Children's Conceptions of Medicine: the Role of Advertising

Thomas S. Robertson, Center for Research on Media & Children, The Wharton School
John R. Rossiter, Center for Research on Media & Children, The Wharton School
Terry C. Gleason, Center for Research on Media & Children, The Wharton School
ABSTRACT - This paper examines three important studies concerning children's orientations toward health and proprietary medicines and the relative impact of advertising on children's beliefs, attitudes, and proprietary medicine usage patterns. The report on the Wharton study is the first data emanating from this recently-completed project.
[ to cite ]:
Thomas S. Robertson, John R. Rossiter, and Terry C. Gleason (1978) ,"Children's Conceptions of Medicine: the Role of Advertising", in NA - Advances in Consumer Research Volume 05, eds. Kent Hunt, Ann Abor, MI : Association for Consumer Research, Pages: 515-517.

Advances in Consumer Research Volume 5, 1978      Pages 515-517


Thomas S. Robertson, Center for Research on Media & Children, The Wharton School

John R. Rossiter, Center for Research on Media & Children, The Wharton School

Terry C. Gleason, Center for Research on Media & Children, The Wharton School

[Funded by National Science Foundation, Grant No. APR76-21130]


This paper examines three important studies concerning children's orientations toward health and proprietary medicines and the relative impact of advertising on children's beliefs, attitudes, and proprietary medicine usage patterns. The report on the Wharton study is the first data emanating from this recently-completed project.


An issue of some consequence in the formulation of broadcast regulatory policy is the impact of medicine advertising on children. This issue was heightened in 1975 by a petition to the Federal Communications Commission emanating from the Attorneys General of nineteen states. In their petition the Attorneys General requested a ban on medicine advertising on television between 6:00 a.m. and 9:00 p.m. (Attorneys, 1975).

The substance of this petition was not so much that exposure to medicine advertising leads children to consume medicine, since consumption is presumably mediated by parents, but that exposure to medicine advertising induces receptive attitudes toward medicine-taking. Concern was voiced that such attitudes later become manifest in increased consumption and perhaps over-consumption, or dependency, on these products in adulthood.

The Attorneys General request for a ban on broadcast advertising of proprietary medicines before 9:00 p.m. was subsequently denied by the Federal Communications Commission on the grounds of lack of evidence for the effects suggested:

"... in the absence of empirical evidence to support the claim of a causal connection (between misuse or abuse of drugs and televised advertisements of over-the-counter drugs) it would be unreasonable and arbitrary for (the Commission) to accept the idea that otherwise lawful advertising should be prohibited" (FCC, 1976).

The purpose of this paper is to review three key studies relevant to the issue of advertising's impact on children's conceptions of medicine. In particular, we will focus on some initial results of a study just completed at Wharton, the objective of which is to provide the missing evidence concerning the link between advertising and medicine conceptions. A point of clarification is in order first, however, concerning children's exposure to proprietary medicine advertising. Industry self-regulation codes prohibit the advertising of medicines on "children's programs" (National Association of Broadcasters, 1974). However, these codes apply only to programs "initially designed for children" --that is, those shows which are concentrated on Saturday mornings. Most children's viewing, however, is to non-children's programs. Among the top fifteen shows viewed by children, only three are shown in time periods covered by the NAB children's codes and 85% of children's viewing is to non-children's programs (Adler, 1977).


There is little systematic research evidence on the effects of medicine advertising on children. Most research concerning advertising and drug products has focused on teenagers, rather than children, and has been concerned with illicit drug use, rather than proprietary medicine use.

The most significant studies to-date are twofold. The first is a study by Milavsky, Pekowsky and Stipp (1975-76) of NBC focusing on television drug advertising and its relationship to proprietary and illicit drug use among teenage boys. The second is a study by Atkin (1975) of Michigan State University, addressing the effects of television advertising on fifth, sixth, and seventh grade children. In this paper we also want to present initial results from the recently-completed Wharton study, which examines the impact of proprietary medicine advertising among third, fifth, and seventh grade children.

The NBC Study

This is a study of teenage boys utilizing a longitudinal panel design. The basic thrust of the research is to relate television viewing to reported usage of proprietary drugs and illicit drugs. The strength of this study is the measurement of television viewing, which is then combined with Broadcast Advertisers Reports, Inc. data which tracks commercials appearing on programs. Thus, the researchers were able to derive a drug advertising exposure measure indicating the total number of advertising seconds for proprietary medicines to which each boy was exposed. This procedure is described in detail by Milavsky, Pekowsky and Stipp (1975-76).

