Do Mothers and Children Share Cereal and Beverage Preferences and Evaluative Criteria?

ABSTRACT - This paper investigates the overlap between mothers and children's brand preferences for 14 cereals and 14 beverages. Multidimensional scaling techniques were used to derive solutions from a paired comparison task of collecting preference judgments. Dimensional weights of both groups were subjected to simple regression analysis. There was very little overlap in the evaluative criteria used by both groups in making preference judgments about cereals and beverages.


Kenneth D. Bahn (1987) ,"Do Mothers and Children Share Cereal and Beverage Preferences and Evaluative Criteria?", in NA - Advances in Consumer Research Volume 14, eds. Melanie Wallendorf and Paul Anderson, Provo, UT : Association for Consumer Research, Pages: 279-282.

Advances in Consumer Research Volume 14, 1987      Pages 279-282


Kenneth D. Bahn, Virginia Polytechnic Institute and State University


This paper investigates the overlap between mothers and children's brand preferences for 14 cereals and 14 beverages. Multidimensional scaling techniques were used to derive solutions from a paired comparison task of collecting preference judgments. Dimensional weights of both groups were subjected to simple regression analysis. There was very little overlap in the evaluative criteria used by both groups in making preference judgments about cereals and beverages.


Relatively little is known about how parents, specifically mothers, influence their children's brand preference formation. Literature provides support for the notion that mothers are the dominant parental influences on the development of children's socialization (Zigler and Child 1969). Within the consumer socialization literature, however, three processes have been identified by which mothers would influence children: (l) children may observe and imitate parental behaviors, (2) parent-child interaction occurring in consumption situations may affect learning, and (3) children may engage in independent consumer behaviors, with some degree of parental guidance (Ward, Wackman, and Wartella 1977a, p. 113). Of these three learning processes, this discussion will primarily focus on the second because it is most likely (other than role modeling) to be involved in mothers teaching their children about brand attributes which, in turn, may be the basis for forming brand preferences. The other two processes involve less verbal interaction between mother and child, and therefore, decrease the likelihood of mothers teaching their children about specific brand attributes.

The interaction between mother and child can be initiated by either component of the dyad. The mother may begin the interaction by discussing an advertisement viewed by both, or the child may initiate the interaction by requesting a specific brand to be purchased, the request perhaps being initiated by exposure to an advertisement. Thus, learning about brands is likely to not only flow from mother to child but also from child to mother.

Parent-Child Interactions

One aspect of the other-child interaction dyad is the degree to which mothers yield to purchase requests from their children. Ewidence suggests that mothers do yield to their child's request for cereals at least 67 percent of the time (Atkin 1975, 1978). Ward and Wackman (1972, 1973) and then Ward, Wackman, and Wartella (1977a, 1977b) studied this phenomenon and found that children across age categories asked for products by brand name at least 60 percent of the time. Further, the percent of time mothers yielded was related to children's cognitive stage. Generally, mothers yield more to children in the concrete-operational stage than to children in the pre-operational stage; 22 percent of the time for pre-operational children and 39 percent of the time for concrete-operational children (Ward, Wackman, and Wartella 1977a).

Ward, Wackman, and Wartella (1977a) not only investigated the mother-child interaction dyad in terms of yielding behavior, but also in terms of how often mother and child discussed television advertisements. A purpose of their investigation was to determine the extent to which mothers teach their children about the intent of commercials, and about products and brands within product classes. The results suggest that 60 percent of the mothers discussed with their children the advertised products and brands and about 85 percent of all comments made by mothers about advertising were negative. The results also indicate that age differences were small; mothers of older children (eight to twelve years old) made a few more comments about television commercials than mothers of younger children (five years old). Thus, to the extent that children learn about brands from mothers, it is likely that television advertising will be presented as a negative influence (even though some advertising may be helpful as a pedagogical tool).

All in all, mothers did not appear to be overly concerned with teaching their children about products and brands. They only were concerned with teaching their children how to be good or effective consumers (Ward, Wackman, and Wartella, 1977a). This study is important because it suggests that parents do have a mediating effect on television exposure and toy request. The present study attempts to further this area of study by determining if mothers have a mediating influence on the type of dimensions which underlie children's brand preferences. Generally, the ewidence on the degree to which mothers influence their children's consumer socialization suggests that the influence of mothers depends on the age of the child.

Therefore, the major purpose of this study is to investigate the influence of mothers via the degree to which mothers' brand preferences appear to predict children's brand preferences. Specifically, the degree to which the dimensions used by mothers in making brand preference judgments overlap with the dimensions used by their children is investigated. It is assumed in this study that children and mothers have a common salient set of dimensions underlying brand preference, but that the specific dimensions used in making a particular preference judgment may differ. This is an assumption that is dealt with in detail in the Results section of the manuscript.


