Celebrity Endorsement: Theory and Experiments with Children

Keywords: Name of author: Title: Institution:

Children, advertising, celebrity endorsement Brian M. Young Dr. University of Exeter School of Psychology Washington Singer Laboratories University of Exeter Exeter EX4 4QG UK +44 (0)1392 264614 +44 (0) 1392 264623 [email protected]

Phone: Fax: E-mail:

Child and Teen Consumption 2006 paper no. 25

Introduction Much research in the field of advertising to children, with a few notable exceptions (John, 1999) has been characterised by a lack of theoretical development as well as a range of different methodologies each with their weaknesses as well as their strengths. The purpose of this paper is to theorise the field using concepts drawn from communication theory and developmental psychology and to report on our1 experimental findings at the University of Exeter in one aspect of advertising to children the use of celebrity endorsement. There are two aspects of promotional communication that are important for an

understanding of advertising in children.

One is an implicit awareness of the basic Although

promotional rule of advertising which is based on a relatively unknown paper by May (1981). Mays paper sits within the tradition of linguistic theory known as pragmatics. there are a few attempts to describe advertising using pragmatic theory, most notably by Forceville (1996) and Tanaka (1994), it is not such a well-trodden route to advertising enlightenment as perhaps semiotics or post-modernist turns provide maybe linguists have their minds on higher things. Within pragmatics there is a particular approach exemplified by the work of Grice (1975), and Sperber and Wilson (1986). They assume that communication between people is governed by certain principles of conduct and that these principles are recognised when they are breached rather than when they are observed. Grices main principle can be paraphrased as: in co-operative conversations people contribute to the ongoing talk by the accepted purpose or direction of the talk exchange in which they are engaged. Grice puts forward several maxims that derive from this principle and they can be summarised by saying that participants in a co-operative exchange should communicate with sincerity, being relevant and clear while providing sufficient information. Of course people dont do this. The point, however, is what happens when a hearer recognises an apparent breach or break with these rules. Inferences are made as to the communicative intent of the speaker. Perhaps a mistake has been made or the lapse is due to external causes such as the presence of empty wine bottles. But more importantly, irony, sarcasm, and other figurative uses of language can be accounted for as deliberate breaches of these rules and the recognition of this by audiences. Although Grice was attempting to characterise communicative conduct in general, this approach can apply to particular categories or genres of communication. May (op. cit.) claimed that there was a particular rule for interested (as opposed to disinterested) communication and that was to treat the
1 I am indebted to my students as always and especially to Sophie Hall, Karen Betts, and Hannah Robins (in chronological order) from 1995-96 who worked with the children in the research.

Child and Teen Consumption 2006 paper no. 25

case that is put as the strongest that can be made. We would suggest that an essential part of the comprehension of the genre of promotional activity is applying this rule. For example advertising candy on TV will present the positive sides of the purchase or consumption experience, and will not mention that candy might contribute to dental caries or obesity. A communicative rule for understanding advertising and other forms of The other aspect of promotional communication is an awareness of the promotion such as self-promotion by people both emerge at about 7 or 8 years of age (Young 2000). discourse structure of advertising where the topic is the brand and the comment is the rest of the promotional communication (Young, 1990). Again we use concepts borrowed from linguistic theory and apply them to audio-visual communication in advertising. Using this basic conceptual equipment, one can then generate hypotheses about celebrity endorsement and its understanding at different ages. We know of no empirical literature on the childs understanding of the basic discourse structure of advertising although summaries of research in the field of advertising to children are widely available. However it seems that at some stage between 3 and 5 years an understanding that spot advertising (on TV usually) is distinct and different from programming emerges2 and the concept of brands in different places including ad-on-TV is mentally represented. The context provides a comment and is assumed to be there on TV ads for fun. There is practically no dissent in the literature about the dominance of the entertainment function in the early years. At some time from about 5 to 8 years of age the child begins to acquire an understanding of the promotional nature of advertising la May and gradually we assume children will infer that the comment serves the function of enhancing the topic. something such as buy the brand or pester Mum to do that. hypothesis. So a celebrity endorser would simply be seen as a fun person by the little ones and indeed Ronald MacDonald, a clown, fits the bill perfectly. Later we anticipate the child beginning to This should emerge at a At least that would be our similar time in middle childhood as children realise advertising is there to get you to do

I have deliberately avoided discussing the changing media landscape but obviously advertising is not now limited to TV spots (if it ever was) and these changes will keep us - as researchers in gainful employment for years to come. Hopefully.