The NBC research findings indicate a positive, but relatively, weak relationship between exposure to proprietary advertising and reported usage of proprietary medicines. This relationship is accentuated in homes where there are many proprietary drugs around the house.

By contrast, the NBC research finds a negative relationship between drug advertising exposure & reported usage of illicit drugs -- a finding similar to that of Hulbert (1974) in research with college students. Illicit drugs, however, are not our concern in this paper, nor is there advertising for illicit drugs.

The Atkin (Michigan State) Study

The Atkin study is based on data collected from a sample of 256 fifth, sixth, and seventh grade children selected from schools in urban, suburban and small town areas in Michigan. The basic thrust of the research is the examination of the relationship between television advertising exposure and a relatively rich set of dependent variables concerning children's beliefs, attitudes, and usage patterns for proprietary medicines.

The strength of the Atkin research is in his conceptualization of a set of interesting mediating variables affecting any proposed advertising exposure-medicine usage relationship. The weakness of the Atkin data, in our opinion, is his measure of "medicine advertising exposure," which was constructed by multiplying amount of viewing to a limited set of television programs by the child's reported frequency of attention to four specific medicine commercials.

Atkin's results may be summarized in terms of the following medicine advertising exposure relationships, all of which are based on sixth-order partial correlations--controlling for grade, sex, social class, scholarship, child's frequency of illness, and parent's approval of medicine use.

Population incidence of illness. Children with high exposure to medicine advertising perceive that people are more often sick (r = .14) and that they more often take medicine (.14).

Belief in medicine. High medicine advertising exposure correlates with the child's belief in the quickness of relief after taking medicine (.10).

Illness anxiety. Children with high exposure to medicine advertising worry more about getting sick (.14).

Approval of medicine. The relationship between medicine advertising exposure and the child's approval of medicine is .12.

Medicine efficacy. Children with high medicine advertising exposure are more likely to report feeling better after taking medicine (.12).

Medicine usage. There is a general lack of relationship between medicine advertising exposure and medicine usage (.03).

In general, these results suggest that medicine exposure does, to a certain extent, influence the child's conceptions of illness and medicine. These relationships tend to be accentuated somewhat among the smarter children (as measured by scholastic performance) and among the higher social status children. Other variables such as age and sex of the child, parental attitudes toward medicine, or the child's frequency of illness all show inconsistent patterns.

Neither the NBC nor the Atkin study begins to settle the issue of proprietary medicine advertising's impact on children's medicine conceptions. A major fault characterizes both studies: the reliability of the measures is not reported. Correlations between unreliable measures can be seriously attenuated. Although both studies employed multi-item measures of medicine advertising exposure which are likely to be quite highly reliable, the reliabilities of the "effects" measures are unknown. The unreliability problem is particularly acute in the Atkin study (the only study with children as subjects) where single-item measures were employed. The consequence of unreliability is to produce lower correlations than might have been obtained with reliable measures. The possibility remains that the impact of medicine advertising on children is substantially larger than previously reported.

The Wharton Study

The Wharton Study was conducted in order to present a more conclusive investigation of children's conceptions of medicine in relation to various socialization factors that may be contributory. Exposure to televised medicine advertising was one of these factors.

The sample consisted of 673 children in approximately equal numbers from the third, fifth and seventh grades (ages 8 to 12). The sample ranged from

disadvantaged to upper-middle class families. Approximately equal numbers of boys and girls were interviewed at each grade level. The medicine advertising exposure measure was constructed in the same manner as in the Milavsky et al. NBC study. A random sample of 35 television shows was taken, excluding school hours and late night programming (past 11:30 p.m.). Children were asked to indicate how often they had watched each program on the last four occasions preceding the interview. Broadcast Advertisers Reports were then consulted to determine the number of medicine commercial minutes appearing on each show over the reporting period. The shows were weighted by the medicine advertising count and by viewing frequency, then summed to derive an overall measure of medicine advertising exposure. Reliability for this measure was 73. [Reliability was computed using Cronbach's coefficient a (Cronbach, 1951). This measures the internal consistency of a series of items used to define a variable (see also Rossiter, 1977). The size of a reflects the degree to which the set of items comprising a test is oriented toward a single underlying dimension.]

The "effects" variables for this study consisted of: 1) Belief in the efficacy of over-the-counter medicines, 2) Affect toward (liking of) over-the-counter medicines, 3) Intention to take over-the-counter medicines when symptoms occur,  4) Request frequency to parents for over-the-counter medicines, and 5) Usage frequency for over-the-counter medicine. Two additional variables comparable to the Atkin study were also examined -- the child's perception of the population's incidence of illness and an illness anxiety measure.

