Lois J Smith [1994]: 'A Content Analysis of Gender Differences in Children's Advertising'

Journal of Broadcasting and Electronic Media, 38(3), 323-37

A Review

Merris Griffiths

A substantial amount of research has been conducted regarding the ways in which children acquire so-called gender appropriate behaviours, and this journal article is particularly relevant to my own research in that it considers the possible influences of television advertising upon a child's perception and subsequent re-enactment of gender role-play. Smith places her argument in the context of social learning theories, discussing the possible influences upon sex-roles in terms of family and peers as well as the media. According to Bandura [1969], social learning theory indicates that children learn their appropriate sex-roles through observing the actions of others, as well as being either rewarded or punished for acting appropriately or inappropriately. That is to say that children learn personality and behaviour patterns through the imitation of their own parents' attitudes and behaviours. It has also been observed by Bandura that children will acquire the patterns of behaviour more rapidly where there is an attractive model whose behaviour is rewarded [Smith, 1994: 324].

Peirce [1989] argues that television teaches children a great deal about sex-typed behaviours, simply because it brings an abundance of readily observable models into the child's own home. However, it is important to acknowledge that television programming is not the only source of gender-example, since advertisements present to children any number of gender-appropriate behaviours that are lavished with many extravagant rewards [Smith, 1994: 324]. Schneider [1989] quotes a particularly worrying statistic in stating that during their viewing, children aged between two and eleven years of age are exposed to approximately 150 to 200 hours of television advertising each year [Smith, 1994: 323]. More often than not, normal social interaction between friends will reinforce what is regarded as appropriate behaviour for boys and girls. The former reinforce one another through playing with action figures, given the chance to be aggressive and win in competitive games, while the latter have fun through playing with dolls, with the opportunity to be loved when they cuddle their toys or stuffed animals [Smith, 1994: 323].

Macklin and Kolbe [1984] argue that the characters seen on television are attractive models for children to emulate because they are often physically beautiful/ handsome, commanding impressive influence over their environments. Meanwhile research by Courtney and Whipple [1983] indicates that children are far more likely to imitate the behaviour of same-sex models than opposite-sex models, either due to reward from parents and peers, or due to more accurate recollection about the actions of their own gender. Furthermore, Bandura suggests that a child's exposure to televised models of behaviour is often more prevalent than their exposure to parental behaviour! Such research evidence only serves to emphasise the need for further investigation into the degree to which children are effected by the advertisements they watch ever day of their lives. Indeed, Courtney and Whipple found a positive correlation between the number of hours of television viewing and sex-typed answers given to subsequent tests, so one could argue that the possible influence of television is too powerful to ignore.

Smith concludes that the effect of children's social learning from advertisements is that they are basically shown how to behave, eventually accepting without question the assumed images as 'real', taking cues regarding appropriate gender behaviour [1994: 325]. Smith recalls the statement made by Erving Goffman [1979], who argues that advertisements show how males and females are different and how they act in relation to one another. Therefore, Smith structures the article in two layers. Firstly, she emphasises the importance of studying the behavioural differences between boys and girls, before secondly examining how advertisements portray the sexes in different ways.

Gender differences in children's reactions to advertisements

The most basic of research has revealed that males and females respond differently to the images they see in television advertisements. Barthel [1988], for example, notes that it is regarded as more effective to target men with advertisements classed as neutral, because males will not use products even vaguely regarded as 'feminine', while women are more likely to use those products normally classed as 'male' [Smith, 1994: 325]. Since these attitudinal features are apparent in adulthood, I feel it is reasonable to presume that such views are likely to be formed, by the viewer, in childhood. Hence, the issues of gender portrayal in advertising remain at the centre of the debate.

Smith refers to an interesting study of child-play conducted by Smetana and Letourneau [1984] that appears to echo Barthel's view of adult purchasing patterns in terms of gendered products, with males preferring male-specific products in contrast with more flexible female attitude. Their study revealed that when boys were playing in all-male groups, they consistently played with male rather than female sex-typed toys. Girls, in contrast, were seen to play with more female-type toys. The most interesting findings emerged when the children were observed playing in mixed-sex groups, since the boys were completely unwilling to play with female-type toys whereas the girls had no objection to playing with toys that would generally be regarded as 'male'. Surely advertising companies much take these attitudes into account when creating their toy advertisements. With these behavioural differences in mind, Smith proceeds in the study of advertisements to assess whether such differences may be observed in terms of gender presentation on the screen.

