Children’S Reactions to Advertising Communication: Multiple Methods, Moderating Variables and Construct Validity Issues

Claude Pecheux, LABACC, Catholic University of Mons
Christian Derbaix, LABACC, Catholic University of Mons
ABSTRACT - This paper investigates 8- to 11-year-olds’ reactions to advertising communication with a specific focus on the construct validity of the scales used to measure the main variables studied. An initial phase of comprehensive scale development was indeed conducted, combining qualitative and quantitative methods in an attempt to provide a richer understanding of the constructs investigated. Next, we ran four experiments designed to study the impact of three moderating variables (felt involvement, mood, time of measurement) on the relationships among the advertising effectiveness indicators (ad attitude, brand attitude, brand beliefs, intent to request the brand, brand choice). Even though the results provide insights on advetising persuasiveness among kids they are only briefly sketched, our objective being more to show the richness of a multi-method approach in a children’s population and to discus several methodological issues such as the appropriateness of using moderating variables when kids are the subjects of the study.
[ to cite ]:
Claude Pecheux and Christian Derbaix (2002) ,"Children’S Reactions to Advertising Communication: Multiple Methods, Moderating Variables and Construct Validity Issues", in NA - Advances in Consumer Research Volume 29, eds. Susan M. Broniarczyk and Kent Nakamoto, Valdosta, GA : Association for Consumer Research, Pages: 531-538.

Advances in Consumer Research Volume 29, 2002     Pages 531-538

CHILDREN’S REACTIONS TO ADVERTISING COMMUNICATION: MULTIPLE METHODS, MODERATING VARIABLES AND CONSTRUCT VALIDITY ISSUES

Claude Pecheux, LABACC, Catholic University of Mons

Christian Derbaix, LABACC, Catholic University of Mons

ABSTRACT -

This paper investigates 8- to 11-year-olds’ reactions to advertising communication with a specific focus on the construct validity of the scales used to measure the main variables studied. An initial phase of comprehensive scale development was indeed conducted, combining qualitative and quantitative methods in an attempt to provide a richer understanding of the constructs investigated. Next, we ran four experiments designed to study the impact of three moderating variables (felt involvement, mood, time of measurement) on the relationships among the advertising effectiveness indicators (ad attitude, brand attitude, brand beliefs, intent to request the brand, brand choice). Even though the results provide insights on advetising persuasiveness among kids they are only briefly sketched, our objective being more to show the richness of a multi-method approach in a children’s population and to discus several methodological issues such as the appropriateness of using moderating variables when kids are the subjects of the study.

INTRODUCTION

In 1991, Harrigan suggested that one of the top priorities of research on children as consumers was the development of adapted measurement tools. Since the 70’s, the topic of Children and Advertising has been recurrently studied, sometimes driven by some public policy concerns. [In Europe, for example, debates are still going on about the need to control advertisements targeting kids.] When advertising effects are scientifically investigated among children, which is seldom the case, this is achieved with the same variables as the ones traditionally used in adults’ advertising effectiveness models: ad attitude, brand attitude, brand beliefs, purchase intent (MacKenzie et al. 1986). In addition, so far, the tests of these models have not included moderating variables (with the exception of brand familiarity; Derbaix and Bree, 1997; Phelps and Hoy 1996). Our research is therefore an attempt to circumvent two limits of prior work in the area: the use of inappropriate measurement scales and the absence of moderators.

BACKGROUND

Debates on advertising and children have started in the early 70s with the belief that advertising directed to children was "unfair". Today, debates on banning television advertising and on children’s reactions to advertising in general have not disappeared. [Several charges are leveled against ads aimed at children: prices are purposefully advertised as being low by stating "only"; the size of certain toys or snack foods appears magnified; fantasy situations are used as appeals, and child=s susceptibility is exploited; desires are aroused that would not otherwise be strong; among others (Kinsey 1987).] However, it is now well supported by research that characterizing children as either manipulated or fully able to understand advertising (the two traditional opposite views about advertising and children; Kapferer 1985) is a too simple picture of the problem. Several authors have underlined the importance of studying children’s reactions to Advertising as well as of learning more about the conditions under which persuasion could occur (Derbaix 1982; Derbaix and Bree 1997; Kapferer 1985; Phelps and Hoy 1996; Moore-Shay and Lutz 2000). According to them, Advertising impacts on children’s behaviors, but it seems that depending on situational and/or individual factors, children will be influenced differently. Therefore, it is important to learn the processes that make children -exposed to an advertisement- buy or prescribe the product or brand advertised. A clear understanding of such processes calls for a scientific and rigorous approach of the problem, using valid measurement tools specifically suited to children.

