Consumer Research Issues At the Federal Trade Commission

ABSTRACT - This paper discusses several theoretical economic issues that are influential in the establishment of FTC consumer protection policy and identifies opportunities for consumer research

Citation:

John E. Calfee, Gary T. Ford, and Thomas J. Maronick (1983) ,"Consumer Research Issues At the Federal Trade Commission", in NA - Advances in Consumer Research Volume 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Abor, MI : Association for Consumer Research, Pages: 263-267.

Advances in Consumer Research Volume 10, 1983      Pages 263-267

CONSUMER RESEARCH ISSUES AT THE FEDERAL TRADE COMMISSION

John E. Calfee, Federal Trade Commission

Gary T. Ford, University of Maryland

Thomas J. Maronick, Federal Trade Commission

[Bureau of Economics, F.T.C.; College of Business and Management, U. of Maryland; and Bureau of Consumer Protection, F.T.C., respectively. The views presented here are the authors' and not necessarily those of any of the F.T.C. Commissioners or other staff.]

ABSTRACT -

This paper discusses several theoretical economic issues that are influential in the establishment of FTC consumer protection policy and identifies opportunities for consumer research

INTRODUCTION

The Federal Trade Commission's "issues agenda" has changed dramatically in the past few years, as the Commission has become more cognizant of the importance of estimating the cost-benefit implications of its decisions. One result of the increased emphasis on cost-benefit analyses and projections has been the growing recognition that such analysis is based, fundamentally, on economic theory. Concomitantly, there has been growing recognition that much of this economic theory, especially in the consumer protection area, has not been subject to empirical investigation. Rather, its validity has been taken for granted, or the theory is so new that it has not yet been tested. Yet, since these theories are central to the resolution of basic policy questions, empirical evidence that supports or refutes them is required by policymakers at the FTC.

Consumer researchers can assist in the resolution of these policy questions in two ways. First, they can provide methods for examining the economic hypotheses that are at the center of the policy debates. Second, they can combine conceptual frameworks drawn from the behavioral sciences with relevant economic theory, in order to test policy issues in a manner that explicitly recognizes the importance of non-economic behavioral factors.

This paper is intended to identify and discuss several of the economic issues that are receiving increasing prominence at the FTC and that could benefit from the insights and expertise of ACR members. The ultimate purpose of the paper is to stimulate consumer research on economic theory and these important policy issues.

The paper is organized as follows. We begin, in the next section, with a review of past uses of consumer research. Following that is a summary (strictly from our own point of view) of what seem to be the most significant recent alterations in how the Commission and its staff view the major issues in consumer protection, with special emphasis on the increasing importance of economic analysis. The fourth section examines a number of specific issues that the Commission is now focusing on and that are likely to involve consumer research. The final section provides suggestions for the type of consumer research that the FTC would find most useful and identifies several attributes of research that can provide valuable input to public policymaking and policy evaluation at the Federal Trade Commission.

PAST USES OF-CONSUMER RESEARCH AT THE FTC

As recently as 1970, the entire annual consumer research budget of the FTC was less than $1,000. Since then, the consumer research budget has grown substantially to encompass annual contract expenditures ranging from $400.000 to $900.000, in addition to including a staff of four full-time professional consumer researchers.

A major reason for this change is legislative insistence that the FTC (ant other Federal agencies), assess the effects of their own action. The FTC Improvement Act of 1975 (PL 93-637); the FTC Improvement Act of 1980 (PL 96-252); and the Regulatory Flexibility Act of 1981 (PL 96-354), included requirements that the FTC examine the impact of its regulatory actions.

The 1980 legislation in particular required the Commission, when promulgating a trade rule, to issue a regulatory analysis that sets out the need for the rule, the alternatives considered by the Commission, and an analysis of the projected benefits and any economic effects of the final rule.

Quite aside from the impetus provided by legislation, FTC staff began to use consumer research as a way to assess the need for enforcement in the marketplace. Thus, the existence and levels of "false beliefs" became a way to estimate deception, and surveys were used to estimate knowledge levels and to detect improper product usage.

In short, consumer research is now used both to detect the incidence and severity of (potential) market failures, as well as to measure the effects of FTC remedies. Generically, the use of consumer research to assess the need for and/or impact of FTC programs, rules, cases and enforcement activities, is refer ed to within the FTC as "impact evaluation".

