Some Issues in Designing Consumer Information Studies in Public Policy

R. Bruce Hutton, University of Denver
Dennis L. McNeill, Arizona State University
William L. Wilkie, University of Florida
ABSTRACT - In order for the consumer researcher to make significant inputs for public policy decisions regarding information provision, he must be aware of the nature of the policy environment and the kind of research that will best fit its needs. This paper addresses several issues of importance in doing policy related research and describes two specific studies in the area of energy information undertaken at the University of Florida's Center for Consumer Research.
[ to cite ]:
R. Bruce Hutton, Dennis L. McNeill, and William L. Wilkie (1978) ,"Some Issues in Designing Consumer Information Studies in Public Policy", in NA - Advances in Consumer Research Volume 05, eds. Kent Hunt, Ann Abor, MI : Association for Consumer Research, Pages: 131-137.

Advances in Consumer Research Volume 5, 1978      Pages 131-137


R. Bruce Hutton, University of Denver

Dennis L. McNeill, Arizona State University

William L. Wilkie, University of Florida


In order for the consumer researcher to make significant inputs for public policy decisions regarding information provision, he must be aware of the nature of the policy environment and the kind of research that will best fit its needs. This paper addresses several issues of importance in doing policy related research and describes two specific studies in the area of energy information undertaken at the University of Florida's Center for Consumer Research.


The content and availability of product information for use in evaluation and decision making by consumers is a major topic of interest and concern among various groups -- marketers, policymakers, consumer educators, and consumer researchers. At issue is the availability and subsequent use of information by which consumers can make more "informed" choices (e.g., objective information representing specific physical features). Programs to provide such information have been created in a broad range of government agencies, including the Environmental Protection Agency, Department of Transportation, Food and Drug Administration, and Federal Trade Commission. Current initiative is reflected in the 1975 Energy Policy and Conservation Act (EPCA) whose basic mandate is to "conserve energy supplies through energy conservation programs . . ." (EPCA, 1975). One of the major priorities of this program is the establishment of labeling rules for major durables, specifically in the area of energy consumption. For the consumer researcher to provide relevant input to such decisions, his research paradigm should reflect the special problems in policy oriented research.

Trade Regulation Rule Influence

Day (1976) notes that we are certain to see an increase in the disclosures of efficiency performance data in the future. The primary vehicle for operationalizing future information provisions will likely be through the Trade Regulation Rule (TRR) power of the Federal Trade Commission (FTC). In effect, TRR's allow a public agency to require particular marketing practices to be performed by all marketers in a given product or service category (Wilkie and Gardner, 1974; Wilkie 1975c). This approach by the FTC will require policy decision inputs at a number of different stages. Wilkie (1975) has delineated six stage process of TRR development:

(1) Selection of product classes.

(2) Identification of relevant product characteristics.

(3) Development of standards and test methods.

(4) Determination of reporting format.

(5) Provisions for dissemination.

(6) Assessment of effectiveness.

The different stages represent different levels of importance for the consumer researcher. Stage 1 is crucial because priority concerns and program objectives are formulated here. Stages 2 and 3 will be primarily undertaken by technical experts. Stages 4, 5 and 6 provide the major opportunities for consumer research.

There are both major and complex questions about the need for TRR's, especially in a cost/benefit sense. Expected benefits from TRR's can be classified into two major types: (1) administrative/legal benefits, and (2) economic/social benefits (Wilkie, Hutton, and McNeill, 1975). TRR's are expected to make the regulatory process more efficient by removing the need for case by case adjudication and to provide marketers with better guidelines for what is expected of them. In the second case, TRR programs should bring the system more towards the idea of the "fully informed consumer" and to give him more objective information for product/brand decisions. At the time of this writing, TRR proposals aimed at increasing consumer information were underway in a number of areas, including: nutritional information in food advertising, OTC drug advertising, used automobiles repair information, standardized product warranties, comparative price disclosures for life insurance policies, detergent cleaning ratings, and energy consumption labels for household appliances. Each of these areas involves millions of consumer purchases and potential for considerable improvements. Consumer research is obviously relevant for program decisions.

