Comparison Processes in Energy Conservation Feedback Effects

ABSTRACT - Theoretical perspectives based upon goal-setting and comparative processes are elucidated and proposed as mediators of the feedback effect found in energy consumption studies. A field experiment designed to test these perspectives revealed that, for natural gas consumption, both goal-setting and activation of comparison processes are necessary conditions in obtaining the feedback effect.


Thomas D. Jensen (1986) ,"Comparison Processes in Energy Conservation Feedback Effects", in NA - Advances in Consumer Research Volume 13, eds. Richard J. Lutz, Provo, UT : Association for Consumer Research, Pages: 486-491.

Advances in Consumer Research Volume 13, 1986      Pages 486-491


Thomas D. Jensen, University of Arkansas


Theoretical perspectives based upon goal-setting and comparative processes are elucidated and proposed as mediators of the feedback effect found in energy consumption studies. A field experiment designed to test these perspectives revealed that, for natural gas consumption, both goal-setting and activation of comparison processes are necessary conditions in obtaining the feedback effect.


Since the onset of the 1973-1974 oil crisis people have been awakened to their reliance upon energy for maintaining or altering their life-styles. Researchers in both the natural and social sciences quickly responded to the crisis via technological and behavioral solutions. Businesses also responded to the challenge through the development of more energy efficient appliances and modes of transportation. Consequently, the vast majority of the early research investigating energy utilization was descriptive in nature and followed by a vast wave of application-oriented research primarily designed to promote energy conservation.

Currently, over a decade later, with consumers having adapted to the relatively higher energy prices and with some energy sources dropping in costs (e.g., oil) the magnitude of published research examining energy consumption has dwindled. However, the current situation offers an opportunity to begin systematic efforts to examine the mediators as well as moderators of various strategies that have been shown to influence energy consumption. Only with a thorough understanding of the parameters and causes of consumption can policies be developed and strategies implemented for dealing with any future shortages.

A variety of intervention strategies, variable manipulations, and research classification schemes have been utilized in examining energy consumption (for reviews see: Becker, Seligman, & Darley 1979; Cone & Hayes 1980; Cook & Berrenberg 1981; Gottlieb & Matre 1976; McClelland & Canter 1981; McDougall, Claxton, Ritchie, & Anderson 1981; Olsen & Goodnight 1977; Shippee 1980 & 1981; Stern & Gardner 1981; Winett & Nietzel 1975; Winett & Kagel 1984: Winkler & Winett 1982). One strategy acknowledged by virtually all of the reviews as being effective in reducing energy consumption is termed the feedback effect.


Basically, the feedback effect entails providing consumers with frequent information (e.g., daily or weekly) concerning their energy consumption. Consumers provided with this consumption feedback have subsequently reduced consumption 10% to 20% when compared to consumers not receiving the feedback (Becker 1978; Becker, Seligman, & Darley 1979; Bittle, Valesano, & Thaler 1979 6 19791980; Hayes & Cone 1977 & 1981; Palmer, Lloyd, & Lloyd 1977; Seaver & Patterson 1976; Seligman, Becker, & Darley 1981; Winett, Kagel, Battalio, & Winkler 1978; Winett, Neale, & Grier 1979) and the effect tends to be maintained even when the frequent feedback is no longer provided and regular monthly utility billing represents the only consumption information. The generalizability of the feedback effect is acknowledged when one notes that it has "been shown to be effective in studies conducted in at least 10 states, with diverse samples of people, during all four seasons, and over a period of a decade" (Winett & Kagel 1986, p. 666).

