Children=S Use of Consumer Heuristics

ABSTRACT - The focus of this paper is to investigate the use of cognitive heuristics by children, compared to adults, as they make consumer-type quality judgements in response to product labels. It is found that 10yr-old boys show almost no use of heuristic processing but that same-age girls and adults do. The research method uses the novel approach of encapsulating negative information within a label heuristic, so that a negative quality attribution shows detailed, elaborative processing while a positive quality attribution shows use of the established ALength is Strength@ heuristic. Policy issues for legislators and marketing implications for business are discussed.


Roger Marshall, Ng Ee Chuan, and Na WoonBong (2002) ,"Children=S Use of Consumer Heuristics", in NA - Advances in Consumer Research Volume 29, eds. Susan M. Broniarczyk and Kent Nakamoto, Valdosta, GA : Association for Consumer Research, Pages: 108-113.

Advances in Consumer Research Volume 29, 2002     Pages 108-113


Roger Marshall, Nanyang Technological University

Ng Ee Chuan, Global Transaction Products

Na WoonBong, Silla University


The focus of this paper is to investigate the use of cognitive heuristics by children, compared to adults, as they make consumer-type quality judgements in response to product labels. It is found that 10yr-old boys show almost no use of heuristic processing but that same-age girls and adults do. The research method uses the novel approach of encapsulating negative information within a label heuristic, so that a negative quality attribution shows detailed, elaborative processing while a positive quality attribution shows use of the established "Length is Strength" heuristic. Policy issues for legislators and marketing implications for business are discussed.


There is a clear moral imperitive to provide critical information to consumers, that is backed up by appropriate legislation in most countries. Consumers with special interests will peruse information carefully and often knowledgeably; so, legislation and morals aside, it is obviously in the best interests of marketers, too, to provide adequate information to these consumer segments. Thus, forinstance, diabetics and others with medical conditions or strongly "green" consumers will carefully check what information is (and is not) given on specific products and behave accordingly. However, there is an old, but well-established, literature that suggests that most people neither read, understand nor remember most label information (Patton 1981; Freiden 1981).

That labels are not often read is not so much because people do not care, but rather because so much decision behavior is automated. When faced with some evaluative task (such as choosing between several brands in a product category), consumers tend to conserve scarce information processing capacity by using mental shortcuts. Taylor and Thompson suggest that "Biases, oversights and shortcuts appear to be intrinsic to the cognitive system rather than introduced by extraneous interfering, affective commitments or needs." (Lloyd and Mayes, 1984, p639). Other cognitive scientists agree that people often form simple, experienced-based rules, or heuristics, to help interpret information quickly and so form judgments. These cognitive heuristics have been described variously as rules people seem to apply to "reduce complex inferential tasks to... simple cognitive operations" (Cervone and Peake, 1986, p493); as "inferential rules of thumb" (Allison, Worth and King, 1990, p801); and as "simple schemas or decision rules" (Axom, Yates and Chaiken, 1987, p30). Examples abound in our daily surroundings. The way in which students believe that the written rather than the spoken word is more authoritative is a familiar example. Experience has presumably taught students that 'if it were not true, then it would not have been committed to paper’.

Kahneman and Tversky (1973) have found two classes of cognitive heuristics commonly used; "representative" and "availability". The former are used to determine "the extent [an object] represents the essential features of it’s parent population" (p163). The latter are used "whenever [a person] estimates frequency or probability by the ease with which instances or associations come to mind" (p174). It is from this latter category that many of the devices used in product labeling are drawn. Associations learned through experience are cued by the use of particular colors, brand logos and illustrations. For instance, illustrations of particular fruits may have exotic or health associations; depictions of buildings and geographical features that carry local or national significance could trigger strong associations of patriotism or loyalty. Probably the most frequently invoked heuristic in marketing is a brand name, which can act as a cue to trigger a set of relevant associations that together form the brand heuristic (Keller, 1991).

Children, however, form a group that is especially vulnerable in this context, particularly within the context of product labelling. Yet, there is no legislation whatsoever in place in order to protect them from either deliberate deception or from simple misunderstanding caused by their lack of experience and knowledge. Thus, a child exposed to words that he or she does not understand, but other consumers might, may well fall back on the use of cognitive heuristics that would lead to a meaning to be drawn that could differ from the intended meaning. Public policy makers could well be interested in the possible need for specific educational or legislative measures that might be required if children do indeed interpret persuasive commercial information in different ways to adults.

