A Longitudinal Study of Factors Affecting Household Energy Expenditures in Canada 1969-1982

ABSTRACT - Data from the Family Expenditure surveys between 1969 and 1982 provide one of the best sources of historical information on the behavior of Canadians throughout the 'energy decade'. Through the application of special regression analysis procedures, the relationship of energy use to demographic and capital stock characteristics is assessed.


Louise A. Heslop (1986) ,"A Longitudinal Study of Factors Affecting Household Energy Expenditures in Canada 1969-1982", in NA - Advances in Consumer Research Volume 13, eds. Richard J. Lutz, Provo, UT : Association for Consumer Research, Pages: 492-497.

Advances in Consumer Research Volume 13, 1986      Pages 492-497


Louise A. Heslop, Carleton University and Statistics Canada

[The views expressed in this paper are the responsibility of the author alone, not of Statistics Canada.]


Data from the Family Expenditure surveys between 1969 and 1982 provide one of the best sources of historical information on the behavior of Canadians throughout the 'energy decade'. Through the application of special regression analysis procedures, the relationship of energy use to demographic and capital stock characteristics is assessed.


Much research has been directed to determining the factors related to and affecting energy use. These efforts have had two main goals. One is to determine who has been most burdened by the rapid increases in energy prices. Policymakers need this information in devising programs of protection for those disadvantaged groups. The second goal of the research is more directly related to conservation programs. By sorting out those factors related to high energy use, the targets for conservation programs can be identified. By determining the characteristics of those who have constrained energy use, the effectiveness of such conservation programs can be more clearly assessed.

This study makes use of the wealth of some existing Statistics Canada data and multivariate analysis procedures to simultaneously examine severAl factors known or postulated to be related to energy use. The data set has particular historical value as it covers the period 1969 1982, spanning the 'energy decade'.

Difficulty of Conducting Research on Energy Use

Studying energy use by household consumers is a very difficult process. Actual energy consumption data is hard to get and is highly suspect in many cases. The problems arise from the fact that most households do not keep long-term information on their ongoing energy purchases. Moreover, since the 1972 'energy crisis' and subsequent pressures on Canadians to conserve, most respondents are cautious to say that they do try to conserve energy and that they have been cutting back on its use. So answers to questions on use are highly reactive. Even worse, it is quite possible that the bias due to self-reporting may be unevenly distributed across the population. Those with higher incomes and educations have repeatedly been found to be more sensitive to energy conservation issues and to report more conservation oriented attitudes. However, energy use has also been consistently related to incomes. Therefore, there might be more under-reporting of energy use among those who on attitude measures are the most conservation-minded but yet are using more energy than those with lesser incomes. Non-reactive and reasonably accurate measures are needed.

Another problem which makes energy research difficult is that it is generally not possible to get comparative data over long periods of time. To assess the impact of the 'energy crisis' at the individual level, data must be available from before and after key events in the history of that crisis period. However, it was only after the beginning of the energy situation that researchers generally became interested in energy consumption and began collecting information on usage rates. Retrieving comparable information from pre-crisis years is very difficult. Certainly respondent recall is not adequate.

In some studies consumption data has been obtained from utility company records (Heslop et al. 1978, McDougall et al. 1979). Such records are excellent sources of detailed data that are not reactive and are historical. However, it has proven very difficult to obtain such data on large samples of respondents. Utility companies require signed releases before they will allow consultation of records on households. Also in many cases records are not kept by suppliers longer than two years. Finally, the number of suppliers of home heating fuel is quite large especially if a large geographical area is being studied. In a recent study (McDougall et al. 1979) electricity utility records could only be obtained on about half and gas and heating oil records on about one-third of respondents who were members of a research panel and, therefore, more likely to co-operate.


