Sex-Role Stereotyping in Television Commercials: A Review and Comparison of Fourteen Studies Done on Five Continents Over 25 Years - Statistical Data Included

Adrian Furnham

Adrian Furnham [1]

Since the pioneering content-analytic study by McArthur and Resko (1975) on sex-role stereotyping of television advertisements in America, many others have used a similar methodology and coding scheme to examine similar stereotypes in their own countries. This study compares and contrasts 14 studies, all using the McArthur and Resko (1975) scheme: 3 from America, 1 each from Australia, Denmark, and France, and 2 and one from Great Britain, Hong Kong, Indonesia, Italy, Kenya, Mexico, and Portugal. Problems of such a comparison are considered: specifically, the equivalence of the channel, the three different time periods, and slight variations in the content categories. Nevertheless, clear patterns arise which attest to the universality of sex-role stereotyping in television commercials.

Over the past 20 years commercial television has spread around the world. Tourists and explorers are often amazed and surprised to find indigenous people in extremely "cut-off" rural settings sitting around watching the television set in their crude dwellings. Nearly every country in the world now has commercial television channels--frequently more than one--and most adults are exposed to thousands of television advertisements annually.

While the power and influence of television on the beliefs and behaviors of viewers remains probably the most important area of research for media researchers, many have concentrated on the portrayal of certain features. Thus violence on television remains one of the most researched and controversial topics (Gunter & Wober, 1988; Lanis & Covell, 1995). Similarly the portrayal of gender on television and the social effects of television on gender-role portrayals have attracted considerable interest (Prakash, 1992), according to a recent review (Gunter, 1995).

Researchers, particularly in the United States, have been interested in the portrayal of men and women in television commercials for over 25 years (Dominick & Rauch, 1972; Maraceck et al., 1978; Liebler & Smith, 1997; O'Donnell & O'Donnell, 1978; Schneider & Schneider, 1979; Suezle, 1970). Sex-role cues in children's commercials have also been studied (Durkin & Nugent, 1998; Furnham, Abramsky, & Gunter, 1997; Smith, 1994; Welch, Huston-Stein, Wright, & Phehal, 1979), as well as sex-role stereotyping on the radio (Furnham & Thomson, 1999; Hurtz & Durkin, 1997). Studies have concentrated on gender-role stereotypes in commercials for particular products such as food (Jaffe & Berger, 1994), beer (Iijima, Hall, & Crum, 1994), and medical products (Leppard, Ogletree, & Wallen, 1993) or particular types of programs such as comedies (Olson & Douglas, 1997) or, specifically, portrayals of people in the work place (Vandeberg & Streckfuss, 1992).

One of the first content-analytic studies of the portrayal of men and women in television commercials was published by McArthur and Resko (1975) on data collected nearly 25 ago, yet studies currently being published follow essentially the same coding categories (Neto & Pinto, 1998). It has proved the model for many subsequent studies (though not all studies of stereotyping on television), and it is the aim of this paper to do a "meta-analytic" review of the research based on the coding scheme. In doing so, it examines the results of 14 studies, from McArthur and Resko (1975) to some still in press.

Bretl and Cantor (1988) reviewed eight studies, all bar one from the 1970s, as well as presenting their own original data. They were particularly interested in studying trends over the 15-year period. At the time they found men and women appeared equally often as central characters but that 90% of all narrators were male. Fewer females were depicted as employed, while women were still more likely than men to be seen in domestic settings, advertising products in the home. The authors concluded 12 years ago that (American) advertisements seemed "to be presenting a less sexist and more equal view of the roles of men and women in society" (p. 607). However, it is uncertain whether this was true in other countries.

At much the same time Lovdal (1989) replicated one specific study (O'Donnell & O'Donnell, 1978). Contrary to the above, she found no significant differences on nearly all of the coding categories examined and concluded that conventional sex-role stereotypes persist in television commercials.

There are numerous problems in attempting to understand compariSons done in different countries over different time periods. All cross-cultural psychologists acknowledge the many problems associated with cross-cultural comparisons (Smith & Bond, 1993). These include stimulus and sample equivalence, the functional and meaning equivalence of particular behaviors, and changes over time if studies are carried out or compared at different points in time. Some stress the subtle uniqueness of culture and the problems with measuring accurately, while others believe that the increasing modernization and Westernization of the world means that all cultures will converge onto a homogenized unity. Some psychologists have taken an etic (based on phonetic) approach that focuses on universals, while others have preferred an emic approach (based on phonemics) which stresses the difference and variety. These issues are particularly problematic for content analysis, which is the method of this study. This is because a content coding category scheme developed in one country at one particular time period is applied to other stimuli resulting from and mirroring a different and unique culture at very different periods of time. There always remains the question of whether the coding scheme is sufficiently robust and sensitive to interpret the many subtle nuances, particularly in the area of sex-role portrayal.

First, an obvious research problem of confounding occurs even when examining data content analyzed in the same country over different time periods. The first is that of channel equivalence. Commercial channels differ enormously in viewing figures, style, and funding. Some are national, others regional; some are aimed at a particular demographic or psychological group which is reflected in the commercials. For instance, British Channel 4 has a larger, younger, better-educated audience overall than New Zealand's TV One, which aims to capture a wider audience. The size, history, and economic status of the country, the number of channels, and the extent to which they deliberately set out to reach select or niche markets mean that it may well be near-impossible to find "equivalent channels" in different countries. Further, the precise audience of the channel is of direct relevance to advertisers, who place their particular commercials with numerous economic demographic and psychographic variables in mind (Gunter & Svennevig, 1987). Doing studies over time using a limited number of channels increases the reliability of finding trends but decreases the generalizability of the findings.

