Gender Stereotyping in Japanese Television:

Gender Stereotyping in Japanese Television:

A Content Analysis of Commercials from 1961-1993

Akira SAKAMOTO Masumi KITOU

Ochanomizu University Ochanomizu University

Mieko TAKAHIRA Nireka ADACHI

Ochanomizu University Ochanomizu University

In T. Sugiman, M. Karasawa, J. H. Liu, & C. Ward (Eds.) (1999)
Progress in Asian social psychology: Theoretical and empirical contributions (vol. 2, pp. 201-212).
Seoul: Kyoyook-kwahak-sa.

Abstract

The contents of Japanese television commercials telecast from 1961 through 1993 were analyzed to examine the frequency of stereotypic portrayals of men and women and to investigate changes in stereotyping over time. The primary results were as follows: (a) men and women in the commercials were generally portrayed in accordance with traditional gender stereotypes and (b) stereotypic portrayals decreased over 33 years as to the central person's argument and the price of product recommended, but increased as to the central person's location, and did not change with respect to occupation, basis of credibility and age. Thus, it was concluded that efforts to reduce traditional gender stereotypes in television commercials in Japan have met with only limited success.

Introduction

Psychologists and sociologists have maintained that the stereotypic portrayals of men and women found in mass communication cultivate and sustain gender stereotypes in society. Research has considered the type and frequency of stereotypic portrayals in the media (cf. Creedon, 1989, 1993; Kato & Tsuganezawa, 1992), and over the past decade gender stereotyping in television commercials has received particular attention

(Bretl & Cantor, 1988; Craig, 1992; Ferrante, Haynes & Kingsley, 1988; Furnham & Bitar, 1993; Furnham & Voli, 1989; Gilly, 1988; Lovdal, 1989; Mazzella, Durkin, Cerini & Buralli, 1992; Mwangi, 1996; Rak & McMullen, 1987; Riffe, Goldson, Saxton & Yu, 1989; Riffe, Place, & Mayo, 1993; Signorielli, McLeod & Healy, 1994; Smith, 1994). Content analysis of television advertisements has revealed that women are portrayed as submissive, powerless and unintelligent while men are portrayed as independent, powerful and intelligent. Women are usually homemakers while men are fully involved in careers. Research has also shown that only young women make appearances in commercials while men of all ages are found there.

Contemporary researchers have examined how stereotypic portrayals have changed in recent years (Bretl & Cantor, 1988; Ferrante et al., 1988; Lovdal, 1989; Macklin & Kolbe, 1984; Schneider & Schneider, 1979; Weigel and Loomis, 1981). Some studies have reported either modest (Schneider & Schneider, 1979) or substantial decrements (Bretl and Cantor, 1988; Ferrente et al., 1988) in stereotyping while others have found no significant changes in the portrayals of men and women over time (Lovdal, 1989; Maklin & Kolbe 1984). In the main, studies that have been conducted over longer time periods appear more likely to report reductions in gender stereotyping.

Although providing useful descriptive information, these studies suffer from noticeable shortcomings. First, some of the longitudinal comparisons have been based on data extracted from different sources, i.e., researchers have compared the results of their own studies with those of other investigators in order to examine changes in stereotypic portrayals over time. It is possible, therefore, that the results of these comparisons have been affected by differences in the selection and classification procedures used for the archival data. Secondly, researchers have analyzed only data from North American and European television commercials. Thirdly, the duration of observations has been limited to a maximum period of 15 years.

Thus, we analyzed the contents of Japanese television commercials that were telecast from 1961 through 1993 to investigate the frequency of stereotypic portrayals of men and women and the changes in these portrayals over time. This analysis was conducted for six dimensions of gender stereotypes: occupation, location, basis of credibility, argument, price of product and age. We regarded the portrayals as being stereotyped when: 1) women were dependents while men were job holders; 2) women were in the home while men were out of the home; 3) women were users of the product they were recommending while men were authorities on the product; 4) women did not explain why the product was good while men provided explanations; 5) women recommended an inexpensive product while men recommended an expensive one; and 6) only young women were found in television commercials while men of all ages were found there. These six dimensions are routinely used in studies of gender stereotyping (Bretl & Cantor, 1988; Dominick & Rauch, 1972; Ferrante et al., 1988; Furnham & Voli, 1989; Lovdal, 1989; McArthur & Resko, 1975; Schneider & Schneider, 1979).

