Cognitive Response to Advertising: the Relation of Child to Adult Models

Bobby J. Calder, Northwestern University
Thomas Robertson, University of Pennsylvania
John Rossiter, University of Pennsylvania
ABSTRACT - It is argued that child models of cognitive response to advertising are not solely constrained adult models. Nor is the relationship between child and adult models completely explained by Piaget's stage construct. Differences are discussed in terms of the child's mediation-al representation, organization, and evaluation of advertising information.
[ to cite ]:
Bobby J. Calder, Thomas Robertson, and John Rossiter (1976) ,"Cognitive Response to Advertising: the Relation of Child to Adult Models", in NA - Advances in Consumer Research Volume 03, eds. Beverlee B. Anderson, Cincinnati, OH : Association for Consumer Research, Pages: 536-538.

Advances in Consumer Research Volume 3, 1976      Pages 536-538


Bobby J. Calder, Northwestern University

Thomas Robertson, University of Pennsylvania

John Rossiter, University of Pennsylvania


It is argued that child models of cognitive response to advertising are not solely constrained adult models. Nor is the relationship between child and adult models completely explained by Piaget's stage construct. Differences are discussed in terms of the child's mediation-al representation, organization, and evaluation of advertising information.

An important theoretical issue in investigating the cognitive response of children to advertising concerns the relationship between child and adult models of cognitive processes. It may be argued that any child model presupposes a well-developed model for adults. A model for children may be viewed as a constrained version of the adult model. The child model simply includes limits in the information-processing abilities of children. The problem with this argument is that it does not go far enough. The purpose of this paper is to contend that it is potentially misleading to characterize the relationship between child and adult models solely as one in which the child model successively comes to approximate the adult model as the child grows older. A more productive view is that, while any child model must allow for the developmental fact that the child does become an adult, the child's cognitive response is both quantitatively and qualitatively different from that of adults.


A child model of cognitive response to advertising must of necessity be dynamic rather than static. Any child model must allow for the process of developmental change. This is usually handled through the construct of developmental stages. The use of the stage construct is illustrated in a study by Rossiter and Robertson (1975) which compared sources of variance in children's ability to comprehend the structure and purpose of television commercials. It was shown that children between the first and fifth grade differ reliably in their ability to perform cognitive operations in understanding commercials. The operations were those of classification, discrimination between commercials and programming; relational thinking, perceiving the relationship between commercial messages and an external sponsor; drawing implications, detecting informative intent and persuasive intent; and transformations, dealing with symbolic claims about products. Of the four operations, transformations were the most difficult. But the significance of this research for illustrating the stage construct is that developmental level, as indexed by age and grade, was shown to account for 40% of the total variance--or three-fourths of the accounted for variance--in children's comprehension. This versus 9% for social environment factors and 4% for viewing time differences within grade.

Age itself, however, is not a stage construct, though it can be a useful surrogate variable for research. The passage of time has no scientific status in and of it- -self. The stage component of a child model must identify the way in which information processing changes over time, both in terms of processing capacity and the abilities of information acquisition and representation. In its most general form, the stage construct calls for a model of the child at any given developmental level. A model whose relation to a comparable model at any other level is well-defined. The best known and most elaborate stage models are, of course, those of Piaget (cf. Piaget and Inhelder, 1969). Most discussions of the stage component of children's consumer information processing models (e.g., Ward, 1974) cite Piaget's work as a point of reference. Nonetheless, it is not completely clear that Piaget's treatment of stages is suited for such models.

It is important to recognize that Piaget has attempted to formulate a general theory of how the child comes to know about the world. The theory deals with operative knowledge, how the child is able to deal with real events. These events may be task skills, operations, rules, concepts, or beliefs. The basic theoretical idea is that knowledge depends on the development of internal cognitive structures. Events are "assimilated" into the cognitive structure and, at the same time, the cognitive structure "accommodates'' to features of the real event. These cognitive structures are not, however, what they are commonly taken for. They are not internalized mediational representations of real events or external information (cf. Furth, 1969). Piaget denies that language is the basis of thought and, moreover, that cognitive structures depend on symbolic or imaginal representations. Piaget's cognitive structures are perhaps best thought of as internal actions. When these actions are integrated into logical systems, they are referred to as operations. By stages Piaget simply means an orderly progression of relatively stable cognitive structures (Furth, 1969). A model at any level is couched in terms of a description of the cognitive structures attained and their relation to the structures of previous levels. At no developmental level are structures considered to be internalized representations.

Note that the Piagetian approach, while extremely interesting, denies the mediational representational view of information processing which is so characteristic of the communications and consumer behavior literature. Indeed, Piaget's approach runs against the historical drift of Western thought in general (Furth, 1969). It may be that Piaget's biological approach and the representational approach are not incompatible, but it seems to us that they are definitely intended for different problems. What one cannot do is impose the notion of a Piagetian stage onto a mediational information processing model. It is not possible to "apply" Piaget's thinking to the typical consumer information processing model. Piaget's cognitive structures are not intended to explain the how and what of the child's cognitive response. These responses may be regulated by structures, or internalized operational schemes, but they are not explained by structural changes. Stages are important only for longer-term, biological maturation.

Consider, for example, the case of perceptual boundedness. This refers to the tendency of children to focus on immediately perceived stimulus features rather than on more delayed or abstracted features. All perceptual bounded-ness does is to describe a limitation characteristic of a given level of development. It does not explain the nature of the child's cognitive response.