Child and Teen Consumption 2006 paper no. 25

understand that the celebrity endorser is there with a specific function to enhance or add value to the brand (although not expressed by the children using these words). Why? Because ads are designed to get people to buy the brands advertised and they are intentional communicative acts. Homer Simpson is not there by accident he has an important role to play. We also know that children from the age of 4 years are capable of understanding that other people have beliefs and intentions that can be different from their own desires and knowledge and that children in middle childhood will commonly make psychological attributions when describing others. However understanding celebrity endorsement is more than just mental growth and developing a deeper understanding. It also requires knowledge of the world of commerce and business and this thread of economic socialisation is as equally important as Vygotskys line of natural development in children (Van der Veer and Valsiner, 1991; p223) Having established a sound theoretical framework before even talking to a child, we then tried to determine the different research factors that it would be necessary to explore when devising a programme of research into celebrity endorsement. and the following are by no means exhaustive: Understanding of fame What are the features of being famous?; do children understand the self as a brand? Levels of representation Are real (e.g. soccer hero) more effective than cartoon (e.g. Homer Simpson) endorsers? Fit between activities of endorser and brand use On what dimensions do children establish congruence3 between endorser and brand?; relevant skill?; level of fame? To what extent do celebrity endorsed brands have perceived added value for children? How can we measure this?; points of indifference and their assessment. This emerges analytically

3 The idea of congruence has been explored in one of the few academically powerful papers in childrens understanding of marketing techniques: see Zhang, S., & Sood, S. (2002). Deep and Surface Cues: Brand Extension Evaluations by Children and Adults. Journal of Consumer Research, 29, 129-141

Child and Teen Consumption 2006 paper no. 25

There is a small literature on celebrity endorsement that further informed our research. Wilson and Weiss (1992) demonstrated how the use of popular characters (whether real or fictional) from childrens programmes may make it difficult for children to distinguish between advertisements and programmes. There is also ample evidence that TV advertising, including celebrity endorsement, plays a dominant role in shaping childrens product preference. Ross et al. (1984) showed how using celebrities to endorse a product could increased childrens preference for that product and their belief that the celebrity was an expert on its subject. Similarly, a study by Atkin (1980) examined 5-7 year olds perception of trade characters from food advertisements that did not feature in surrounding programs and found that children actually believed characters like Fred Flintstone & Barney Rubble knew about breakfast cereals and accepted them as credible sources of nutritional information. They found that this finding was stronger for heavy viewers of television. Likewise, Bandyopadhyay, Kindra and Sharp (2001) showed how children feel validated in their choice of product when a celebrity endorses it. Empirical work Next, we turn to the research evidence that has emerged from our own work at the University of Exeter. As this is in progress, conclusions at the moment are tentative. Preliminary work in 1995 at Exeter established that there was a developmental trend in understanding celebrity endorsement. Ninety children aged between 6 and 11 years were presented with two posters advertising a new chocolate bar called Boom. One was endorsed by a famous person that the child had first chosen as having seen before from a range of 10 well-known celebrities and the other was endorsed by an unknown (to the child) colleague of the experimenter. They were then asked why the famous person would be on an ad like this and why or why not they think this is a good ad. We found that a tripartite scale (which had been used in this field since the pioneering work of Ward et al., 1977) could be used. Responses were categorised and scored as showing no (e.g. because hes funny), partial (because hes famous), and full (because hell make more people buy Boom) understanding of celebrity endorsement. It was concluded that a full understanding of celebrity endorsement only emerged in 10-11-year-olds and the majority of younger children were not able to show this complete understanding. In another study from that year 55 children were presented with four cans of a (hypothetical) new soft drink called Zest. All the cans were covered in Zest livery. Two

