Final Report: Decision Making and Valuation for Environmental Policy

EPA Grant Number: R830988
Title: Decision Making and Valuation for Environmental Policy
Investigators: Teisl, Mario F. , Rubin, Jonathan
Institution: University of Maine
EPA Project Officer: Hahn, Intaek
Project Period: September 1, 2003 through August 31, 2006 (Extended to August 31, 2007)
Project Amount: $399,979
RFA: Market Mechanisms and Incentives for Environmental Management (2002) RFA Text |  Recipients Lists
Research Category: Economics and Decision Sciences


To implement and study the effects of an eco-information (labeling and marketing) program for vehicles sold in Maine (under the auspices of a voluntary agreement between the State’s Department of Environmental Protection, the Maine Automobile Dealers Association and the Natural Resources Council of Maine).

Summary/Accomplishments (Outputs/Outcomes):

Six focus groups were held in Maine during the fall of 2003; the groups were used to develop and evaluate messages used as part of the information program, and to identify consumer perceptions of vehicle emissions in Maine. The latter information informed the marketing strategy while also identifying consumer misperceptions that needed to be quantitatively documented with the baseline mail survey.

A pre-marketing (baseline) mail survey was designed and administered during the summer of 2004 to a random sample of 1,148 Maine adults who had registered vehicles in Maine; the response rate was 60 percent and respondents were representative of the state. We used the survey data for several purposes. First, the data provided the baseline level of perceptions, attitudes and behaviors that helped guide the design of the eco-information program. Second, the baseline survey responses were compared to responses from the post-marketing survey to establish whether the information program was successful. Third, the data allowed us to empirically model whether, and how, eco-information impacts vehicle choice. Because the focus group research indicated that vehicle buying is a two-stage process (consumers first choose a class of vehicle; e.g., car, SUV, truck, then choose a vehicle from the class) we used a two-stage, nested vehicle choice question to measure consumers’ willingness to alter their purchases based on changes in a vehicle’s emissions attributes (criteria pollutants and global warming gases). This allowed us to determine whether consumers’ buying decisions are impacted at the class or vehicle level. Finally, the data allowed us to empirically model how eco-information is processed and the factors impacting this information processing.

Pre-marketing (baseline) vehicle sales data, as proxied by new vehicle registrations was collected (registration data covers all vehicles registered in Maine during the 12-month period from July 1, 2003 to May 31, 2004). This baseline data will be compared to the post-marketing sales data to establish whether the eco-information program was successful in altering behavior.

Eco-information program - There were two main parts to the eco-information program. One part focused on providing information to vehicle shoppers at the point-of-purchase (new vehicle dealerships). Although providing some educational benefit, the primary purpose of the eco-labeling was to provide information to improve consumers’ ability to make cross-product comparisons. This dealer-based information consisted of brochures explaining the program, and placement of eco-labels in the window of qualifying vehicles. Participation of individual dealers was voluntary so one aspect of the research was to measure the level of participation among dealers.

The second part of the eco-information program focused on educating Mainers about Maine’s air quality, its link to motor vehicles and to heighten awareness of the program. The first component was the program’s website that provided detailed information about vehicles and their contribution to air quality problems. In addition, the program included newspaper and radio advertisements that provided eco-messages, information about the program and the program’s website. The eco-marketing portion of the program started on January 31st and ran continuously until June 13th in several Maine newspapers and radio stations. The marketing was only performed in the southwestern (treatment) region of the state; the remainder of the state served as a control.

A post-marketing mail survey was administered during the summer of 2005 to a random sample of 1,163 Maine adults who had registered vehicles in Maine; the response rate was 64 percent. Again, respondents were similar to the characteristics of the Maine adult population. We analyzed changes in survey responses to the two mail surveys across the samples of individuals residing in and out of the marketing treatment area. This allowed us to identify the impacts of the eco-information program on consumer knowledge, awareness, perceptions and preferences.