The Family Expenditure Survey data base offers many advantages for studies of energy use under these circumstances. Detailed expenditures on a national sample of households were collected in 1969, 1972, 1974, 1976, 1978 and 1982. So the time period covered precedes the 1973 "crisis" and includes an early post-1973 measure, then one preceding the large price increases in 1978-79 and one following them, during a time of moderating and stabilizing prices. Sample sizes are typically several thousand and the sample is nationally representative. There is no problem arising out of selection into the sample because of personal interest in the subject matter. The information on energy use can be considered as non-reactive. The study is not focussed on energy use itself. Respondents are asked about energy expenditures just as they are asked about all other expenditures. So the respondent is likely to be less sensitive to give the "right" answers.

The data set also includes a wealth of other information which can be linked to energy use, including expenditures on other categories of goods and services, demographic information on family composition, age, income and education of the head, and also capital goods information such as dwelling type, size and location, and appliance ownership. The use of this information along with energy expenditures allows for the profiling of the lifestyle of the energy user and the factors associated with energy use .

The data includes energy expenditure information in dollar terms, but not in quantity terms. Where accurate price information is available, conversions are possible. However, in order to make these conversions, within-year price changes make it necessary to do some rather rough interpolations of proportions of the total year's energy which may be consumed at different times of the year. This would be a highly judgmental process and so was not undertaken in this project. So this study is of energy expenditures as one measure of energy use.

The data set is not without its problems, however. There have been many changes over the years in how some categories of expenditures have been handled in reporting. For this study, these categories were manipulated where necessary to provide for a consistency in categorization across the 13-year period. As far as possible the 1982 categorization format was used. [For detailed information on the construction of the data sets used in this study, contact the Social and Economic Studies Division of Statistics Canada.]

The study is a recall study, but considerable care is taken to aid recall, and householders can check personal records for assistance. The data is collected in the first 3 months of the year following the year under study to minimize recall difficulties. Therefore, errors are likely to be relatively small and random.

Data on direct energy use is only available from those households who pay directly for their energy. Apartment dwellers by and large and especially during the earlier periods under study were on bulk meters. Therefore, how much they actually pay for their in-home energy cannot be determined as their rent payments include some unknown proportion for energy. Some studies have arbitrarily assigned a certain proportion of the rent to be the measure of energy use. However, this has not been done here. Where the procedure has been used, the judgments as to the rental proportion going to energy have been vaguely and arbitrarily set as a fixed proportion of rents. There appears to be no justification for this technique, except to increase sample size. Furthermore, renters usually can do very little to control energy use in their dwelling and have no incentive to do so. They are, therefore, likely to be far less sensitive to energy price increases and efforts to encourage conservation. It should also be kept in mind that such renters are atypical of the entire population. They are likely to have lower incomes and be younger than average or older with small families. So, because accurate inChome energy consumption on such apartment renters is not available and because they are a less important set of energy consumers in terms of the amount they consume and as targets for conservation activities, they have been removed from the sample. Their energy levels-especially within the home are likely to be lower than average.

Also rejected from the sample were part-year families, who lost or gained members or who were formed in the year to ensure that such major within-year changes did not affect the ability of models used in the study to explain variances in expenditures. Therefore, only full-year households who were owners and renters of houses - detached, semi-detached, row, town- throughout the year and who made payments directly for their energy were included in the samPle.

Finally the data set used in this study contains data only from the eight major urban areas used in all six surveys for consistency. Although it does restrict the conclusions to city-dwelling Canadians, this is the vast majority of Canadians. It also allows for the inclusion of climate factors in energy use analysis as discussed later.

Two main measures of direct energy use are used in this study -- in-home energy and transportation energy for the personal use of the car or truck. The in-home energy expenditures measure was created by combining heating and electricity information. This was done because there was no feasible way to determine what proportion of an electrically-heated home's electricity was used for home heating and how much was used for lighting and cooking. Similar problems arise in the case of other fuels which have multiple uses.

In this study then in-home energy and car and truck fuel expenditures are examined. Factors known or postulated to be related to energy use are combined and used in regression models to determine if they are useful in explaining these expenditures in any or all of the years for which data is available. The results can be usefully applied to meeting the two goals of such research, identifying who has been most affected by and who has responded to the rapid energy price increases of the 1970's.