Child and Teen Consumption 2006 paper no. 25

were 150ml, 1 was 330ml, and 1 was 440ml. One of the 150ml cans showed Bart Simpson on it and the other didnt. Using a forced choice paradigm with two alternatives, it was found that about half the children said their friends would pay more for the endorsed 150ml can than the non-endorsed 150ml can, and a half said they would pay the same amount with no age differences. the endorsed can. However when the endorsed 150ml can was compared with the 330ml or the 440ml one then practically all the children said their friends would pay less for We can conclude from this experiment that any added value that is perceived to accrue from the celebrity endorsement is not matched by an increase in quantity. This experiment raises basic questions about the meaningfulness of the concept of added value by celebrity endorsement with children. For example we are working on a more graduated method for simulating an increase in size of the non-celebrity endorsed alternative in order to measure a point of indifference. But theoretically there is not really any way of knowing whether an inability to ascertain points of indifference in children below a certain age reflects a cognitive immaturity in the child or a lack of understanding. In 2006 we attempted to design and validate a scale suitable for assessing the childs understanding of celebrity endorsement, using a set of nine statements. These were response scaled using a 5-point Likert scale (Strongly agree, agree, dont know, disagree, strongly disagree) which was also accompanied by 5 associated stars differing in size; the largest denoting strongly agree and the smallest strongly disagree. The 9 statements are given in order of presentation in Table 1.
Table 1. Order of presentation, attitude statement and rationale for selection. Order of presentation 1 2 3 4 5 6 7 8 9 Statement The celebrity is used for fun The celebrity is used to grab attention The celebrity is used because he likes the product The celebrity was used to make you like the product The celebrity was used because he was paid The celebrity was used to make you buy the product The celebrity was used to get you to remember the product The celebrity was used because he knows about the product Celebrities always tell the truth Theoretical rationale Ads entertain Ads get your attention Literal endorsement Detached analysis of ad - affect Detached analysis of ad actors interest Detached analysis of ad consumer purchase Detached analysis of ad consumer recollection Source credibility Source authority

Child and Teen Consumption 2006 paper no. 25

In addition each child was given a standard set of four items designed to assess understanding of TV commercials; what is a TV commercial?; why are commercials shown on TV; what do commercials try to do?; what is the difference between a TV commercial and a TV programme? (see Ward et al., op. cit.). These measured the childs understanding of advertising intent. Two groups of children were used. There were 43 5-6-year-olds and 71 10-11-year-olds. In addition a small convenience sample of 11 adult students was chosen to establish an adult benchmark. Eight versions of the same poster advertising a new cookie jar bar were designed. Four of the posters contained a cartoon celebrity endorser chosen for their popularity amongst both age groups (either Bart Simpson, Lisa Simpson, SpongeBob SquarePants or Wilma Flintstone) and four contained a famous human celebrity endorser (either David Beckham, Robbie Williams, Victoria Beckham or Kylie Minogue). The real life celebrities were selected on the basis that they cater for two distinct categories of fame: pop and sports stars and that it was possible to obtain each character with a similar facial expression and front-on camera shot. There were two male and two female endorsers in each category in case any gender specific preference was made. In addition the eight characters were displayed separately without the product on their own laminated, white cards (each measuring 10x11cm). These were presented prior to the adverts being shown to enable the children to pick their favourites, one from the cartoon set and one from the human set. Two manipulation check questions were used to test recognition of the endorser (any participant who failed to answer the manipulation check questions was excluded from the analysis). Each child was then asked which of the two posters [cartoon or real with their favourite cartoon and real as the two alternatives] would make you want to buy the cookie jar bar more? and that was then used as the poster for the child to respond to. Results for each group are given in Table 2.