Purpose of this Paper

The design and conduct of research for public policy needs has been a focus of attention at the University of Florida Center for Consumer Research. A series of experiments have been conducted which deal specifically with the area of consumer product information. In designing these studies, the authors have encountered several important issues of likely interest to both the academic researcher and policymaker. The purpose of this paper is to examine these issues, show how they have influenced two particular studies dealing with consumer impact of "new" product information, and to provide some general suggestions for others' consideration of this research area.


In any research project, decisions will have to be made regarding the focus of the research, design issues, data collection, and analysis. In policy related research, while some issues will be unique to the public policy field, the majority of decisions to be made will be of the same type as any research project. However, the final approach the researcher takes will be different in many cases because he or she is addressing the research to policy needs. The subsequent policy-related approach carries with it a series of issues and problems, the combination of which is not faced in the normal research program. Major areas of concern for such policy related research are presented in Table 1.

The Role of Policy Objectives

In any research setting, it is crucial for the researcher to have the focus of the study clearly in mind. In the case of research applied to policy needs, an understanding of policy objectives is a necessary condition to defining the problem and establishing the reasons for investigation.

Wilkie, et. al., (1975) state that vague or unspecified objectives by the policymaker can easily allow differing research thrusts, designs, and measurements supposedly aimed at the same policy decisions. The risk is that some of these will be inappropriate for the policy or program mandate. Consider, for example, the fundamental problem of determining dependent variables for studying the effect of information on consumers. A natural tendency would be to choose brand choice or at least attitude and intentions as appropriate variables. But these may not be appropriate criteria. For example, if the goal of the recent energy program is to reduce demand, this would not be reflected in brand shifts as much as model shifts. This point is also made by Bettman (1975) who states that the choice of criterion variables is a function of intent of the information provision. If the intent is that the information be used in a specific manner, attitudes and behavior become more important relative to recall, knowledge, and awareness. If the intent is that consumers should be aided in perceiving and processing the appropriate information but no commitment concerning their use of it, recall, knowledge, and awareness variables become crucial.



Once the researcher understands the basic problem, he must define for himself the rationale for investigating the area (e.g., timely, theoretical significance, practical significance, instrument development, clarification of existing perspectives, stream of research). Based on the reason(s), the decision to be made is whether the focus of the research will be problem or theory oriented. Wilkie (1975) addresses this issue in his discussion of basic vs. applied research. The criteria for choosing may be established as:


Once the researcher has decided these fundamental questions, more specific issues arise concerning questions of appropriate stimuli, units of analysis, and criteria.

Should the researcher choose to focus on the programs, the stimulus chosen for study should meet a set of criteria based on policy needs. The product class chosen would be a function of: (1) the state of knowledge of the consuming public and subsequent sample, (2) characteristics of product use, and (3) the presence of within product class variance. That is, the product class should show variability along the dimension to which the program is directed.

The unit of analysis in addressing policy questions is usually the individual as opposed to the product. Some experiments done in the information provision area focus on shifts in choice (e.g., Houston, 1972; Isakson and Mauizi, 1973; Russo,, 1975); therefore, the unit of analysis is necessarily brand, which does not measure information use directly. For most information programs however, policy-oriented research will focus on the individual consumer level at early stages.

In the past, criteria for evaluating the effects of information have tended to be market share or brand switching behavior. The appropriateness of utilizing these measures alone in evaluating information provision can be questioned, especially when one recognizes that the consumer is faced with a new environment and brief exposure to the information. In such a limited exposure, the use of brand choice places a great deal of responsibility on this level of impact. As with other marketing research, it is more advisable to measure the effects of information at different levels of response using multiple dependent measures (Day, 1976; Heeler and Ray, 1972). Multiple dependent measures will most likely require multiple tasks, with emphasis on realistic content and order.

Research Design

The choice of the appropriate experimental design is, of course, dependent on a number of factors. However, in policy related research, stress is placed on external validity.

The key, then, is to make the experiment more valid externally without losing internal validity. Only two experimental designs meet the criteria for internal validity and control at least one source of external invalidity -- Solomon four-group design and posttest-only control group design (Campbell and Stanley, 1963). Because of the cost involved in running the Solomon-four, the posttest only is usually preferred. Both of these designs control for the possible interaction of testing and X (i.e., the experimental treatment). However, two other interactions are also a problem. It is possible that any effects demonstrated hold only for that unique population from which the groups were selected (interaction of selection and X). In order to reduce this problem the researcher should increase the number and types of groups used. The other interaction involves the effects of recent events on subject responses (interaction of history and X). For example, an experiment on energy information done immediately following the presentation of President Carter's energy program might produce a responsiveness to X not to be found on other occasions. Consequently, it is desirable for policy research to lead, not lag, the phenomena under study.