Regression Techniques for Use with the Survey of Consumer Expenditures

The Family Expenditure Surveys use samples drawn from the Labor Force Area Sampling Frame. The sampling procedure is of the stratified cluster sampling type. It also involves disproportionate sampling of smaller cities in order to ensure the reliability of city estimates. In developing total estimates, individual records are weighted to restore to each city its appropriate importance in the total and compensating for any variability in response rates between cities as well as variable sampling ratios.

Because of the cluster sampling procedures used, the assumptions of independence of observations made for ordinary regression analyses are violated. Kish and Frankel (1970) suggest that with large samples the regression coefficients themselves that arise from the use of normal regression procedures of data from complex (correlated) samples tend to approach a normal distribution about the corresponding parameters. It is, however, the variance estimates which are more problematic. With the usual regression procedures they are likely to be underestimated, thus affecting the validity of tests of statistical significance.

Several modified regression techniques have been suggested and used to deal with these problems including balanced repeated replications, jackknife, Taylor linearization and, more recently, bootstrap procedures. The first two tent to be very time consuming and expensive operations because of the high number of iterations required. The Taylor linearization procedure has been chosen as an usually effective and less costly alternative.

The problem of variance estimation for the regression coefficients is solved "by finding a linear approximation to the estimator of the coefficients. This approximation is then used to derive an approximation of the variance" (Holt 1977, p. 5). The linearization technique used is the Taylor series expansion of the estimator. [3For more detail on the Taylor linearization procedure, see Frankel (1971) and Holt (1977).] The regression results discussed in the following sections have used this procedure with sample weights included as derived from the original data set.


The regression models used in this study have the following variables:

Dependent variables

1. Dollar expenditures on in-home energy and gasoline for private car and truck operation as explained in an earlier section.

2. In-home energy and personal transportation energy per person or per capita in the household.

3. In-home energy per room in the dwelling per 1000 degree days. [Degree-days are determined as the number of days of temperatures below 18 celsius. The information on degree-days for each year for the cities used in the study was obtained from Environment Canada and added to the expenditures data set.]

These last two measures are extremely useful measures over time of efficiencies in energy use. Schipper and Ketoff (1981) note the importance of focusing on intensity of energy use in order to understand what is actually occurring in energy use over time and to tease out the effects of conservation efforts that may be hidden in simple consumption data.

Independent variables

1. Total expenditures- This variable is being used as a substitute for the income effect frequently found to be an important explanatory variable. At the aggregate level Schipper and Ketoff (1981) note that higher average incomes of a population are associated with higher energy usage. At the individual level family income has been positively associated with income energy use (see, for example, McDougall, Ritchie, Claxton 1981) and also with energy conservation efforts involving capital investments (see, for example, Decision Research 1982; Pitts, Willenborg and Sherrill 1981; Tienda and Aborampah 1981).

Total expenditures has several advantages over income as an independent variable here. The expenditure data collected for the Family Expenditure Survey is collected in a careful and-detailed way, but less emphasis is placet on the income data. From a theoretical perspective, using total expenditures allows expenditures on categories of goods and services to be viewed within the perspective of total allocations of monies to this function Moreover, there is no concern then as to whether monies for these expenditures were derived from current income or from dissavings. The current flow of resources for expenditures is the basis of analysis. Of course, in using this variable in the regressions, expenditures on the energy source under study were subtracted from total expenditures to eliminate the situation of using the energy expenditures on both sides of the equation.

2. Stage of family life cycle - Energy is used to carry out activities in the household, rather than as a direct end in itself. Therefore, it is to be expected that families of different composition in terms of age of head and number and ages of children would have different energy needs for the different range of activities in which they must be involved. For example, the use of the family car has been repeatedly linked to the presence and ages of children (McDougall, Ritchie, Claxton 1981; Pitts, Willenborg and Sherrill 1981: and Zimmerman 1981).