Despite a common agreement about the widespread existence of gender stereotypes in advertising, studies have produced mixed results concerning changes in stereotypic portrayals over time (Bretl & Cantor, 1988; Ferrante et al., 1988; Lovdal, 1989; Macklin & Kolbe, 1984; Schneider & Schneider, 1979; Weigel and Loomis, 1981). As these investigations have largely examined the same dimensions of gender stereotypes, we believe the mixed findings may be attributed, at least in part, to the varying periods of observation across studies. It appears that decrements in stereotypic portrayals are more frequently found when studies examine commercials over longer periods of time.

We also believe that stereotypic portrayals decreased in American television commercials, at least for the period of the early 1970's through the late 1980's, as the status of American women improved (Schneider & Schneider, 1979). The status of women in Japan has also changed in recent years. For example, the number of women in employment has greatly increased over the past three decades. The Ministry of Labor in Japan (Women's and Young Workers' Bureau, 1996) reported that 9,130,000 women were employed in 1965, compared to 20,480,000 in 1995. The percentage of women with tertiary educational qualifications has also increased. The ratios of students who entered universities or colleges were 5.5% for women and 14.9% for men in 1960. In 1989, however, the percentage of women entering colleges and universities (36.8%) became even greater than the percentage of men (35.8%; Statistics Bureau, Management and Coordination Agency of Japan, 1990). In addition, women's social influence has increased in recent years. Women have gained more powerful social positions, and they have begun to exert more influence on decisions about important issues (Morioka & Mochizuki, 1983).

It seems that each of these changes could lead to decrements in stereotypic portrayals in Japanese television commercials. For example, changes in women's employment status are likely to be associated with a decrease in stereotypic portrayals with respect to occupation, location and basis of credibility. Changes in education status are also likely to be linked to decrements in stereotypic portrayals, at least in connection with the argument bias. In addition, changes in social status and influence may lead to a reduction in stereotypic portrayals relating to product price; however, we have noted no specific social trends that may be predictive of a decrease in the age-related stereotypes of women. Indeed, none of the previous studies have shown decrements in this domain of stereotyping (Ferrante et al., 1988; Schneider & Schneider, 1979).

Thus, we predicted that the stereotypic portrayals of men and women in Japanese television commercials would have decreased with respect to occupation, location, basis of credibility, argument, and price of product from 1961 through 1993, although this decrease would not be found for the age dimension.

Method

Television Commercials

Since 1961, ACC (All Japan Radio and Television Commercials Confederation) has annuallygiven awards to excellent television commercials. ACC is an association of five Japanese organizations in the field of commercials: JAA (Japan Advertisers' Association, Inc.), JAAA (Japan Advertising Agencies Association), ABA (Association Broad Advertising Agency), JAC (Japan Association of TV Commercial Production

Companies), and NAB (National Association of Commercial Broadcasters in Japan). Since the members of the judging committees for the awards are the representatives of these organizations, the judgments are thought to be fair and impartial, and consequently, these awards are regarded as being the most prestigious in Japan (Noda, 1979). Criteria for the judgments are shown in the Appendix.

ACC gave awards to 4276 commercials between 1961 and 1993 (approximately 130 commercials per year). We were not able to view these commercials because the videotapes on which they were recorded are now missing. However, the contents of all the commercials are described in detail with sentences and photographs in the Annual of the All Japan Radio and Television Commercial Confederation issued by ACC. Therefore, were able to examine the contents of the commercials and whether they have changed between 1961 and 1993 by analyzing the descriptions in this book.

Classification of Commercials

Based on the description of each awarded commercial, we first judged who was the most central figure among people appearing in the commercial. This was judged from the amount of words, the size of images and the appearance time. Next, we assessed specific characteristics of the central person and classified the person along the following dimensions:

Gender. We first judged whether the central person was a man or a woman.

Occupation. We classified the person as: 1) job holder; 2) dependent (e.g., parent, family member, housewife, partner); or 3) other/ uncertain.

Location. We classified the person as being: 1) in the home (e.g., kitchen, bathroom, other rooms, garden); 2) out of the home (e.g., office, workplace, store, restaurant, station); or 3) other/ uncertain.

Credibility basis. We classified the person as being: 1) an authority on the product; 2) its user; or 3) other/ uncertain. It is probable that Dominick and Rauch (1972) first used the occupation, location and credibility dimensions in research on gender stereotyping in television commercials; these dimensions have been commonly employed in subsequent research (Bretl & Cantor, 1988; Ferrente et al., 1988; Lovdal, 1989; McArthur & Resko, 1975; Schneider & Schneider, 1979).