Second there is the problem of sample equivalence. It has been demonstrated that there is a time-of-day effect in commercials (Harris & Stobart, 1986), because of the different patterns of viewing of different groups in the population. Thus in most countries children watch early-morning and early-evening television, and homemakers, retired and unemployed people, and shift workers watch daytime television, while a wider general population sample watches in the evening and on weekends (Gunter, 1995). However, the precise audience mix does not differ from channel to channel and from country to country. Thus, if study 1 samples predominantly daytime and study 2 predominantly nighttime commercials, any differences may be attributable to sampling rather than actual differences. It is also known that the advertisements for particular products differ according to time of day, day of week, and season of year as a function of what is known about the audience (Furnham et al., 1997; Furnham & Skae, 1997; Gunter, 1995). Thus toy advertisements peak in the morning, whereas cars, insurance, alcohol, and food are more likely to occur in evening programs. This pattern is typical for Western-developed economics but there are subtle differences. Further, some countries ban advertisements for toys, alcohol, or other specific products (Furnham, 1993), increasing problems for equivalence.

Third, there is the problem of content categories. While most researchers have followed McArthur and Resko (1975), many have altered or added to the content categories for their own purposes. This renders many comparisons impossible unless, in some instances, certain categories are collapsed. However, it is a testament to the original work of McArthur and Resko (1975) and a number of studies done in the decade after theirs was published (Manstead & McCulloch, 1981; Harris & Stobart, 1986; Livingstone and Green, 1986), that so few changes and adaptations have been made despite cross-cultural work being done. This suggests that the original categories were comprehensive. The most difficult problem is that of national differences, which are, for many, of most interest. All the above problems occur such that it is difficult to attribute any difference to culture rather than methodological differences.

Nevertheless, it is interesting and important to look for patterns in the data available. There is probably no such thing as a typical commercial. They differ in length enormously from country to country, from 5 to over 300 sec. They may or may not have actors in scenes or voice-overs, and they usually, but not always, mention the product's name and usually describe its benefits.

This study focuses on the portrayal of gender roles in television commercials in different countries, comparing current research studies with a watershed study done over 25 years ago. It attempts a comprehensive review of all content analytic studies that followed the basic categories set out by McArthur and Resko (1975) 25 years ago (Lanis & Covell, 1995; Leppard et aL, 1993; Moore, 1992). This review attempts a comprehensive international analysis of all recent studies following the McArthur and Resko (1975) coding scheme in an attempt to look for both trends and cross-national differences and similarities in sex-role portrayal in television advertisements.

METHOD

The criteria for selecting a study were the following. First, using Psychological Abstracts and Psychlit, an attempt was made to find all television advertisement content analyses based predominantly on the McArthur and Resko (1975) system published in psychology and communication journals. In the original paper eight characteristics of the central figure were coded: sex, basis for credibility, role, location, arguments given on behalf of a product, rewards offered or repeated for using a product, punishments threatened or incurred for not using a product, and type of product advertised. Coding reliabilities among three coders varied from 66% agreement to 91% agreement. Next it was decided to restrict studies to the best studies from each country (number of advertisements coded), including if there were more than one. According to these criteria there were two from the United States and two from Great Britain. Another criterion was to obtain studies of comparable quality and method from as many different countries as possible to represent work done in Africa, America, Asia, Australia, and Europe. Fourteen studies were selected for analysis.

Coding procedures of the different studies vary, thus only similar attributes and categories were chosen for comparison. A careful review of the 14 different studies selected suggested that five coding categories had to be reanalyzed or collapsed into the following categories.

1. Role

Autonomous: worker, professional, celebrity, and interviewer/narrator.

2. Location

Outdoor: includes restaurant/bars, business school, and outdoors away from home.

3. Age

In some studies, those younger than 35 years old are considered young, while 35-65 is middle-aged, and over 65 years old is old.

4. Argument

Scientific: arguments contained or purported to contain factual evidence concerning the product.

Nonscientific: consists simply of opinions or testimonials in favor of the product.

5. Product type

Domestic: a combination of body, home, and food products.

Away: products used in places other than at home.

Home and away: products can be used at home and elsewhere.

Auto: includes automobiles, auto accessories, and related products.

Sports: any product or service connected with sports.

Some of the studies had used all 10 categories but others had omitted certain categories for various reasons detailed in each paper. Nevertheless, all data on every paper were included, for maximum understanding.

RESULTS

Table I shows the 14 studies which form the basis for the comparison. Table II presents all the comparative data considered in this study, indicating, when appropriate, where significant differences occurred.

Frequency of Male and Female Central Figures

Mode of Presentation

America. Craig (1992) showed that in daytime commercials the primary visual characters were female (60%) but this dropped to 48% in the evening and to 20% on the weekend. Bretl and Cantor (1988) found that 91% of the narrators (i.e., voice-over) were males, which was a result close to other studies done in the 1970s.

Results are consistent across different countries--males are often presented as voice-overs, while women are more likely to be presented visually.

United Kingdom. The first study using "mode of presentation" as an attribute was in the United Kingdom, where Manstead and McCulloch (1981) found that males were more likely to be voice-overs (67.2%) and females were more often presented visually (92.4%).

Europe. Such disproportionate portrayal was also found in Europe in Italy (Furnham & Voli, 1989), where 70% of males were voice-overs and 78.8% of females were presented visually; in Denmark and France (Furnham, Babitzkow, & Uguccioni, 1999) more advertisements were dominated by visually presented central figures, however, males were still more likely (40.2 and 42.9%, respectively) than females to act as voice-overs, while females were often depicted visually (78.1 and 57.1% respectively). In Portugal, only 10% of the voice-overs were done by women and 90% were presented visually (Neto & Pinto, 1998).

Australia. In Australia, Mazzella (1992) showed that though advertisements were dominated by visually presented central figures, females were more often depicted visually (86%) than males (52%), while men were more likely to do voice-overs (48%) than women (14%).

Asia. Sex-role stereotyping is found in the mode of presentation: in Hong Kong, 75.4% of females were depicted visually, while 67.1% of males acted as voice-overs (Furnham, Mak, & Tanidjojo, 1999). In Indonesia, female central figures were often presented visually (66%) and male central figures were more likely to do voice-overs (52.3%).