There is another side to this coin, too; the issue of children’s use of heuristics in their response to label information holds interest for marketers. Children are rapidly becoming a consumer force to be reckoned with (Liebeck, 1994; Crowell and Hsieh, 1995) and marketers are interested in designing product packages and labels to suit children’s cognitive style.

This is the focus of the present research; to investigate the use of cognitive heuristics by children using information on product labels in a consumer decision situation. We are interested to know when chilren commence to use heuristics, so that recommendations can be made to both marketers (about the type of information than is appropriate on labels designed for children’s use) and to governments or consumer groups interested in protecting consumers in general, and children in specific (about potential legislation that could be advantageous to these groups).


The use of cognitive heuristics indicates some sophistication and experience in the use of language, therefore it seems to follow that the use of heuristics should develop in accordance to the cognitive ability of a developing child. Consequently, child development theories offer a guide to our expectations with regard to when we would expect children to begin adopting heuristic modes of thinking. Piaget’s Theory of Cognitive Development suggests that children go through specific, identifiable developmental stages; including the sensorimotor, preoperational, concrete and formal operations stages. Children in the formal operations stage of development are between 12 to 15 years of age and any further cognitive development is claimed by Piaget to be unlikely; it also appears that many cognitive gender differences noted in children tend to even out during this time, as boys "catch up" to their sisters (Maccoby and Jacklin, 1974). In the penultimate development stage, known as the concrete operational stage, a child begins to think logically and develops certain fundamental logic rules. During this time, children overcome many of the errors that characterize their thinking in the prior, preoperational, stage and, in a sense, are free from the total reliance on perception and intuition which is characteristic of the even earlier sensorimotor and preoperational children (Lefrancois, 1990). However, the group of children aged between 7 to 11, in the concrete operational stage, are not yet quite cognitively mature, but are in a transitional stage between reliance on perception and intuition, and formal or logical thinking (Hetherington and Parke 1993). Thus we would expect that the use of cognitive heuristics will increase in children as they grow towards cognitive maturity, around the age of 10 to 12 years. Our first hypothesis simply draws upon this basic idea, that a child’s maturity parallels his or her use of heuristics, including consumer heuristics.

H1: Children in the concrete operational stage will not make as much use of (consumer) heuristics as cognitively mature adults.

A second hypothesis, related to the first, is based on the premise that if children do follow the expected cognitive developmental patterns, then it is at least possible that girls will display use of heuristics to a greater extent than same-age boys (until the formal operations stage has been passed, at least). It is well-documented that girls do display a different intellectual maturation pattern to boys, particularly in respect to the acquisition of language skills, which seems relevant in the present context (Kimble, Garmezy and Ziler, 1984; Rebok, 1987; Maccoby and Jacklin, 1974). Consumer heuristics, in particular, may be developed more quickly in girls, in that they may relate to shopping behavior more than boys. It is the subjective opinion of the authors that it seems more likely that girls will have more shopping experience than boys even in these enlightened times, and therefore may exhibit a stronger tendency to use cognitive shortcuts to assist them when making consumer decisions.

H2: Girls who have not reached the formal operations stage of cognitive development will display greater use of (consumer) heuristics than boys of the same age.


Choice of consumer heuristic

The overall research design concept is to expose respondents of varying ages to products bearing labels containing a heuristic, and then ask them to make a quality judgment (a typical consumer use of heuristics) concerning the product. A difficulty immediately arises in that a positive attribution from the label will result in a more positive judgment of the product but will not tell us whether or not the respondent was using a heuristic or an elaborative, detailed method of processing information. If a recall test of the information detail is used, it may be memory that is tested rather than information processing method. However, one availability heuristic that has been noted, but not yet rigorously investigated, offers a solution. The "length is strength" (LIS) heuristic is the quality associations heuristic that is cued merely through the presentation of a quantity of written label information regardless of the actual informational content (Patton, 1981; Robertson and Marshall, 1987).