The other major concern for the researcher in this area involves demand characteristics. For the policy researcher two factors are especially relevant. First, the subject is placed in a changed information environment. Consequently, exposure to new concepts may cue him in terms of the focus of the study. Second, the experiment set up to have the subject engage in multiple tasks is open to demand characteristics if one task cues subject response on other tasks in ways not intended by the researcher. Consequently, the flow of tasks and measures is a primary consideration for the researcher. For an excellent summary of demand characteristics in lab experiments and how to deal with them, see Sawyer (1975).

Data Collection

Focus group interviews can play an important part in the development of the policy research format, with primary purposes being to: (1) Explore the current state of consumer knowledge and attitudes toward the stimulus; (2) Obtain a better understanding of the consumer language used in thinking about the stimulus. The results of the focus group interviews will provide important information for the development of scales for subject response. Scale construction should be a function of the population under a study and not the desired analysis procedures. Obviously scales and tasks should be set up so they can be understood by all subjects. Unfortunately, this may not always be possible, especially when dealing with some segments of the population. In order to minimize this problem, scales should reflect as much as possible the language of the consuming population and real world choices.

Other important questions in the research design involve sampling procedures. In order to provide policy guidance, it is obviously desirable to use subjects which reflect a realistic population of consumers for the program at issue. Lab experiments will normally not be feasible with probability samples, but neither are they normally very useful with college sophomores. One guide might be to use real consumers, attempting to include a broad range of demographics within the sample.

One economical (time-wise) approach is to solicit "intact groups" such as social, religious, and educational organizations. (This will, however, often require with-in-session assignments to different conditions, which will impact on the design and operationalization of the study itself.) If this approach is taken, the researcher should vary the nature and demographic character of the groups chosen, use as many groups as possible, and have no single group dominate the sample.

Real consumers, while providing more relevant data, do present some formidable design problems when compared to college students. Special care must be taken to prevent difficulties the subjects might encounter with certain abstract concepts and with a potential reactive response to being placed in a "testing" mode. These problems require careful development of instructions and tasks and attention by those involved in the conduct of the experiment. Some guidelines we have developed based on recent experiences include:

* Provide complete assurance regarding the anonymity of subject responses. This will help reduce the feeling of being tested and graded.

* Researcher should be ready to accept the fact that some segments of the population just may not be able to do all tasks. This may reflect the study's demands rather than the program being studied; it may be necessary to do a separate study for these segments.

* Non-responses are a fact of life. Some subjects will refuse to make a "best estimate" or "guess" because they feel that would be dishonest or a form of cheating.

* It may be beneficial to provide the subjects with familiar surroundings (i.e., go to their church, club, etc.). Since the researcher is now out of his familiar setting, he should be ready to meet a variety of unforeseen occurrences. Examples these authors encountered include a subject's breaking into tears, a blind subject, mothers missing directions because of a 5 year old's nature call, and confrontation with an inebriated by-stander. Dealing with these kinds of factors certainly approaches art more than science.

Data Analysis

The relationship between the issues being discussed is not one of independence. Naturally, many of the factors discussed previously will affect the data analysis stage (e.g., non-responses, scale construction, etc.). However, there are several specific points that should be advanced:

* The type of people and responses they might normally make will influence the analysis procedure. For example, in the construction of appropriate scales, the result may, in the interest of realistic alternatives, be a nonmetric, as opposed to metric scale.

* There is a need to determine in advance the accepted level of significance. Typically is set at the "sacred" .05 or .01 level. In fact, a number of criteria should be used to determine the appropriate level (Labovitz, 1970). In dealing with issues and samples of the paper above, less stringent significance levels may be justified. In any case, it is helpful to report the actual level attained and overtly provide the researcher's judgment as to the meaning of the obtained result.

