Research Issues, Empirical Findings, and Public Policy Implications For Energy Labeling

R. Bruce Hutton, University of Denver
Dennis L. McNeill, University of Denver
ABSTRACT - Canada and the U.S. have recently developed, with inputs from consumer research, energy labeling programs for major home appliances. This paper provides an overview of current energy labeling research, analyzes the public policy implications of the research for existing and future labeling efforts, and discusses issues for future consideration.
[ to cite ]:
R. Bruce Hutton and Dennis L. McNeill (1980) ,"Research Issues, Empirical Findings, and Public Policy Implications For Energy Labeling", in NA - Advances in Consumer Research Volume 07, eds. Jerry C. Olson, Ann Abor, MI : Association for Consumer Research, Pages: 283-289.

Advances in Consumer Research Volume 7, 1980     Pages 283-289

RESEARCH ISSUES, EMPIRICAL FINDINGS, AND PUBLIC POLICY IMPLICATIONS FOR ENERGY LABELING

R. Bruce Hutton, University of Denver

Dennis L. McNeill, University of Denver

ABSTRACT -

Canada and the U.S. have recently developed, with inputs from consumer research, energy labeling programs for major home appliances. This paper provides an overview of current energy labeling research, analyzes the public policy implications of the research for existing and future labeling efforts, and discusses issues for future consideration.

INTRODUCTION

In the next year energy labels will become a permanent part of the product information available on major home appliances in the United States. Meanwhile, Canada already has an energy label for refrigerators and is planning to expand sequentially into other product classes in the future. In both programs consumer research has played a role in the decisions concerned with the label. The goal of this paper is to bring this body of research together in order to summarize the research issues and empirical findings. Further, the discussion will present those issues which have been neglected by the current body of energy labeling research as well as implications for future empirical work. This paper is done in the spirit that the placement of labels on appliances is part of an ongoing decision to provide the best label rather than the culmination of a final decision for the energy label. Consequently, there is a role for future research as well as an evaluation of past research efforts.

THE LABELING PROGRAMS

U.S. Labeling Program

The labeling program in the U.S. began with the Voluntary Labeling Program in the Department of Commerce which was a result of a Presidential directive to label the energy use of major home appliances. This program provided a label for room air conditioners and had proposals for refrigerators and hot water heater labels in the advanced stages. The program's only in-place labels--the EER (energy efficiency ratio) for room air conditioners--were said to result in favorable first year results (McGuire and Vadelund, 1975).

The voluntary program was replaced by the Energy Policy and Conservation Act (EPCA) which made the labels a mandatory part of product information. The responsibilities for the labels rested with two government agencies. The Federal Trade Commission (FTC) was responsible for label design, while the Department of Energy (DOE) was required to develop testing standards and educational materials. This process has proceeded slowly and it is now expected that within six months from Commission approval the labels will be found on appliances.

It is instructive to observe the labels as proposed by the FTC staff in a recent staff report (FTC Staff Report, 1979). The label, even in its simplest form, is complex. Common characteristics for all labels are the name or corporation producing the model, the product type and model, Energyguide (all in caps) heading, a disclosure of energy consumption, a comparative range for similar products, and a section (usually in the form of a matrix) to allow the reader to calculate their personal energy cost. The labels are yellow with bold black type and are approximately 5 x 8 inches in size. The staff proposed two ways to disclose energy consumption data. The first method is in annual dollars of energy use using a standard (average) cost per unit of energy to derive the consumption figures. Because climate control devices (air conditioners, furnaces, etc.) vary in energy use according to geography and climate, the staff proposed an energy efficiency index for these products. The energy efficiency index (the higher the index the better) represents a transformation of the energy used in the DOE specified test. Thus, in the U.S. the labels will have a standard display, but the disclosure of energy use will vary by product.

Canada's ENERGUIDE Program

In an attempt to decrease energy consumption through more energy efficient consumer purchase decisions, the Canadian government has developed an energy consumption labeling program for selected major home appliances. The specific objectives of the labeling program (ENERGUIDE), outlined by the Honorable Warren Allmand, Minister of Consumer and Corporate Affairs, Canada in a May 16, 1978 news release are: (1) to permit consumers to select, among otherwise comparable models, refrigerators that consume the least amount of energy; (2) to enable the consumer to realize a monetary saving by using energy consumption as a purchase criterion; and (3) to encourage manufacturers to continue to improve the energy consumption of their appliances.