Impact Evaluation at the FTC

The basic model for assessing the impact of FTC rules and cases in the marketplace is a "before-after" study design (Cook and Campbell, 1975), with the promulgation of a rule or litigation of a case as the experimental treatment between the before and after measurements. Using this motel, the FTC has conducted several large-scale field studies that measured consumer experiences immediately prior to promulgation of a Trade Rule by the Commission. In addition, as is discussed below, the Commission has sponsored a number of smaller scale laboratory experiments.

Baseline Studies, which refer to the large-scale studies conducted before a rule is promulgated, have used mail or telephone survey techniques to elicit statements about experiences from respondents concerning such areas as: (1) purchase of major appliances (to assess the role of energy information in the purchase decision), (2) purchase of home insulation (to assess consumer knowledge of and use of R-Value information in their purchase decision), (3) arranging funerals (to determine the extent to which consumers sought ant/or received information regarding specific aspects of the funeral, including prices and optional services), and (4) the purchase of a used car (to assess the extent to which consumers shopped before purchase, the information provided by the seller, and the consumers' after-sale repair experience with the car).

Each of the above studies is to be followed by a second similar study to investigate any changes in knowledge, attitudes and behavior. that may have occurred since the promulgation of the rule. Thus, the "before-after" nature of the design. Until these follow-up studies are conducted, however, the baseline study only includes correlational data. The baseline studies are also limited because they employ retrospective self-reports to Rather the data.

Because of these potential design weaknesses, especially in analyzing purchases that the consumer is likely to have difficulty evaluating, consumer responses to a mail or telephone survey are evaluated independently in a second type of study used by the Commission. This approach was used in a study of housing defects, where home owners responded to a telephone survey on problems with their new homes. A subset of the homes of these respondents were then inspected by qualified building inspectors to determine actual effects (including both those reported by respondents and those identified by the inspector but not detect ed by the homeowner).

Similarly, in a study of contact lens wearers, a sample of respondents was asked to come to one of eighteen central examination sites where the fit of their lenses was examined by an optometrist, an opthamalogist, and an optician. The independent findings from the three providers were then compared and related to the source of the lenses, the price paid for them, and the behavior (use and care) of the wearer. The findings are to be used by the Commission in d awing conclusions regarding the relationship between price, quality and marketing practices of the various provider groups in the market.

A third type of study undertaken by the FTC involved time-series analysis, using samples of consumers surveyed periodically over a sixteen-month period. The purpose of this study was to measure the change in attitudes and beliefs regarding the ability of Listerine to cure colts and more throats. The "experimental treatment" being tested was a year-long corrective advertising campaign mandated by the Commission of the makers of Listerine. This study differed from other impact evaluation studies in that it measured attitudes and beliefs, rather than behavior, and because it took into account incremental and absolute change as a result of the corrective ad campaign.

Besides these large (and expensive) field studies, the FTC has also sponsored a number of controlled laboratory experiments, both to investigate the effects of potentially deceptive advertisements and to estimate the likely impact of alternative remedies. Generally, although there is an attempt to built as much external validity into the studies as is possible in a lab study, the primary dependent variables are knowledge levels, attitudes and beliefs, rather than behavior.

For example, in the still pending FTC antacids rulemaking proceeding the Commission sponsored an experiment to test the efficacy of embedding in televised ads warnings about inappropriate use of antacids. Subjects were shown advertisements for Alka-Seltzer, with alternative warning messages embedded in them as treatments. The primary dependent variables were changes in knowledge levels and behavioral intentions (Houston and Rothschild 1978).

In another study, the Commission sponsored an experiment to determine whether varying the words describing the terms of a product warranty would influence respondents' product evaluations and intention to purchase. The study involved print ads, a mall intercept format with adult consumers, and three products. Behavioral change could not be investigated in this experiment since it involved an "after only" design with a single observation.

A review of the above and other impact evaluations leads to three conclusions. First, the large scale baseline studies investigate knowledge levels, attitudes, beliefs and behavior. Since no follow-up studies have as yet been conducted, these surveys can only provide correlational data. Second, many of the laboratory experiments have investigated changes in knowledge, beliefs and attitudes. They have generally not concentrated on testing economic theory (except in an indirect fashion, i.e., markets work better when consumers have full and/or accurate information). Third, neither the laboratory research nor the field studies have generally attempted to investigate the complete cost-benefit implications of potential Commission-mandated remedies. For example, the research has not focused on the cost to consumers of false beliefs.