Credibility Basis

America. In 1975, McArthur and Resko found that 84% of female central figures were depicted as product users, compared to 30% of their male counterparts, while only 14% of females were presented as the authority, compared to 70% of males. Such figures did not decline significantly in 1988, when Bretl and Cantor revealed that 74% of females were still acting as product users in advertisements, and males were more likely to be authoritative central figures (44%) than their female counterparts (26%).

United Kingdom. Manstead and McCulloch (1981) showed that 81.2% females were portrayed as product users, in contrast to 22.4% of males; on the other hand, males comprised a larger proportion as the authority (77.7%) than females (14.8%). Furnham and Skae (1997) found that most of the advertisements consisted of authoritative central figures, and both sexes were equally represented as the authority (83.5% for men, 84.6% for women).

Europe. In Italy, 69.7% of females were product users and 77.6% of males portrayed as the authority. In France and Denmark, most central figures were product users (more than 60% of both males and females); however, females were still more likely than males to be the users (79.7% in Denmark and 86.7% in France) and males were slightly more often the authority (37.9% in Denmark and 33.3% in France). In Portugal 77.5% of the male central characters were depicted as authorities, whereas only 29% of female characters were depicted as authorities.

Mexico. In Mexico (Gilly, 1988), most central figures were depicted as the users of advertised products, yet females were more frequently portrayed as the users (83.7%) compared to males (51.2%) and males were more likely to be the authority (48.8%) than females (16.3%).

Australia. The Australian study (Mazzella, Durkin, Cerini, & Buralli, 1992) showed that females were significantly represented more often as product users (68%), while males were usually portrayed as authoritative central figures (68%).

Asia. This is a study with data from Hong Kong, where men dominated as the authority (76.4%) in the current study, while women were more often the product users (78.5%). Such differences were as significant in Indonesia; 73.8% of males were presented as authoritative central figures while 57.4% of females were delineated as users (Furnham et al. 1999).

Roles

America. Bretl and Cantor (1988) showed that women were more likely to have a dependent role (53%) compared to men (27%). On the contrary, professional roles were more often presented by men (21%) than women (11%). Craig (1992) found males more likely to be portrayed as celebrities and professionals and females as interviewer/demonstrator, parent/spouse, or sex object/model in daytime television. Males were less likely than females to be portrayed as spouse or parent during the day (17 vs 32%) and also in the evening (9 vs 21%).

United Kingdom. Manstead and McCulloch (1981) demonstrated that males often had autonomous roles (89.9%), while females frequently had dependent roles (74.1%). This stereotype was less strong after 16 years, when Furnham and Skae (1997) found that most central figures turned out to be interviewer/narrator. Still, more women were assigned dependent roles (13.5%) compared with men (0%), while men tended to be interviewer/ narrator (80.3%) more often than women (60.4%). However, more women were depicted as professionals (26.1%) than men (19.7%).

Europe. A significant difference was found in Italy, where females were persistently portrayed as having dependent roles (66.7%) and males were more likely to have professional roles (48.6%). Men were also more often the interviewer/narrator (25.2%) than women (10.6%). Sex-role stereotyping in France was stronger than in Denmark: French advertisements showed 63.3% of women with dependent roles and 57.5% of men as professionals; in Denmark, most central figures were depicted as having dependent roles and stereotyping was less strong but still existed--82.8% of women had dependent roles, compared to 66.7% of men. In Portugal 60.3% of males, but only 16% of females, were portrayed as interviewer/narrator, whereas 6.4% of males but 50% of females were portrayed in a dependent role.

Mexico. Mexican advertisements showed more females with dependent roles (46/6%) compared with males (30.4%), and males appeared more often in autonomous roles (51%) than females (23.4%).

Australia. Mazzella et al. (1992) found a significant difference in roles of males and females in Australia: both genders were represented more as autonomous central figures; however, men were more likely than women to have autonomous roles (85%), while women were more likely than men to have dependent roles (44%).

Asia. Women were consistently shown with dependent roles (55.4%) more often than men (19.3%) in Hong Kong, while men were more likely to be interviewer/narrator (65/7%) than women (18.5%), and professional roles were also associated with men more (7.9%) than women (1.5%). In Indonesia, most commercials consisted of interviewers/narrators as central figures; however, men were still more often (84.6%) portrayed as interviewer/narrator than women (72.2%), with more women portrayed in dependent roles (16.7%) than men (0%).

Location

America. Bretl and Cantor (1988) found that males were significantly more often presented outdoors (53%), while females were frequently depicted at home (61%). Craig (1992) found clear time-of-day effects with respect to setting. There were no sex differences in the evening but during the day women primary characters were more likely than men to be portrayed indoors (kitchen, bathroom, other room), while men were likely to be portrayed in outdoor or business locations.

United Kingdom. Manstead and McCulloch (1981) revealed a significant difference in location between portrayals of male and female central figures: females were more likely to be shown at home (38%) than males (7.3%), while males tended to be portrayed in various locations other than the home (83.6%) more often than females (52.2%). Furnham and Skae (1997) did not find a significant difference in portrayal of location.

Europe. Furnham and Voli (1989) showed that in Italian advertisements females were depicted more often at home (34.8%) than males (22.4%), while males were slightly more often to be shown in occupational settings (18.7%) than females (15.1%). Neither Denmark nor France showed a significant difference in portrayals of central figures in different locations. However, in Portugal, Neto and Pinto (1998) found significant differences, with proportionally more females (35%) being portrayed in the home than males (12.7%) and more males being portrayed in occupational settings (37.7%) than females (22.0%).

Mexico. The Mexican commercials did not show a significant difference between males and females across different locations. Both sexes were equally likely to be portrayed at home, at work, and elsewhere.