Research Design

Children in two age groups, four to five and eight to nine and their mothers were chosen to perform the task of 28 paired-comparisons (David 1963) on stimulus sets of 14 beverages and 14 cereals. Subjects were chosen in such a manner as to maintain representation across gender, age, and two levels of socio-economic status within a large western metropolitan area. School districts were contacted and approval was sought. Letters were then sent out to parents soliciting the cooperation of themselves and their child. This process resulted in achieving a sample size of 104 children and their mothers.

Stimulus Sets

Cereals and beverages were chosen according to the most popular in terms of gross sales in the geographical area for which the study was conducted. Photographs were taken of each stimulus and then laminated on a 3x5 card for easy access. In the situation where a beverage is displayed in both bottle or can, the photograph of the beverage was taken with both containers. An effort was also made so that of the 14 cereals and 14 beverages chosen, 3 were considered to be adult brands and 11 were considered to be children's brands.


Mothers and children both performed paired comparisons on 28 sets of stimuli for both product categories. This method was adopted by David (1963) which reduces the general task of 91 comparisons when using 14 stimuli per set. In this linked-paired comparison method, each stimuli and each pair of stimuli have an equal chance of appearing across 13 subjects. Older concrete-operational children also performed an attribute rating task across the 14 cereals and beverages. These results were used in the analysis of dimension overlap between mothers and their children.

Cognitive Stage

In this study, age was not used as a surrogate for cognitive stage. All children in the study received 3 test (Inhelder and Piaget 1964), which is part of a battery of test to determine cognitive stage. A criterion of two out three would classify a particular child as either preoperational of concrete-operational. Data were collected from 52 children in each age group. The results of the test to determine cognitive stage produced 47 children in the pre-operational stage and 57 children in the concrete-operational stage.

Data Analysis

Testing the overlap of dimensions that underlie brand preference formation between mothers and their children was accomplished in three steps. First, individual MDPREF (Chang and Carroll 1969) solutions for children and for mothers were obtained. This resulted in separate solutions for pre-operational and concrete-operational children and a three dimensional solution for mothers. Second, the number of dimensions in the solutions for both children and parents were determined. Third, simple regression analyses matching dimensional weights for each dimension between mothers and children were performed to determine the relationship between mothers' dimensional weights and their children's dimensional weights. Simple regression was appropriate because there was only one independent and one dependent measure matching individual dimensions from mother's and children's space solutions. It was also used because it was the most straightforward analysis for determining if values on a dimension for mothers predict the values on the same dimension for children. This outcome helped to detect parental influence (i.e., the evaluation of the experience between the mother and child dyad) in the processes of children's brand discrimination and brand preference formation.


Subject Preference Spaces - Mothers and Children

Analyzing the subject spaces (i.e., individual preference vectors) between mothers and children via regression analysis describes the relationship (if any) between the types of dimensions used by these groups of subjects in preference. Specifically, mothers' preference vectors were regressed on children's preference vectors.

Before regression analysis was performed, an MDPREF solution for mothers was generated. Mothers' cereal preferences were represented in a three dimensional space as suggested by FITA2B correlational analysis (Schonemann and Carroll 1970). This analysis was performed using a split-half procedure. The correlations between dimensions of the split-half analysis were .93, .67, .58, for dimensions one, two and three, respectively. Dimension one explained 22% of the variance, dimension two explained 17% and dimension three explained 14%. Dimension four only explained an additional 8% and therefore was not included in this analysis. The fit between the data and the vector preference model was .35 in one dimension, .53 in two dimensions and .61 in three dimensions.

In the case of children, FITA2B analysis between the dimensional weights by cognitive stage produced a zero correlation for a three dimensional solution. Thus, 5 distinct and separate dimensions were the results of MDPREF for childrens' brand preferences. For pre-operational children, the variance explained by the first dimension was 43%, and the second dimension explained 39%. The correlation of fit between the data and the vector model was .32 with one dimension, and .50 with two dimensions which represents an average fit. Considering concrete-operational children, the-first 3 dimensions explained 73% and the fit between the data and the model was .88, .51 and .61 for the three dimensions, respectively.

With respect to preference judgments made for cereals by mothers and pre-operational children, the results of regression analysis (presented as correlations of all the possible combinations between mothers' and children's brand preferences) reveal that a statistically significant relationship between the preference spaces for these two groups of subjects does not exist (results are presented in Table 1). This finding suggests that the dimensions used for making preference judgments for cereals were different between mothers and pre-operational children.