RECENT EVOLUTION OF THE FTC'S APPROACH TO CONSUMER PROTECTION

In the 1960's, the FTC was criticized for confining its activities to minutia while the vast majority of potentially deceptive or fraudulent consumer transactions remained essentially unregulated. Whether this was true or not, in the 1970's the Commission increased its activities substantially. This brought in its trail criticisms of a different kind: the Commission has been said to bring major cases for trivial reasons, to give inadequate recognition to the costs to business and society at large of its cases and regulations, to use inaccurate estimation of the benefits of its actions, and to provide inadequate substantiation for its major initiatives (Miller 1982).

These criticisms more or less amounted to an appeal to consider economic efficiency as a standard for judging FTC actions. The result (a trend that began well before the 1980 elections, as noted in the discussion of legislative mandates) has been a gradual shift in emphasis toward more economic analysis in consumer protection activities. This trend is evident in public statements by Commissioners and staff, in the wording of Commission complaints and administrative law judge opinions, in the economic and other reports issued by the staff, and perhaps most obviously, in the creation and funding of the Impact Evaluation group whose purpose as described above is mainly to assess the need for and results of major Commission interventions.

This trend has been enhanced and symbolized by the appointment of James Miller as the first economist to act as Chairman of the FTC. Miller's appointments include an attorney with a PhD in economics as director of the Bureau of Competition, a scholar in the field of law and economics as director of the Bureau of Consumer Protection, and several trained economists, some of whom are also attorneys, as advisors to the director of Consumer Protection.

These appointees are encouraging a distinctly economic viewpoint in all Commission actions, in the sense that the overall costs and benefits to consumers are generally regarded as the most important considerations in determining most Commission actions. Thus the Bureau of Consumer Protection is in the process of reviewing a number of potential trade rules, and is doing 80 using explicitly economic criteria that require an empirical base for any significant market intervention. Attorneys in many areas of case-by-case intervention are actively encouraged to use "protocols" or general guidelines which, again, refer explicitly to economic cost-benefit criteria that require empirical support.

A result is greater emphasis on the magnitude of consumer harm that can be prevented by cases and rules, and on the magnitude of costs and benefits associated with remedies. This in turn has brought a generally increased importance for measurement of market outcomes whether the outcomes are diversion of sales from deceptive advertising, increases (or decreases) in the cost of credit available to persons protected by restrictions on creditor actions, or myriad other reasons for, or probable effects of, Commission action.

CURRENT ISSUES IN CONSUMER PROTECTION AT THE FTC

At a conceptual level, one may divide future consumer research-related issues at the FTC into two groups: those that arise from older, almost traditional concerns in FTC cases, and those that come to light as one explores new economic insights into the kind of market transactions that the FTC regulates.

Traditional Consumer Research Issues

Traditional consumer research-related issues abound at the FTC, the only change being their greater prominence in recent years. These issues have always been present, (Wilkie and Gartner 1974), but for one reason or another they have in the past received (in our opinion, at least) inadequate attention from Commission staff and consequently, too little work in the consumer research profession. Consider advertising cases. There is a sense in which Commission cases depend crucially on what might be called (from the standpoint of contractual law) "reliance" -- i.e., the degree to which consumers actually depended upon the claims in the advertisements (presumably to their detriment). So long as the Commission takes reliance for granted (which it legally can, and of ten toes), not much need be known about what consumers actually do when they meet certain advertising claims. But as reliance becomes an important and controvertible issue, the empirical demands escalate. To what extent to consumers discount various kinds of advertising claims? How does "reliance" depend upon the specificity of the claim? The nature of the claim (search qualities versus experience or "credence" qualities)? The reputation of the seller? The reputation of the medium (New Yorker versus...we hesitate to say)? What is the influence of specific product guarantees, or general guarantees such as "Satisfaction Guaranteed" slogans of many retailers?

A quite different area of Commission intervention is product defects. Here the Commission is concerned with the kind of problems that are customarily covered by warranties - so long as the warranties last. What, if anything, is to be tone about "unusual" problems that pop up after the warranty expires? What if the seller knew of the problems, but failed to disclose them? What if the seller knowingly fail d to disclose the problems, but the costs of disclosure would have been greater than the benefits? These considerations raise a host of questions that in theory at least, could be answered by the methods of consumer research. What to consumers really "expect"? How do consumers alter their behavior when they know that they will be reimbursed by the seller for any Problems?