Advertising effectiveness among adults

Two variables are generally considered as indicators of advertising effectiveness: the ad attitude (Aad) ["The evaluative judgment of an ad during or immediately after exposure to this ad" (Derbaix 1995).] and the brand attitude (Ab). ["Psychological tendency that is expressed by evaluating a particular brand with some degree of favor or disfavor", adapted from Eagly and Chaiken (1993).] Since 1981, a lot of pieces of research have been devoted to the relations linking Aad to other variables. In their literature reviews, Brown and Stayman (1992) as well as Muehling and McCann (1993) give evidence of the core nature of the Aad variable. In addition, according to Brown and Stayman (1992), no less than 37 studies have proven the impact of Aad on Ab, and this relationship seems to be true for different experimental conditions. Actually, past research has strongly supported the Aad-Ab link (Gardner 1985; Homer 1990; Moore and Hutchinson 1983; Park and Young 1986). The most significant contribution concerning the study of the relationships between Aad, Ab and other related constructs is due to MacKenzie and his co-authors’ programmatic research (Lutz et al. 1983; MacKenzie et al. 1986; MacKenzie and Lutz 1989). These authors have hypothesized four competing views concerning the causal relationships among Aad and other measures of advertising effectivenss. The tests performed on these four models demonstrated the superiority of the Dual Mediation Hypothesis that specifies two influences of Aad: a direct effect on Ab and an indirect effect on Ab through Cb (even though Cb Bbeliefs about the brand- did not reveal the anticipated effect on Ab). Among the relationships investigated, the link between Aad and Ab is the strongest. Let us note that the products used by MacKenzie et al. (1986) in order to test these models were low-involving which could explain why the Aad-Ab relationship dominates. Involvement has indeed often been proposed as a moderator of the relationships between advertising effectiveness indicators (Lutz 1985; Muehling and Laczniak 1988, 1992; Rossiter and Percy 1984) even though the authors sometimes refer to different types of involvement (advertising message involvement, felt involvement, enduring involvement). Broadly speaking, involvement can be defined as "the degree of personal relevance or importance" (Park and Young 1986) and refers, in this context, to the motivational aspect of message processing (Petty and Cacioppo 1981) which partly determines the type of processing that will be observed. A high level of involvement is often associated to a central processing of the message (brand beliefs¦brand attitude) while low involvement would lead to a peripheral processing (ad attitude¦brand attitude).

Advertising effectiveness among children

To our knowledge, only a few studies investigated the relationships between ad and brand attitudes, brand beliefs and purchase intent in children. The first one is due to Phelps and Hoy (1996) who provide an initial exploration of the Aad-Ab-PI (Purchase Intention) relationships under different measurement timing conditions and after controlling for prior brand attitude. They showed that Aad significantly affects Ab both for familiar and unfamiliar brands even though the relationship was stronger in case of novel brands. A significant relationship was observed between brand attitude and purchase intent, this relationship being stronger in case of familiar brands (which might reflect past experience in the formation of brand attitude for familiar brands). A direct Aad-PI relationship was also observed although weakly significant. Finally, the manipulation of time of measurement did not highlight any significant differences in the relationships studied.

In a study whose aim was to investigate the impact of verbal as well as non-verbal affective reactions (measured through facial expressions), Derbaix and Bree (1997) found that the evaluative judgments of the ad elements of executionBbut not of the ad argumentsBare instrumental in shaping children’s Aad and Ab. The link between Aad and Ab was significant both for familiar and unfamiliar brands. Moreover, verbal affective reactions, especially the positive ones, are important predictors of Aad and Ab. Of particular interest are the results obtained for brand beliefs. Indeed, Derbaix and Bree (1997) tried to measure them but observed poor results and concluded that among children, brand beliefs might not be really formed and might be generated by the question supposed to measure them.

Moore-Shay and Lutz (2000) tested how children of different ages perceived the relationship between the advertisements they see and the products they consume. The results revealed differences between the two age groups considered (7- and 8- versus 10- and 11-year-olds) about the ability of children to integrate several informational inputs (for example: a product trial and an advertisement). They showed that among older children, those exposed only to product trial differ in their brands beliefs and attitudes from those exposed to advertising prior to product trial while young children do not exhibit such differences. Regardless of age, when advertising is the sole source of information (for example in the case of unknown brands), Aad has a direct impact on Ab. When ad exposure precedes a product trial, Aad has a direct impact on Ab for young children and older children exhibit a direct Aad-Ab link as wel as an indirect link (via Cb). When a product trial precedes ad exposure, Aad’s impact on Ab is lesser (but still significant). Let us note that the authors made no distinction between the children’s involvement levels when testing the different routes to persuasion.