Child and Teen Consumption 2006 paper no. 25 Table 2: Entries are mean (standard deviations) of group response on each attitude statement (1 = strongly agree; 5 = strongly disagree). Significance based on one-way ANOVA with students removed. Order of presentation 1 2 3 Statement 5-6 group (n = 43) 1.63 (1.16) 3.42 (1.56) 2.07 (1.39) 1.93 (1.24) 2.60 (1.50) 1.93 (1.24) 2.02 (1.44) 2.40 (1.51) 2.58 (1.53) 9-10 group (n = 71) 2.38 (1.02) 1.87 (0.83) 2.97 (1.12) 2.15 (0.95) 3.04 (1.30) 2.00 (1.03) 2.44 (0.97) 3.37 (1.05) 3.75 (1.09) Student s (n = 11) 3.09 (1.38) 1.18 (0.41) 4.36 (0.67) 1.73 (0.91) 1.64 (1.21) 1.45 (0.52) 1.73 (1.10) 4.36 (0.67) 4.64 (0.51) Significance

The celebrity is used for fun The celebrity is used to grab attention The celebrity is used because he likes the product The celebrity was used to make you like the product The celebrity was used because he was paid The celebrity was used to make you buy the product The celebrity was used to get you to remember the product The celebrity was used because he knows about the product Celebrities always tell the truth

** ** **

NS

5 6

NS NS

NS

**

**

Each child was asked which do you think is the main reason after responding to the 9 statements. Results are shown in Table 3.

Child and Teen Consumption 2006 paper no. 25

Table 3: Entries are number of respondents selecting that reason as the main reason for the celebrity being on the poster. (Column totals do not add up to group size because of missing values) Order of presentation 1 2 3 Statement The celebrity is used for fun The celebrity is used to grab attention The celebrity is used because he likes the product The celebrity was used to make you like the product The celebrity was used because he was paid The celebrity was used to make you buy the product The celebrity was used to get you to remember the product The celebrity was used because he knows about the product Celebrities always tell the truth 5-6 group (n = 43) 16 0 2 9-10 group (n = 71) 9 27 2 Students (n = 11) 0 1 0

5 6

1 3

5 16

2 5

Both of these tables confirm that the entertainment function of advertising is seen as less important as a reason for this particular advertising technique of celebrity endorsement with the older group of children. This result confirms what is already known about the childs understanding of advertising intent at a young age that it there to entertain (John, 1999) although we know of no literature that has explored the use of rhetorical techniques and reported this finding. The older children recognise that the celebrity endorser is there not because he likes the product (younger children differ from older children in Table 2) but because he is trying to grab your attention (Tables 2 and 3) and wants you to buy the product (Table 3). They recognise the function of rhetoric although the do not always see the distal4 function as dominant over the proximal one. There is some evidence using the Likert scale methodology that younger children are more likely to think that celebrities actually like the advertised brand whereas older children are more likely to claim that they
4

I have used these terms distal and proximal to distinguish respectively between the bottom line goal of advertising which is purchase and one of the immediate goals to get attention.

Child and Teen Consumption 2006 paper no. 25

know more about the brand than younger children (Table 3). The well-known increase in scepticism with advertising with increasing age (Mangleburg and Bristol, 1998) is confirmed here (Table 3) as older children are less likely to choose celebrities always tell the truth as the main reason. Children were asked why they thought the celebrity was on the poster and replies were content analysed. Comprehension levels (1, 2 and 3) were assigned to each response. Level 1 indicated little or no understanding of celebrity endorsement, level 2 meant some understanding and level 3 was assigned if the child had a full understanding of celebrity endorsement. Each child then received a score from 1 to 3. Judgements were made independently by each researcher and a significant inter-rater reliability measure (Pearson r= 0.82) was obtained. Examples of each level of understanding of celebrity endorsement are given in Table 4.
Table 4: Levels of assigned to each child in response to question; why do you think the celebrity is on the poster? Understanding of celebrity endorsement Level 1 Level 2 Level 3 Example makes it funnier / dont know because people like him its saying if you want to be like David Beckham do this or eat that

The four items used to assess the childs understanding of advertising intent had a reliability coefficient (Cronbachs alpha) of 0.86 and we can conclude that they are scaleable on a criterion of internal consistency i.e. they are measuring the same construct. Consequently a total score was computed and correlated with the understanding of celebrity endorsement measure described in Table 4, giving a Pearson r of 0.61 (p < 0.01). We can conclude that understanding advertising intent and understanding celebrity endorsement are related parts of a general advertising literacy which increases with age. One-way analyses of variance on the data showed, as expected, a significant effect of age on both measures in the expected direction.