Content analysis

In brief, content analysis is the study of the frequency with which certain identifiable elements in a given set of advertisements occur. For reasons of clarity, Smith analyses the content of advertisements according to a series of four different hypotheses. These will be listed below to include a synopsis of the issues raised:

Hypothesis One: More single-sex advertisements will be positioned towards boys than girls [Smith, 1994: 326]

Smith cites a number of studies concerned with the frequency of gender representation in advertisements [1994: 326]. For example, O'Kelly [1974], in examining seven hours of children's television programmes and their accompanying advertisements, found that 67% of all the characters shown were male. Similar discoveries were made by Doolittle and Pepper [1975]. They studied Saturday morning advertisements and found that 87% of the single-gender advertisements were male only. The advertisements using only female models were almost exclusively for girl-oriented products. A more recent study by Macklin and Kolbe [1984], seen to echo a similar analysis by Vernon in 1975, found that there were almost three times as many male as female-oriented advertisements aimed at children. Hence, the issue of product positioning is an important factor in terms gender representation.

Hypothesis Two: The sex of the narrators' voices will correspond to the gender positioning of advertisements [Smith, 1994: 327]

It has been acknowledged a well known fact that male voice-overs have occurred more frequently as announcers or narrators than have female voices, and a number of studies have been conducted in this field. Courtney and Whipple [1983], in their review of a number of content analyses, discovered a far higher frequency of male announcers. Indeed, these male announcers even featured in advertisements targeted exclusively at women. Bretl and Cantor [1988] confirmed these findings in their own content analysis of adult advertisements, concluding that there were more male than female narrators. They also add a further dimension to the situation by noting that there was a relationship between the sex of the primary characters and the sex of the narrator. That is to say, a female announcer is more likely to narrate advertisements directed at women, comprised of women characters.

Since my own research is primarily concerned with the construction of advertisements directed at children, I find the results of a 1979 study by Welch, Huston-Stein, Wright and Plehal extremely interesting and relevant. They found that in many cases, even if the product was regarded as 'female', most girl-positioned advertisements used male narrators and voice-overs. This raises many questions about how advertising agencies are more inclined to employ male voices as a more convincing recommendation and validation for the quality of their product, hence perpetuating the notion that 'men know best'. It may even be that female products are positioned as being inferior to products aimed exclusively at men. Doolittle and Pepper [1975] found that female announcers voices occurred only for girls' toys, seeming to imply that girls' products are not impressive enough to warrant the recommendation of a man. Macklin and Kolbe [1984] also found that, of female-oriented advertisements, 42.8% used female-only narrators, compared with 61.6% of male-oriented advertisements using male-only narrators.

Hypothesis Three: Advertisements positioned towards girls will show models engaged in more passive and fewer physical and antisocial activities than those models shown in boys' advertisements [Smith, 1994: 328]

Perhaps one of the most effective ways to assess the level of activity in an advertisement is through an analysis of content. Indeed, such studies have often shown a pronounced difference in the levels of activity for male and female actors in children's advertisements, and the results are noted on an active-passive scale [Smith, 1994: 327]. One of the most controversial aspects of physical activity in advertisements is that of violence. Welch et al. [1979] studied the aggressive behaviour displayed in child characters taken from a sample of sixty morning-time toy advertisements and the results are concerning. They discovered that aggression was found to be limited almost exclusively to advertisements aimed at boys, with these advertisements also containing higher levels of action and movement than in those aimed at girls.

This result is comparable with an earlier study of content by O'Kelly [1974], in which it was discovered that the most common activities for girls were traditional, such as playing house, cooking and so on. Only 10% of the girls portrayed were seen to be active. In turn, the boys also acted out traditional gender roles such as playing football and fishing. Indeed, the situation changed very little from the 1970s to the 1980s, since the 1983 study by Courtney and Whipple also found traditional gender roles to be the norm. Verna [1975] found that girls were more passive and less aggressive and competitive than boys, while Macklin and Kolbe [1984] found that 16.6% of boy oriented advertisements contained aggressive behaviour.

Hypothesis Four: Girls' advertisements will use more in-home settings; boys' advertisements will use more out-of-home settings [Smith, 1994: 329]

Of the research highlighting this issue, the results have been as one might predict. Bretl and Cantor [1988] found that in advertisements aimed at adults, males appeared away from the home setting far more than females did. This result echoed that of an earlier study by Schneider and Schneider [1979], who discovered that 34% of women appeared in the home, with only 22% of men in a similar setting. It would be reasonable to presume that similar representations are to be found in advertisements aimed at children.