Assam and Bucklin (1973) measured several aspects of perceived product quality in response to the provision of nutritional information. They found a decrease and then an increase in perceived quality across four levels of information. Patton (1981) and Freiden (1981) both found a monotonic, positive relationship between product preferences and increasing amounts of nutritional information. In fact, Patton explicitly noted that:

"...(consumers) might make purchase decisions largely on the basis of the amount of information available...although increased amounts of information may have a neutral or even dysfunctional effect on the decision process, amount of information may still override content of information as a decision criteron, the result being a greater likelihood that some consumers will purchase a brand that provides more information than it’s competitors." (Patton 1981, p93).

A little later, Robertson and Marshall (1987) demonstrated the use of the heuristic. All these studies, however, dealt with adding more positive information to the label which increased the quality perception of the labeled good. On the face of it, however, there seems no reason why negative information should not be substituted for positive to yield the same result. If this is indeed the case, then this heuristic will serve very well within the present research context. Increasing positive judgements in response to more negative information would indicate that the information has being perceived heuristically as a (positive) information chunk rather than as a list of (negative) items of information. Negative judgments of extra information, on the other hand, will indicate situations where elaborate, non-heuristic information processing of the negative information took place rather than heuristic processing.

The question of involvement level complicates the research a little. Earlier research suggests that heuristics can be invoked in both a high and low involvement setting, but that a heuristic response is more often found under conditions of low motivation and/or low ability to process message content (Eagly and Chaiken, 1984). Involvement is usually defined as the level of personal importance and/or interest evoked by a stimulus. In a purchase situation that commands only minimal involvement, devices used in product labeling B such as a brand logo or color scheme B may provide a sufficient cue for purchase action (Allison, Worth and King, 1990). On the other hand, under high risk situations, consumers are likely to engage in complex information search and evaluation (Schiffman and Kanuk 1991). It has already been suggested, above, that in very high involvement situations, such as "green" consumers, or those with an illness that required dietry care, it would be expected that information would be perused with care. Apart from such extreme circumstances, though, it seems probable (to the authors, at least,) that the LIS heuristic should operate no matter what the involvement level as previous research, already quoted, has shown quite clearly that consumers rarely understand label information even if they notice it B thus a heuristic is likely to be invoked.

This contention is further supported by the work of Petty and Cacioppo (1983), who suggest that an elaborative information processing is only used when the respondent has the the motivation and the ability to process it B the work quoted previously suggests that the ability is often not present no matter what the involvement level. This may be particularly true not only in the more obvious case of high technology products but also where children are the label-information audience.

Thus, amount of negative information (control plus two levels), involvement (two levels), Gender and Age (two levels each) are manipulated in a simple research design to test the effect of increased negative label information upon quality judgements within a consumer judgement environment. Responses were collected by interface survey methods.


The overall sample consists of 104 respondents, 52 school children and 52 undergraduats, all collected on a convenience basis. The first sub-set of the sample is composed of 26 male and 26 female schoolchildren B this was the group selected to test whether the LIS heuristic is developed in parallel with general cognitive development and life experience. The choice of age group for the child sample was made in the light of Piaget’s Theory about stages of cognitive development. The 52 children finally selected were 10 year-olds from three better-performing classes in a public primary school, thus minimizing any bias that might have arisen from using children with differing language ability. Care was also taken to pick schools with a history of academic excellence, so as to approximate the intellectual potential of their undergraduate counterparts.



A convenience sample was also drawn from a nearby university, to represent a group of cognitively mature people. Business undergraduates were intercepted and interviewed along university corridors and reading rooms; again, 26 males and 26 females were interviewed.



Product mock-ups were made for a fictitious, new brand of ice cream. This product was chosen as it was felt that both children and adults were equally familiar with the product category. Identical tubs of ice cream were built and given a fictitious brand name so that there would be no prior brand image involved. The three mock-up ice cream tubs each had separate, and increasing, levels of negative information available on the labels.

A separate preliminary test, run on ten adult subjects from the university, had been conducted earlier to determine the connotation of each individual ingredient to be used as a negative information item. This list was then made into three different labels, each to be used on one tub of ice cream. The labels were made up such that no negative information was included in the lowest information level, which was designed to act as base-level, control, information. However, as information levels increase, so does the declared content of negative substances.

