Public Policy Issues in Advertising

Citation:

Debra L. Scammon (1989) ,"Public Policy Issues in Advertising", in NA - Advances in Consumer Research Volume 16, eds. Thomas K. Srull, Provo, UT : Association for Consumer Research, Pages: 231-233.

Advances in Consumer Research Volume 16, 1989      Pages 231-233

PUBLIC POLICY ISSUES IN ADVERTISING

Debra L. Scammon, University of Utah

INTRODUCTION

In order to focus on commonalities across the three papers included in this session on public policy issues in advertising my comments will address broad and general policy issues that the papers evoked in my thinking rather than more specific dimensions of the work presented in each paper. My comments pertain to disclosure remedies, how they should be formulated, to whom they should be targeted, and how they should be evaluated. Each paper has something to say about at least one of these topics though I grant that might not have been the intention of the authors.

PRESTON/RICHARDS

The Preston/Richards paper raises the important questions of whether and at what cost miscomprehension of advertising can be reduced. Their experiment includes two straight-forward "corrections" designed to reduce miscomprehension without simultaneously reducing the level of other information conveyed:

1. elimination of allegedly false claim, and

2. affirmative disclosure of information to clarify an allegedly implied claim.

They interpret their results as suggesting that ads can easily be rewritten to reduce miscomprehension without significantly affecting consumers' perceptions about other useful product attributes.

Overall, the study makes a valid point that such refinement is possible. But, in its simplicity their study fails to address some other issues inherent in Craswell's proposition (1985).

First, an aspect of Craswell's argument is that advertising has different impacts on different consumers -- that is, the same ad claims may convey truthful and useful information to some consumers yet convey false beliefs to others. This reflects a distributional problem common to economic analysis. The point is that any ad, and any FTC-required modification of an ad, will have uneven effects across the population. This implies that rather than examining the impact of an ad on a whole population, an attempt should be made to more finely segment the population and examine effects in several (all) segments. In the case of Nalgecin, for example, those consumers wishing to avoid side effects of aspirin may react differently to ad claims than consumers seeking fast pain relief even when presented with the same ad. Preston and Richards' study does nothing to help determine the existence of distributional effects.

Secondly, the FTC's pattern in remedies has been to require an advertiser to "cease and desist" from making an allegedly deceptive claim and/or to affirmatively disclose information designed to clarify an otherwise allegedly ambiguous claim. Preston and Richards rewrite their ads to include both modifications. A stronger evaluation would weigh the impact of each type of change separately and also look for interaction effects. Such an analysis may reveal, for example, that eliminating the allegedly deceptive claim is responsible for most of the reduction in miscomprehension. When dealing with broadcast messages particularly, but all ads in concept, time is money. If the advertiser is required to include an affirmative disclosure, the cost will undoubtedly be passed on to the consumer. This cost is just as important as the cost of eliminating potentially useful information that Preston and Richards focus on. If the desired result could be accomplished simply by eliminating the allegedly deceptive claims this would be a lower-cost alterative for advertisers. The marginal effect of each type of correction is really the critical issue.

A third point relates to the question Preston and Richards asked in their experiment:

"did the ad make you feel that it was trying to convey this claim to you?"

I wonder if this is the relevant question. It seems to me this falls in the same category as several other "semantic" distinctions under FTC rules. For example, for ads to be found to be deceptive the FTC doesn't have to show actual deceptive impact but rather the ad must have a tendency to deceive. It is not critical to determine advertisers' intent to deceive but instead focus is on capacity to deceive consumers. Thus, I wonder whether respondents' perceptions that the ad was designed to intentionally convey certain beliefs is important. It seems that the likelihood, or the fact, that respondents hold certain beliefs after exposure to the ad is the more relevant question.

These points are intended to suggest that a more rigorous analysis could be employed to answer the question that Preston and Richards set out to address. Such an analysis should include:

1. examination of distributional effects through analyses of impacts upon various market segments;

2. evaluation of both the individual, as well as the interaction, effects of removal of claims and addition of clarifications in ads; and

3. assessment of pre- and post-exposure beliefs about claims (rather than belief in advertisers' intention to convey message).

A final comment relates to Preston and Richards' observation that "there may be greater costs involved in achieving, such masterful obscurely than in merely writing plainly." This raises the issue of the advertiser's strategic objectives and opens up a broader cost/benefit question from the advertiser's point of view To determine whether obscurity is worth the pains, one must know the objectives of the ad and bc able to assess its impact on other consumer beliefs about product attributes and company reputation (e.g. the halo affect) as well as on such things as consumer confidence, trust. etc. The point is, an obscure ad may be intentionally obscure not to mask a particular product attribute but rather to enhance consumer generalization to other product benefits. Such things as imagery, visualization, or even hemispheral lateralization may be achieved at the expense of literal clarity. One must assume that if, as Preston and Richards suggest, advertisers "are on notice" regarding the FTC's view of certain claims they must feel that the benefits obtainable from use of such claims are well worth the risks of detection and prosecution.