* In the case of multiple tasks and measures, analysis should be done in light of other responses. For example, a complexity of information measure was taken on a standard seven point bipolar scale. The results showed no difference across groups with all groups judging it "simple." A clear case of easy to understand information? Maybe not. Subsequent tasks involving the use of the information showed the groups acted differently, and suggesting that many subjects, in fact, may not have recognized the complexity of the information and only dealt with it in a superficial manner. Additional items and analyses can be included to address these possibilities.

Issues Specific to Energy Labeling

In light of recent developments (e.g., CIA report on energy resources, President Carter's energy program, development of a Department of Energy), issues of the effects of energy labeling of appliances on consumer response have become particularly relevant.


The Voluntary Labeling Program resulted from a Presidential directive in April 1973. This program was established within the Department of Commerce to deal exclusively with major energy consuming household appliances. Its general purposes were:

(1) Encourage manufacturers (including private brand labelers) to voluntarily provide consumers with information concerning the energy efficiency or energy consumption of major durables.

(2) Encourage consumers to utilize the information when evaluating products by providing, at the point of sale, energy information presented in a uniform manner and readily understandable in order to facilitate product comparisons.

Early success included the fact that 24 manufacturers representing an estimated 95% of room air conditioner sales in the U.S. were participating in the program as of 1975 and generally favorable consumer attitudes were exhibited toward such labels. In 1976 this voluntary program was abolished by the Energy Policy Conservation Act, and was replaced by a mandatory labeling program under the primary jurisdictions of the FEA and FTC.

The Energy Policy and Conservation Act (EPCA) passed in December 1975, reflects a broad range of programs. Included as important thrusts are efforts on product testing, labeling, energy standards, and consumer education. The Act requires that labels disclose, for each model, (1) the estimated annual operating cost and (2) the range of alternative's estimated annual operating costs. In addition the FEA has the power to prescribe energy efficiency standards to any type or class of covered products. However, it is stipulated that resulting energy savings must outweigh any increase in price or maintenance expense, lessening of product performance, or negative effects on competition in order to be implemented. Finally, the FEA will also carry out a program to educate and encourage consumers to use such information in conserving energy.


The purpose of this research was to address and examine certain consumer information processing issues within the Energy Act's formal programs of standardized energy consumption disclosure. The issues surrounding the energy labels are important, in that the form and content of the labels may well effect the impact of the program on consumers. This research focused on two central questions:

(1)Will the presence of energy consumption data on labels attached to major home appliances be likely to assist the consumer in purchase related activities?

(2) Is the reporting format required under the Energy Act the best way to disclose energy consumption?

The first question is important from a program evaluation perspective, and, in addition, will provide a frame of reference for the investigation of the format question. The impact of the disclosure, given the early stages of program development and the timing of this research, should provide insights into the potential impact of the program and whether it appears wise to proceed.

The second area of interest is the format of the disclosure, an area which appears to have been infrequently researched. Particular issues of interest regarding the energy labeling format were:

* Unit of Measurement -- The required disclosure is in dollars. The unit of measurement question is important in light of program goals. For this study the units of measurement under investigation were the dollars required by the EPCA and the kilowatt per hour (KWH) alternative.

* Comparative Disclosure -- The EPCA requires a new dimension of product performance when it requires comparative energy use for similar products. Effect of this additional information was investigated.

* Time Period for energy use computation -- The required time period is in annual terms. The issue is one of magnitude, for the time period used can inflate the model consumption figures. For this study a comparison of monthly and yearly time periods was conducted.

* Degree or Specificity of disclosure -- A crucial aspect of the disclosure of energy use is the ability of the consumer to identify those attributes of the products which account for energy use. The impacts of explicit -- by feature --disclosure of energy use were also assessed in the study.

Energy Use Data/Policy Research Issues

The combinations of each of these issues required too many experimental cells for the limited resources of the study. Rather than eliminate controllable decision variables, it was decided that certain combinations should be dropped, these determined in light of that combination's realistic chances of actually being used by the program. Readers will note that this decision incurs costs in terms of our subsequent ability to disentangle main effects of the format variables mentioned above. On the other hand, it does allow us some assessment of more variables than would otherwise have been possible. Six conditions were run in the study, ranging from control, with no energy use information, to annual dollars with a range, broken down by feature. The design was completely randomized, after-only, with a control condition and five disclosure conditions.