Africa. The Kenyan study (Mwangi, 1996) suggested that females were consistently portrayed at home (79.4%), in contrast to males (20/.6%), while males were frequently depicted at work, in leisure activities, and in places other than the home.

Asia. In Hong Kong (Furnham et al., 1999), women were more often portrayed at home (36.9%) than men (21.4%), while men were more likely to be in occupational settings (9.3%) than women (6.2%). The Indonesian study showed stronger stereotyping, as more women were depicted in home settings (48.1%) than men (10.8%), while men were frequently associated with leisure settings (83.1%) than women (48.1%).

Age

America. Neither of the American studies attempted a clear classification of the age of the central characters.

United Kingdom. Furnham and Skae (1997) showed that most central figures were middle-aged; however, women were still more likely to be younger than men (28.1 vs. 5.7%), while male central figures were more frequently middle-aged (88.6%) than females (71.9%).

Europe. French advertisements showed significant differences between the ages of central figures. Female central characters tended to be young (63.3%), while most male central figures were middle-aged (49.4%). There were also more old males (12.6%) than old females (7.1%). Danish commercials were stereotyped but to a lesser extent: 59.4% of female central figures were young, while 57.5% of male central figures were middle-aged; more old males (11.5%) were shown than old females (7.8%). In Portugal 56.0% of the female figures were coded as young, while only 13.7% of the male figures were coded as young. On the other hand, 81.9% of male characters were coded as middle-aged, while only 40.0% of female characters were coded as middle-aged.

Mexico. Gilly (1988) revealed a consistent pattern toward sex-role stereotype in terms of age: 70.8% of females were young, compared with 20.6% of males, while 72.5% of male central figures were middle-aged, in contrast to 23.3% of females.

Australia. In Australia, most female central characters appeared to be younger (62%), while most male central figures tended to be middle-aged (51%).

Asia. Significant differences were found in both Hong Kong and Indonesia: female central figures had a higher likelihood of being young (56.9 and 74.1%, respectively), while male central figures were frequently middle-aged (77.1 and 61.5%, respectively). Both studies did not find old female central figures and found 1.5% old male central figures.

Argument

America. Only McArthur and Resko (1975) demonstrated that females were significantly more likely than males to make no argument. Bretl and Cantor (1988) found no such difference between males and females in terms of the arguments they made.

United Kingdom. Manstead and McCulloch (1981) found females significantly more likely not to argue (63.4%) than males (19.8%). Males were more likely to make both scientific and nonscientific arguments than females. In 1997, no differences were found between arguments made by the two genders.

Europe. Furnham and Voli (1989) found sex-role stereotyping in terms of argument in Italy: females were more often associated with no arguments (53%), while males were often associated with factual arguments (43.9%) than females (25.8%). In France and Denmark, most advertisements gave neither a factual argument nor opinion; however, French commercials had more females giving opinions (23.5%) than males (10.3%) and more males giving factual argument (13.8%) than females (5.1%). No significant difference was found in Denmark. In Portugal, Neto and Pinto (1998) found that twice as many male (23.5%) as female (11.0%) central characters presented factual arguments, while half as many males (22.5%) as females (44.0%) provided no argument in favor of the product.

Australia. In Australia most central figures tended to give opinions; however, males were more often associated with factual arguments (19%) than females (5%), with females more often making no arguments (29%) than males (13%).

Asia. In Hong Kong no differences were found between types of argument. On the other hand, Indonesian commercials showed more females giving opinions (74.1%) and more males giving factual arguments (52.3%).

Reward Type

United Kingdom. Manstead and McCulloch (1981) found very significant differences in that males were often associated with rewards that were practical (74%) and more females depicted with rewards aimed at self-enhancement (40.2%). In 1997 the difference was significant, but to a lesser extent: females were more likely to be associated with self-enhancing rewards (24%) than males (12.7%), but males frequently were portrayed with pleasurable rewards (44.7%) than females (19.7%). However, women were more often shown with practical rewards (56.3%) than men (42.6%).

Europe. Results in Italy indicated that females were more often depicted with socially approved rewards (36.4%) than males (15.9%), and females were also slightly more, frequently portrayed with practical rewards (24.2%) than males (21.5%). On the contrary, men were more likely to be associated with self-enhancing rewards (28.9%) compared with women (24.2%), and men were also shown more with pleasurable rewards (16.8%) than women (4.5%).

In Denmark, significant differences were found between different reward types: males were more often portrayed with pleasurable rewards (50.6%), while females were frequently associated with self-enhancing rewards (42.2%). In Portugal less than a tenth (10.3%) of males, but over a third (37.0%) of females gave self-enhancement as a reward type. On the other hand, half (50%) of males but less than a third (30%) of females provided a pleasure/other reward type explanation for their advertised product or service.

Australia. Mazzella et al. (1992) obtained significant results in Australia, showing that females were more often shown with socially approved and self-enhancing rewards (26 and 18%, respectively) than males (14% in both cases). On the other hand, male central figures were frequently associated with practical rewards (51%) compared with female central figures (30%).

Asia. Results in Hong Kong revealed significant differences: women were highly likely to be portrayed with self-enhancing rewards (60%), while men were more likely than women to be shown with rewards that are practical (17.1 vs. 12.3%) and pleasurable (26.4 vs. 20%).

The Indonesian study also showed that females were more likely to be associated with self-enhancing rewards (59.3%), while men were more often portrayed with rewards that are pleasurable (67.7%).

Product Type

America. McArthur and Resko (1975) found that females were more likely to advertise domestic products (76.1%) compared with their male counterparts (56.5%). Bretl and Cantor (1988) showed a significant difference between males and females over types of products: females were more frequently shown with domestic products (86%) than males (66%), while males were more often selling products that are not used at home (17%) than females (6%). Craig (1992, p. 205) found dramatic differences as a consequence of when the advertisement was portrayed: "Men were most likely to be primary characters in food or body commercials during day time and prime time, but equally likely to be primary characters in automotive or business product/service ads on weekends."