3. House characteristics - These have been found to be the most useful in explaining energy use in the home as well as energy conservation activities (Decision Research 1982; McDougall, Ritchie, Claxton 1981; Tienda and Aborampah 1981; Verhallen and Raaij 1978). Number of rooms in the dwelling and whether the house was a single-detached, semi-detached, rowhouse or duplex could be identified from the FAMEX data.

4. City - In Canada there is a dramatic difference among the regions in the need for energy to maintain comfortable house temperatures. The heating degree day information for each of the cities for each of the years under study was obtained and added to the data set as a way of accounting for these differing needs. Also the eight different cities are in seven different provinces representing all the different regions of Canada. Each province differs in the price charged for the major forms of energy because of differences in distances from sources of energy, differences in provincial taxes, and, in the case of provincially-owned or controlled utility companies, differences in legislated rates and rate structures. So city variables hold useful information relating to the supply of energy, the demand for energy and its price. Since provincial factors are 80 important the two cities within the province of Ontario, Toronto and Ottawa have been combined for analysis.

5. Education - This variable has been found to be related to energy conservation activities (Decision Research 1982; Pitts, Willenborg and Sherrill 1981). It also can serve as a proxy measure of socio-economic status which has also been linked to the holding of socially conscious attitudes.

The regression models then take the form

Ei = a + b1 (TE-Ei) + b2 (S')+ b3 (Rooms) + b4 (H') + b5 (C') + b6 (E') (1)


a,bi = parameter estimates.

Ei = expenditures on the energy in the home or in the ear

TE-Ei = total expenditures for the household minus the expenditure for the energy under study

S' = dummy variable set for stage of the family life cycle where

Stage 1 = married couple with head under 45 years of age and no children

Stage 2 = married couple with head under 45 years of age and at least one child under 5 years of age

Stage 3 = THE OMITTED REFERENCE CONDITION, married couple with head under 45 years of age and at least one child 5-16 years of age

Stage 4 = married couple with head 45-64 years of age or over with at least one child 5-16 years of age

Stage 5 = married couple with head 45-64 gears of age with no children under 16 years of age

Stage 6 = married couple with head 65 years of age or over

Stage 7 = unattached individual under 45 years of age

Stage 8 = unattached individual 45-64 years of age

Stage 9 = unattached individual 65 years of age or older

Rooms = number of rooms in the household dwelling unit

H' = dummy variable set for the type of dwelling unit with single-detached duelling being the OMITTED REFERENCE CONDITION

S = a semi-detached duelling

R = a rowhouse

D = a duplex unit

(NOTE: Information on this variable set was not collected in 1969.)

C' = a dummy variable set for cities where Toronto and Ottawa (Ontario cities) combined are the OMITTED REFERENCE CONDITION

SJ = St. John's, Newfoundland

H = Halifax, Nova Scotia

M = Montreal, Quebec

W = Winnipeg, Manitoba

E = Edmonton, Alberta

V = Vancouver, British Columbia

E' = dummy variable set for education where

Grade 8 or less completed is the OMITTED REFERENCE CONDITION

PLS - partial secondary level completed

CS - completed secondary level

PS - non-university post-secondary education, partial or complete

PU - partial university

CU - university degree completed

The entire list of independent variables was used in the models involving the dependent variables of in-hone energy expenditures and in-home energy expenditures per capita. For the in-home energy expenditures per room per 1000 degree days, the number of rooms and the city variables were dropped from the list of independent variables.

For the car and truck fuel expenditures the house characteristics were, of course, not included as independent variables.


Tables 1 and 2 summarize the results of the testing of the regression models. Presented in these tables are all the significant variables in the models with the numerical values given for the estimates of the intercept terms and the total expenditures variable parameters. Also given are the adjusted R-squared values for the models.



In-Home Energy

In-home energy expenditure variances were well explained by the model elements . All the adjusted R-squares are significant and most are reasonably high. The adjusted R-squared values were higher for the dollar expenditures and the dollar per capita measures than for the dollar per room per degree day indicating the fundamental importance of climate, prices of energy and house size in understanding energy consumption. Also for the first two measures there is a definite decline in the model's ability to explain the variance over time suggesting that individual factors not included in the model may be accounting for more and more of the differences among households over time. For the dollars per room per degree day measure the R-square increases somewhat in the later period as more of the variables in the model (stage of family life cycle and education) become significant.