Each "effects" variable was measured by five self-rated items. These items comprised five categories of over-the-counter medicines: headache remedies, stomach remedies, cold remedies, cough remedies, and vitamin preparations. Reliability coefficients computed over the five items for each variable were: Beliefs .81, Affect .81, Intentions .75, Requests .76, Usage .52, Population Incidence .59, and Illness Anxiety .84.

Results. The findings are shown in Table 1, including the zero-order correlations and the fifth-order partial correlations. The variables controlled for (generally similar to Atkin) are grade, sex, social class, child's illness level, and parental supervision of medicine use.

In general, the magnitude of relationships is similar to Atkin, suggesting a somewhat tenuous association between medicine advertising and the several "effects" variables. The results may be summarized as follows:

Medicine advertising exposure shows a zero-order correlation of .11 (significant at .01 level) with beliefs in the efficacy of medicine but only .04 at the fifth-order level.

The relationship builds slightly when medicine advertising exposure is related to affect toward medicine (.14 zero-order and .07 fifth-order partial correlations).

The relationship improves again when medicine advertising exposure is related to intent to take medicines when ill (.17 zero-order and .10 fifth-order).

The strongest relationship holds between medicine advertising exposure and reported requests for medicine when ill (.24 zero-order and .18 fifth-order).

When medicine advertising exposure is related to usage, the zero-order correlation is only .07 and the fifth-order .00.

The relationship between advertising and the child's perception of the population incidence of illness is .00 (zero-order) and -.10 (fifth-order).

Finally, the medicine advertising exposure/illness anxiety relationship is .22 (zero-order) and .13 (fifth-order).



Interestingly, there is an absence of an advertising/ usage relationship. However, requests for medicines to parents represents the critical response measure for children's consumer behavior. This is because parents execute most purchasing for children (especially for medicines) and requests, therefore, should be regarded as the "response" over which the child has control. Analysis of the advertising exposure/requests relationship indicates that it is moderate (.18 fifth-order) and highly significant in a statistical sense (.001 level).

Conclusions.  What have we learned at this early stage of our analysis? Generally, television advertising of proprietary medicines has a rather moderate effect on children's beliefs, affect, intentions and request behavior. Interestingly, the effect increases as we move from beliefs to requests (fifth-order partials of .04 versus .18). There is no obvious impact of television advertising on actual usage of medicines. Apparently, usage behavior is a function of other factors, most particularly illness level and parental mediation.

In order to understand children's conceptions of medicines, television advertising as an explanatory factor must be placed in the context of a multi-mediational model of effects. The examination of such an enriched model will comprise the next phase of our analysis. In particular, we wish to understand the family mediation process and the role of multiple information sources in affecting children's conceptions of medicine.


Richard Adler, et al., The Effects of Television Advertising on Children. A report submitted to the National Science Foundation, April, 1977.

Charles K. Atkin, The Effects of Television Advertising on Children: Survey of Pre-Adolescent's Responses to Television Commercials. A report submitted to the Office of Child Development, H.E.W., July, 1975.

Attorneys General, "Petition of the Attorneys General to Promulgate a Rule Restricting the Advertising of Over-The-Counter Drugs." Before the Federal Communications Commission, Washington, D.C., August, 1975.

Broadcast Advertisers Reports, Inc., "Network TV: Brand Product and Parent Company Schedule Detail and Expenditure Estimates for the Weeks Ending February 20,27, March 6,13, 1977." New York: BAR, 1977.

L. J. Cronbach, "Coefficient Alpha and the Internal Structure of Tests," Psychometrika, Vol. 16, September, 1951, 297-334.

Federal Communications Commission, "News Report No. 14747," December 10, 1976-B.

James Hulbert, "Applying Buyer Behavior Analysis to Social Problems: The Case of Drug Use," in Proceedings of the American Marketing Association. ed. Ronald C. Curhan, 1974, 289-292.

J. Ronald Milavsky, Berton Pekowsky, and Horst Stipp, "TV Drug Advertising and Proprietary and Illicit Drug Use Among Teenage Boys," Public Opinion Quarterly, Volume 39, Winter 1975-76, 457-481.

NAB, "Children's TV Advertising Statement of Principles," Code News, The National Association of Broadcasters, Volume 7, No. 6, June 1974.

John R. Rossiter, "Reliability of a Short Test Measuring Children's Attitudes Toward TV Commercials," Journal of Consumer Research, Volume 3, March, 1977, 179-184.