The ENERGUIDE label specifies the number of kilowatt hours (kWh) per month consumed by a particular model when tested in accordance with Canadian Standards Association (CSA) standards. The label itself is round, five inches in diameter, with the term ENERGUIDE in bold letters across the middle, and with the information presented in both English and French.

The first label requirement, for refrigerators, was introduced in October 1978. Labels were attached to all refrigerators at the point of manufacture. Labels appear on the inside wall of the door. Other appliances scheduled for labeling between 1979 and 1981 include freezers, ranges, clothes washers, clothes dryers, dishwashers, and air conditioners.

In each of the above programs research has played a role in the label decisions. The following section will describe the relevant research issues and findings.

DO THE LABELS REFLECT THE RESEARCH?

Empirical Results

Table 1 provides a description of each of the empirical studies that have addressed issues within the energy labeling programs.

Essentially, the research efforts have focused on three major components of the energy label decision. Consumer knowledge and attitudes, an examination of label format issues, and an evaluation of the impact of the labels on consumer decision making.

TABLE 1

A SUMMARY OF THE RESEARCH DIRECTED TO ENERGY LABELING ISSUES

Those studies designed to investigate consumer knowledge and attitudes have been exploratory research (Heit, et al., 1974, Contemporary Research Centre 1977, Response Analysis 1977). Results from this research would lead policy makers to a simple label which provides cost information (rather than kWh (Contemporary Research Centre 1977) or an index (Response Analysis 1977). In fact the Heit, et al. (1974) study indicated consumers did not want the comparative range on the label apparently due to the increased complexity that dimension brings to the label format. Additionally, this type of research has found attitudinal support for the provision of labels and the potential for use of the labels by consumers. Since this research is exploratory in nature it is not surprising that the ultimate label design does not reflect these considerations. This, however, would be justified only if the more controlled studies contradicted the consumers' self-reported desires regarding the label.

The experimental studies have comprised the bulk of the empirical findings on energy labels and for the most part the issues dealt with are subtle changes in the information which is likely to result in subtle differences in label effects (e.g. the provision of monthly or yearly data).

Similar to the effort expended in the earlier Voluntary Labeling Program several studies addressed the effect of changes in the mode of presenting energy data. These studies (Chestnut 1976, Response Analysis 1977, McNeill and Wilkie 1979, and Anderson and Claxton 1979) dealt with various ways to present energy consumption data. Chestnut investigated the impact of verbal vs. numeric presentations of energy consumption data on recall. The inadequate subject population, choice of product class, and lack of realistic presentation of the energy data rendered this study only marginally useful for policy decisions. Additionally, the verbal mode may not be able to satisfactorily represent small but significant absolute differences in energy use between products. The study by Response Analysis (1977) tested several subtle changes in format on attitudes and judgments. These results showed no overall effect of the format changes but due to the idiosyncrasies of some products' energy use there appeared to be some format X product interactions. McNeill and Wilkie (1979) tested 5 alternative label formats and found some effects on comparative judgments due to the presence of the comparative range of energy use. Otherwise the format variables provided no differential effect. Anderson and Claxton (1979) tested the energy use presentation in dollars vs. kilowatt hours. None of the differences in sales between the conditions was significant although the performance of the conditions was not uniform. The impact of the dollars format, which was found in the post hoc survey, indicated that this format heightened the importance of energy use in consumer decisions.

From the format studies the results indicate that the manipulations either do not affect the dependent measures in each test dramatically, that the wrong dependent measures were used, that the effects of format changes is an effect over time, or that the format variables do not have dramatic impact on label effects. Unfortunately, all four of these reasons are likely. There would seem to be a need for more finely tuned dependent measures, and longer term studies in order to eliminate the format variables from consideration in label decisions.

The final type of study has been evaluation oriented and tested the labels against some no-label-control condition (Worrall 1976, McNeill and Wilkie 1979, Anderson and Claxton 1979, Hutton and Wilkie 1979). A rather striking result emerges from all of these studies. It appears that the energy label as prescribed will not alone change consumer choice behavior. Most studies show some consistent effects in initial judgments, perceptions, evaluations, learning, etc., but indicate no significant impact on sales (or choice as operationalized). Two studies go beyond the label itself to test whether or not new information could bolster the effect of the label information. Worrall (1976) tested the label with substantial education and persuasion against both a control (no energy data) and label-only conditions. The results indicated that the label alone would not work, but that the label plus education resulted in significantly lower size and energy consumption of the models chosen. Hutton and Wilkie (1979) tested the concept of life cycle cost (LCC) against a control condition and one that provided yearly energy cost. The results indicated some superiority of both the energy use data and LCC over the control with the energy use data alone providing no effect on behavior. However, the LCC condition, which provides energy data over the life of the product, produced significant behavior impact over both the control and yearly energy use conditions.