The intercept tenn can be thought of as a minimum amount of expenditure for the household as defined by the set of reference or baseline conditions for all the dummy variables. In this case it would be a family in Stage 3 of the family life cycle (male heat under 45, one or more children S-16 years of age) living in Ontario in a single-detached house and the male of the household head has less than a high school education. The amount of this minimum has been increasingly dramatically over the thirteen-year period. For in-home energy dollar expenditures it jumped sharply between 1969 and 1972, then more slowly but consistently until 1976, then took a large leap to the 1978 value and finally levelled off. The total increase over the period was 404%. For dollars per capita the general pattern was the same with an overall increase of 1123%. For dollars per room per degree day this increase was more gradual throughout the early period until after 1978 with it took a very large jump with the final percentage increase being 256% for the period. The ouch larger increase for the per capita expenditures likely reflects the basic changes in family structure that occurred during this period due to such demographic changes as the decline in the number of children per family, the maturing of the baby-boom generation, and rising divorce rates. All these lead to drops in average family size. Further the population of elderly families continuing to maintain their own homes is increasing. The smaller increase in the expenditures per room per degree day may indicate that the Canadian population is making some real efforts to curtail energy use which may not be as obvious when only total expenditures are examined.

The household's total budget has a consistent significant positive effect only on the total energy expenditure measure. The trend in the size of the parameter estimates for total expenditures seem to follow the trends in the real prices (adjusted for income increases) of energy in Canada which declined in the early 1970's, rose in the mid-70's and then levelled off or moved down slightly towards the last part of the period under study. Total expenditures are not significant for the other two energy expenditure measures until prices started upward. Those with wore money to spend then were the ones who could continue despite these price increases to spend more intensively, i.e., more per person and more per room per degree day. They clearly were more able to absorb the price increases, whereas those with more limited resources had to find ways to hold the line. So it would appear that as prices of energy increased those with larger budgets spent more and made significantly greater gains in their use of energy.

The stage of the family life cycle variable set is very interesting in what it reveals across the different measures of energy use and the different years. For in-home energy expenditures it appears that most families spend alike and do not differ substantially from the reference condition (Stage 3). In the early periods, the smaller younger families spent less. During the later periods with energy prices much higher, the elderly couples spent more in 1978 and the young singles spent less in 1978 and 1982. However, in terms of expenditures per capita the economies of scale of the full nest families (Stage 3 and 4) become obvious. All other households spent more in every year. Finally, again for dollars per room per degree day there is little difference. The full-nest families may spend less per person but they usually occupy larger homes and do not use the energy any more or less intensively in these rooms. The only trend seen in this last measure of in-home consumption of note is the evidence that the elderly are spending more in later periods. These households are more likely to be in older homes, less well-insulated with older, less efficient oil-fired heating systems (see Heslop 1985a). For a variety of social and economic reasons they seem to be less able to make the necessary moves to improve their efficiency of energy use and must bear the costs resulting.

House characteristics are very important as would be expected from the previous research. Number of rooms is significant for four of the sis periods when relative energy prices are highest for dollar expenditures and all years for per capita expenditures. Rowhousing seems to be the only consistent winner over detached single houses for all measures for all years. The other multiple unit dwellings are in some periods but are not in others, especially the later ones. This would suggest that perhaps those in single detached dwellings have been improving the energy efficiency of their dwellings.

The set of city variables indicate the hardship experienced by those who are more isolated from energy supplies and are burdened with high prices. The cities in the Eastern Maritime provinces spent significantly greater amounts on total in-home energy and on a per person in the household basis. Those in the West near the Alberta oil supplies or in Quebec with its abundance of hydroelectric power spent less. Also those blessed with more temperate weather on the West coast usually, but not always spent less.