One can take an even larger view of Commission intervention, and inquire into the overall costs and benefits of entire programs. Again, advertising is an example. Some ten years ago, the Commission enhanced its traditional prosecution of fraudulent and deceptive advertising by adding the requirement that firms have in hand a "reasonable basis" for, or "substantiation" of specific product claim. Some commentators feel that this program has been a complete success, greatly improving the quality (ant perhaps quantity) of information available to consumers through advertising. But little is known of the real costs and benefits of this approach. Do consumers really believe that virtually all specific claims are in fact founded upon "competent and reliable" tests, as the Commission has argued? Is the possible cost (in terms of bringing cases in which deception may not in fact have occurred) of prosecuting ad substantiation cases smaller than the benefits that may flow from the Commission's being able to dispense with producing convincing evidence of injurious deception?

Similar questions may be asked in connection with the Commission's Guides (such as the one that applies to the advertising of automobile fuel economy) and Trade Rules (e.g., the Franchising Rule).

Now let us turn to issues that derive primarily from the more purely economic analysis that Chairman Miller and others have urged upon the Commission (ant that to some extent the Commission and its staff have undertaken voluntarily). Of course there is no perfectly clear line of demarcation between what is important here and the issues just described. But we believe that there is a reasonably clear intuitive difference between the two groups .

Issues Raised by "Newer" Economic Analysis

Most FTC actions center on problems in markets with imperfect information- buyers who do not understand the workings of the products they buy, sellers who do not know how their customers will treat their products, and so on. In the past decade or two economists have devoted increasing attention to problems of information.

Product warranties are a case in point, and a good introduction to some of the issues involved. Warranties serve two distinct purposes: as a signal of what level of product quality the buyer may rely upon (this is a purely informational role), and as an insurance contract. Both these roles are of general interest, as are a number of other information-related issues. What follows here is a sampling of the interaction between FTC regulation, economic theories of information, and consumer research.

Information and the Free Rider Problem. Products and information about products are conceptually distinct entities, and theoretically could be sold separately. In a perfectly functioning market consumers could buy at competitive prices the information they need to make intelligent product choices, and then buy their favored products. Of course the market cannot work that way. Information is subject to a notorious set of marketing difficulties. A major problem is that information, once sold, can be passed on to other consumers without anything being paid to its originator. Because of this problem, producers of information cannot recoup full payment for their efforts and thus may have little incentive to produce information that is highly valuable to consumers. The free rider problem pervades all kinds of information markets including, for example, the market for purely intellectual work.

Advertising presents several important special cases of this general problem. A product may have a valuable feature which is little known to consumers, but is present in competing brands. Both advertisers and regulators face a dilemma. If one brand advertises this valuable quality for itself, the advertiser runs the risk of being accused by the FTC of purveying (perhaps implicitly) a false claim of uniqueness. If he modifies the at suitably so as to assure that consumers realize that the quality is actually common, the advertiser has created his own free rider problem and may lose more than he gains from running the at. The regulator's problem derives from the advertiser's. The FTC wishes to prevent deception, but wishes also to increase valuable information for consumers. To the extent that this kind of free rider problem exists, the regulator faces conflicting goals.

A very similar problem arises with what we might call "partial messages", that is, advertisements whose meaning depends for accuracy upon information derived from a source other than the advertisement. Such advertisements can obviously be misleading. But they can also yield great efficiencies. The regulatory dilemma is the same as with implied uniqueness claims; the prevention of deception can also suppress valuable information.

The theme here is that markets have their own way of overcoming information problems. But these market reactions often involve off-setting effects, and intelligent regulation depends upon an assessment of the these effects. Again, we have a need for quantification of consumer beliefs, intentions, etc., in response to various forms of communication

Market Signals. The existence of one product attribute can signal something about other attributes of the same product. To take a seemingly irrelevant example, possession of a college degree tells a potential employer not only that the applicant knows something about his major subject, but also that he is capable of the self-discipline necessary to complete a degree. Indeed, one can easily imagine obtaining a degree mainly to provide this kind of evidence to employers, and only incidentally for the learning experience itself (Spence 1973).