A post hoc test was conducted to ensure that the final labels were perceived, by both the children and the adults, in the manner they were designed to be. A sample of 10 year-old children and university students (15 male and 15 female in each case) were individually asked to rate how they would feel about eating some food containing each of the 26 items featured on the labels. A 5-point scale was anchored by "very happy" and "very unhappy". Paired comparison t-tests between the means of the items used on each label, (reported in Table 1) show that separation in the required direction was achieved. In addition, the mean value of the items added to the low level label were compared to the mean of the low level label items, and the extra items on the high information level labels was compared to the medium level information. This shows that the increased information was really perceived as negative by both children and adults when attention is drawn to it’s nature.


A simple, one-page questionnaire was designed. Ratings of the quality of each tub of ice cream were collected from the adults on a single, seven-point semantic differential scale. Subjects were then asked two questions designed to assess their level of involvement with the judgment process, one question phrased positively and the other negatively.

Children were asked to express their quality perception by using a non-numeric measurement device. A cardboard strip, with a center-marker dividing a colored portion from an uncolored one, was threaded through a backing card. This strip can then be pulled through the backing board to move the mid-point toward one end or the other. Several backing cards were made up with suitable scale anchors printed on them to measure the desirable attributes ("High Quality"/Low Quality, to measure quality perception; "Do not care"/"Care" and "Important"/Unimportant to measure involvement levels). The reverse of the backing card had a scale marked from which the researcher could read off the judgment value. This method was favored instead of a standard seven-point scale as the investigators felt that it would facilitate comprehensibility and lower the extreme response tendencies of the children. No difficulty was experienced in administering either instrument.




Each of the 104 subjects was interviewed alone, with one interviewer showing the ice cream tubs while the other administered the survey. In order to eliminate any sequential bias, the order in which tubs were exposed was randomized through the use of a die. Under the high involvement manipulation, subjects were allowed to handle and view the tubs as they were brought out in the predetermined, random order. Respondends in this experimental condition were allowed as much time as needed to inspect the stimulus, but were only allowed one exposure to each tub. Subjects in the low involvement manipulation were only allowed a five second exposure to each tub. Experimenters rotated each tub once to give subjects a view of the entire tub, after which the tub was put out of view. It is recognized that this method of manipulating involvement is somewhat unorthodox, but it is well-documented that when insufficient time is available for elaboration of the stimulus message then a low-involvement processing strategy is evoked.

After being shown the experimental stimuli, the undergraduate respondents were asked to fill out the questionnaire by which their quality assessments and involvement levels were captured. The same process was repeated with the sample of 10 year-olds, except that the cardboard slide was used to help them capture quality and involvement assessments.



Scale reliability for the involvement measurement was assessed by correlating the two items; the correlation coefficient of .56 is highly significant (p<.001). Consequently a single item representing involvement was calculated by taking a mean of the two items. A t-test of the high and low involvement conditions shows that separation was achieved (t = 28.36, p <.001, Mean(low)= 3.45, Mean(high)= 5.84, where the larger number implies higher involvement).

An analysis of variance was then conducted in which Quality Perception was the dependent variable and Information Level, Involvement, Gender and Age were the independent. The output of this analysis is displayed in Table 2.

Age is highly significant, as expected, and Information Level approaches significance. This latter relationship does not matter in a sense, as it is really only of interest in the interaction situation. Although the results for Involvement do approach significance, it appears that the involvement level over the measured range, as expected, is actually immaterial to the judgment results in this instance. No three-way interactions are significant, but there are two sets of two-way interactions noted, the first being between Age and Information Level and the second Gender and Information Level.


Inspection of the means displayed in Table 3 reveals the relationship between the first set of variables, Age and Information Level. This suggests that the sample of children did not respond to the heuristic cue whilst the adult, undergraduate sample did, as quality judgments for the children seem to drop as negative information levels increase, whereas the undergraduates show increasing quality perception with increasing information. Thus Hypothesis 1 is supported.


The second interactive effect, also apparent in the data in Table 3, reveals yet more detail. It can be seen that both males and females experience an overall increase in quality perception in response to increasing levels of informtion, although there is some difference in the functions in that males have a tendency to experience a drop in quality perception when exposed to medium information levels. This treatment however, involves looking at gender differences without separating the two different age samples. In order to have a clearer picture of the cause of the gender differences, quality perceptions of the two age groups, shown in numbers in Table 3, are also shown diagramatically in Figure 1.