The product class chosen, refrigerator-freezers, is an interesting (and relatively straightforward) one from the program perspective. The product occupies a prominent place in the home and accounts for a significant proportion of average household energy use. The use of the product is relatively constant across households, reducing the variation in actual consumption as a function of household and geography. In addition, the variability in energy utilization is predominantly associated with features, rather than only brand, choice.

In determining the tasks and measures, the program's objectives played an important role. The goals of this particular program are overtly to reduce the demand for energy. One research focus thus should be on consumers' behaviors re: energy. However, as noted earlier, features are more significant than brands here, such that research tasks had to be altered from the normal stress on brand preference or choice. To assist in more detailed decisions, focus groups were conducted, providing inputs for the language and attributes included in stimulus development.

Table 2 charts the conduct of the experiment. The purpose of the experimental tasks was to allow for a complete profile of energy use impact through multiple dependent measures, while the subjects encountered the information in realistic situations. This set of tasks and measures took slightly over one hour; a fuller description is available in McNeill (1977).

Essentially, the tasks reflected three stages of information exposure and use. The initial exposure required the subject to acquire the information (ala Jacoby's "information board") by pulling tabs attached to four individual model displays. (Models had been designed to allow research inferences as to how well the information was being communicated and used.) In this way, the subjects learned about the models and the researcher could follow "where the subject's eyes go" when evaluating the model. The subjects were then asked to evaluate the models based on unaided recall of the information they had acquired. The second level of information exposure was a second visit to the same models, now allowing external memory while pursuing the evaluation tasks. The final level of information effect was a unique operationalization of the choice dimension. This requested the subject to "build" their own personal model of refrigerator-freezer feature by feature, with a number of subsidiary measures taken afterward. This series of experiment tasks provided responses to the information across a wide range of impacts, allowing a reasonably comprehensive analysis of effects. Results and conclusions are detailed in McNeill (1977).


Given the comprehensive design of this study with its multiple measures and tasks, a detailed explanation of findings is beyond the scope of this paper. Overall, the results of this study should prove useful to both policymakers and consumer researchers. The presence of energy consumption data had a significant impact on consumer's information processing responses. After the initial shopping task (involving information acquisition) the product preferences, overall impressions, and energy use judgments of the experimental subjects were in the hypothesized direction. This was also found in evaluative judgments in the presence of the information on the second shopping trip task. Essentially, the nature of the results shows that the higher the energy use the lower the model evaluative judgments. This is supportive of the energy use impact potential in light of both the intent and content of the disclosure.

The second question under study concerned the format of the information and the relationship of format to consumer information processing response. The format variable which produced differential response was the comparative disclosure. Given that the goal of this disclosure was to provide more effective relative judgments of model performance, a set of specifically designed scales were developed. The comparative judgments reflected the comparative disclosures (in the case of relatively high energy use) on the likelihood of finding a better value, the worth of additional shopping, and purchase recommendations. While the other format variables did not produce significantly different responses, there were no format variables which appeared to hinder the consumer response in terms of the goals of the disclosure. The format question, however, is an important one and must be investigated from other perspectives in order to make completely profile this decision for the policymaker.


While present policy consideration is focused on the format questions discussed above, it may be that a more comprehensive information scheme is much more meaningful. This index, termed "life cycle cost," provides a framework for the summary presentation of three product dimensions -- price, energy, and service costs (M.I.T., 1974a). Life cycle cost (LCC) is defined as the discounted sum of all dollars paid for an average product during its useful life (i.e., purchase price + energy cost + service cost). In essence, LCC provides the consumer with an organized and consistent way of processing all product cost information by incorporating the three cost dimensions within a consistent time frame (i.e., average product life) and common units of measurement (i.e., dollars).

Our study was aimed at exploring the impact of LCC, as a new information form, on consumers. Specifically, the study compared the existing consumer information environment -- with its emphasis on purchase price -- to a new environment providing LCC data. In addition, the study stressed multiple tasks and measures aimed at more comprehensive evaluation of information effects, attempting to account for the fact that such a program would not have been encountered before.