Finally, Roedder et al. (1983) examined conditions in which children are likely to make consistent attitude-behavior choices in response to a television commercial. Two experiments conducted among children from 7 to 13 revealed differences between younger and older children due to more limited cognitive abilities of the younger ones. Indeed, in the behavioral context of choosing an alternative, the older children appeared to rely on their attitudes toward the various alternatives and to select the preferred alternative regardless of advertising while younger children appeared to base their choice on their evaluation of the advertised product and not on a comparison of choice alternatives.

CONCEPTUAL FRAMEWORK AND HYPOTHESES

[As already explained, the objective of this paper is not to stress on the conceptual aspects of this research but well on the methodological ones. We will therefore not present all the hypotheses we tested but only a few of them in order to give the reader an idea of what this research was about.]

In this research, in order to extend previous work, the same advertising effectiveness indicators were used: ad attitude (Aad), brand attitude (Ab), brand beliefs (Cb), intent to request the brand (IR) [Working with children who are often prescriptors, the "intent to request" variable was preferred to the "intent to purchase".] and brand choice (BC). Variables likely to moderate the relationships between these five indicators of advertising effectiveness were then included.

Moderator 1: Involvement

Involvement appears to be essential to understand most consumption behaviors. In this respect, Dussart (1983) states that "in consumer behavior oriented research one has to systematically measure the respondent involvement level before testing any other assumption". In this research, we worked with one type of involvement appearing as particularly important in a persuasion context: Felt involvement defined as "an individual’s cognitive effort expanded during processing of ad content". Actually, felt involvement has two types of antecedents: situational and enduring (Celsi and Olson 1988; Houston and Rotschild 1978).

One of our hypotheses proposed that the Dual Mediation Hypothesis (a direct link between Aad and Ab and an indirect link through brand beliefs) would be verified in situations of high felt involvement while a direct Aad-Ab link only would be observed in case of low felt involvement. Indeed, Moore-Shay and Lutz (2000) showed that both Aad and Cb (beliefs) were determinants of brand attitude among kids. This is contradictory to Derbaix and Bree (1997) findings who observed that the beliefs construct was not understood at all by kids. Nevertheless, none of these studies did consider the level of felt involvement of the kids. As suggested and verified by MacKenzie et al. (1986), Muehling and Laczniak (1988, 1992), Park and Young (1986) among adults, involvement moderates the relationships between these constructs. For high involvement receivers, both brand beliefs and Aad will influence brand attitude. However, when receivers’ involvement is low, the focus of attention is not likely to be the advertised message. As a result, message-based processing should be minimal and fewer beliefs and cognitions should be stored in memory. Low-involved receivers will try to minimize their processing while making an evaluative judgement of a brand (Gardner 1985; Park and Young 1986; Wright, 1973). We propose a similar pattern of relationships for children.

Moderator 2: Time

Consumers typically do not purchase a product immediately after exposure to advertising, there is usually a time interval separating advertising exposure from purchase. As stressed by Chattopadhyay and Nedungadi (1990), very fe studies have investigated the stability of ad attitude or how the Aad-Ab relationship evolves over time. Indeed, experiments in this stream of research are often designed to occur in one session, which means that the Aad and Ab measures are taken almost simultaneously. In addition to leading to an overestimation of the relationship between these two constructs, such an approach builds up a rather artificial testing situation. As a consequence, we incorporated a temporal dimension in our study, taking into account two different timing conditions: in the first one, all measures took place during the same session (time 1) while in the second timing condition, the brand-related measures (brand attitude, intent to request and brand choice) were delayed (time 2). We therefore expected the relationships between the different advertising effectiveness indicators to be moderated by the time elapsing between the measures. Nevertheless, even though we could expect some decline in the relationships investigated, we anticipate that the Aad-Ab relationship will endure. This hypothesis is based on several previous children research’s findings showing that the Aad-Ab relationship is a strong one (Derbaix and Bree 1997; Moore-Shay and Lutz 2000; Phelps and Hoy 1996).

Moderator 3: Mood

"Mood is a general feeling state that is not directed toward any particular target. Moods tend to be thought of as pervasive, low-level affective states" (Clark and Isen 1982; Isen 1984; Schwarz and Clore 1988). Mood is generally depicted as mild, generalized, diffuse and transient affective state. In general, mood are elicited by: after effects of emotions; organismic conditions such as illness, fatigue; general environmental conditions and side-effects of activities (Bagozzi et al. 1999).