We mention here two kinds of possible market signals whose role may be important in FTC operations. The first, already suggested, is product warranties, which raise a number of interesting signalling questions. What is signaled by the existence of an optional extended warranty or service contract (which consumers may not purchase but still may pay attention to in their purchase decisions)? Do consumers look at the terms of an extended warranty and infer anything concrete about the product itself? Do they, for example, consider the price of an extended contract for automobiles as an indication of expected repair costs? Do they compare warranties that come with competing products as a means for comparing the products themselves? (If they do, that is an argument against a policy requiring that all warranties be more or less identical.) Do consumers think that stores that advertise "satisfaction guaranteed" are more likely to provide satisfaction?

Casual evidence suggests that in at least some situations, these things may be important to consumers. Pennzoil has recently advertised vigorously that their warranty (against oil-related engine failures) is more complete than their competitors' ("No one else does", the announcer intones in a manner to suggest that the implications are inexpressibly significant). Pennzoil clearly undertook this campaign with the idea that consumers will believe the implicit signal associated with more complete warranty coverage. But little is known with certainty in this area. More consumer research could be interesting.

A second and perhaps less obvious area is advertising claims. A seller who makes a "puffy" claim is providing an indirect as well as (perhaps) a direct message. He is telling potential customers that he wishes to be compared on the attribute in question, and that he has invited other firms to challenge his claim if their products are superior. Thus advertising claims can act as signals. Again, do consumers use them? and if they do, how does this relate to the assessment of consumer injury and benefits from various kinds of advertising? This potentially rich area for investigation has barely been touched, and it requires the special skills of consumer researchers

Lemons Markets. In 1970, George Akerlof published an article in which he described the "market for lemons". The idea arises from markets with the following characteristics: product quality varies, buyers know only the average quality of what is for sale and have no way on knowing the quality of individual items, and individual sellers cannot disclose true product quality in a way that buyers will believe. The result is a little like the adverse selection problem (discussed next), but sellers rather than buyers drop out of the market. Sellers of above-average quality products -used cars are the classic if misplaced example -- cannot get more than the average price, and so pull their product off the market. Average quality drops, so does the price, another group of sellers drop out, and the process continues until only the lowest quality (the "lemons") are left. Everybody loses and everybody wishes there were a way for potential buyers to know the true quality of what is for sale.

Little is known about these kinds of markets beyond the fact that opportunities for their existence are numerous. One may argue that advertising presents a moderately "sour" lemons market, in the sense that consumers may systematically discount many advertising claims, thus raising difficulties for advertisers whose claims really are literally true. The same may be true for warranties: people expect middling service and they get it, while the seller who wishes to provide exceptional service (for a price) has no way to convey credibly his intention. Consumer research could shed light on, for example, the preconditions for such markets to exist. When do consumers systematically discount claims? How and when can advertisers overcome these problems?

Moral Hazard and Adverse Selection. Economists have noted for some years certain problems inherent in any attempt to sell insurance. People who buy insurance contracts may have an incentive to behave differently and perhaps less efficiently than they would have behaved without the insurance. Owners of medical insurance, for example, may stay in hospitals longer than necessary, lose contact lenses regularly, or find psychoanalysis downright appealing. The term "moral hazard" has been coined to refer to this phenomenon.

Closely related is the problem of "adverse selection". This refer-s not to a change in the behavior of individuals but to a change in the makeup of groups that will be attracted to an insurance contract. Again, medical insurance is a good example. A policy may be offered at an actuarily fair rate for the population at large. But individuals differ in their expectations of how much medical care they will need. The healthiest may find the policy unappealing and drop out. The average required medical care for those remaining in the plan will rise, and the premiums increase accordingly. Then a new group will have an incentive to drop out, and the process may continue until either the plan administrators find a way to vary the premiums more precisely in accordance with expected health, or the rules for dropping out are made more restrictive ("group" coverage being an example), or only the sickest are insured (at high premiums).

It is well established that the problems of moral hazard and adverse selection can destroy apparently sound insurance plans. The same principles apply to all kinds of insurance including some that are not called insurance at all, product warranties being a prominent example. The difficulties become even more pronounced when the very existence of a warranty is obscure. This can happen when the FTC extends seller liability for problems, as the Commission has often done in orders arising from cases.