Implications

A question arises whether these studies can tell the policy maker whether the label will achieve its prescribed goal of reducing the energy consumption of consumer choices. From the data above it appears that either the labels alone will not work, that the label effect on behavior is one that takes time (a greater time than that found in the test situation), or that the labels tested do not contain the right information. It is likely that, in part, all of these alternatives are true to some degree.

Given the complexity of the selling situation, it is unlikely that the label can overcome some of the more salient characteristics of consumer choice. This is a problem since many of the "important" product attributes carry with them additional energy consumption. It is probably an unfair burden to place both motivation/ persuasion, and information responsibilities on the label. From a policy perspective the decisions appear to be provision of the most information rather than the most persuasive information.

The possibility that label effects may be delayed until sometime after label introduction is quite real. While the consumers may desire energy information, a first time limited exposure is not likely to elicit the dramatic effects when time and learning are necessary for consumer response. Even the Anderson and Claxton (1979) study which took place in the field over 6 weeks may not have been long enough to allow the label effects to take hold.

In addition, it is likely that label effects may interact with product class. If this is true the dominance of refrigerator freezers as experimental stimulus may preclude conclusions as to label effectiveness.

Reviewing the issues addressed leaves the conclusion that some of the more important issues in ultimate label effectiveness may have been ignored.

OTHER LABELING RESEARCH ISSUES

In the case of the energy labeling programs in both the U.S. and Canada, it is clear that the intent of the policy maker is to go beyond that of information provision and into the more difficult area of influencing consumer purchase decisions. Within this expanded role, there are a number of issues that existing labeling research either has not addressed or has not dealt with in sufficient detail to make the kind of conclusions necessary for effective policy decision input.

In his report on consumer's use of product information and informational impacts, Wilkie (1975) discusses several "problem zones" which are likely to impact both the design and outcome of an information program. No one study can hope to address all aspects of the consumers' interface with new product information in a changed environment. However, there must be a recognition of the possible consequences of these other variables in terms of results and conclusions of research designed to assist in forming the labeling program and evaluating its effectiveness. Following is a brief discussion of five variables which have received little, if any, empirical attention and whose impacts on an energy labeling program are likely to be significant.

Very little research has focused directly on motivational factors designed to get people to read the label and consequently use it in their decision making framework. Some research has dealt with the graphics of the label design (e.g., size, color, etc.), but not much has been done to determine what will cause people to attend to the information once they have been exposed to the label in a realistic setting.

In looking at the results of existing studies, some light is shed on why motivation may be a real problem in energy labeling. As Wilkie (1978) points out, one thread that tends to run through the literature on motivation is the importance of anticipations or expectations on consumer purchase behavior. That is, the consumer looks ahead to assess whether or not the information will be useful. Usefulness can be broken down into two factors: (1) useful in terms of the information stimulus itself and (2) in terms of the consumers' capabilities to handle the information. In the first instance, several of the studies have shown that energy has not been considered an important factor in appliance purchases either because they don't believe there is an energy crisis, that appliances vary in energy use, that appliances use significant amounts of energy. These kinds of perceptions were addressed in two of the existing studies. Anderson and Claxton (1979) incorporated a salesperson treatment aimed at motivating consumers through the sales force to attend to energy dimensions. A lack of control probably accounts for the ambivalent results here. Worrall's (1976) studies indicated that the effects of a strong educational condition in addition to the labels are positive. Both studies are important steps in the right direction, but further research is needed with more realistic educational treatments, more controlled sales force participation, and longer time periods for assessment.

The second factor deals with consumers' internal capabilities to handle the information. It is likely that the emergence of a new objective information dimension such as energy cost is likely to increase, not decrease uncertainty for the consumer regarding product evaluation. This is especially true in the U.S. where the appliance labels are very imposing with a large amount of information included. For Canada the problem is not so much with the amount of information as with the way the information is presented. Kilowatt hours are an unfamiliar dimension to many consumers and may be hard to trade-off against other attributes such as price. Tests in which subjects are exposed to the labels in experimental short term situations are not likely to shed light on true motivational properties of the labels since subjects' motivation has been artificially heightened (e.g., Response Analysis 1977, McNeill and Wilkie 1979, Hutton and Wilkie 1979, Chestnut 1976). More realistic field studies such as that attempted by Anderson and Claxton need to be done in this area.