United Kingdom. Results from Furnham and Skae (1997) suggested that women were more frequently portrayed with body products (35.45) as opposed to men (10.5%); however, men were shown more often with food products (43.9%), and they were more likely to sell auto/sports products (7%) than women (2.1%).

Europe. The Italian study revealed a significant difference in product types between men and women: women were highly likely to be shown with body products (50%) in contrast to men (30.8%); however, males were more likely than females to advertise food products (30.8 vs. 18.1%) and they were more likely to sell auto/sports products (27.1 vs. 18.1%).

Results in France demonstrated that women were frequently depicted with body and home products (24.5 and 20.4%, respectively) as opposed to men (2.3 vs. 6.9%, respectively). However, men were often advertising food products and auto/sports products (35.6 and 55.2%, respectively). In the Portuguese study few males (7.8%) advertised body products (30.0% for females), while few females (5.0 vs. 20.6% for males) advertised auto/sport products. Females also appeared in food ads (33.0%) more often than males (22.5%). Results in Denmark did not show any significant differences.

Australia. In Australia, female central figures were more likely to be shown with food products (42%) compared with male central figures (28%), and females were also more often depicted with body products (17%) than males (10%); on the other hand, males were more often portrayed with auto/sports (28%) products compared to females (19%).

Asia. The Hong Kong study indicated that more females were depicted with body and home products (35.4 and 15.4%, respectively) compared to their male counterparts (20.7 and 9.3%, respectively). However, male central figures advertised slightly more food products (30.7%) than female central figures (27.7%).

Results in Indonesia revealed significant differences, suggesting that females were more likely to sell body products (59.3%) than males (32.3%) and slightly more likely to advertise home products (11.2 vs. 7.7%) and food products (20.4 vs. 16.9%); however, men were more frequently associated with auto/sports products (15.4%) than women (0%).

Background (Setting)

United Kingdom. Furnham and Skae (1997) showed no significant differences between males and females against various backgrounds.

Europe. In Italy, Furnham and Voli (1989) demonstrated that males were more likely to be shown against a female background (25.2%) (i.e., home) than were females (10.6%), and more likely than women to be shown against a male background (i.e., work) (13.1 vs. 9.1%). On the other hand, females were more likely to be depicted with a background of children (9.1%) compared with males (2.8%). Furnham et al. (1999) showed a significant difference in France, where males were more often portrayed against a female background (16.1%) as opposed to females (5.1%), while women were more often shown against a male background than their male counterparts (17.4 vs. 14.9%). Females were also more likely to be seen with children. No significant differences were found in Denmark. There were dramatically significant differences in Portugal. Half the females (56.0%) were portrayed against a female background (compared to 3.9% of the male characters), while just under half of the male central characters (44.1) were portrayed agai nst a male background (compared to 4.0% for female characters).

Australia. Mazzella et al. (1992) showed no significant differences between males and females against different backgrounds.

Asia. Females were more likely than males to be shown against a female background in Hong Kong (21.5 vs. 13.6%); also, they were shown more often with children (12.3 vs. 9.3%). On the contrary, males were often depicted against a male background and a mixed background (23.6% and 25.7% respectively). No significant differences were found in the Indonesian study.

End Comment

United Kingdom. Furnham and Skae (1997) indicated that although most central figures made an end comment in an advertisement, males were still more likely to make an end comment (82%) compared to females (60.4%).

Europe. The Italian study demonstrated that males were three times more likely than females to make an end comment (68.2 vs. 22.7%), while females often did not make an end comment (77.2%). Results from Denmark showed that most central figures made an end comment; however, males more frequently made an end comment (93.1%) compared to females (78.1%). Commercials in France tended not to have central figures making end comments, so results were not significant. Results from the Portuguese content analysis was very clear: when the advertisement did have an end comment, males were much more likely to make it (59.6 vs 13.01%).

Asia. Results in Hong Kong were highly significant: men frequently made an end comment (63.6%) as opposed to women (36.9%). In Indonesia, results were not significant. Both females and males were as likely to give an end comment.

DISCUSSION

Table II indicates that sex-role stereotyping tends to be surprisingly consistent across different countries during these 25 years. Perhaps the most consistent difference across the analyses was sex difference in the mode of presentation. Mode of presentation sex-role effects occur in 9 of 11 studies: males are consistently more likely voice-overs of an advertisement, while females are often visually portrayed.

Two patterns appear from the results of the content category "credibility." The first one shows that males are frequently shown as the authoritative central figures, whereas females are more likely to be users of the advertised products: McArthur and Resko (1975) and Manstead and McCulloch (1981) showed this pattern in the study of America and the United Kingdom respectively, while studies in Italy, Australia, Hong Kong, Indonesia, and Portugal also showed such a pattern. The second pattern refers to advertisements that comprise mostly central figures as users of a product: although both males and females appear more as users, men are still relatively more likely than women to act as authoritative central figures, whereas women tend to be depicted relatively more as users. Studies in America by Bretl and Cantor (1988), in Mexico by Gilly (1988), in Denmark and France by Furnham et al. (1999), and in Portugal by Neto and Pinto (1998) all showed the second pattern. These two patterns serve one central claim: m ales are more likely than females to act as the authority of a product.

The roles played by male and female characters showed two patterns: the first pattern showed that males are frequently portrayed as interviewers or professionals while females are more likely to have dependent roles. Studies in the United States, the United Kingdom, Mexico, Italy, France, Hong Kong, and Portugal all show such patterns. The second pattern shows that one particular category is often dominated by both sexes, but still men are more often depicted in autonomous roles as opposed to women; studies done in the United Kingdom by Furnham and Skae (1997), in Australia, and in Indonesia indicate that though most central figures are professionals (interviewer/narrator), men are more likely to be portrayed as professionals, while women are more often depicted as dependent. In Denmark, most central characters are shown as dependent; however, females are relatively more often shown as dependent and males are more often professionals. These two patterns suggest that despite the different proportions of autonomous or dependent roles the advertisements have, males often dominat e as interviewer or professional expert (including celebrity), while females are more likely to be confined to dependent roles (parent, spouse, homemaker).