These problems raise questions for which consumer research methods are particularly well-suited. Often we need information on how consumers would behave if (for example) warranty terms were generally different from what they are nov. Or, what would or could consumers do to protect themselves from economic harm if certain protective measures are eliminated? There is 8 consistent need for non-market data designed to answers questions about market situations that have not yet arisen, or that do exist but not in a manner that can yield useable data.

CONSUMER RESEARCH OPPORTUNITIES

The set of issues discussed above were included for two reasons. First, a better understanding of each is required by the staff and Commissioners at the FTC. Additionally, many of these issues are likely to be unfamiliar to consumer researchers. Second, well-designed consumer research can be of significant assistance in determining the validity of the assumptions that underlie the economic and behavioral theories associated with ad substantiation, moral hazard, free riders, and other inadequately researched policies and hypotheses. The following paragraphs provide suggestions on the nature and types of consumer research that would assist in the resolution of these policy issues.

The extent to which consumers rely on ant/or discount advertising messages, and the sources and nature of the expectations they develop about product performance is at the heart of much of the FTC's advertising regulatory activity. Market signalling and lemons markets hypotheses, as well as the advertising substantiation doctrine, are based on assumptions about reliance and expectations. Well-designed consumer research, i.e., studies exhibiting high degrees of both construct and internal validity, can provide important information about these issues.

Moral hazard and adverse selection are essentially market segmentation problems. Consumer research can help FTC staff learn more about these issues in two ways. First, research can begin to provide insight into the levels of attributes at which consumers become willing to engage in behavior that is harmful to both them and society. The classic, if extreme, example concerns the dollar amount at which individuals will commit suicide for the sake of life insurance payments. Thus, consumer research can provide information about the tradeoffs that can be expected with different levels of insurance, guarantees and warranties. Second, consumer research can provide empirical evidence about the size or incidence of the adverse selection segments. That is, what percentage of contact lens insurance purchasers will act carelessly because of the insurance? How does the size of these adverse selection segments vary with different levels of coverage?

The examples of consumer research opportunities provided above can be investigated with the traditional social science research methodologies of experiments and surveys. To increase the probability that the resulting findings are valuable, ideally the research should allow the Commission to evaluate the assumptions or theory underlying the test, and should include measures of behavior and measures of the cost/benefit relationships uncovered.

For example, suppose an experiment is conducted to examine the nature and degree of "discounting" that consumers engage in when faced with an ad for a new product with qualities that cannot be observed before purchase. The measures in such an experiment should allow determination of the amount of discounting, the appropriateness of the discounting, beliefs engendered, attribute importance, and behavior. Included in the behavioral measures should be estimates of both the Type 1 and Type 2 errors from discounting, including estimates of the opportunity cost associated wit 1 bans on certain advertising claims. The FTC should ideally be able to decide whether intervention is warranted or whether consumers can adequately interpret the advertising messages. Of course, this ideal study may be well beyond the normal resources of most academicians. However, the example should serve to indicate that the costs and benefits from regulatory action or inaction are paramount.

In this paper we have attempted to provide consumer research professionals with our assessment of several of the most important consumer protection policy issues facing the FTC. The economic theory issues selected require much more empirical evidence and can all be investigated with consumer research procedures. The issues are important and unresolved; we invite consumer researchers to Participate in the debate.

REFERENCES

Akerlof, George A, (1970), "The Market for 'Lemons': Quality Uncertainty and the Market Mechanism," Quarterly Journal of Economics. 29 (August) 488-500.

Cook, Thomas D., and Donald T. Campbell (197-5), "The Design and Conduct of Quasi-Experiments and True Experiments in Field Settings," in Handbook of Industrial and Organizational Research. M. D. Dunnette, ed. Chicago: Rand-McNally, 223-326.

Houston and Rothschild (1978), "An Experimental Investigation of the Efficacy of Warning Messages in Antacid Television Commercials," report prepared for the Federal Trade Commission.

Miller, James C. (1982), "Prepared Statement before the Committee on Commerce, Science and Transportation, Consumer Subcommittee, United States Senate, July 22.

Spence, Michael (1973), "Job Market Signalling", Quarterly Journal of Economics. 87, 355-374.

Wilkie, William L., and Davit M. Gardner (1974), "The Role of Marketing Research in Public Policy Decision Making", Journal of Marketing. 38 (January) 38-47.

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