The sample

Having identified the areas Smith regards as most important, she proceeds to recount the methods used to collect her sample. The sample consisted of advertisements recorded during a typical week of children's television programming. She emphasises that, with the sample taken from the second week of February, she has avoided the excessive number of toy advertisements characterising the pre-Christmas period, while at the same time still within the winter months that characterise heavier viewing figures for children [according to Nieben Media Research, 1992] [Smith, 1994: 329]. The times in which the taping of advertisements occurred included after-school hours on weekdays [essentially 3p.m. to 5.30p.m.] and Saturday mornings [7a.m. to 11p.m.]. Smith further adds that the number of times each advertisement appeared was noted, but the actual study did not include any repetitions. The study also focused on those advertisements aimed at a single sex, excluding the analysis of advertisements that may be classed as uni-sex or neutral; a detail that I intend to include in my own study, so as to keep my conclusions focused on the gender issues in question.

The coding procedure

Smith warns that attempting to define gender-positioning in advertisements is challenging, working on the notion that such advertisements included real and/or animated characters of only one sex [Smith, 1994: 330]. Smith catalogues a range of other definitions suggested by fellow researchers in the field. For example, Bretl and Cantor [1988] defined the gender-positioning of an advertisement according to whether the primary characters in that advertisement were male or female, with the primary characters identified as having the greatest amount of screen exposure. Macklin and Kolbe [1984] considered advertisements in a similar way, observing whether they were primarily male or female oriented according to the prevalence of one or another sex on the screen, or: "...when those characters were involved in the major product-related activity" [Smith, 1994: 330]. As far as research with children is concerned, Ruble et al. [1981] found that children are more likely to model their behaviour directly from advertisements using only one sex.

The actual coding of variables was achieved by using one male coder and one female coder. It is noted that, in the event of disagreement, they viewed the advertisements together to arrive at consensus. The variables the coders were required to measure are listed as follows [Smith, 1994: 330]:


The results of the content analysis were discussed in terms of whether they proved or disproved the hypotheses established by Smith. In support of hypothesis one, she found that more advertisements were aimed at boys than at girls, in terms of the number of advertisements using boys and in terms of the type of products advertised. Of the eighty two different advertisements studied, 27 [or 32.9%] were aimed at girls, while 55 [or 67.1%] were aimed at boys [Smith, 1994: 331]. In addition, the toys classed as either male or female were differentiated in terms of traditional sex roles. The second hypothesis was also confirmed as correct, in that a definite correspondence existed between the sex of the narrator and the intended position of the advertisement. Of the twenty two advertisements for girls making use of a narrator, only two used female voices exclusively. In contrast, those advertisements for boys used only masculine voices.

The question of activity levels in hypothesis three resulted in mixed support, since there was no real marked difference between the sexes. It was found that boys and girls both performed passive activities, in proportion to the total number of actions, on an approximately equal level. A reasonable proportion of girl characters were seen to be physical, but only boys were involved in antisocial behaviour [Smith, 1994: 332]. In terms of setting, it was revealed that girls do indeed spend a significant amount of time in the home. Boys were seen mainly away from home. The coders agreed it was necessary to add the category of fantasy setting, because make-believe surroundings were very common in boys' advertisements. The out-of-home settings for girls were more limited and 'ordinary' [Smith, 1994: 333]. Of those settings classed as fantasy for girls, there were two typical features - they were coloured pink, with a sort of simple cotton-candy background. The boys, however, were shown in far more adventurous fantasy situations, with Smith giving the example of being drawn into video games [p.334].

It also seems that advertisers have taken heed to the results of studies on social learning theory [as discussed earlier], in that they are seen to take the safest route of audience address even in children's advertisements. They sell their products using either a mixture of boys and girls in their advertisements, or simply just boys. Smith concludes that, of all the variables considered in her study, the question of setting was probably the most gender stereotyped. In contrast, the most obvious area of non-stereotyped behaviour was seen in the assessment of physical activity pictured in some advertisements for girls. As Mayers and Valentine [1979] observe, television is in a position to allow children to fantasise about a variety of roles in their lives, but if advertisements only show traditional sex roles then they will impose a limit on the range of experiences that children can try [Smith, 1994: 335]. I hope that my own research will raise similar issues on the portrayal of sex stereotyping in advertisements. If children do indeed take what they see on television as examples of appropriate behaviour, then the role of advertising is central to their developing perception of gender, both in terms of themselves and others.