Argument. We classified the person according to the argument presented to recommend the product. We used the following three categories: 1) factual; 2) opinion; and 3) none. McArthur and Resko (1975) were most likely the first to examine this dimension although subsequent studies have also used it (Bretl & Cantor, 1988).

Price of product. We classified the person according the price of the product recommended. We used three categories: 1) over 5,000 yen (about 50 dollars); 2) up to 5000 yen; and 3) other/ uncertain. Furnham and Voli (1989) were probably the first to include this dimension; consequently, its use is relatively new.

Age. We classified the person as being: 1) very young (an infant or an elementary school child); 2) young (13-29 years); 3) old (over 30 years); and 4) uncertain. We then combined the first and fourth categories. Since Dominick and Rauch used this dimension in 1972, a number of other researchers have also included it in their studies (Ferrante et al., 1988; Schneider & Schneider, 1979).

If we were unable to identify a central person, we omitted the commercial from further analysis. In addition, we did not analyze the commercials in which people were absent. We eventually classified 2727 central persons which were drawn from the same number of commercials.

Classifications

One judge classified the 2727 persons. In order to determine the reliability of the classification, we randomly selected five commercials per year and had another judge classify the central person appearing in these commercials. Over 90% of the classifications by the two judges were in agreement for gender, occupation, location, argument, price of product and age. The correspondence rate of the credibility dimension was 82.5%, but still acceptable. We therefore concluded that the classification system was reliable.

Results

We divided the years of 1961 - 1993 into three time periods: (1) Period 1 (from 1961 through 1971); (2) Period 2 (from 1972 through 1982); and (3) Period 3 (from 1983 through 1993).

Gender. The percentages of male and female central persons as a function of the periods are presented in Table 1. Men appeared in the Japanese television commercials a little more often across periods than women (54% and 46% in the "All" column). This tendency has not changed over time (X2=1.49, df=2, n=2716, n.s.).

Occupation. The percentages of each occupation group are presented in Table 2. We examined the effects of (a) gender (men and women), (b) periods (Period 1, Period 2, and Period 3), and (c) the interaction of Gender x Periods on occupations using a log-linear model analysis. The main effect of gender was highly significant (X2=302.77, df=2, n=2716, p<.0001). This indicates that the distribution of occupations was different between men and women across time periods. In fact, more men were found in the job holder group than women (76% and 24% in the "All") while more women were found in the dependent group than men (67% and 33%, respectively). The interaction effect of Gender x Periods was also significant (X2=14.95, df=4, n=2716, p<.01), but the percentages of "job holders" and "dependents" remained fairly stable. The significant interaction effect was mainly caused by the temporal change in the percentages of the "other" category, as is found in Chi squared tests for each category (Table 2). Thus, the present study did not support the results of Bretl and Cantor (1988) and Ferrante et al. (1988) which showed decreases in stereotypic portrayals of occupations.

Location. As presented in Table 3, more men were shown out of the home than women (57% and 43% in the "All" column) while more women were in the home than men (51% and 49%). In fact, the two-way log-linear model analysis revealed that the effect of gender on location was significant (X2=9.05, df=2, n=2716, p<.01). The interaction effect was also significant (X2=9.05, df=2, n=2716, p<.01), and we actually found that the percentage of men in the home decreased (54%, 53%, and 44% for Periods 1, 2, and 3, respectively). This indicates that the stereotypic portrayals have recently increased, rather than decreased, as to the location of the central person. This finding differs from the results of previous studies which showed that stereotypic portrayals had either decreased (Bretl & Cantor, 1988; Ferrante et al., 1988) or had remained unchanged (Lovdal, 1989).

Basis of credibility. As is shown in the "All" column of Table 4, men appeared as authorities (71%) more frequently than users (29%) while womenwere portrayed more often as users (54%) than as authorities (46%). The log-linear model analysis revealed that the effect of gender on the basis of credibility was highly significant (X2=25.24, df=2, n=2716, p<.0001). In addition, the differences between men and women remained stable; the interaction effect was not significant (X2=8.35, df=4, n=2716, .05<p<.10), and this result differed from findings by Bretl and Cantor(1988).

Argument. As is shown in the "All" column, men presented factual statements (60%) and opinions (60%) more often than not (52%) while women were more likely to withhold statements (52%) than to present factual analyses (40%) or render opinions (40%). The log-linear model analysis revealed that the effect of gender was highly significant (X2=13.00, df=2, n=2716, p<.01), again supporting male-female differences. However, the interaction effect was also significant (X2=19.86, df=4, n=2716, p<.001), indicating that the differences disappeared over time. This result is consistent with Bretl and Cantor (1988).