A variety of parameters have been identified which moderate the feedback effect and are useful in providing insights into the possible mediators that are active. First, studies have shown that cost feedback as opposed to or in conjunction with usage feedback (e.g., kilowatts of electricity) is more effective in reducing consumption (e.g., Palmer et al. 1977). Furthermore, the superiority of cost feedback tends to be more pronounced with cumulative as opposed to discreet feedback for a limited time period (Bittle et al. 1979-80). This superiority of cost feedback is consistent with the long-term effect of pricing on demand from an economic perspective (i.e., Winkler & Winett 1982). Second, externally-generated feedback (e.g., supplied by the researcher or utility company) tends to result in less usage than self-generated feedback (e.g., generated by the consumer) although some reduction is found when consumers self-monitor their consumption (e.g., Winett et al. 1979; Pallak & Cummings 1976). Third, individual feedback is more effective at reducing consumption than group feedback such as for a neighborhood or region (e.g., Winett et al. 1978-79). Fourth, as suggested by some authors (e.g., Winett & Kagel 1984), increasing the frequency of providing the feedback may result in increased energy savings. For example, daily feedback may be more effective than weekly feedback which in turn may be more effective than monthly feedback as currently supplied by most utility companies. Unfortunately, no studies have been conducted which have systematically examined the effects of the frequency of feedback on consumption; most studies have utilized daily feedback and the suggestion for increasing the frequency of the feedback as been via extrapolations from other studies using weekly feedback. Also, as noted by Shippee (1980) and Stern and Gardner (1981) the costs of providing daily feedback may make it impractical. Furthermore, other studies have shown that even weekly feedback can result in a 24% reduction in consumption (i.e., Bittle et al. 1978) or, in the case of another study, simplifying the format and content of monthly utility bills resulted in a 10% reduction in consumption relative to normal billing format and content (i.e., Hayes & Cone 1981); reduction figures similar to those found for daily feedback. These studies suggest some caution in attempting to generalize about the effects of the frequency of feedback. Fifth, some studies have shown that a public commitment to saving energy as opposed to a private commitment results in a greater reduction in consumption levels (i.e., Pallak & Cummings 1976; Pallak, Cook, & Sullivan 1980; Seaver & Patterson 1976). Finally, other studies have demonstrated that providing consumers with a difficult goal for reducing their energy consumption results in lowered consumption rates when compared to rates for consumers provided with a relatively easy goal (e.g., Becker 1978).

An examination of the above parameters suggests that for feedback to be effective at lowering energy consumption consumers must have some standard against which to evaluate their performance (Becker 1978; Becker et al. 1979; Seligman et al. 1981; Cone & Hayes 1980). Consumers must be able to "evaluate how well or poorly they are doing with respect to their desired level of conservation" (Becker et al. 1979, p. 52). This interpretation entails three necessary conditions. First, consumers must have some standard or goal in evaluating their consumption. This standard or goal may entail previous usage levels, experimenter suggested reduction goals, or budget maintenance goals. Second, consumers must engage in the comparison process in comparing present consumption levels or behaviors with the standards or goals. Third, the information utilized in setting goals or standards and comparing with present consumption levels or behaviors must be understandable and germane. This interpretation suggests that an understanding of the feedback effect and, hence, identification of the mediators, is based upon when comparisons will be invoked, what is being compared, and the prediction of behaviors following the comparison process. Festinger's (1954) social comparison theory provides an "established core of confirmed theory" (Singer 1981, p. 171) for testing the veracity of the comparison process in explaining the feedback effect.

Social comparison theory postulates that individuals have a drive to evaluate their opinions and abilities. In the absence of objective physical standards upon which to make these evaluations, individuals will utilize the opinions and abilities of others in assessing their opinions and abilities. Festinger (1954) also predicted that individuals would prefer to make these evaluations, along relevant dimensions, with similar as opposed to dissimilar others. Given that an individual's opinions and/or abilities are different from others', the individuals can either (a) attempt to persuade others to adopt his or her opinion or ability, (b) modify his or her opinion or ability to match the others, or (c) reject the others' abilities or opinions for use in making comparisons due to the dissimilarity along some dimensions.

The ability to wisely use or conserve energy and/or an individual's opinion of himself as being a prudent consumer or energy conscious consumer justifies the examination of energy utilization from a social comparison theory perspective. In the absence of any objective standards (e.g., experimenter provided, RCS audit) against which to judge their energy consumption, consumers may use information concerning similar others' consumption levels in assessing their own performance. For example, an individual may engage in the comparative process by asking a neighbor "how much was your electric bill?" A study by Pallak et al. (1976; 1980) also supports this not ion in finding that consumers receiving both usage feedback and the average usage figures for all individuals in the study conserved more energy than did consumers receiving only the usage feedback.