Education effects are very interesting. They do not contribute much to the understanding of total expenditures. However, several education variables do enter for the other two measures in later years. Post-secondary education is associated with higher per capita consumption between 1972 and 1978 but then is not, suggesting that those with higher education made disproportionately heavy use of energy on a per capita basis until its price rose dramatically in 1978-79. Moreover, with regards to expenditures per room per degree day those with the highest education used significantly less after the first 'crisis' period. This may suggest that those with higher education levels could and did respond more effectively to information about energy conservation and incentives to conserve energy or switch to less expensive sources.

Car and Truck Fuel Expenditures

For expenditures on fuel for personal private transportation (see Table 2), again the models used are significant in explaining variance in both expenditure measures across all the years. However, the adjusted R-squares are much lower than they were for in-home energy expenditures. This is undoubtedly due to the absence from the models of vehicle characteristic measures (gross vehicle weight and total engine displacement) which usually explain most of the variances in usage.

The intercept terms take a slight drop between 1969 and 1972 and these gradually increase until there is a large jump between 1978 and 1982. A-Rain the trend reflects the trends in real energy price increases. The dollar expenditures intercept term increases by 430% and the dollar per capita term by 469% over the thirteen-year period. Again, as for in-home energy, there appears to be a greater increase in per capita expenditures suggesting that car use is increasing despite gasoline price increases. Demographic shifts may explain much of the change. The period was marked by a rapid increase in the number of young drivers who are traditionally heavy users of cars and by an increase use of cars by older married couples (see Heslop 1985b).

Total expenditures again are very significant in accounting for individual differences in both expenditure measures. Those with larger budgets spent more in all years. The value of the parameter estimate increases generally over time and suggests that the 'income' sensitivity of expenditures for personal transportation is increasing.

The stages of the family life cycle that are important differ greatly between the two measures. In gross expenditures it is clear that older households spent less.

Also those households with middle-aged heads and older children still at home spent more during the middle 1970's but not thereafter. This would suggest that these heavy users responded to the need to cut back. In terms of per capita consumption the results look more like those for in-home per capita expenditures. Almost all family types consumed more per capita than the full nest families. Only the unattached individual elderly consistently did not. Even elderly married couples frequently consumed like younger households indicating recent growth of interest in driving among this group.

City location is not as consistent a predictor of personal transportation fuel expenditures as it is for in-home energy. Essentially it seems to capture here the effects of provincial variations in fuel prices. The Maritimers to not appear to be so burdened in this case. Since driving is somewhat more discretionary than home heating, they may prefer to drive less in the face of their higher gasoline prices. Montrealers appear to have lost any advantages they had in earlier periods which allowed them to have lower expenditures before 1982. Indeed, in 1982 the average price of gasoline was higher in Montreal than in all of the other seven cities (Statistics Canada 1983).

Finally again the education variable set is interesting. During the early periods those with some or all of secondary school or a non-university post-secondary education spent more than those with less than Grate 9. However, those with a university education did not. By 1982 those with post-secondary education of any kind spent significantly less than the reference group. Again there appears to be an educational achievement effect in conservation efforts resulting in lower energy expenditures.


The Family Expenditures Survey data on energy expenditures can be a very informative source of historical information on the experiences of households and their responses to the events of the 'energy decade'.

Those households in the Eastern Maritime provinces have been particularly burdened with high costs of in-home energy. Those households with higher incomes spent more and used energy more intensively. Smaller households used less total energy in the home, but families in the full nest stages used less per capita. Some multiple unit dwellings, especially rowhousing, used less inhome energy than did single-detached houses and larger houses used more.

Further there is evidence to suggest that the more highly educated have been more responsive. Also those living in single family dwellings appear to have made more efforts to improve efficiencies of in-home energy use. Finally there are indications that the elderly may be falling behind in efficiency of energy use in the home and so are incurring larger than average expenditure burdens.