Observation of website activity - We examined the level of activity on the program’s website before, during and after the marketing program to determine whether the program had any impact.

Post marketing dealer observation - To determine the level of dealer participation, we had several student employees visit new vehicle dealers. Visits to dealers were done toward the end of the marketing treatment. Of the 134 eligible dealers, 105 were visited for a 78 percent visitation rate. During the visit the students indicated they were interested in purchasing a vehicle and were interested in an environmentally friendly vehicle. Students would mention the Maine Clean Car Campaign by name, and indicate they had learned of the program via radio or newspaper ads. Students recorded whether the dealership displayed the eco-labels on qualifying vehicles, and whether the marketing brochure was available (displayed in the showroom or provided by a salesperson when requested). Students also recorded whether the salesperson knew about the program’s website. Finally, the students recorded qualitative information about the apparent level of knowledge exhibited by the salesperson.

Post-marketing vehicle sales data was collected from May 31 2004 to May 31, 2008. In total we will have about 43 months of vehicle registration data (19 months before the marketing effort, five months during the marketing effort and 19 months after the marketing effort). The quasi-experimental design of the marketing will allow us to use the market data to isolate any changes caused by the program.


Focus groups - Mainers placed a low level of importance on emissions when purchasing a vehicle because participants felt air pollution was not a problem in Maine, that vehicle emissions were not a major contributor and that all vehicles ‘pollute about the same’. This indicated that the program needed to educate consumers about the magnitude of the air pollution problem, help them understand that air emissions can vary greatly across vehicles, and that vehicles contribute a significant amount of air emissions.

Many participants get their vehicle information from on-line sources, vehicle-related publications and friends/relatives; dealerships were often not visited until late in the car-buying process and were used only to gather monetary and ‘experiential’ information. Hence, a limitation of a localized eco-marketing effort is that some important information channels (e.g., nationally distributed vehicle publications) are largely unavailable. Many participants were in favor of an eco-labeling program for new vehicles sold in Maine; however, many questioned the likelihood of success because they assumed that relatively few vehicles would obtain an eco-friendly rating. When told many vehicles across a variety of vehicle classes would meet Maine’s standard, many participants thought this would bode well for the program’s success.

Participants preferred labels that were the clearest and simplest at conveying the eco-message. Although logos alone were seen as inadequate, most participants preferred graphical approaches to facilitate vehicle choice. Many participants felt detailed information should only be provided off-label (on a website or in brochures). Most participants liked the idea of including a reference value in addition to the value presented for the specific vehicle. However they preferred the reference to be based upon vehicles within the same class of vehicle rather than being based upon all vehicles. This was because participants felt most people shop for a type of vehicle and would like to know how the vehicle they are considering buying rates relative to other close substitutes.

Baseline survey results - Analysis of the survey data supported the focus group findings that Mainers underestimate the importance of vehicles’ contribution to air pollution and think that all vehicles make the same amount of pollution. Many respondents perceived lower emission vehicles to have lower performance and to be more expensive. On a positive note, most respondents find emissions information important and indicate they wanted environmental information readily available. However, even if the eco-labeling program is extremely successful in altering perceptions and behavior, the full impact of the program would take several years to measure because most respondents hold on to their vehicles for several years and only about half of the respondents buy new vehicles (as opposed to pre-owned).

We find an individual’s use of a vehicle is the most important determinant of class choice; other class-specific attributes (including emissions information) had no significant impact. However, vehicle choice was positively impacted by global warming information but not by criteria pollution information. This suggests that consumers do value the benefits of more envi­ronmentally benign vehicles (at least with respect to global warming gases) and that consumer purchases may support an eco-labeled market.

With respect to how people process eco-information, we find that both how the information is presented and the individuals’ characteristics (e.g., psychological and perceptual factors) are important in influencing stated behavior. The results from the modeling also provided us with the prime messages for the eco-marketing campaign. For example, one message has to increase people’s perception that what they buy matters (i.e., that vehicles are significantly different in their environmental characteristics).