The source of the male-female differences becomes clear once the two are separated. Both adult females and males actually display a similar (heuristic) information processing strategy, with quality perception rising with increasing (negative) information levels. This observation is supported by conducting an ANOVA on the adult responses, with Perceived Quality as the dependent variable and Information Level and Gender the independents. Although Information Level shows a strong main effect (F = 23.222, p <.001), the effect for gender is non-significant, as are the 2-Way interactions. The children’s plots, on the other hand, show that quality perceptions of girls fluctuate around the 4.30 mark through all information levels, while boys display a negative monotonic relationship over increasing negative information levels. In this case there is a significant main effect for information level (F = 4.76, p = .010) but the gender effect is non-significant. The interaction between gender and information level, however, is significant (F = 4.277, p = .016). This data offers support for Hypothesis 2.





The above analysis uses mean scores of the various groups, and this does hide some information. A scan of individual responses was undertaken, noting whether each child respondent demonstrated a pattern of falling quality perceptions (indicating use of an elaborative model) or a rising quality perception in response to increasing levels of information (indicating use of a heuristic processing strategy). Results show that a significant number of the girls were actually displaying a LIS heuristic response while others were not, giving rise to an overall, mean flat line response, whereas the overwhelming majority of the boys were using an elaborative, non-heuristic strategy. Of the 26 girls, 12 displayed a clear use of the LIS heuristic, 11 displayed use of an elaborative technique, while the remaining 3 had near flat-line responses. Of the boys, on the other hand, only 3 show a clear use of heuristics, 2 show a flat-line response and all the rest show strong evidence of elaborative processing.


The purpose of this work is to investigate the development of the use of heuristic information processing in children in a consumer decision context. The results clearly confirm the presence of heuristic processing in the adult sample, but the children’s responses shows that 10 year-old boys display a strong negative response to more negative information, while almost half of the same-age girl respondents were already using more sophisticated, heuristic processing strategies. Thus both research hypothese are supported.

Involvement had no great importance here, probably as the overall level of involvement stimulated in the respondents was not as high as it might be for someone for whom a similar decision in real life might result in an illness or, perhaps, a breach of personal values. That is, the measured involvement was probably over a fairly low absolute range of involvement (which may wel be typical of most label-scanning activities in real situations).

Of course, only one heuristic has been used here, "Length is Strength". Whether or not the work can be extended to other labeling situations, other information situations and other heuristics cannot be judged from this data. Of particular interest would be work on the use of brand as a heuristic cue among children, to see if a similar pattern of usage emerged in that situation as was evident in the present research. Of even more interest might be to consider the extent to which puffery is believed by children, when they are analyzing the material in an elaborative manner.


The most interesting public policy implication comes from peripheral information gleaned in the research process - the fact that the Length is Strength heuristic seems valid for negative label information has some fairly odd ramifications. Public policy makers should bear in mind that any response to calls of consumer groups for adequate provision of information on product labels should be made in the realization that higher informational content alone may often serve as a heuristic cue to quality, unless appropriate educational measures are also undertaken. The spectre that is raised by this statement is that a consumer may be induced into a higher quality perception of a product merely by increasing the level of information content on the product label regardless of whether the information is positive or negative!

That children under the age of 10 do not use heuristics to a large degree might also be considered by legislators, who may consider that special care should be enforced to give non-technical information (able to be comprehended by the child consumer) and usage and benefit claims that are honest, rather than mere puffery based around association to some current child fad hero.


To the marketer, there are interesting implications for advertising and packaging communications. Whilst adults do seem (probably under most but very high involvement conditions, at least,) to attribute quality to amount of information no matter what the content of the actual information, children do not. Full disclosure of information that may have negative connotations may not hurt the quality perceptions of adults, but may affect the attributions of children. Furthermore, if children of less that 10 years of age do not commonly use heuristic processing, the speed of presentation of information chunks may have to be slower than for adults. The use of brand names (a very common form of consumer heuristic) may even be in doubt for young children, without, at least, n extensive educational program encapsulated in the promotion program of the marketer.