Price of Product. As is shown in the "All" column of Table 6, men were associated with products that cost "over 5000 yen" (59%) more frequently than those "up to 5000 yen" (51%) while the reverse was true for women (49% and 41%, respectively). This indicates that men recommended expensive products rather than cheap ones while women recommended cheap products rather than expensive ones. The log-linear model analysis revealed that the effect of gender was highly significant (X2=15.55, df=2, n=2716, p<.0001). The interaction effect was also significant (X2=15.60, df=4, n=2716, p<.01) and showed that gender differences decreased over time.

Age. As is shown in the "All" column of Table 7, more men were found in the old group than women (76% and 24%) while more women were found in the young group than men (79% and 21%). The log-linear model analysis revealed that the main effect of gender was highly significant (X2=477.92, df=2, n=2716, p<.0001). The interaction was not significant, and the present study showed that the gender difference had not changed over time. This is in line with previous findings by Ferrante et al. (1988) and

Schneider and Schneider (1979).

Discussion

Consistent with earlier international studies, men and women appearing in Japanese commercials are portrayed in terms of traditional gender stereotypes. That is, women in these commercials are more likely to be young, dependents, in the home and users of the products. They are also more likely to recommend inexpensive products without the support of factual arguments. Men, on the other hand, are more likely to be old, job holders, out of the home and authorities on the products. They are also more likely to explain why the products are good and to recommend expensive items.

Long-term changes in such stereotyped portrayals depended on the aspects. We predicted that the stereotyped portrayals would have decrease for the aspects of occupation, location, and credibility basis, due to a change in women's real occupation status, but the present study did not support this prediction. The stereotyped portrayals even increased for the location. On the other hand, the predictions for the aspects of argument and price of product, which were made, based on changes in women's education status and social power respectively, were supported. The results showed the decreases of stereotyped portrayals for these aspects. As for age, the prediction of no changes was supported.

Thus, we found a discrepancy in the results between the aspects related to occupation status and the ones related to education status and social power. This discrepancy might be due to differences among recent impressions of women's occupation status, education status, and social power. Women have recently become more highly educated and perhaps more powerful at least in the home than men. As for occupation, however, men have still much higher status than women, although women's status has greatly increased. It might be that recent television commercial producers have been more greatly impressed with the height of their education status and the strength of their social power than with the height of women's occupation status, and consequently they have used women who were intelligent and powerful, if they have not used those who had an occupation. Future research is necessary to examine this explanation and other possibilities.

Overall our results have shown that stereotypic portrayals are still found in Japanese television commercials. While some aspects of stereotyping have declined (e.g., argument, product price), others have become more pronounced (e.g., location). Therefore, we cannot conclude that stereotypic portrayals in Japanese television commercials have markedly decreased over the past three decades.

A number of American studies have shown that stereotypic gender portrayals in television commercials have declined substantially (Bretl & Cantor, 1988; Ferrante et al., 1988). The present study, in contrast, demonstrated only modest reductions despite the extended period of data analysis. These findings suggest that changes in stereotypic portrayals of men and women in television commercials occur more slowly in Japan than in the United States and that efforts to reduce traditional gender stereotypes in Japan have met with limited success.

Finally, it should be acknowledged that the commercials used in the present study are not representative of all Japanese television advertisements. While this might be considered a limitation, we suggest that the commercials provide appropriate research materials. First, the awarded commercials have been identified as those which "reflect Japanese culture" and provide "windows" to the future (see Appendix). This highlights their social relevance to daily life in Japan. Secondly, the commercials awarded by ACC have been extensively broadcast. This suggests greater social influence through high exposure, i.e., these commercials are more likely to precipitate social consequences than infrequently broadcast commercials. Although the commercials may not be representative of Japanese television advertisements, they have social significance which makes them appropriate for research purposes.

In conclusion, traditional stereotypic portrayals of men and women in Japanese television commercials have not substantially decreased from 1961 to 1993. Nor do they accurately reflect contemporary social trends in Japan. As these traditional portrayals can negatively affect attitudes toward women, greater efforts to reduce gender stereotyping in current Japanese television commercials are recommended.

Author Note

We would like to thank Mr. James Cranford and anonymous reviewers for their helpful comments on a draft of this paper.

Correspondence concerning this article should be addressed to Akira Sakamoto, Department of Psychology, Ochanomizu University, 2-1-1 Otsuka, Bunkyo-ku, Tokyo 112-8610, Japan. Electronic mail may be sent via Internet to hyperlink mailto: sakamoto@li.ocha.ac.jp.