A second area of concern is the issue of system capacity. A number of research studies have addressed this issue under the concept of information overload. The main point being that humans are subject to finite limits in terms of their abilities to process information in any given time. The result of overload leads to some type of dysfunctional consequences (e.g., irrational choice) or a complete suspension of processing activity. Very few studies have addressed this issue directly and none has done so in a realistic consumer purchase setting. Response Analysis (1979) included a complex information condition but did not incorporate measures for the overload type of condition. Hutton and Wilkie (1979) used self report measures of helpfulness and complexity of energy information.

These modest attempts only serve to point out the need for more detailed research in this area. Research in realistic settings needs to be done on both abilities of consumers to process the label information alone, and on their abilities to handle the information in conjunction with other product and environmental attributes that may impinge on the situation. Finally, results from this type of research can possibly be used to segment consumers in terms of levels of sophistication in the use of energy information.

A third issue concerns attention to consumers' existing knowledge and predispositions. Only one study, Worrall (1974), has attempted to take this factor into consideration in accounting for labeling impacts. This area of inquiry is basically concerned with three concepts: (1) consumers' conceptual structures used to guide information processing, (2) knowledge about the product class and alternatives, and (3) consumers' predispositions, including attitudes and preferences (Wilkie, 1978). While a great deal of consumer research has been devoted to predispositions regarding energy and energy related products, very little attention has focused on the other two. It is especially appropriate for this type of research to be undertaken in the context of energy labeling research because of the unique features of the product category (i.e., durable goods).

A fourth issue of concern is system invariance which refers to consumers' changing conceptual structures as a result of the availability and use of the energy labels. None of the existing research directly assesses this factor, although it is certainly a long run goal of the programs. An initial test of this type of issue is in the Hutton and Wilkie (1979) study which focuses on changing the consumers' conceptual structure for product cost from the unidimensional cost equal to price to a more multidimensional structure (i.e., cost equal price plus energy and service expenses).

Probably the most serious issue here, and one in which none of the research addresses adequately, is the impact of "learning" on label effects. Although not all of the studies were conducted at single points in time, not enough time was given in any study to truly assess the impact of the label in light of learning. Consequently, all the studies may tend to understate the ultimate impact of the label since there is no explicit accounting for changes in consumer knowledge and predispositions.

A final problem to be faced is that of individual differences. In the existing research little attention has been paid to this important issue. This area raises difficult problems for the policy maker trying to design a label and for researchers attempting to assess effects. For example, who should the policy maker attempt to design the label for? Given that there is no "average" consumer, what segment should the program be intended to reach--the sophisticated consumer most likely to use the information, the less sophisticated but needier population? Should there be more than one label? The researcher is faced with an equally difficult problem in assessment, because of the lack of precise criterion measures. The best example of this problem is "choice quality." Optimal choices rarely exist, and who can say what the best choice for an individual is? In the case of energy, one is tempted to say the best choice is the one that uses less energy. This unfortunately does not take other consumer criteria into account, the interactions between factors such as price and energy, or variables such as family size which affect energy related decisions. Consequently, the risk is that the design of the label to fit the average consumer actually fits no one consumer well. While attention to individual differences introduces complexity into the decision it also possesses the potential for enhancing label effects.

The above issues are difficult ones for both the policy maker and the researcher, and lead to implications for future research efforts. These future efforts would take on the more complex issues in a programmatic framework to provide for a better policy decision.

RECOMMENDATIONS AND CONCLUSIONS

Energy labels are already either a fact of life (Canada) or about to be a fact (U.S.) for consumers involved in purchasing major home appliances. However, rather than negating the need for further research in the area, there are some critical research needs still present. The preceding section on Research Issues deals with some specific gaps in current knowledge. Beyond this, however, there are two specific types of research that need to be engaged in.