A slight relationship did exist between mothers' preference judgments for cereals and their concrete operational children's preference judgments. Of the nine possible combinations of correlating two, three dimensional spaces, a statistically significant relationship existed for matching dimension two of the mothers' space with dimension two of the children's space, R = -.384, significant at the .05 level. The negative direction of this correlation i plies an inverse relationship on this dimension. The inverse relationship on dimension two suggests that preferences for concrete-operational children area based on the attributes unhealthy, children's cereal, liking the cartoon character, while mothers' preferences are based on the perceived attributes of healthy, adult's cereal, disliking the cartoon character. Because only one of fifteen correlations was significant, it is concluded that a very weak relationship exist between the dimensions (evaluative criteria) used by mothers and children in their cereal preferences.

Subjecting mothers' beverage preference scores to MDPREF, produced a three dimensional solution. This was based on FITA2B analysis where the dimension correlations between split-half runs were .87, .84, .61 for one, two, and three dimensions, respectively. Dimension one explained 24% of the variance, dimension two 16%, and dimension three 15%. Dimension four explained only 9% of the variance and was not included in this analysis. Examining the fit between the data and the vector model produced the following result; chi-square was .35, .44, and .59 in one, two, and three dimensions, respectively.

Children's brand preferences for beverages were captured in a combined two dimensional solution. Correlations were .78 and .95 for dimension one and two, respectively. Seventy-eight percent of the variance was explained in the combined cognitive stage two dimensional solution. The goodness of fit between the beverage preference data and the vector model in two dimension was .38 and .51, respectively.

Upon performing correlating analysis of the beverage preference data, it was determined that a non-significant relationship exists between mothers' and children's preference solutions (refer to Table l for the nonsignificant R's). This result suggests that children and mothers make preference judgments for beverages based on different dimensions.

Stimulus-Spaces - Mothers and Children

Table 2 presents the correlations between mothers' preference stimulus space and their children's preference and perceptual stimulus spaces by product category. The purpose of this analysis was to determine the relationship (if any) between mothers' brand preferences and their children's brand preferences and brand perceptions. The correlations were determined by using the Schoneman and Carroll FITA2B analysis incorporating stimulus weights from each of the separate ALSCAL and MDPREF solutions. These correlations were used to assess the overall fit between mothers' and children's spaces by rotating one space to another space via a least squares fit rotation. Generally, results from these correlations imply that mothers' preference judgments are more closely related to their children's preference judgments for beverages than for perceptual judgments of beverages. This result implies that mothers and children are likely to prefer the same beverages. That is, the projections of the stimuli (beverages) onto the dimensions were similar. Although these projections appear to be similar, the product moment correlation between the preference scores (means) for the two groups was .12, nonsignificant at the .05 level. This statistic indicates a low correlation between the beverages preferred by children and their mothers. However, for cereals, mothers' preference judgments were more similar to their children's perceptual judgments than their children's preference judgments. Comparing children's cereal preference scores with their mothers' preference scores produced a product moment correlation of .42, nonsignificant at the .05 level. One explanation for this inconsistency is the measure of analysis. In one case, stimulus projections and dimensions are being compared, and in the other case, raw preference scores (means) are being compared.


Mother-Child Dimension Overlap - Subject Spaces

Although many researchers outside of marketing have stressed that mothers have an identifiable influence on their children's cognitive abilities, this study has failed to demonstrate that mothers exert much influence over children's brand preference formation. Children appear to develop their own criteria for preferring brands of cereals and beverages that do not correspond with their mother's criteria.



There was an inverse relationship between attribute weights of dimension two (unhealthy, child's cereal, liking the cartoon character) for mothers and concrete-operational children with respect to the cereal preference judgments. The data suggest that the more favorable concrete-operational children were to the attributes comprising the dimension, the less likely mothers were in using these attributes as a basis for cereal preference. No relationship existed between mothers' dimensions and concrete-operational children's dimensions used in beverage preferences. The least amount of similarity in dimension overlap occurred between mothers and preoperational children. There were no similar dimensions between mothers and pre-operational children. In summary, the overlap between mothers and concrete-operational children was at best, only slight. Thus, there is not enough ewidence to support the proposition that mothers exert much brand preference influence on the dimensions used by their children for making brand preference judgments. In order to determine if other family members influence brand preference, an investigation of fathers' or siblings' influence on the dimensions used in children's brand preference may have been useful, but was not considered in this study. These findings, although exploratory, to suggest that mothers do not appear to teach (indirectly via observation) their children about brands within these two products categories with respect to why some brands may be preferred over others.