For location, the patterns between countries are more complex. There are two general patterns. First, females are more often portrayed at home while males are more frequently portrayed during leisure/outdoor. Studies such as those by Bretl and Cantor (1988) in the United States, by Mwangi (1996) in Kenya, and in Hong Kong display this pattern. Although studies in the United Kingdom, Mexico and France did not find significant differences, they do show a similar pattern. The second pattern also shows women as frequently depicted at home, while men are shown in various "other" locations: Manstead and McCulloch (1981) show this pattern in the United Kingdom. It was also found in Australia, Italy, Indonesia, and Portugal. The two patterns, in general, suggest that women are more often confined to a domestic setting, while men are shown in diverse locations. Location is still a significant predictor of gender stereotyping, although results from different countries vary slightly. Interestingly "occupation" is found not to be a good indicator of gender stereotyping, as some studies show a similar proportion of men and women portrayed in occupational settings (e.g. Furnham & Skae, 1997; Gilly, 1988; Manstead & McCulloch, 1981).

Age is often one of the best indicators of sex-role stereotyping. Although studies define "young," "middle-age," and "old" on slightly different scales, a prevalent picture is indicated: females are consistently shown as younger than males. Most studies show that central figures are dominated by middle-aged males and young females. Only one study in the United Kingdom (Furnham & Skae, 1997) shows central figures dominated by middle-aged males and females.

The content category "reward types" showed many (9 of 14) genderrole effects. The general pattern is that males are shown to be associated with pleasurable rewards, while females advertise more products that yield social/self-enhancement. All nine studies show that females are associated more often with social approval and/or self-enhancement. The category "practical" is not a strong indicator of sex stereotyping, as different countries tend to show a diverse pattern. Some show that males are associated more often with practical rewards (e.g., Manstead & McCulloch, 1981), while others show the reverse pattern (Furnham & Voli, 1989).

Advertisements that sell home and body products are consistently shown to be female-dominated: nine studies that include product type indicate this pattern. Males, on the other hand, are more frequently shown selling automobiles/sports products than females. For "food," different countries show various patterns. One can conclude that females are more likely to advertise products that are used at home (such as domestic and body products), whereas males are often shown with products "away from home."

Two patterns are found in the "end comment." First, males more frequently offer an end comment in an advertisement, whereas females frequently do not give any end comment (Furnham & Voli, 1989). Second, most central figures appear to make an end comment, yet males are relatively more likely to give the comment as opposed to females (e.g., Furnham & Skae, 1997). The only exception is France, which demonstrates a reverse pattern. On the whole, end comment is still highly indicative of sex-stereotyping.

Two attributes, however, seem to show less sex-role stereotyping. The picture is particularly complicated in the "argument" category. There is no single consistent pattern across different countries in the past 25 years. Earlier studies of sex stereotyping in television commercials show that females are more likely than males not to make an argument for the product (McArthur & Resko, 1975; Manstead & McCulloch, 1981). The study in Italy and Portugal also suggests that males more often make an argument (no matter factual or opinion) compared to females. In Australia most central figures prefer to give an opinion; while males are relatively more often making a factual argument, females tend not to make any argument. Studies in France and Denmark reveal that most central figures appear not to make an argument in an advertisement. In Britain, Furnham and Skae (1997) showed a reverse pattern which is not statistically significant: females make more factual arguments compared to males. The only study which shows a significant sex stereotyping is Indonesia, where males more frequently give a factual argument and females give an opinion. Since different studies denote various results, it is unlikely to conclude that argument is a good indicator of sex stereotyping.

Finally, it is difficult to draw any conclusions from the "background" factor. Some results show that males are more often depicted against a female background, whereas females are shown against a male background [e.g., in France (Furnham et al., 1999); others show that males are more likely to be shown against a male background and females against a female one (e.g., Furnham & Skae, 1997; Neto & Pinto, 1998); the Italian study even indicates that males are more frequently depicted against a female, male and mixed background compared to females. The only consistent pattern shown is that females are often shown against a children's background.

Few studies have attempted to delineate the precise circumstances (i.e., types of commercials) in which men and women are more likely to be depicted more equally. Furnham and Skae (1997) did note a number of specific incidences of sex-role stereotype reversal, where men are portrayed as the dependent, perhaps unintelligent user of a product and women were clearly portrayed as independent, professional, and authorities on the product benefits. Two of these products were for domestic cleaners and two for food, and nearly always humor was involved. Indeed it may be that currently the only way advertisers are happy to see sex-role reversal or even sex-role equivalence is by using humorous settings and scripts that imply that it is unusual.

However, as Furnham and Skae (1997) in Britain and Ferrante, Haynes, and Kingsley (1988) in America have noted in their small-scale comparative studies, there do appear to be small, but stable and detectable trends in gender-role advertising. Women are being depicted less as dependent and in a domestic setting, which no doubt reflects global statistics on the increase in the number of working women. However, some of the sex differences seem impervious to change such as the fact that, on average, female central characters nearly always seem to be considerably younger than their male counterparts.

Studies done in more traditional societies show, not surprisingly, that sex-role stereotyping is stronger (Neto & Pinto, 1998). Examining commercials in both Hong Kong and Indonesia, Furnham et al. (1999) found sex-role differences in 90% of the coding categories. This was even greater than was found in the original studies done in America over 25 years ago. Those two Asian countries differ significantly in size, history, geography, and religion but, like many Asian countries, have been slower to adopt ideas of sexual equality popular in the west (Williams & Best, 1982). This may well be true of most Asian countries and fits with the data on sex-role stereotyping.