It is rather easy to justify the importance of studying mood and children. In the context of the exploration of relationships between advertising effectiveness indicators, taking into account the child’s mood is important because it could influence his/her attitude toward commercials or toward advertised brands. From the stream of research on mood the congruence bias that consumers’ moods exert on the evaluation of stimuli is one of the most robust results. Considering mood allows us to study the affective state generated by the ad context and the brand context. Mood is more often depicted as an antecedent of an ad than as a reaction to an ad (as emotions are). Therefore, mood can be used as a potential moderator. In that respect, we hypothesized the Aad-Ab relationship to be stronger in case of positive mood than when the child’s mood is neutral or negative. Since past research has shown that positive mood leads to more favorable evaluations than neutral or negative mood (positive mood biases evaluations in a positive direction: Bower et al. 1981; Forgas 1991; Isen et al. 1978; Petty et al. 1991), we could anticipate that children in a positive mood will exhibit more favorable attitudes toward the ad and the brand and therefore that the relationship between these two attitudes is more likely to be strong than in the case of neutral mood. Moreover, there seems to be a general agreement that positive mood increases persuasion, at least when argument quality is constant (MacKie and Worth 1991).

In addition, a particular hypothesis regarding the explanatory role of mood was also stated. Indeed, for children, mood might be an input in the evaluation and decision processes. The "How do I feel about it" heuristic (Schwarz and Clore 1988) hypothesizing that individuals might simplify complex evaluation tasks referring to their mood at the time of judgment is a shortcut particularly suited to children, for whom the task at hand appears very often complex and for whom the use of a variety of attributes in a decision process has a very low probability to occur (Derbaix and Pecheux 1999). Rather than basing their evaluations on a piecemeal analysis of the available information, childrenBas "cognitive misers" (Taylor 1981)Bconsult their mood as a salient source of relevant information. Mood effects might indeed be more pronounced (Schwarz and Clore 1988): the more complex the judgmental task, the less cognitively developed the subjects, the less accessible other information, the higher the time pressure, and the more burdensome the judgment. ThusBespecially for childrenBmood states might serve informative functions through simplifying complex tasks. Mood effects have indeed a greater impact when evaluations are ambiguous and complex than when they are clear-cut and simple. In sum, we proposed that mood could be, especially when involvement is low, an explanatory variable of brand attitude.

SCALES DEVELOPMENT

Once the conceptual framework had been set up and before testing the relationships hypothesized, scales were developed in order to measure the key variables of this research (ad attitude, brand attitude, involvement [The scale developed aimed at measuring enduring involvement in the product class or, in other words, enduring antecedents of felt involvement. The other variables included in our research were not the subject of such a long development procedure simply because they were measured by short scales (one or two items) successfully used in previous studies with children (situational involvement, intent to request the brand). As far as brand beliefs are concerned, they were directly based on the content of the advertisement (since the brand was new).] and mood). As already stressed, this scale development phase was much more than one step among the others of this research, it was indeed a central concern of our work. Because building appropriate measurement scales, i.e., suited to children is far from being limited to a problem of wording, we refrain from translating for children scales developed for adults.

Actually, our objective was to make sure that the tools used would measure appropriately the variables studied, in other words, to use scales that were both reliable and valid. To achieve this goal, we followed a "revised" Churchill’s paradigm (1979), executing all the steps recommended by Churchill in order to provide reliable and valid scales and thereby guaranteeing the construct validity of the variables studied (Derbaix and Pecheux 2000). For each scale, there was an emphasis on qualitative techniques (depth interviews with children, focus groups with children, parents and teachers) to generate items. We realized that the qualitative steps were much more important than suggested by Churchill. For example, while developing the brand attitude scale, we first conducted focus groups of mothers and teachers (about 20 adults studied), we then performed an interpretative study of 13 children interviewed twice at home, followed during a period of 6 months and observed while shopping. Finally, 49 children were interviewed by pairs (with their best friend). These three types of qualitative studies provided a rich material to generate items.

Numerous data collections were conducted for each scale (including test-retest studies) with the objectives of purifying the measure, checking the reliability of the scale and its dimensions, and verifying the various forms of validity (discriminant validity, predictive validity). During these data collection phases, each scale was administered using at least two different methods (Likert, Smiling Faces, Semantic Differential) and for different stimuli (different brands, product categories, ads, etc.).

What we learned from the scale development phase

This comprehensive scale development procedure enables us to pinpoint 7 interesting methodological issues:

1) For the kids, the definition of the construct has to be considered as temporary until the exploratory phase is completed. A thorough analysis of all the verbatims could lead to refine the starting definition. Otherwise stated, the construct will be clearly defined on the basis of the dimensions, items grounded in the data of the exploratory phase which has to be much more important than the one classically undertaken for adults. Moreover, if the construct has several meanings, the one most often given by adults is not necessarily the one usually assigned by cildren.