Stimulus Spaces

Comparing the stimulus spaces (i.e., the dimension weights) for mothers' preferences with children's preferences and perceptions produced a unique finding. The greatest similarity occurred between mothers' preferences for beverages and children's preferences for beverages, and mothers' preferences for cereals and children's cereal perceptions. From this finding, the following inference can be mate; similarity between brand perceptions and preference judgments made by mothers and children is product class specific. Although mothers and children have a tendency to prefer the same beverages, they use different dimensions in making these preference judgments about beverages. For cereals, mothers and children do not prefer the same cereals nor do they use the same dimensions to make cereal preference judgments. One explanation for this is that beverages may not carry as strong an image as cereals regarding those that are considered to children's beverages and those that are considered to be adult beverages (excluding liquor). This would imply that the same beverages could be preferred between mothers and children. Another explanation is that within the product class beverage, some of the brands were merely differences in flavor and not actually brand differences. This limitation could contribute to a possible perception that beverages are more similar than are cereals. Less brand variation implies less preference discrimination between mothers and children.

Mothers' preference judgments (i.e., location of cereals in a preference space) for cereals were more highly correlated with pre-operational and concrete-operational children's cereal perceptions (i.e., location of cereals in a perception space) than with pre-operational or concrete-operational children's cereal preferences. This finding suggests that mothers' preference judgments for cereals are similar to their children's brand perceptions about cereals, implying a relationship exists between mothers' preferences and children's perceptions. A possible explanation for this is the traditional view of cognition preceding affect. Since adults are presumed to be more cognitively organized than children in either of the two stages tested, brand perceptions are stronger and brand evaluations (affect) have been molded into brand preferences. With pre-operational and concrete-operational children, brand perceptions are still forming and evaluations and preferences are even less stable. Thus, from a cognitive development perspective, it is possible for preferences of a group of mothers perceived to be more cognitively organized to correlate with the perceptions of a group of children. This explanation is limited in that some lack of aggregate correspondence toes not necessarily mean lack of individual correspondence. Thus, on an individual basis, mother's cereal preference could coincide with their child' 8 cereal preferences, but in the aggregate, do not. At best, this is only speculative as to the nature of this finding and it is still unclear as to why the finding is product specific.

Future research should consider the influence of fathers as well as siblings on children's brand preference formation. Another variable which may prove to be fruitful in understanding children's brand preference formation is birth order. Differences attributed to by gender and socio-economic status may provide even greater insight into children' s preference formation.


Atkin, Charles (1975), Affects of Television Advertising on Children-Survey of Children's and Mothers' Responses to Television Commercials. Technical Report, Department of Communication, Michigan State University.

Atkin, Charles (1978), "Observation of parent-child interaction in supermarket decision-making," Journal of Marketing. 42, 41-45.

Chang, Jih and J. Douglas (1969), How to Use MDPRE. A Computer Program For Multidimensional Analysis of Preference Data. Unpublished Manuscript, Bell Laboratories.

David, H. (1963), The Method of Paired Comparisons, New York: Hafner Publishing Company.

Inhelder, Barbel and Jean Piaget (1964), "The early growth of logic in the child. New York: Horton.

Robertson, Thomas and John Rossiter (1977), "Children's response to commercials," Journal of Communicates 27, 10 1- 106.

Schonemann, Paul and J. Douglas Carroll (1970), "Fitting One Matrix to Another Under Choice of a Central Delation and a Rigid Motion," Psychometrika, 35 (June), 245-255.

Ward, Scott and Daniel Wackman (1972), "Children influence purchase attempts and parental yielding," Journal of Marketing Research, 9, 316-319.

Ward, Scott and Daniel Wackman (1973), "Children's information processing of television advertising," New Models for Mass communication Research. P. Clarke (ed.), Beret Hills: Sage Publication, 119-146.

Ward, Scott, Daniel Wackman and Ellen Wartella (1977a), How Children Learn to Buy, New York: Sage Publication.

Ward, Scott, Daniel Wackman and Ellen Wartella (1977b), "The development of consumer information processing skills: integrating cognitive development and family interaction theories," in Advances In Consumer Research, ed. W. D. Pearault, Atlanta, Georgia: Association for Consumer Research, 166-167.

Zigler, E. and I. Child (1969), "Socialization," in The Handbook of Social Psychology. 3, eds. C. Linzey and E. Aronson. Reading. MA: Addison Wesley, Publishers, 480-490.



Kenneth D. Bahn, Virginia Polytechnic Institute and State University


NA - Advances in Consumer Research Volume 14 | 1987

Share Proceeding

Featured papers

See More


Making Sense of Spontaneity: In-The-Moment Decisions Promote More Meaningful Experiences

Jacqueline R. Rifkin, Duke University, USA
Keisha Cutright, Duke University, USA

Read More


Slow and Steady versus Fast and Furious: The Effect of Speed on Decision Making

Ellie Kyung, Dartmouth College, USA
Yael Shani-Feinstein, Ben Gurion University, Israel
Jacob Goldenberg, IDC

Read More


K9. Measuring Internet Slang Style in Advertisement: Scale Development and Validation

Shixiong Liu, Shenzhen University
Yi Wu, Tsinghua University
Wu Gong, Shenzhen University

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.