M.I.T. Report

Background for, and development of, the LCC concept is described in the monograph "Productivity of Servicing Consumer Durable Products," authored at M.I.T. under an NSF-RANN grant. The report asserts that consumers fail to recognize energy costs in appliances, do not seek out energy and service related information during product evaluation, and instead opt for increasing use convenience at the expense of increasing energy consumption. Since higher prices may be necessary for the introduction of energy conserving design changes, a shift in consumer perceptions of cost from sole reliance on initial purchase price to a long-run total (LCC) cost appears necessary for the market to operate so as to reduce energy consumption. The nature of presumed consumer impacts from LCC information is presented in Wilkie and Hutton (1977).



LCC Policy Research Issues

The combination of a new conceptual framework for cost and a policy orientation provided this study with a number of challenging issues. First, the introduction of a new concept such as LCC provides the consumer with a different environment in which to operate. Multiple tasks were needed to reflect different levels of consumer response expected within the proposed new environment. Index figures for LCC had to be realistic, which required technical advice and a series of product cost calculations on refrigerator-freezers, which were chosen for this study for much the same reasons as noted earlier. Also, major concern was again with factors in the product that relate to cost (i.e., features). In the case of refrigerator-freezers this would include size, frost free, automatic ice maker, power saving switches, etc. Stimuli are identified as models (combinations of features) since they reflect the different cost combinations more so than brand by brand comparisons of the same model types.

The research design was a posttest-only control group design. Besides the LCC condition, an energy/year condition was included to reflect the EPCA mandated program. Both experimental groups represent a changed environment from today. Intact groups were used, so random assignment to conditions was done within groups instead of among the total subject population.

A flow chart of the experiment is given on the bottom of Table 2. Design of the tasks and measures involved some difficult trade-offs among policy concerns, research control, analytical needs, and subject's capabilities and motivations. Some examples: (1) how many models to present? (2) whether, and how, to introduce budget constraints, (3) whether scales should be balanced versus skewed when dealing with arguably positive or negative concepts, (4) how to assess basic capabilities (e.g., computation skills) without overtly testing, (5) how to represent negative costs (savings), (6) how to introduce this new information together with some education on it, while minimizing demand characteristics, and (7) determination of appropriate criteria by which to reach summary conclusions, given the large number of measures. Details of these decisions are available in Hutton (1977).


Once again, given the extensive design of the study with its multiple measures and tasks, a detailed explanation of findings is beyond the scope of this paper. For a complete analysis of results see Hutton (1977). In general, however, the strongest results were seen in tasks reflecting levels of consumer response in a more cognitive as opposed to behavioral sense. The study provides support for the provision of more comprehensive objective information. By letting control group responses represent what consumers know going into a buying situation and LCC responses what is learned, the differences indicate that consumers do learn from the presentation of such information. In fact, control group responses show a significant underestimation of the magnitude of operating costs. There is also evidence that what is learned by the consumer is more than a specific cost figure but rather a different conceptual view of product cost from the one dimensional association with price toward a multidimensional long run concept.

From a policy standpoint, it is important to know the impact of a proposed provision along various dimensions of utilization. Can they utilize the information correctly or does it result in confusion, misinterpretation, or ignoring of the information package? Findings along these lines are equivocal. Consumers do utilize some aspects of the information, even lower educated ones, but not without certain negative consequences. In addition, there is some indication that consumers will evaluate models and features which conserve energy more favorably and energy consuming features less favorably with the availability of energy cost data. In terms of purchase behavior, it appears that subjects are willing to pay for energy savings but not at the expense of certain conveniences and not without actual cost data for price/energy saving trade-offs. Overall, the evidence shows that consumers' existing knowledge in the area of energy and its relation to products and other costs is sadly inadequate, but the provision of objective information of the type explored in this study holds promise as a valuable decision making tool for consumers.


The public policy environment provides a unique opportunity for consumer researchers, if he or She is aware of factors that distinguish policy-oriented research from that done for other purposes. In particular, four "generalizations" may be useful for empirical researchers considering research in this area:

* Policy objectives are crucial in determining the research approach. In the short run, emphasis is necessarily placed on problems solving. But, in the long run, research addressing broader issues of CIP is necessary for a fuller understanding of consumer information use as well as for assessing the effects of information.

* Experiments show particular promise as a vehicle for both research approaches.