The first involves the evaluation of existing labeling programs. Focus here should be on multiple levels of consumer response. Since it is likely that the full impact of any program will only occur over time, total reliance on behavioral shifts may well place too great a burden on the label in its early stages. Concentration should be on early levels of consumer response initially (e.g., knowledge, attitudes, preferences, behavioral intentions). In addition research needs to focus on label impacts in settings other than choice situations. Since durables are purchased relatively infrequently, much of the impact of the labels and consumers' exposure to them will occur while the consumer is not in a purchase behavior mode. Finally careful attention should be given to evaluation methodologies. Evaluating effectiveness in cases where the label is placed in the environment all at once presents severe problems for researchers (e.g., it may preclude the use of a proper control group). New methodologies need to be developed to account for such problems.

A second type of research involves efforts to improve on existing labels. It is not at all clear that the existing labels contain the "best" information for motivating or assisting consumers in making purchase decisions. Questions remain to be answered in format development, differential impacts across consumer groups and the interface of the label with education and promotional programs on the range of conservation behaviors relevant to label effects. Research addressing these issues must pay careful attention to the research environment in order to provide useful information for the policy maker (e.g., samples must, at a minimum, reflect the audience the program is aimed at).

Two final points need to be made concerning energy labeling research. The first is a cautionary note. Given that the objective of current labeling programs goes beyond information provision into consumer influence, at some point the ultimate criteria for success or failure will have to be behavioral shifts. However, there is a real danger in determining the cause of a lack of such shifts. It may, of course, mean that the labeling program simply won't work. If it is determined that this is the case, a strong look must be given in a cost/benefit sense to continuing the program.

The danger, however, is that this conclusion will be reached too quickly without adequate study of alternative reasons for disappointing results. The failure may be due to the fact that the label contains the wrong information, or that it's presented in a complex and confusing manner, or that appropriate support materials (e.g., education, promotion, sales training, etc.) have not been supplied or effectively operationalized, or some combination of all these. It's only when these possibilities have been examined closely and the program has been in force over an adequate length of time can suppositions as to the efficacy of an energy labeling program be made with any degree of confidence.

Finally, a note concerning the interface between the policy maker and the researcher. First, there is certainly a need for the policy maker to take advantage of consumer research findings in label formulation. However, if one looks at the existing labels and the current research findings, one is hard pressed to see a good match here. On the other hand, the public policy researcher has a real, and important, responsibility to provide research that is reliable, valid, and useful for policy decision making. In the case of energy labeling, it appears that both the policy maker and the public policy researcher have considerable room for improvement in order to meet the best interests of the consumer.

REFERENCES

Anderson, C. Dennis and Claxton, John D. (1979), Impact on Consumer Refrigerator Purchases of Energy Consumption Information at Point of Sale, Canada, University of Manitoba.

Brown, Gary L., Hiet, Robert L. and Harron, M. Dean (1975), A Study to Evaluate NRS-Developed Labels for Water Heaters, McClean, Virginia: Human Sciences Research, Inc.

Contemporary Research Centre Limited (1977), "Report of a Qualitative Study on Energy Efficiency Label Designs," prepared for the Department of Consumer and Corporate Affairs, Canada.

Federal Trade Commission (1979), Labeling and Advertising of Consumer Appliances, Final Staff Report and Recommended Rule, Washington, D.C.: FTC.

Hiet, Robert L., Harron, M. Dean, and Worrall, Jay W. (1974), Research Project to Develop Energy Information Labels for Rome Appliances, McClean, Virginia: Human Sciences Research, Inc.

Hutton, R. Bruce and Wilkie, William L. (1979), "Life Cycle Cost: A New Form of Consumer Information," working paper #21, Center for Consumer Research, University of Florida, Gainesville, Florida.

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

McNeill, Dennis L. and Wilkie, William L. (1979), "Public Policy and Consumer Information: Impacts of the New Energy Labels," Journal of Consumer Research, 6, 1-11.

Response Analysis (1977), Communication Effectiveness of Energy Consumption Labels for Major Appliances: A Study Conducted for FTC, Princeton, New Jersey: Response Analysis.

Wilkie, William L. (1975), How Consumers Use Product Information: Assessment of Research in Relation to National Needs, Washington, D.C.: Government Printing Office.

Wilkie, William L. (1978), "Consumer Information Processing Issues for Public Policy," in Wm. Michael Denney and Robert T. Lurid, Eds. Research for Consumer Policy, National Science Foundation, 88-117.

Worrall, Jay W. (1976), A Study of the Effectiveness of Energy Use Labeling as A Device to Increase the Efficiency of New Appliances, McClean, Virginia: Human Sciences Research, Inc.

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