The expenditure increases of households has been rapid and dramatic. The overall pattern of dollar expenditures for energy indicates little effort to control increases as they generally parallel price increases. However, a closer look at the data through the use of intensity of use measures (per capita and per room per degree day) give more evidence that efforts to counteract the price increases with restraints on the growth of energy use have occurred. Demographic and economic trends have masked some of these efforts so that they are not obvious in the gross measures.


Decision Research Ltd. (1982), An Initial Evaluation of the Canada Oil Substitution Program: Converter and Nonconverter Profiles, Ottawa: Consumer and Corporate Affairs.

Frankel, Martin Richard (1971), An Empirical Investigation of Some Properties of Multivariate Statistical Estimates from Complex Samples, unpublished PhD thesis, University of Michigan.

Heslop, Louise A., Morin, Lori and Cousineau, Amy (1978), "Consciousness in Energy Conservation Behavior: An Explatory Study, Journal of Consumer Research, 8,3, 299-305.

Heslop, Louise A. (1985a), Energy Expenditures of Canadians 1969-1982: The Elderly and Energy Use, Ottawa: Social and Economic Studies Division, Statistics Canada.

Heslop, Louise A. (1985b), An Analysis of Expenditure Patterns of the Elderly: Cohorts - Going Through Life Together, Ottawa: Social and Economic Studies Division, Statistics Canada.

Holt, Mary Margaret O. (1977), Drawing Inferences with Regression Models in Sample Surveys, Unpublished Master's Thesis, Department of Mathematics, Graduate School, Duke University, North Carolina.

Kish, Leslie, and Frankel, Martin R. (1970), "Balanced Repeated Replications for Standard Errors , Journal of the American Statistical Association, 65-331, 1071-1094.

McDougall, Gordon H.G., Ritchie, J.R. Brent, and Claxton, John D. (1981), Energy Consumption and Conservation Pat- terns in Canadian Households, Ottawa: Consumer and Corporate Affairs Canada.

Pitts, Robert E , Willenborg, John F., and Sherrell, Daniel L. (1981), "Increasing Gasoline Prices" in J. Claxton et al. (ed.), Consumer and Energy Conservation, New York: Praeger.

Schipper, Lee and Ketoff, Lee (1981), Residential Energy End Use: Developing an International Data Base" in J. Claxton et al. (ed.) Consumers and Energy Conservation, New York: Praeger.

Statistics Canada (1983), Consumer Price Index, Catalogue Number 62-001. Ottawa, Supply and Services Canada.

Tienda, Marta and Aborampah, O.M. (1981), "Energy Related Adoptions in Low-Income Non-Metropolitan Wisconsin Counties", Journal of Consumer Research, 8,3.

Verhallen, Theo M.M. and Raaij, W. Fred (1981), "Household Behavior in the Use of Natural Gas for Heating", Journal of Consumer Research, 8,3.

Zimmerman, Carol A. (1981), Household Travel Patterns by Life Cycle Stage" in J. Claxton et al. (ed.) Consumers and Energy Conservation, New York: Praeger.



Louise A. Heslop, Carleton University and Statistics Canada


NA - Advances in Consumer Research Volume 13 | 1986

Share Proceeding

Featured papers

See More


Can “Related Articles” Correct Misperceptions from False Information on Social Media?

Yu Ding, Columbia University, USA
Mira Mayrhofer, University of Vienna
Gita Venkataramani Johar, Columbia University, USA

Read More


Data... the 'Hard' & 'Soft' of it: Impact of Embodied Metaphors on Attitude Strength

Sunaina Shrivastava, University of Iowa, USA
Gaurav Jain, Rensselaer Polytechnic Institute
JaeHwan Kwon, Baylor University
Dhananjay Nayakankuppam, University of Iowa, USA

Read More


The Self-Perception Connection: Why Consumers Devalue Unattractive Produce

Lauren Grewal, Dartmouth College, USA
Jillian Hmurovic, University of Pittsburgh, USA
Cait Lamberton, University of Pittsburgh, USA
Rebecca Walker Reczek, Ohio State University, 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.