Recently Calder, Robertson, and Rossiter (1975) have outlined a range of possible non-Piagetian differences between child and adult models. To elaborate further on the differences between child and adult models, as we see them, we will focus here on the child's comprehension, retention, and evaluation of advertising information.

Although our knowledge of what advertising information is salient to children is far from complete, an even more challenging question concerns how such information is represented. What are the basic products of the child's comprehension process? There are at least three types of memory codes, or forms in which information may be stored. Children's research should not follow the pattern of adult research in ignoring all but the linguistic coding of information, although this will certainly be a temptation because of the dominance of questionnaire and rating scale techniques. Considerable emphasis must be placed on enactive (motor) and imagery codes. The presence of enactive codes is perhaps best seen in early development in the form of imitative modeling. Gewirtz and Stingle (1968), for instance, argue that imitation involves learning from a model those behaviors that work--that yield the child what he wants. Verbal imitation also seems to precede linguistic understanding (McNeill, 1970). Children may thus be observed to repeat slogans in a purely enactive, "meaning-less" way (Wells, 1965).

The importance of motor codes probably decreases with development. The child gradually relies more heavily on imagery codes. These differ from symbolic codes in that imagery representation is essentially "stimulus faithful" whereas symbolic representation bears only an arbitrary relationship to external stimulus features. We may speculate that visual imagery plays an important role in children's product related behavior. Take, for example, a child accompanying a parent to a supermarket or toy department. The child might only have information about a cereal or toy in the form of a visual image obtained from a television commercial, but this image may yield a "match" when the child actually sees the product. This familiarity effect could engender preference indications such as pointing.

Imagery occurs in other forms too. Representation in the echoic mode is possible. As already noted, children often repeat brand names, slogans, and jingles from advertisements. This may reflect largely the echoic storage of information. Musical encoding of jingles and themes provides an interesting case. Nonmusical verbal codes are subject to considerable interference from conversation and require specific abilities. But echoic codes are subject to very little interference and involve storage only of the acoustic pattern. Retrieval of this echoic code may influence purchase preference or at least suggest a preference to parents. Again, such a process may not be tapped with verbal research instruments.

In addressing the role of nonverbal stored information we do not mean to suggest that enactive or image codes exclude or necessarily replace linguistic codes. For adults, though imagery may be quite important, much information is probably stored in parallel fashion. Our hypothesis is that children differ from adults in the extent to which these parallel systems are developed. The child's use of linguistic codes is greatly hampered by psycholinguistic factors.

While memory codes are the most important issue for comprehension, the organization of coded information in memory is the central issue for retention and evaluation. Evidence indicates that the dimensions the child uses to organize information stored in memory changes with development. Rossi and Wittrock (1971), for example, investigated changes in children ranging from 2 to 5 years in mental age. Younger children were more likely to use phonetic relationships (rhyming) in memory organization whereas usage of conceptual categories increased with age. We might also hypothesize that the use of simple, linear belief structures predominate with younger children, the use of hierarchical structures perhaps increasing with age. This is especially plausible considering the memory codes employed by children. In any case, the child's capacity to retain information, especially from the media, no doubt depends greatly on his ability to organize the information.

With adult models, evaluation is also thought to depend on how product information is organized. Many evaluation models for adults are best thought of in terms of information integration (Calder, 1975). The notion that evaluation entails some combinatorial process, however, becomes quite tenuous with children. Evaluation for children may be contingent on isolated bits of information which the child happens to be focusing on. There may be no attempt to integrate the stored information. This would explain why children's preferences are probably more unstable than adults! It would lead one to expect recency effects for the presentation of product information too. In short, the summative approach to modeling evaluation cannot be applied uncritically to children.


In conclusion, we have alluded to a number of possible differences between child and adult cognitive response to advertising. While all differences may be related to developmental processes, they are not tied, at least in any well-defined way, to the Piagetian construct of stages. They are defined in terms of the child's mediational representation, organization, and evaluation of advertising information.

In addition to suggesting areas for children's research, differences between child and adult models may serve to draw attention to problems which have not yet been recognized in the area of adult consumer information processing. The child and adult areas are complementary. Research on children must be carried out within a theoretical framework emphasizing all aspects of the relationship between child and adult models.


Calder, B. "The Cognitive Foundations of Attitudes: Some Implications for Multi-Attribute Models" in M. Schlinger, ed., Advances in Consumer Research, Vol. 2, (University of Illinois at Chicago Circle, 1975, 261-268).

Calder, B., Robertson, T. and Rossiter, J. "Children's Consumer Information Processing," Communication Research, Vol. 2 (July, 1975), 307-316.

Furth, H. Planet and Knowledge (Englewood Cliffs, N.J.: Prentice-Hall, 1969).

Gewirtz, J. and Stingle, K. "The Learning of Generalized Imitation as the Basis for Identification," Psychological Review, Vol. 75 (1968), 374-379.

McNeill, D. The Acquisition of Language (New York: Harper & Row, 1970).

Piaget, J. and Inhelder, B. The Psychology of the Child (New York: Basic Books, 1969).

Rossi, S. and Wittrock, M. "Developmental Shifts in Verbal Recall between Mental Ages Two and Five," Child Development, Vol. 42 (1971), 333-338.

Rossiter, J. and Robertson, T. "Canonical Analysis of Developmental, Social, and Experiential Factors in Children's Comprehension of Television Advertising," Journal of Genetic Psychology, (1975, forthcoming).

Wells, W. "Communicating with Children," Journal of Advertising Research, Vol. 5 (1965), 2-14.