10

Child and Teen Consumption 2006 paper no. 25

Discussion We suggest the following developmental trajectory based on theory and already published research evidence. By the age of two years many brands are recognised with pleasure by children, independently of the context in which they occur whether on TV, in the home, or in a supermarket. Any advertising context is understood as having only one function to The prototypical example of this is Ronald entertain. If the context of the brand-in-the-ad includes a celebrity endorsing the product, then that person will be seen as entertaining. MacDonald, a clown whose only existence is in the world of that brand. At some point in middle childhood (usually 7-8 years of age), the child begins to understand the enhancement effect of the context using the promotional rule. From this time an understanding of the role of celebrity endorsement develops and celebrities are perceived as being there not just to entertain but as having been chosen to promote the product. In conclusion I have described some recent research undertaken in the School of Psychology at the University of Exeter on childrens understanding of celebrity endorsement. The applied importance of such research cannot be underestimated as celebrity endorsement has been frequently mentioned as a significant issue in the advertising and marketing to children debate. I hope I have convinced my colleagues that it is equally important to embed this research is a sophisticated set of theories of communication and psychology.

11

Child and Teen Consumption 2006 paper no. 25

References Atkin, C.K. (1980). Effects of television advertising on children. In E. L. Palmer & A. Dorr (Eds), Children and the faces of television, teaching violence, selling. (pp. 287-306). New York: Academic Press. Bandyopadhyay, S., Kindra, G. and Sharp, L. (2001). Is television advertising good for children? Areas of concern and policy implications. International Journal of Advertising 20 (1), 89-116. Forceville, C. (1996). Pictorial Metaphor in Advertising. London: Routledge. Grice, H. P. (1975). Logic and conversation. In P. Cole and J. Morgan (Eds.), Syntax and Semantics 3: Speech Acts. New York, NY: Academic Press. Pp.41-58. John, D. (1999). Consumer socialization of children: a retrospective look at twenty-five years of research. Journal of Consumer Research, 26, 183-213. Mangleburg, T. F. and Bristol, T. (1998). Socialization and adolescents' scepticism toward advertising. Journal of Advertising 27(3), 11-21. May, J. D. (1981). Practical reasoning: extracting useful information from partial informants. Journal of Pragmatics, 5, 45-59. Ross, R.P., Campbell, T., Wright, J.C., Huston, A.C., Rice, M.L., & Turk, P. (1984). When celebrities talk, children listen: An experimental analysis of childrens responses to TV ads with celebrity endorsement. Journal of Applied Developmental Psychology, 5, 185-202. Sperber, D. and Wilson, D. (1986). Relevance: Communication and Cognition. Oxford: Blackwell Tanaka, K. (1994). Advertising Language: A Pragmatic Approach to Advertisements in Britain and Japan. London: Routledge. Van der Veer, R. And Valsiner, J. (1991). Understanding Vygotsky: A Quest for Synthesis. Oxford: Blackwell. Ward, S., Wackman, D., and Wartella, E. (1977). How Children Learn to Buy. Beverly Hills CA: Sage. Wilson, B.J., & Weiss, A.J. (1992). Developmental differences in childrens reactions to a toy advertisement linked to a toy-based cartoon. Journal of Broadcasting & Electronic Media, 36, 371-394. Young, B. M. (1990). Television Advertising and Children. Oxford: Oxford University Press. Young, B. M. (2000). The childs understanding of promotional communication. International Journal of Advertising and Marketing to Children, 2(3), 191-203.

12