Given the previous energy related research, the discussion of social comparison theory up to this point, and the Pallak et al. (1976; 1980) study, a number of predict ions can be made. First, consumers provided with usage and cost feedback which does not facilitate the setting of standards nor the activation of a comparison process should conserve less energy than consumers provided with the same feedback plus facilitator. More specifically, consumers provided with current usage feedback plus previous usage figures should be more likely to engage in a self-comparison process using the previous figures as a standard and, hence, reduce their consumption more than consumers provided with only current usage feedback. Second, if consumers actually desire to evaluate their performance in conserving energy or their opinion about themselves as related to consumption, those consumers provided with feedback allowing them to make social comparisons should alter their consumption in the direction of the comparison point more than if they received self-comparison feedback. Consumers provided with their current usage figures and the current usage figures of their neighbors should adjust their consumption levels using their neighbors' consumption levels as a standard. These hypotheses suggest that consumers would be utilizing either their past consumption levels or others' consumption levels as an implied goal or standard in judging their own consumption and making adjustments in consumption according to the standards. The change in consumption should be greater under social-comparison as opposed to self-comparison feedback since consumers could justify changes in their own consumption as being due to temperature fluctuations.

The information that one receives for comparative purposes obviously plays an important role. From Becker's (1978) study, one would predict that when an individual uses past consumption levels as a standard or goal in comparing current consumption levels, the higher the discrepancy the more the individual would attempt to conserve energy, assuming previous levels were lower than current levels of consumption. In other words, if previous consumption levels are utilized as a standard in a self-comparison situation, the greater the discrepancy with current levels of consumption the greater the change in future consumption levels. However, although this replication of Becker's findings should be found for self-comparison feedback, social comparison theory postulates the opposite effect for social-comparison feedback. Specifically, hypotheses III of social comparison theory suggests that as the discrepancy between one's ability/opinions and those of others increases, the probability of the comparison being made decreases and, hence, the tendency to alter behavior decrease. Therefore, as the discrepancy between an individual's consumption level and others' consumption level increase, less attempts should be mate by that individual to alter their consumption level. Hence, an interaction should be found such that for self-comparison feedback a more difficult goal should result in more change in consumption while for social-comparison feedback a more difficult goal should result in less change in consumption when compared to a relatively easy goal.



A local natural gas utility company submitted the names and addresses of 173 natural gas consumers to the author. These consumers were selected from the company's records on the basis of being within the same billing cycle, residing within similarly constructed homes located in one of three subdivisions, and with the restrictions that all homes were equipped with natural gas heating furnaces and water heaters. The majority of the homes selected were either single-level brick or wood-frame homes and occupied by two adults and either one or two children.

Initially, the consumers received a letter from the Vice-President of the local utility company explaining the consumers' potential involvement in the study and emphasizing the fact that participation was voluntary. Subsequently, the consumers were contacted in person and were asked to participate in a natural resource utilization study to be conducted over an extended period of time, allow access to their usage records of natural gas, and to give their permission for their utility meters to be read on a weekly basis. Of the 165 consumers that were contacted, 122 agreed to participate (84.1%) and signed consent and release forms. After examining previous usage records and subsequent records prior to the onset of experimental treatments, 14 consumers were dropped from the study due to floor effects in consumption; these consumers did, however, receive weekly usage feedback. Hence. 108 consumers were eligible for assignment to treatment conditions.


A 3 X 2 X 4 two-between, one-within design was used manipulating the type of comparison feedback (self, social, or both), the implied goal as indicated by the feedback (13% or 23% reduction in usage), and feedback trials (first, second, third, and fourth week). A trailer control condition was also included in the design in which participants received only usage feedback.

All of the participants received usage feedback consisting of the amount of natural gas used during the previous week in thousands of cubic feet (MCF) and the estimated cost. The cost figure was computed by estimating the monthly consumption, applying the appropriate billing rate structure, and deriving the weekly cost figure. Participants in the trailer control condition received only this usage feedback (cost & MCF consumed).