This study compared similar content analytic studies with each other and with important earlier studies. The analysis is a step toward understanding the general picture of sex-role stereotyping in television commercials: in the 1970s the commercials typically show men as authoritative and knowledgeable, whereas women are confined at home. Such pictures have not changed much over the 25 years, although for some attributes men and women are depicted more equally. However, the study by Furnham and Skae (1997) does show a change in this pattern; advertisements are much less sex stereotyped than previous European studies. This may imply a decline in sex-role stereotyping in the Western world, in particular, Britain. Gunter (1995, p. 50) noted in his recent critical review of many studies looking at gender-role portrayals in advertisements directed to adults and to children that, from the late 1980s onward, advertisements did seem less sex-role stereotyped: "More advertisements emerged featuring women in central, independent roles, assuming greater degrees of control over the immediate situation in which they were depicted and more generally over their own lives. As yet, this new pattern in gender-role portrayals has been visible in advertisements aimed at adults, while recent research has indicated that advertisements aimed at children have remained as gender-stereotyped as ever." Current studies in Asia suggest that sex stereotyping in television commercials is not declining there and is much stronger. than in Europe (Furnham et al., 1999). This indicates that the general pattern over the world is not identical; though in the Western world sex stereotyping appears to decline, this is not the case in Asia and possibly not in Africa (Mwangi, 1996). Some Asian and African countries are not nearly as technologically advanced as Western countries and comparatively new to television. Thus South Africa, the most developed country in Africa, has had television for barely 25 years. The fact that many have more stereotypical gender-role portrayal than in the West may be another indication of their lagging behind the Western world in issues with respect to gender equity. Perhaps future studies in this area could look at the relationship between cultural values (individualism--collectivism, masculinity--femininity) and sex-role stereotyping. This would allow one to develop and test quite specific hypotheses about where and why differences in sex role-stereotyped television commercials would occur.

(1.) To whom correspondence should be addressed at Department of Psychology, University College London, 26 Bedford Way, London WC1, Great Britain. Fax: 020 7436 4276. e-mail: ucjtsaf@ucl.ac.uk.

REFERENCES

Bretl, D. J., & Cantor, J. (1988). The portrayal of men and women in U.S. television commercials: A recent content analysis and trends over 15 years. Sex Roles, 18, 595-609.

Bryant, J., & Zillman, D. (1994). Media effects. Hillsdale, NJ: Lawrence Erlbaum.

Craig, R. (1992). The effect of television day pay part on gender portrayals in television commercials: A content analysis. Sex Roles, 26, 197-211.

Dominick, R. J. (1979). The portrayal of women in prime time, 1953-1977. Sex Roles, 5, 405-411.

Dominick, R. J., & Rauch, G. (1972). The image of women in network TV commercials. Journal of Broadcasting, 17, 259-266.

Durkin, K. (1986). Sex roles and the mass media. In D. Hargreaves & A. M. Colley (Eds.), The psychology of sex roles. (pp. 201-214). London: Harper & Row.

Durkin, K., & Nugent, B. (1998). Kindergarten children's gender-role expectations for television actors. Sex Roles, 38, 387-402.

Embree, A. (1970). Media Images. I: Madison Avenue brainwashing--The facts. In R. Morgan (Ed.), Sisterhood is powerful: An anthology of writings from the Women's Liberation Movement. (pp. 175-191) New York: Vintage Books.

Ferrante, C., Haynes, A., & Kingsley, S. (1988). Images of women in television advertising. Journal of Broadcasting and Electronic Media, 32, 231-237.

Frank, R. E., & Greenberg, M. G. (1980). The public's use of television: Who watches and why. Beverley Hills, CA: Sage.

Furnham, A. (1993). Reaching for the counter: The new child consumers. London: Social Affairs Unit.

Furnham, A., & Bitar, N. (1993). The stereotyped portrayal of men and women in British television advertisements. Sex Roles, 29, 297-310.

Furnham, A., & Farragher, E. (1999). A cross-cultural content analysis of sex-role stereotyping in television advertisements.: A comparison between Britain and New Zealand. Journal of Broadcasting and Electronics Media, in press.

Furnham, A., & Skae, E. (1997). Portrayals of men and women in British television advertisements. European Psychologist, 2, 14-51.

Furnham, A., & Thomson, L. (1999). Gender role stereotyping in advertisements on two British radio stations. Sex Roles, 40, 153-161.

Furnham, A., & Voli, V. (1989). Gender stereotyping in Italian television advertisements. Journal of Broadcasting and Electric Media, 33, 175-185.

Furnham, A., Abramsky, S., & Gunter, B. (1997). A cross-cultural analysis of children's television advertisements. Sex Roles, 37, 91-99.

Furnham, A., Mak, T., & Tanidjojo, L. (1998). An Asian perspective on the portrayal of men and women in television advertisements: Studies from Hong Kong and Indonesian television. Journal of Applied Social Psychology, in press.

Furnham, A., Babitzkow, M., & Uguccioni, S. (1999). Gender stereotyping in television advertisements: A comparative study of French and Danish television. Genetic, Social and General Psychology Monography, in press.

Gilly, M. C. (1988). Gender roles in advertising: A comparison of television advertisements in Australia, Mexico, and the United States. Journal of Marketing, 52, 75-85.

Gunter, B. (1995). Television and gender representation. London: John Libby.

Gunter, B., & Svennevig, M. (1987). Behind and in front of the screen. London: John Libbey.

Gunter, B., & Wober, M. (1983). Violence on television: What the viewers think. London: John Libbey.

Harris, R. J. (1994). The impact of sexually Explicit media. In J. Bryant & D. Zillman (Eds.), Media effects: Advances in theory and research. (pp. 273-314). Hillsdale, NJ: Erlbaum.

Harris, P. R., & Stobart, J. (1986). Sex-role stereotyping in British television advertisements at different times of the day: An extension and refinement of Manstead & McCulloch (1981). British Journal of Social Psychology, 25, 155-164.