2) As far as the selected items are concerned, a minor modification in the framing couldBfor children- lead to drastic changes in their answers. Seemingly redundant items for adults may look quite different for children (i.e., it’s fun versus it’s great).

3) In order to get the child’s cooperation, it is essential to show how important his/her answers are by using in the framing of the questions: "I", "ME", (in the items).

4) This framing has to take into account one of the objectives of building a scale: to use it in as many contexts (i.e., with as many brands, product classes, ) as possible. So, let us avoid too specific (idiosyncratic) items.

5) Do not forget (as stated by Bree (1991) and Greenleaf et al. (1999)) that a four-point format is the one to be selected when working with children to the extent that they are unable to discriminate more than 4 categories (in general) in their evaluative processes. At this level Bon the basis of our numerous data collections- let us recommend the use of Likert or Semantic Differential scales but not of the Smiling Faces format especially confusing when coupled with negatively formulated items.

6) One more time, let us stress the importance of the exploratory qualitative phase. Indeed, comprehensive qualitative research in the very beginning of the investigation could lessen the risk of "mere measurement effect" (i.e., the mere fact of questioning creates cognitions that did not previously exist) during subsequent quantitative phases.

7) From a practical point of view and coming back one more time to the exploratory stage, it seems very useful to interview kids in a large variety of contexts where they feel comfortable (schools, kids’ clubs, sport meetings, ) in order to avoid too situational (context-dependent) reactions.

TABLE 1

STEPS FOLLOWED TO DEVELOP THE SCALES

EXPERIMENTS

Once reliable and valid scales were obtained by this approach (the four scales developed have been published), several experiments were designed in order to test the main research questions while extending previous work (Derbaix and Bree 1997; Moore-Shay and Lutz 2000; Roedder et al. 1983). Three experiments were conducted with the objective of testing the impact of a single moderating variable (involvement, mood and time) at a time on the relationship between ad attitude and brand attitude. Elaborating on the results of these three studies, a fourth experiment was designed in which the three moderators were investigated simultaneously (n=181) [A 2*2*2 mixed experimental design was used.], thus enabling the study of interaction effects among the moderators. In addition, for this last study, more "behavioral" advertising effectiveness indicators were added (intent to request the brand and brand choice). For this final experiment, instructions were used (on a between-subjects basis) to elicit a high or a low level of involvement in the advertising exposure situation (situational antecedents in combination with different levels of involvement in the product class created different levels of felt involvement, Celsi and Olson 1988). Either a neutral or positive mood was induced by the TV program in which the advertisements were embedded (on a between-subjects basis), and the time of measurement of the brand measures was manipulated on a within-subjects basis (all children were assigned to an immediate measurement condition (time 1) and a delayed condition (time 2) for which the brand measures were repeated). While the first three studies used both familiar and unfamiliar brands from several product categories (sport shoes, sodas, snacks), the test brand (a brand of snacks) for the final experiment was unfamiliar (this brand was not available in the country where the experiment took place).

TABLE 2

SUMMARY OF THE SCALES DEVELOPMENT

RESULTS OF THE EXPERIMENTS AND DISCUSSION

The first conclusion that can be drawn from the results is that the attitude toward the ad is a powerful explanatory variable of brand attitude among children. This relationship is stable whatever the involvement level, the age (chronological and cognitive) of the child, the mood induced, the level of product class knowledge and the timing condition (or more precisely for this last moderator, the Aad-Ab relationship endures when brand attitude is measured at a later point in time). This is consistent with previous research among children (Derbaix and Bree 1997; Moore-Shay and Lutz 2000; Phelps and Hoy 1996) even though none of these studies included the moderators we considered. In addition, different measurement formats were used for Aad and Ab in order to make sure that the relationship between these two constructs is not an artificial one, a "measurement artifact". Moreover, the different tests performed on these two constructs (during the scale development phase, i.e., exploratory and confirmatory factor analyses) revealed that Aad and Ab were really different constructs, though related. In addition, the relationships between ad and brand attitudes and the two behavioral variables (intent to request the brand and brand choice) also appeared as highly significant.

The second obvious result of this research is that contrary to Moore-Shay and Lutz (2000) but consistent with Derbaix and Bree (1997) who showed that brand beliefs do not seem to be formed by children, this construct was almost never explanatory of brand attitude. Even when kids were specifically instructed to pay attention to the message and to what was said in it (high situational involvement manipulation), brand beliefs were almost never used by children to form a brand attitude. Our results mean that children do not form strongly held beliefs after one or two exposures to an advertisement presenting a new brand (working with a known brand might have led to slightly different results). In conclusion, the cognitive route of brand attitude formation (the one linking Aad to Ab through brand beliefs) is not observed among children even when they are motivated to process the ad main arguments and therefore brand information.