Participants in the comparison feedback conditions were provided with information indicating an implied goal with which to compare their usage of natural gas. The implied goal was manipulated by showing the participant that the comparison information indicated either a 13% or 23% discrepancy between their own present usage and past usage or between their present usage and the use of a group of comparison others. The participant was always indicated as having consumed more natural gas than shown in the comparison feedback. The participant received only the usage and cost figures for use in making comparisons; the percentage difference was not indicated. Participants receiving this false feedback were provided with comparison information for natural gas consumed either by the participant during the same week of the immediately preceding year (self-comparison feedback), similar households during the same week (social-comparison feedback), or both (self-comparison plus social-comparison feedback). For example, the self-comparison feedback indicated to the participants that they consumed either 13% or 23% less natural gas during the same week of the previous year.

For the participants receiving both self-comparison and social-comparison feedback, the self-comparison feedback was presented first for half of the participants and last for the other half of the participants within the 13% and 23% implied-goal conditions. Because of the possibility of arousing the suspicion of the participants receiving both types of comparison feedback if the figures were exactly equivalent, 12% and 14% reduction figures were used for the 13% implied-goal conditions for the first and second type of feedback presented on a single feedback trial while for the 23% implied-goal conditions, 22% and 24% reduction figures were used. Feedback was provided to the participants by mailing a single post card containing the appropriate feedback on the same day their utility meters were read.


After obtaining permission from the participants for their involvement in the study and for their usage records to be released from the utility company, the participants were not contacted in person at any time in the future. The participants natural gas usage records were obtained from the utility company for December, 1981, and the average weekly consumption rate for each participant was determined. Ten participants were randomly withdrawn prior to any subsequent assignment to the experimental conditions. These participants were withheld for the purposes of replacing any of the other participants assigned to the treatment conditions who might withdraw from the study or move from the area prior to the presentation of the feedback. However, none of the participants assigned to the treatment conditions withdrew from the study. Hence, the withheld participants were included in the usage feedback condition. Using the average weekly consumption rate during December as a baseline, a ranking of the 98 remaining participants was made according to consumption rate and they were assigned to each of the seven conditions using a block-randomized procedure. With' the inclusion of the ten participants withheld earlier for replacement purposes, 24 participants were included in the usage feedback condition while 14 participants were assigned to each of the other conditions.

Beginning in the third week of January, 1982, participants' natural gas utility meters were read and recorded once a week for six weeks. Meters were read on the same weekday and at approximately the same time each day. Five research assistants unaware of the purposes of the study were trained by the natural gas company to read the meters. Periodically, a sixth separately trained research assistant read a subset of the meters in order to obtain reliability estimates. An overall reliability estimate of r - .96 was obtained.

The last week in January was used as the baseline level to be used for the designation of half of the participants in each condition as high-level users and the other half of the participants as low-level users. Also, this week was used as the baseline level for subsequent analyses. During this week no feedback was presented; the MCF usage was used for subsequent feedback. Beginning in the first week of February, 1982, the participants received via a mailed post card the appropriate feedback for the particular condition to which they were assigned. Feedback was provided for four weeks.


As a manipulation check, an analysis of variance was conducted on the amount of natural gas consumed during the last week of January for the type of comparison feedback (self, social, or both) X implied goal (13% or 23%) X usage level (high or low) matrix. The only significant effect was the main effect of usage level, F(1,72) - 79.17, p < .0001, with means of 3.45 MCF and 4.91 MCF for low-level and high-level users, respectively. No other effects approached statistical significance, p > .10.

The dependent measure employed for all subsequent analyses investigating the change in consumption due to feedback was the percent consumed relative to baseline levels. Because of the lower limits involved in using the percent consumed relative to baseline as the dependent measure, a square root transformation was employed. Means reported in the text, however, have been retransformed into the percentage figures.


A four-factor analysis of variance with one repeated measure was conducted on the type of comparison feedback X implied goal X usage level X feedback trials matrix. A marginally significant comparison feedback X implied goal interaction was obtained, F(2,72) = 2.74, p = .07. The means contributing to this interaction are presented in Table 1.



Tests of simple main effects revealed a significant difference in percent consumed between participants receiving social-comparison feedback indicating a 23% implied goal and participants receiving social-comparison feedback indicating a 13% implied goal, F(1,24) s 7.96, p < .01. For the social-comparison feedback, whereas participants receiving the 23% implied goal decreased their consumption by 7% relative to baseline levels, participants receiving the 13% implied goal increased their consumption relative to baseline by 8% during the same time period. None of the other simple main effects approached acceptable levels of statistical significance, p < .10.