Allison, S. T., Worth L. T., & King, M. W. (1990), "Group Decisions as Social Inference Heuristics", Journal of Personality and Social Psychology, 58, 5(May), 801-811.

Assam, E. H. & Bucklin, L. P. (1973), "Nutrition Labeling for Canned Goods: A Study of Consumer Response", Journal of Marketing, 37(April), 32-37.

Axom, D. Yates, S. & Chaiken S. (1987), "Audience Response as a Heuristic Cue in Persuasion", Journal of Personality and Social Psychology, 53, 1(January), 30-40.

Cervone, D. & Peake, P. K. (1986), "Anchoring Efficiency and Action: The Influence of Judgmental Heuristics of Self-efficacy Judgments and Behavior", Journal of Personality and Social Psychology, 50, 3(March), 492-501.

Crowell, T. and Hsieh, D., "Little Emperors", Asiaweek (December 1, 1995), 44-50.

Eagly, A. H. & Chaiken, S. (1984), "Cognitive Theories of Persuasion", Advances in Experimental Social Psychology, 17, 267-359.

Freiden, J. B. (1981), "The Effect of Nutrition Information on Brand Rating: Testing for a Non-Use Benefit", The Journal of Consumer Affairs, 15 (Summer), 106-114.

Hetherington, E. M. and Parke, R. D.(1993), Children Psychology: A Contemporary Viewpoint, 4th Edition. New York; Singapore: McGraw Hill.

Kahneman, D. & Tversky, A. (1973), "On the Psychology of Prediction", Psychological Review, 80, 237-251.

Kahneman, D., Slovic, P. & Tversky, A. (1982), Judgment Under Uncertainty: Heuristics and Biases, New York; Cambridge University Press.

Kimble, G. A., N. Garmezy & E. Zigler (1984), Principles of Psychology (6th Ed.). John Wiley & Sons, New York.

Keller, K. L. (1991), "Cue Compatibility and Framing in Advertising", Journal of Marketing Research, 27(Feb), 42-57.

Liebeck, Laura 1994, "Billions at stake in growing kids market", Discount Store News, Feb 7, Vol 33 (3), 41-43

Lefrancois, G. R. (1990). The Lifespan (3rd Ed.), Wadsworth Publishing Company, New York.

Maccoby E. E. & C. N. Jacklin (1974), The Psychology of Sex Differences, Stanford University Press, California.

Patton, W. E. (1981), "Quantity of Information and Information Display Type as Predictors of Consumer Choice of Product Brands", The Journal of Consumer Affairs, 15 ( Summer), 92-105.

Petty, Richard E and Cacioppo, J. T. (1983), "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement", Journal of Consumer Research, 49 (Winter), 89-98.

Rebok, G. W. (1987), Life-Span Cognitive Development, Holt, Rhinehart and Wilson, New York.

Robertson, K. R. & Marshall, R. (1987), "Amount of Label Information Effects on Perceived Product Quality", International Journal of Advertising, 6, 3, 199-205.

Whan P. C. & Parker, L. V. (1981), "Familiarity and Its Impact on Consumer Decision Biases and Heuristics", Journal of Consumer Research, 8(September), 223-230.



Roger Marshall, Nanyang Technological University
Ng Ee Chuan, Global Transaction Products
Na WoonBong, Silla University


NA - Advances in Consumer Research Volume 29 | 2002

Share Proceeding

Featured papers

See More


F2. Can Stricter Ethical Standards Increase Tolerance for Ethical Misconduct?

Olya Bullard, University of Winnipeg
Sara Penner, University of Manitoba, Canada
Kelley Main, University of Manitoba, Canada

Read More


Resolving Humorous Incongruity in Advertising Facilitates Impressions of Firm Competence

*Chi Hoang, Norwegian School of Management, Norway
Klemens Knoferle, Norwegian School of Management, Norway
Luk Warlop, Norwegian School of Management, Norway

Read More


“It’s Not You, It’s Me”: How Corporate Social Responsibility Decreases Customer Citizenship Behavior

Sofia Batista Ferraz, EAESP-FGV
Andres Rodriguez Veloso, University of São Paulo, Brazil
Diogo Hildebrand, Baruch College, USA

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.