Then we checked for moderators’ effects. The results revealed that involvement does not appear as a powerful moderator of the relationships investigated among children. Children do not seem to exhibit different ways of processing information depending on their involvement level. The highly cognitive processing assumed in case of high involvement does not seem to occur among children. They seem to form evaluations on a more affective basis.

The lack of a moderating effect of involvement is perhaps due to the fact that children never exhibit different levels of felt involvementBwhich was the case in our experimentBor because different levels of felt involvement do not have the same consequences as in adult populations (in terms of ad and brand attitudes formation and their relationships). Finally, we should not put all the blame on involvement. Of course, involvement did not moderate persuasion as we expected but the problem might come from the other determinants (ability and opportunity to process the message) needed to follow a central route.

Similarly, we did not detect a significant impact of the second moderating variable: mood. Indeed, contrary to what has been shown in the adult literature, mood does not moderate persuasion. The link between ad and brand attitudes is not stronger in case of positive mood. To the extent:

B that brand beliefs are not taken into account in the formation of a brand attitude indicating thus a very poor cognitive treatment of the message arguments,

B that involvement does not moderate the relationship between Aad, Ab and other advertising effectiveness indicators (intent to request and brand choice), the processing route being therefore essentially an affective one,

there is no room for a significant moderating effect of mood on the relationships involving Aad, Ab and Cb. This means that being in a positive mood (as opposed to a neutral mood) does not lead, contrary to what is expected in theory, to a reduced cognitive treatment of the message.

However, our results reveal that mood (at time of brand evaluation) is an explanatory variable of the delayed brand attitude whatever the involvement level. Thus, among children, the best way to represent mood’s impact is as an explanatory variable of brand attitude. The "How do I feel about it ?" heuristic (Schwarz and Clore 1988) seems to be particularly adapted to this age group. Schwarz and Clore (1988) proposed that mood might be the informational input to an evaluation, especially when subjects are low cognitively developed, when the task is difficult and when respondents are low motivated in the task.

The timing of measurement of the different constructs and especially of Aad, Ab, IR and BC was also considered. Similar to Hutchinson and Moore (1985), we observed some decrease in the strength of the relationships studied. Nevertheless, our results revealed that the relationships investigated were still significant in the delayed conditions. In other words, an ad attitude measured on a given day impacts on the brand attitude measured the same day but also on the brand attitude measured the day after. In addition, when explaining the delayed brand attitude by the first (immediate) brand attitude and by Aad, the first explanatory variable is by far the most significant. This seems to indicate that brand attitude expressed at time 1 (immediate brand attitude) has been fixed and is well alive 24 hours later ! Moreover, we believe that if the attitude toward the brand were not measured at time 1, it would be almost impossible to measure brand attitude at time 2 (in the absence of the brand) [Let us stress one more time that the brand was novel and that kids were only presented twice (at time 1 only) with an advertisement for this brand.] and therefore difficult to investigate the relationship between Aad (time 1) and the delayed brand attitude (time 2). In the case of children and when working with a novel brand, there is therefore a clear need of "fixing" brand attitude when the attitude object is first met.

The weak results obtained for the moderating variables raise important questions such as: under what conditions are analyses dealing with moderators appropriate ? What other moderators seem relevant to investigate in children’s populations ? Would the moderators we selected have an impact on other relationships ? We strongly believe that these issues could be the starting point of further research in the area.

CONCLUSION

A highly significant result emerges from these experiments. The impact of Aad on Ab was indeed very clear and this result can be expanded Bas we showed- to other (or more) behavioral variables (the intent to request the brand and the brand choice). Moreover, we are very confident in the robustness of this result as well as in the lack of impact of our moderators to the extent that we worked with reliable and valid scales we built in order to progress in our understanding of the persuasion process of the child. In addition, some of the results suggest new methodological issues in the context of designing research on children.

REFERENCES

Bagozzi, Richard P., Mahesh Gopinath and Prashanth U. Nyer (1999), "The role of emotions in Marketing", Journal of the Academy of Marketing Science, 27 (2), 184-206.

Bower, Gordon H., Stephen G. Gilligan and K. P. Monteiro (1981), "Selectivity of learning caused by affective states", Journal of Experimental Psychology, 110, 451-473.

Bree, Jodl (1991), "Quelques problFmes de choix d’Tchelles pour mesurer les attitudes chez les enfants", Recherche et Applications en Marketing, 6(4), 27-58.