BOZINOFF, ET AL.

The Bozinoff, et al. paper utilizing data from the Canadian Gallup Poll represents just the sort of segment analysis suggested by recognition of the distributional effects of advertising I mentioned with regard to the Preston/Richards paper. Their study emphasizes that one must identify precisely the segment of the market one wishes to influence and then must design ads that appeal to the motivations of that (those) segment(s). They segment their sample into teens in various stages of adoption/addiction to drugs and/or alcohol then test the effects of ads designed for teens in the trial or pre-adoption addiction stages. Their findings confirm differential impact across segments with non-users generally responding more favorably.

Of course, what this study does not address is the costs/benefits of a preventive campaign targeted to non- or trial-users compared to the costs/benefits of a campaign targeted to heavy users or relapsed addicts (that is, cessation or rehabilitation campaigns). This latter question is a more macro issue and one which would necessitate examination of the successes of counseling and treatment programs targeted to users as well as the impacts of mass media campaigns targeted to non- or trial-users. The preventive campaign tested in the Bozinoff study is an example of a policy fostering prevention rather than cure and rehabilitation. Incidently, it is also an example of a social orientation valuing youth. This orientation may stem from conscious economic analysis (e.g., the potential income stream of a 14 year old compared to a 50 year old) or from cultural mores (e.g., youth signifies vigor, beauty, etc.)

From a methodological point of view, the Bozinoff study suffers from the same criticism that the authors level at previous surveys of drug/alcohol use; that is, one of the key independent measures is a self-report of an illegal behavior. It is also a highly subjective measure (e.g., light, moderate and heavy use are self-defined and perceptions are likely to change with use experience).

There are also several details about the survey method and content that, if reported in the paper, might shed light on the findings:

- was length of use measured? how long had these kids been using drugs/alcohol (adaptation effect);

- was motivation for use measured? were respondents asked why they did drugs or alcohol?;

- were social networks and their importance measured? who were respondents' heros?

More important that these data that are not reported however is the failure to use more insightful analyses. Why stop with a pair-wise comparison of proportions of respondents in two stages of use behavior? Though the authors do not elaborate, I presume the independent measures of recall, evaluation, and effectiveness were scale responses. Thus, regression, discriminant analysis, probit or logit analysis could be used to tease out some of the more interesting implications from the data. Specifically the curvi-linear relationships for recall that suggest that light users are more likely to recall ads than either non- or heavier users of alcohol are ripe for some further explanation. What other factors are at play here?

POPPER/MURRAY

The Popper/Murray study raises a couple of important questions for policy makers. First their study involved evaluating the comparative performance of four alternative disclosure formats. As has been observed in countless other studies, the only significant differences were between disclosures and the control group; there was little difference between the alternative disclosures tested.

The obvious conclusion from this study, and many others of a similar ilk, is that the FTC is not terribly good at designing effective messages. Bill Wilkie recognized this way back in 1970 in the Hawaiian Punch case in which he encouraged the FTC to use an innovative approach to designing the remedy -- what has become known as a performance standard. This approach basically says "we will tell you what to do but not how to do it; you decide how to do it and show us you have done it."

This seems to me to be the solution to the "r.o significant difference" problem. Advertisers are, and very well should be, more effective than the FTC in designing communications.

A second question for policy analysis is what is the appropriate population when evaluating potential impact of a proposed remedy. In this case the population included the target group but an (assumed) highly sensitive and capable subset of this group. Popper and Murray argue that this is wise as it provides a conservative estimate of potential impact. That is, if it doesn't work with this group it can't be expected to work with any group. This is analogous to test marketing in Salt Lake City. If a product is successful there, it will be successful anywhere

In summary these three papers dealing with public policy issues in advertising raise some important issues:

-the need to evaluate the distributional effects of a proposed advertising remedy;

-the need to isolate the marginal contribution from individual elements of a proposed message;

-the need to take a broader perspective and compare not simply one ad to another but also one approach to another, e.g. mass media targeted to non-users vs. group counseling targeted to moderate users; and

-the advisability of imposing a format standard rather than allowing advertisers to demonstrate performance.

Methodologically these studies also leave us thinking about the need for more rigorous analytical methods that might help answer some of the more naughty questions in policy implementation.

REFERENCES

Craswell, Richard (1985), "Interpreting Deceptive Advertising," Boston University Law Review, 65(4), 660-732.

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