Hawkins, D. J., & Coney, K. A. (1976). Advertizing and Differential sex roles in contemporary American society. Journal of the Academy of Marketing Science, 4, 418-428.

Hurtz, W., & Durkin, K. (1997). Gender role stereotyping in Australian radio commercials. Sex Roles, 36, 103-114.

Iijima Hall, C., & Crum, M. (1994). Women and "body-isms" in television beer commercials. Sex Roles, 31, 329-337.

Jaffe, L., & Berger, P. (1994). The effect of modern female sex-role portrayals on advertising effectiveness. Journal of Advertising Research, 231, 32-39.

Jennings, J., Geis, F., & Brown, V. (1982). Influence of television commercials on women's self-confidence and independence judgement. Journal of Personality and Social Psychology, 38, 203-210.

Lanis, K., & Covell, K. (1995). Images of women in advertisements: Effects on attitudes related to sexual aggression. Sex Roles, 32, 639-649.

Leppard, W., Ogletree, S., & Wallen, E. (1993). Gender stereotyping in medical advertising: Much ado about something? Sex Roles, 29, 829-839.

Liebler, C., & Smith, S. (1997). Tracking gender differences: A comparative analysis of network correspondents and their sources. Journal of Broadcasting and Electron Media, 41, 58-63.

Livingstone, S., & Green, G. (1986). Television advertisements and the portrayal of gender. British Journal of Social Psychology, 25, 149-154.

Looney, G. (1971). Television and the child: What can be done? Position paper presented to Section on Child Development, the American Academy of Paediatrics, Chicago, IL.

Lovdal, L. (1989). Gender roles messages in television commercials: An update. Sex Roles, 21, 715-724.

Manstead, A. S. R., & McCulloch, C. (1981). Sex-role stereotyping in British television advertisements. British Journal of Social Psychology, 20, 171-180.

Maracek, J., Piliavin, J., Fitzsimmons, E., Krogh, E., Leader, E., & Trudell, B. (1978). Women as TV experts: The voice of authority. Journal of Communications, 28, 159-168.

Mazzella, C., Durkin, K., Cerini, E., & Buralli, P. (1992). Sex role stereotyping in Australian television advertisements. Sex Roles, 26, 243-259.

McArthur, L. Z., & Resko, B. G. (1975). The portrayal of men and women in American television commercials. Journal of Social Psychology, 97, 209-220.

Millum, T. (1975). Images of women: Advertising in women's magazines. London: Chatto & Windus.

Mischel, W. A. (1966). A social learning view of sex differences in behaviour. In E. E. Maccoby (Ed.), The development of sex differences. (pp. 56-81). Stanford, CA: Stanford University Press.

Moore, M. (1992). The family as portrayed on prime-time television, 1947-1990: Structure and characteristics. Sex Roles, 26, 41-61.

Mwangi, M. (1996). Gender roles portrayed in Kenyan television commercials. Sex Roles, 34, 205-214.

Neto, F., & Pinto, I. (1998). Gender stereotypes in Portuguese television advertisements. Sex Roles, 39, 153-164.

O'Donnell, W., & O'Donnell, K. (1978). Update: Gender role message in TV commercials. Journal of Communication, 28, 156-158.

Olson, B., & Douglas, W. (1997). The family on television: Evaluation of gender-roles in situation comedy. Sex Roles, 36, 409-419.

Pracash, V. (1992). Sex roles and advertising effectiveness. Journal of Advertising Research, 21, 43-53.

Schneider, K. C., & Schneider, S. B. (1979). Trends in sex roles in television commercials. Journal of Marketing, 43, 79-84.

Smith, L. (1994). A content analysis of gender differences in children's advertising. Journal of Broadcasting and Electronic Media, 38, 323-337.

Smith, P., & Bond, M. (1993). Social psychology across cultures. Hemel Hempstead, Hertfordshire: Harvester Wheatsheaf.

Suezle, M. (1970). Women in labour. Trans-action, 8, 50-58.

Vande Berg, L., & Streckfuss, D. (1992). Prime-time television's portrayal of women and the world of work: A demographic profile. Journal of Broadcasting and Electronic Media, 36, 195-208.

Welch, R., Huston-Stein, A., Wright, T., & Phelial, R. (1979). Subtle sex-role cues in children's commercials. Journal of Communication, 29, 202-209.

Williams, J., & Best, D. (1982). Measuring sex stereotypes: A thirty nation study. Beverly Hills, CA: Sage

Zajonc, R. (1968). Attitudinal effects of mere exposure. Journal of Personality and Social Psychology, 9, 1-29.

   Content Analysis of Television Commercials Across Different Countries
                              (1975-1997)
                                        Month & year
         Reference          Sample size  of sample     Source of sample
McArthur & Resko (1975)         199     Spring 1971    America
Bretl & Cantor (1988)           397     Jan.-Feb. 1985 America
Manstead & McCulloch (1981)     170     July 1979      United Kingdom
Furnham & Skae (1997)           109     Nov. 1995      United Kingdom
Gilly (1988)                    204     Feb. 1985      Mexico
Furnham & Voli (1989) [a]       175     Sept. 1987     Italy
Mazzella et al. (1992)          281     Aug. 1989      Australia
Craig (1992) [b]               2209     Jan. 1990      America
Mwangi (1996)                   105     July-Aug. 1996 Kenya
Furnham et al. (1999)           151     Nov. 1995      Denmark
Furnham et al. (1999)           211     Oct. 1995      France
Furnham et al. (1999)           119     Summer 1997    Indonesia
                                175     Summer 1997    Hong Kong
Neto & Punto (1998)             304     June 1996      Portugal
(a.)Only the evening sample is chosen for analysis
because of inability to combine all three times of day.
(b.)The results of this study are not tabulated because of
the way they were reported in the paper but are fully
discussed in the text.

COPYRIGHT 1999 Plenum Publishing Corporation
COPYRIGHT 2000 Gale Group