Brown, Stephen A. and Douglas M. Stayman (1992), "Antecedents and consequences of attitude toward the ad: a meta-analysis", Journal of Consumer Research, 19(1), 34-51.

Celsi, Richard L. and Jerry C. Olson (1988), "The role of involvement on attention and comprehension processes", Journal of Consumer Research, 15, 210-224.

Chattopadhyay, Amitava and Prakash Nedungadi (1990), "Ad affect, brand attitude and choice: the moderating roles of delay and involvement", in Advances in Consumer Research, eds. Marvin E. Goldberg, Gerald Gorn and Richard W. Pollay, vol 17, Provo, UT: Association for Consumer Research, 619-620.

Churchill, Gilbert A. (1979), "A Paradigm for Developing better measures of Marketing Constructs", Journal of Marketing Research, 14(1), 64-73.

Clark, Margaret and Alice Isen (1982), "Toward Understanding the Relationship Between Feeling States and Social Behavior", in Cognitive Social Psychology, eds. Albert Hastorf and Alice Isen, New York: Elsevier/North Holland, 73-108.

Derbaix, Christian (1982), "L’enfant, la communication publicitaire et la hiTrarchie des effets", Revue Frantaise du Marketing, 89, 31-47.

Derbaix, Christian M.(1995), "The impact of affective reactions on the attitude towards the ad and on the attitude towards the brand: a step towards ecological validity", Journal of Marketing Research, vol 32, 470-479.

Derbaix, Christian M. and Jodl Bree (1997), "The impact of children’s affective reactions elicited by commercials on attitude toward the advertisement and the brand", International Journal of Research in Marketing, 14, 207-229.

Derbaix, Christian M. and Claude Pecheux (1997), "L’implication et l’enfant: proposition d’une Tchelle de mesure", Recherche et Applications en Marketing, vol 12, n¦ 1, 47-70.

Derbaix, Christian M. and Claude Pecheux (1999), "Mood and Children: Proposition of a measurement scale", Journal of Economic Psychology, vol 20, n¦ 5, 571-591.

Derbaix, Christian M. and Claude Pecheux (2000), "Des outils pour comprendre l’enfant-consommateur: Bilan de 5 annTes de recherche", proceedings of XVIFme CongrFs de l’Association Frantaise du Marketing, eds. Richard Michon, Jean-Charles Chebat and Frantois Colbert, Ecole des Hautes Etudes Commerciales, MontrTal, 11-25.

Derbaix, Christian M., Sarah Blondeau and Claude Pecheux (1999), "l’enfant et l’attitude envers l’annonce publicitaire: mise en uvre du construit et prTcisions conceptuelles", Recherche et Applications en Marketing, vol 14, n¦ 3, 23-39.

Dussart, Christian (1983). "La prise de dTcision: forte ou faible implication ?" , in Comportement du Consommateur et stratTgie de Marketing, ed. McGraw Hill, 448-470.

Eagly, Alice H. and Shelly Chaiken (1993), The Psychology of Attitudes, Forth Worth, Harcourt Brace Jovanovitch.

Forgas, Joseph (1991), Emotion and Scial Judgments, Oxford, Pergamon.

Gardner, Meryl P. (1985), "Does attitude toward the ad affect brand attitude under a brand evaluation set ?", Journal of Marketing Research, 22, 192-198.

Greenleaf, Eric A, Barbara Bickart and Eric A. Yorkston (1999), "How response styles weaken correlations from rating scales surveys", Working Paper No. 99-5, New York University, Leonard N. Stern School of Business.

Harrigan, Judy A. (1991), "Children’s Research: Where it’s been, where it is going", in Advances in Consumer Research, vol 18, eds. Rebecca H. Holman and Michael R. Solomon, Provo UT: Association for Consumer Research, 11-17.

Homer, Pamela M. (1990), "The mediating role of attitude toward the ad: some additional evidence", Journal of Marketing Research, 22, 192-198.

Houston, Michael J. and Michael L. Rothschild (1978), "Conceptual and methodological perspectives on involvement", in Research Frontiers in Marketing: Dialogues and Directions, ed. Subhash C. Jain, Chicago: American Marketing Association, 184-187.

Isen, Alice M.(1984), "Toward understanding the role of affect in cognition", in Handbook of Social Cognition, eds. Robert S. Wyer Jr and Thomas K. Srull, Hillsdale, N.J: Lawrence Erlbaum, 179-236.

Isen, Alice M., Thomas Shalker, Margaret Clark, and Lynn Karp (1978), "Affect, Accessibility of Material in Memory, and Behavior: A Cognitive Loop ?", Journal of Personality and Social Psychology, 36 (January), 1-12.