Website activity - There was a relatively strong increase in website activity once the newspaper and radio advertisements began. Interestingly, the level of website activity was maintained for about six months then it began to increase – this was occurring months after cessation of radio and newspaper advertising. Apparently, the website continued to attract attention and it is currently unclear why this occurred. It could be that ‘word-of-mouth’ advertising (either by previous website users or by sales personnel at dealerships) began generating its own stream of new visitors. However, we have noticed that a percentage of our visitors are from outside the state (in fact, some of our visitors are from outside the US). This may indicate that news of the website may be spreading from among eco-conscious web-users, or may be simply an artifact of how websites are ranked during keyword searches.

Results from the dealer observations - Few dealerships actively participated (displayed the eco-stickers on qualifying vehicles or had brochures available). Sales personnel knowledge of the eco-information program was also low and few salespeople seemed willing to assist the customer. This raises the policy recommendation that eco-communication programs should give more consideration on how to involve the retail sales force (or alternatively, make such information mandatory).

Comparing responses across the pre and post marketing surveys, we find that 15 of 16 different awareness, behavioral, and attitudinal variables move in the desired direction (significantly so for eight variables) for individuals exposed to the eco-marketing treatment. These individuals were significantly more likely to recognize the eco-label, a minimum requirement of the eco-marketing program. Additionally, we find these individuals had a more pessimistic view of Maine’s current level of air quality and were more likely to view current air pollution control laws as weak. Movements in these variables should increase the effectiveness of the eco-information program. We also find that individuals exposed to the eco-marketing placed a greater faith on the state’s abilities to protect the environment and in other people’s willingness to pay for environmental protection.

That the other “correct” effects of the marketing program are insignificant does not necessarily indicate a general ineffectiveness of the program. Analyzing responses to any information program is similar to analyzing a dose-response function. The potential magnitude of the effect is related to the size of the ‘dose’. On several fronts, our marketing effort was of a relatively low dose: the entire program cost less than $125,000, did not use other available media (e.g., television) and ran for only about four and a half months. In addition, the low level of participation among dealers and the low level of awareness and knowledge among sales personnel likely limited the overall impact of program. The fact that we find correct impacts in all but one variable suggests a stronger marketing effort would be associated with more significant, positive responses.

Journal Articles on this Report : 2 Displayed | Download in RIS Format

Other project views: All 22 publications 4 publications in selected types All 2 journal articles
Type Citation Project Document Sources
Journal Article Noblet CL, Teisl MF, Rubin J. Factors affecting consumer assessment of eco-labeled vehicles. Transportation Research Part D: Transport and Environment 2006;11(6):422-431. R830988 (Final)
  • Abstract: Science Direct Abstract
  • Other: Science Direct PDF
  • Journal Article Teisl MF, Rubin J, Noblet CL. Non-dirty dancing? Interactions between eco-labels and consumers. Journal of Economic Psychology 2008;29(2):140-159. R830988 (Final)
  • Abstract: Science Direct Abstract
  • Other: LEF PDF
  • Supplemental Keywords:

    air, global climate, mobile sources, VOC, nitrogen oxides, public policy, conjoint analysis, psychological, social science, modeling, transportation, preferences,, RFA, Economic, Social, & Behavioral Science Research Program, Scientific Discipline, POLLUTION PREVENTION, Energy, Economics and Business, decision-making, Market mechanisms, Social Science, Economics & Decision Making, eco-efficiency, market incentives, consumer perception, decision making, environmental decision making, green vehicles, green products, market valuation models, socioeconomics, environmental policy, eco-labeling, energy efficiency, consumer behavior, behavior change, environmental marketing, cost effectiveness

    Relevant Websites: Exit

    Progress and Final Reports:

    Original Abstract
  • 2004 Progress Report
  • 2005
  • 2006