Significant main effects for the usage level factor. F(1,72) - 2.81, p < .10, and trials, F(3,216) - 641.21, p < .001, were qualified by a significant usage level X trials interaction, F(3,216) - 2.8%, p < .05. Simple main effects revealed that, during Trials 1, 2, and 3, low-level consumers altered their usage relative to baseline (Xs = 1.31, 1.41, .88, .79, respectively) more than did high-level consumers (Xs = 1.19, 1.31, .79, respectively). No significant differences between usage levels were present during Trial 4 (low-level users X = .60, high-level users X = .59).

Comparison Versus Usage Feedback

Using a pooled mean square error term, Dunnett's t-tests were calculated testing the differences between the means for the usage feedback condition and the comparison feedback X implied goal conditions using computations outlined by Winer (1971, p. 470). These analyses revealed that participants receiving social-comparison feedback indicating a 23% implied goal reduced their consumption relative to baseline (X = .93), while participants receiving only the usage feedback increased their consumption relative to baseline levels (X = 1.03), t(94) = 1.82, p < .05. Hence, the addition of the social-comparison feedback indicating a difficult goal resulted in a 10% decrease in energy consumption when compared to the usage feedback. No differences were found between the other feedback conditions and usage feedback condition.

Because of the significant usage level X trials interaction found for the participants receiving comparison feedback, it was deemed appropriate to compare the means involved in this interaction with the corresponding means of the same interaction for participants receiving usage feedback using pairwise Dunnett's t-tests. Again, a pooled mean square error term was employed. None of these comparisons revealed any significant differences, D > .10.

Comparisons Involving Expected Usage

In order to assess any differences between the percent consumed relative to baseline consumption due to the various types of feedback in the present study and the percent of natural gas consumption that would be expected due to temperature fluctuations, it was necessary to derive the percent change in heating degree days relative baseline for each of the respective feedback trials. Basically, heating degree days are the number of degrees that the average temperature per day is less than 65 degrees F. (18.33 C.). To obtain the number of heating degree days over a set period of time (i.e., seven days in the present study), the number of heating degree days are summed. Degree days, whet her heating or cooling, are the units of measurement commonly used by utility companies to predict energy consumption during specific periods of time and have been used in previous studies (e.g., Becker & Seligman, 1978).

Heating degree days were obtained from the National Weather Service for the area in which the study was conducted and the percent of heating degree days was calculated for the treatment time periods relative to the baseline time period. These estimates of expected usage due to changes in heating degree days were felt to reflect the changes in energy consumption due to temperature fluctuations. Using the square root transformation, the expected change in energy usage tue to heating degree days was compared to the changes in energy usage found in the present study using Dunnett's t-tests for comparisons involving a single control group (i.e., expected change in usage tue to temperature changes) and multiple experimental groups (i.e., means involved in the comparison feedback X implied goal conditions and usage feedback condition). However, caution should be exercised in interpreting the results of these analyses. First, these analyses assume that if a control group had actually been included in the study the mean percent consumed would have been equal to that predicted by changes in the number of heating degree days. This assumption may be erroneous in that changes in degree days may not be related to actual changes in natural gas consumption in a one-to-one relationship. Also, the slopes and intercepts may be different. Similarly, the analyses assume that energy usage can be perfectly predicted by changes in temperature, an assumption that is unlikely. Second, using the percent of heating degree days during the experimental periods as compared to baseline levels of degree days in creating a control group for comparison purposes, an assumption must be made that if a control group had actually been included in the study, the obtained variance would have been homogeneous with the variances obtained for the participants receiving the experimental manipulations.

In order to minimize some of the possible violations of the assumptions mentioned above in testing the changes in consumption by participants in the study with the expected percent change in energy usage predicted from heating degree days, the corresponding mean square error terms and number of subjects considered in the immediately preceding analyses were employed as the denominator in calculating the Dunnett's t-statistic. Using this procedure actually results in a more conservative estimate of t than assuming that a control group was actually included in the study (i.e., specifying a number of subjects for the expected change scores).