Kapferer, Jean-Nodl. (1985), L’Enfant et la PublicitT, Dunod.

Kinsey, Joanna (1987), "The use of children in advertising and the impact of advertising aimed at children", International Journal of Advertising, 6, 169-177.

Lutz, Richard J.(1985), "Affective and cognitive antecedents of attitude toward the advertisement: a conceptual framework", in Psychological Processes and Advertising Effects: Theory, Research and Application, eds. Linda F. Alwitt, and Andrew A. Mitchel, Hillsdale, N.J., Lawrence Erlbaum Associates, 45-63.

Lutz, Richard J., Scott B. MacKenzie and George E. Belch (1983), "Attitude toward the ad as a mediator of advertising effectiveness: determinants and consequences", in Advances in Consumer Research, 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Arbor: Association for Consumer Research, 532-539.

MacKenzie, Scott B. and Richard J. Lutz (1989), "An empirical examination of the structural antecedents of attitude toward the ad as a mediator of advertising effectiveness", Journal of Marketing, 53, 130-143.

MacKenzie, Scott B., Richard J. Lutz and George E. Belch (1986), "The role of attitude toward the ad as a mediator of advertising effectiveness: a test of competing explanations", Journal of Marketing Research, 23, 130-143.

MacKie, Diane and L. Worth (1991), "Feeling good but not thinking straight: the impact of positive mood on persuasion", in Emotions and Social Judgments, ed. Joseph P. Forgas, Elmsford, NY: Pergamon Press.

Moore, Danny L. and J. Wesley Hutchinson (1983), "The effects of ad affect on advertising effectiveness", in Advances in Consumer Research, vol 11, eds. Richard P. Bagozzi and Alice M. Tybout, 526-531.

Moore-Shay, Elizabeth S. and Richard J. Lutz (2000), "Children, advertising and product experiences: a multimethod inquiry", Journal of Consumer Research, 27 (1), 31-48.

Muehling, Darrel D. and Russell N. Laczniak (1988), "Advertising’s immediate and delayed influence in brand attitudes: considerations across message involvement levels", Journal of Advertising, 17(4), 23-34.

Muehling, Darrel D. and Russell N. Laczniak (1992), "An examination of factors mediating and moderating advertising’s effects on rand attitude formation", Journal of Current Issues and Research in Advertising, 14(1), 23-34.

Muehling, Darrel D. and Michelle McCann, M. (1993), "Attitude toward the ad: a review", Journal of Current Issues and Research in Advertising, 15(2), 25-57.

Park, C.Whan and Stacey M. Young (1986), "Consumer response to television commercials: the impact of involvement and background music on brand attitude formation", Journal of Marketing Research, 23, 11-24.

Pecheux, Claude and Christian Derbaix (1999), "Children’s attitude toward the brand: a new measurement scale", Journal of Advertising Research, vol 39, n¦4, 19-27.

Petty, Richard E. and John T. Cacioppo (1981), "Issue involvement as a moderator of the effects on attitude of advertising content and context", in Advances in Consumer Research, 8, ed. Kent B. Monroe, Ann Arbor: Association for Consumer Research, 20-24.

Petty, Richard E., F. Gleicher and S.M. Baker (1991), "Multiple roles for affect in persuasion", in Emotion and Social Judgments, ed. Joseph Forgas, Oxford, Pergamon, 181-200.

Phelps, Joseph E. and Mariea G. Hoy (1996), "The Aad-AB-PI relationship in children: the impact of brand familiarity and measurement timing", Psychology and Marketing, 13 (1), 77-105.

Roedder, Deborah L., Brian Sternthal and Bobby Calder (1983), "Attitude-Behavior consistency in Children’s responses to television advertising", Journal of Marketing Research, 20 (4), 337-349.

Rossiter, John R. and Larry Percy (1984), "Advertising Communication Models", in Advances in Consumer Research, eds. Elizabeth C. Hirschman and Morris B. Holbrook, vol. 12, Ann Arbor: Association for Consumer Research, 510-524.

Schwarz Norbert and Gerald L. Clore (1988), "How do I feel about it ? The Informative function of affective states", in Affect, Cognition and Social Behavior, eds. Klaus Fielder and Joseph Forgas, Toronto: Hogrefe International, 44-62.

Taylor, Shelley (1981), "The interface of cognitive and social psychology", in Cognition, Social Behavior and the Environment, ed. Harvey, H, Erlbaum, Hillsdale.

Wright, Peter L. (1973), "The cognitive processes mediating acceptance of advertising", Journal of Marketing Research, 10, 53-62.

----------------------------------------