        Republican
                    Administrations   Administrations

Attractiveness          .252               .118
Goodness                .589 *             .606 *
Aggressiveness          .008              -.040
Intelligence            .443 *             .443
Friendliness            .582 *             .302
Romantic activity       .331               .175
Outcome in film         .248              -.254

* Denotes p < .01

ENDNOTES

(1.) We acknowledge that an Alpha of .41 is below most researchers' standard of acceptability. We therefore advise the reader to interpret with caution findings that pertain to the dimension of friendliness.

REFERENCES

Case, C. (1996). The ultimate movie thesaurus. New York: Henry Holt and Co.

Cones, J. W. (1997). What's really going on in Hollywood. Marina del Rey, CA: Rivas Canyon Press.

Cones, J. W., & Jaeger, J. (n.d.) Retrieved August 9, 2002, from http://www.homevideo.net/firm.

Lichter, S. R., Rothman, S., & Lichter, L. S. (1986). The media elite. Bethesda, MD: Adler & Adler.

Macionis, J. J. (1999). Sociology (7th ed.). New Jersey: Prentice Hall.

McIntosh, W. D., Smith, S. M., Bazzini, D. G., & Mills, P. S. (1999). Alcohol in the movies: Characteristics of drinkers and nondrinkers in films from 1940 to 1989. Journal of Applied Social Psychology, 29, 1191-1199.

Mills, C. W. (1956). The power elite. New York: Oxford University Press.

Powers, S., Rothman, D. J., & Rothman, S. (1996). Hollywood's America: Social and political themes in motion pictures. Boulder, CO: Westview Press.

Prindle, D. F., & Endersby, J. W. (1993). Hollywood liberalism. Social Science Quarterly, 74, 137-149.

Rothman, S., Powers, S., & Rothman, D. (1993). Feminism in films. Society, 30, 66-72.

Smith, S. M., McIntosh, W. D., & Bazzini, D. G. (1999). Are the beautiful good in Hollywood? An investigation of the beauty-and-goodness stereotype on film. Basic and Applied Social Psychology, 21, 69-80.

Vallone, R. P., Ross, L., & Lepper, M. R. (1985). The hostile media phenomenon: Biased perception and perceptions of media bias in coverage of the Beirut massacre. Journal of Personality and Social Psychology, 49, 577-585.