The results of the Dunnett's tests indicated that participants receiving the social-comparison feedback indicating a 23% goal reduced their consumption relative to baseline (X = .93) differentially than would be expected from changes in heating degree days (X = 1.02), t(94) = 1.67, p < .05, one-tailed test. This result can be interpreted, with the cautions mentioned above, as demonstrating that participants receiving social-comparison feedback indicating a difficult goal used 9% less natural gas than would be expected by changes in the number of heating degree days. None of the other comparisons approached statistical significance.


The effects of feedback found in the present study can not be adequately explained by either the goal-setting perspective (Becker 1978; Seligman et al. 1981) nor by a social comparison perspective (Festinger 1954) when these perspectives are considered independently. However, when these perspectives are integrated, the present data suggest multiple necessary causes. From the goal-setting perspective, it seems clear that individuals will attempt more effective energy conservation efforts when a difficult, as opposed to a relatively easy goal, is imposed. This goal-setting effect is qualified, however, by the comparison processes that are activated via the feedback. Feedback about others' energy usage patterns appear to be more effective in promoting energy conservation than feedback about past personal consumption levels or than feedback simply stating current consumption levels. However, contrary to hypothesis III of social comparison theory and consistent with the goal setting perspective, the social-comparison feedback in the present study was only effective when consumers received information indicating that others were using considerably less energy than the consumers. Consumers were able to utilize information about others' energy consumption levels in evaluating their own consumption levels and in adopting personal energy usage goals. Hence, the present study tents to support an integrated goal setting and comparative process interpreted ion of the feedback effect.

A number of issues can, however, be raised concerning the two perspectives for explaining the feedback effect and, specifically, the lack of any main effects. From the goal-setting perspective, the absence of a goal main effect may have been tue to the source of the goal, the commitment to the goal, or to the credibility of the information provided. In previous studies examining consumption goals (e.g., Becker 1978), the goal has been set by the experimenter. In the present study, the goal was merely implied to the consumers via the feedback. Hence, the goal effect found in previous studies may have been due to demand characteristics or, simply, to the overall saliency of the goal itself. Consumers may not have been as committed to the goal in the present study as when they made a "public" statement to try to achieve a goal in previous studies. The lack of the goal main effect could also be due to consumers justifying their consumption levels since outside temperature was not controlled for in the feedback; a factor impacting upon the credibility of the feedback (Winett & Kagel 1984). Similarly, previous studies have provided feedback concerning performance relative to the goal. However, in the present study no attempts were made to provide information to the consumers indicating performance relative to the goal. In fact, in the present study the discrepancy between current and previous usage levels remained constant throughout the study. However, the plausibility of this factor in contributing to the lack of a main effect is weakened when one notes the absence of any goal X trials interaction. Finally, the absence of a goal effect may be due in part to the generalizability of the finding from electricity consumed for home cooling (Becker 1978) to natural gas usage for home heating. Some ancillary evidence for this proposition exist from a study by Winett and Nietzel (1975) which found rebates to be effective in lowering electricity consumption but not for natural gas usage when both energy sources were used for home heating. Also, the ability to detect significant differences due to measurement sensitivity for kilowatts of electricity as compared to thousands of cubic feet of natural gas may have contributed to the null finding.

As with the absence of a goal main effect, measurement sensitivity may have hampered the ability to detect a comparison feedback main effect. In other words, the insensitivity of MCF (thousand of cubic feet) as opposed to KWH (kilowatts) may have caused any differences between social-comparison and self-comparison feedback to be undetected. An additional problem for social comparison theory involves hypothesis III. From hypothesis III of social comparison theory it was predicted that when a relatively large discrepancy between others' consumption levels and the individuals consumption level were portrayed via the feedback, the individual would be less likely to engage in the comparison process and, hence, alter their subsequent consumption. However, it was found that under the larger discrepancy (i.e., goal) individuals did alter their consumption. This effect may be tue to the 23% goal not being perceived as a difficult goal or as not indicating "different" others. Alternatively, individuals receiving the social-comparison feedback implying a 13% goal may not have perceived the difference to be discrepant from their own consumption levels. In other words, either the 23% goal may not have been discrepant enough, the 13% goal may have not been perceived to be discrepant, or both factors may have contributed to the lack of an effect. Future research examining the difference thresholds for comparative feedback is warranted.