* For the policy-oriented researcher, the design is very much a function of the population under study and the problem to be addressed.

* Task environment, measures, and stimulus vehicles are particularly important because of the stress on external validity. Careful attention should be paid to all three in the interest of realism and to counter potential demand characteristics.

Anyone taking these seriously will find ample challenge in terms of research creativity, as well as having the opportunity to contribute something useful to society, our profession, or both.


James R. Bettman, "Issues in Designing Consumer Information Environments," Journal of Consumer Research, 2(December 1975), 169-177.

James R. Bettman and Pradeep Kakkar, "Effects of Information Presentation Format on Consumer Information Acquisition Strategies," Journal of Consumer Research, 3(March 1977), 233-240.

Donald T. Campbell and Julian C. Stanley, Experimental and Quasi-Experimental Designs for Research, (Chicago: Rand McNally, 1963).

George S. Day, "Assessing the Effects of Information Disclosure Requirements," Journal of Marketing, 40 (April 1976), 45-52.

R. M. Heeler and M. L. Ray, "Measures Validation in Marketing,'' Journal of Marketing Research, 9(November 1972), 361-370.

Michael J. Houston, "The Effect of Unit Pricing on Choices of Brand Size in Economic Shopping Behavior," Journal of Marketing, 36(July 1972), 51-54.

R. Bruce Hutton, "The Impact on the Processing of New Information for Durable Goods," unpublished Ph.D. dissertation, University of Florida, 1977.

Hans R. Isakson and Alex R. Maurizi, "The Consumer Economics of Unit Pricing," Journal of Marketing Research, 10(August 1973), 227-235.

Fred N. Kerlinger, Foundations of Behavioral Research, 2nd Ed., (New York: Holt, Rinehart, and Winston, Inc., 1973).

Sanford I. Labovitz, "Criteria for Selecting a Significance Level: A Note on the Sacredness of .05," in D. P. Forcese and S. Richer, eds., Stages of Social Research: Contemporary Perspectives, (New Jersey: Prentice-Hall, Inc., 1970).

B. J. McGuire and E. A. Vadelund, Voluntary Labeling Program for Household Appliances and Equipment to Effect Energy Consumption: Annual Report for Calendar Year 1974, Center for Consumer Technology, Institute for Applied Technology, National Bureau of Standards, Washington, D.C., February 1975.

Dennis L. McNeill, "An Investigation of Issues Surrounding the Disclosure of Energy Use Data on Major Home Appliances," unpublished Ph.D. dissertation, University of Florida, 1977.

J. Edward Russo, Gene Krieser, and Salley Miyashita, "An Effective Display of Unit Price Information," Journal of Marketing, 39(April 1975), 11-19.

Alan G. Sawyer, "Demand Artifacts in Laboratory Experiments in Consumer Research," Journal of Consumer Research, 1(March 1975), 20-30.

Alan G. Sawyer, The Energy Policy and Conservation Act, Public Law 94-163, 1975.

Alan G. Sawyer, The M.I.T. Report: Consumer Appliances, the Real Cost, RANN Document Center, National Science Foundation, Washington, D.C., 1974.

Alan G. Sawyer, The Productivity of Servicing Consumer Durable Products, The Center for Policy Alternatives, Massachusetts Institute of Technology (M.I.T.) with Charles Stark Draper Laboratory, Inc., Report No. CPA-74-4, 1974a.

William L. Wilkie, "Analysis of Effects of Information Load," Journal of Marketing Research, 11(November 1974).

William L. Wilkie, How Consumers Use Product Information. Report prepared for National Science Foundation, Grant No. GI-42037, Washington, D.C.: Government Printing Office, 1975a.

William L. Wilkie, "Consumer Information Processing Research: Product Labeling," Paper Series; Marketing Science Institute, 1975b.

William L. Wilkie, "New Perspectives for Consumer Information Processing Research," Communication Research, 2 (July 1975c), 216-231.

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

William L. Wilkie and R. Bruce Hutton, "Life Cycle Cost: Impact on Consumer Response to New Product Information," in progress, 1977.

William L. Wilkie, R. Bruce Hutton, and Dennis L. McNeill, "Application of Consumer Information Processing Research to Public Policy," Southern Marketing Association Conference Proceedings, November 1975.