It was predicted that the combined self- and social-comparison feedback would increase consumers perceptions of the similarity between themselves and others, leading to more conservation attempts. However, at least for the 23% implied goal, when provided with both types of comparison feedback the self-comparison feedback negated any beneficial effects of social-comparison feedback. It is feasible that when both types of feedback were presented to consumers the similarity between their previous usage levels and others' current usage levels may have been used as indicating or emphasizing a dissimilarity between the other individuals and themselves. Alternatively, the amount of information presented for the combined comparison condition may have acted as a buffer in not initiating or facilitating the comparative process.

On a more general level, the results of the present study underscore the importance of examining the parameters and mediators involved in energy utilization from a process perspective. The underlying processes involved must be examined empirically and used in making policy decisions. For example, the results of the Pallak et al. (1976; 1980) study and the present study seem to suggest that providing consumers with comparative information and, specifically, social comparison information on utility bills may effect subsequent consumption. However, if one acknowledges that if an average consumption figure was provided to consumers, the information would indicate a higher rate of consumption for half of the recipients and a lower rate of consumption for half of the recipients and, hence, any beneficial effects could be negated; lower level users may cease any efforts to conserve. Additional systematic research is warranted examining the feedback effect and testing alternative theoretical perspectives in field settings in order to further identify and understand the energy consumption process.


Becker, L. J. (1978), "Joint Effect of Feedback and Goal Setting on Performance: A Field Study of Residential Energy Conservation," Journal of Applied Psychology, 63, 428-433.

Becker, L. J., Seligman, C., & Darley, J. (1979), "Psychological Strategies to Reduce Energy Consumption: Project Summary Report," Center for Energy and Environmental Studies, Princeton University: Report PU/CEES 90.

Bittle, R. G., Thaler, G., & Valesano, R. (1978), 'the Effects of Weekly Cumulative Cost Feedback on Residential Electricity Consumption," Unpublished manuscript, Anna Mental Health and Development Center, Anna, Illinois.

Bittle, R. G., Valesano, R., & Thaler, G. (1979), "The Effects of Daily Cost Feedback on Residential Electricity Consumption," Behavior Modification, 3, 187-202.

Bittle, R. G., Valesano, R. M., & Thaler, G. (1979-80), "The Effects of Daily Feedback on Residential Electricity Usage as a Function of Usage Level and Type of Feedback Information," Journal of Environmental Systems, 9, 275-287.

Cone, J. D. & Hayes, S. C. (1978), Environmental Problems/Behavioral Solutions. Monterey, Calif.: Brooks/Cole Publishing Co.

Cook, S. W. & Berrenberg, J. L. (1981), "Approaches to Encouraging Conservation Behavior: A Review and Conceptual Framework," The Journal of Social Issues, 37, 73-107.

Festinger, L. A. (1954), "Theory of Social Comparison Processes," Human Relations, 7, 117-140.

Gottlieb, D. & Matre, M. (1976), "Conceptions of Energy Shortages and Energy Conserving Behavior," Social Science Quarterly, 57, 421-429.

Hayes, S. C. & Cone, J. D. (1977), "Reducing Residential Electrical Energy Use: Payments, Information, and Feedback," Journal of Applied Behavior Analysis, 10, 424-435.

Hayes, S. C., & Cone, J. D. (1981), "The Reduction of Residential Energy Consumption through Simple Monthly Feedback," Journal of Applied Behavior Analysts, 14 (Spring), 81-88.

McClelland, L. & Canter, a. (1981), "Psychological Research on Energy Conservation: Context, Approaches, and Methods," In A. Baum & J. Singer (Eds.), Advances in Environmental Psychology: Energy Conservation: Psychological Perspectives, (Vol. 3). Hillsdale, N. J.: Lawrence Erlbaum Associates, Inc.

McDougall, G. a. G., Claxton, J. D., Ritchie, J. R. B., & Anderson, C. D. (1981), "Consumer Energy Research: A Review," Journal of Consumer Research, 8 (December), 343-354.

Olsen, M. E. L Goodnight, J. A. (1977), "Social Aspects of Energy Conservation," Northwest Energy Policy Project, Battelle Human Affairs Research Centers.

Pallak, M. S., Cook. D. A., & Sullivan, J. J. (1976), "The Effects of Comparison Feedback on Electricity Usage," Unpublished manuscript, University of Iowa.

Pallak, M. S. , Cook, D. A., & Sullivan, J. J. (1980), "Commitment and Energy Conservation," In L. Bickman (Et.), Applied Social Psychology Annual, Beverly Hills, Calif.: Sage Publications, Inc., 1, 235-254.

Pallak, M. S. & Cummings, W. (1976), "Commitment and Voluntary Energy Conservation," Personality and Social Psychology Bulletin, 2, 27-30.

Palmer, M. H., Lloyd, M. E., & Lloyd, K. E. (1977), "An Experimental Analysis of Electricity Conservation Procedures," Journal of Applied Behavior Analysis, 10, 665-671.

Seaver, W. B. & Patterson, A. H. (1976) "Decreasing Fuel-Oil Consumption through Feedback and Social Commendation," Journal of Applied Behavior Analysis, 9. 147-152.

Seligman, C., Becker, L., & Darley, J. (1981), "Encouraging Residential Energy Conservation through Feedback," In A. Baum & J. Singer (Eds.), Advances in Environmental Psychology: Energy Conservation: Psychological Perspectives, (Vol. 3). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.

Seligman, C., Kriss, M., Darley, J. M., Fazio, R. a. , Becker, L. J., & Pryor, J. B. (1979), "Predicting Summer Energy Consumption from Homeowners' Attitudes," Journal of Applied Social Psychology, 9, 70-90.

Shippee, G. E. (1980), "Energy Consumption and Conservation Psychology: A Review and Conceptual Analysis," Environmental Management, 4, 297-314.

Shippee, G. E. (1981), "Energy Conservation and the Role of the Social Sciences," Energy Policy, March, 32-38.

Singer, J. (1980), "Social Comparison: The Process of Self-Evaluation," In L. Festinger (Et.), Retrospections on Social Psychology. New York: Oxford University Press.

Stern, P. C., & Gardner, G. T. (1981), "Psychological Research and Energy Policy," American Psychologist, 36, 329-342.

Winer, B. J. (1971), Statistical Principles in Experimental Design, New York: McGraw-Hill Book Co.

Winett, R. A., Kagel, J. a., Battalio, R. C., & Winkler, R. C. (1978), "Effects of Monetary Rebates, Feedback, and Information on Residential Electricity Conservation," Journal of Applied Psychology, 63, 73-80.

Winett, R. A. & Kagel, J. a. (1984), "Effects of Information Presentation Format on Resource Use in Field Studies," Journal of Consumer Research, 11 (September), 655-667.

Winett, R. A., Neale, M. S., & Grier, R. C. (1979), "The Effects of Self-Monitoring and Feedback on Residential Electricity Consumption: Winter," Journal of Applied Behavior Analysis, 12, 173-184.

Winett, R. A., Neale, M. S., Williams, K., Yokley, J.,& Kauder, i. (1978-79), 'the Effects of Feedback on Residential Electricity Consumption: Three replications," Journal of Environmental Systems, 8, 217-233.

Winett, R. A.& Nietzel, M. I. (1975), "Behavioral Ecology: Contingency Management of Consumer Energy Use," American Journal of Community Psychology, 3, 123-133.

Winkler, R. C. & Winett, R. A. (1982), "Behavioral Interventions in Resource Conservation: A Systems Approach Based on Behavioral Economics," American Psychologist, 37, 421-435.



Thomas D. Jensen, University of Arkansas


NA - Advances in Consumer Research Volume 13 | 1986

Share Proceeding

Featured papers

See More


E12. Green versus Premium Choice and Feelings of Pride

Cecilia Souto Maior, Federal University of Paraná
Danielle Mantovani, Federal University of Paraná
Rafael Demczuk, Federal University of Paraná

Read More


Inside Out: Product Essence is Perceived to be Concentrated in the Center of a Group of Products

Kunter Gunasti, Washington State University, USA
Noah VanBergen, University of Cincinnati, USA
Caglar Irmak, University of Miami, USA

Read More


O3. The Effect of Numeric Information on Product Evaluation

Zhen Yang, Drexel University, USA
Yanliu Huang, Drexel University, USA
Dengfeng Yan, University of Texas at San Antonio, USA

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.