A Comparative Study of Energy Consumption and Conservation Across Family Life Cycle

ABSTRACT - Household energy consumption is examined in the specific context of home heating expenditures. Analysis of home heat costs across stages of the family life cycle reveals a distinctly different pattern than previously found when total household energy costs were examined. In contrast to the curvilinear pattern where expenditures peak during middle stages of the life cycle, these data show a positive linear relationship where higher costs are associated with elderly consumers. The research focus is expanded to include heat-related conservation behavior as well as expenditures in an effort to better understand implications for policy makers and-marketers.

Citation:

Cynthia J. Frey and Duncan G. LaBay (1983) ,"A Comparative Study of Energy Consumption and Conservation Across Family Life Cycle", in NA - Advances in Consumer Research Volume 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Abor, MI : Association for Consumer Research, Pages: 641-646.

Advances in Consumer Research Volume 10, 1983      Pages 641-646

A COMPARATIVE STUDY OF ENERGY CONSUMPTION AND CONSERVATION ACROSS FAMILY LIFE CYCLE

Cynthia J. Frey, Boston College

Duncan G. LaBay, University of New Hampshire

ABSTRACT -

Household energy consumption is examined in the specific context of home heating expenditures. Analysis of home heat costs across stages of the family life cycle reveals a distinctly different pattern than previously found when total household energy costs were examined. In contrast to the curvilinear pattern where expenditures peak during middle stages of the life cycle, these data show a positive linear relationship where higher costs are associated with elderly consumers. The research focus is expanded to include heat-related conservation behavior as well as expenditures in an effort to better understand implications for policy makers and-marketers.

INTRODUCTION

According to reports from the Department of Energy, energy consumption in the average home could be slashed by 60% through the use of conservation methods and investment in new products to improve the efficiency of consumption. Toward this end, DOE (1980) survey results show some sort of conservation-related equipment or insulating material was added to 50% of the housing units eligible for Federal energy tax credits in 1977-78. Homeowners in 1980 claimed tax credits on an estimated $4 billion worth of investments in energy saving products primarily related to home heating. Analysts suggest that this investment rate could reach the $30 billion level per year by 1990 (Business Week 1981). As the result of consumer efforts from 1973-1980, DOE estimates that energy consumption per household has been reduced an average of 12: (Forbes 1982).

While many policy decision makers would agree that demand for energy should be reduced, not all households are equally capable of changing their consumption patterns or investing in energy saving products. Determining where the biggest gains in improved efficiency of energy usage are likely to occur is a necessary first step in the development of public conservation programs as well as the marketing of conservation-related products and services. The study of conservation behavior on the part of consumers needs to be examined within the framework of consumption patterns if one is to develop an understanding of the impact of behavioral change.

Thus far the study of consumer behavior related to energy consumption and conservation has been highly fragmented. The focus has tended to be on either consumption patterns or conservation behaviors with little effort directed at understanding the relationship between these two fields of study. The purpose of this paper is to examine both energy consumption and conservation patterns in the specific context of home heat usage.

RESEARCH BACKGROUND

The appropriate unit of analysis for the study of both home heating consumption and subsequent heat-related conservation behavior is the household. While households can be categorized on any number of dimensions including income, number of members, and relative age of members, a framework for study which presents a composite of several measures related to consumption behavior is family life cycle. In a comprehensive study, Wells and Gubar (1966) documented distinct differences in consumption patterns associated with various stages of development in the family unit. Since that time, several modified formulations of the life cycle concept have been proposed (Murphy and Staples 1979; Norton 1974; Duvall 1971). The Murphy and Staples version has been the only one to generate much support in the field due to its ability to classify larger percentages of households. While the specific category descriptions of the family life cycle may have been modified to encompass a greater variety of family life styles, the premise that household needs and priorities change depending on stages of family development remains a sound one.

Application of the family life cycle concept to the study of energy consumption behavior has occurred on limited basis. Morrison and Gladhart (1976) propose that families in child-rearing stages use more residential energy than families without children at either earlier or later stages of the life cycle. Unfortunately, no quantitative analysis is presented to substantiate their findings. Support for the Morrison and Gladhart contention, however, can be found in Fritzsche (1981). Fritzsche examines household energy expenditures by family life cycle stage while controlling for income and number of family members in the labor force using multiple classification analysis (MCA). A curvilinear pattern showing peak family energy expenditures occurring during the child-rearing years is documented using both the Wells and Gubar and Murphy and Staples versions of family life cycle. While relative expenditure patterns are very similar across life cycle stages in both versions, the advantage of the Murphy and Staples model is its greater ability to classify respondents with only 3% unclassifiable compared to 10% for the Wells and Gubar model.

Fritzsche concludes that family life cycle is a meaningful construct for examining energy consumption patterns and proposes that energy conservation programs be directed at high expenditure segments. Two shortcomings are apparent in this reasoning. First, while family life cycle appears to be a valuable predictive measure, no comparison of relative explanatory power is made with other demographic measures.

Of even greater importance, however, is the statement regarding policy implications. Fritzsche's focus on total household energy expenditures, including energy used for home heating, hot water, lights and appliances, and gasoline, assumes a unidimensional concept of energy usage when a dollar's worth of heat has the same consumer utility as a dollar's worth of gasoline. Research by McDougall, Claxton, and Ritchie (1981) suggests that households make energy trade-offs between usage in the home for heat or electricity and usage for transportation. High consumption of one energy source does not necessarily imply high consumption of all other sources. From Fritzsche's analysis of total energy consumption it would be impossible to determine the relative amounts consumed in various usage contexts and direct programs toward specific consumer needs.

Within the context of energy usage to satisfy specific needs. work be Zimmerman (1981) evaluated household demand for transportation energy across the family life cycle. Using the Murphy and Staples model, Zimmerman found gasoline consumption to peak during middle life cycle stages where children were in their late teens. Once out of the work force, household consumption declined drastically. If this finding is widespread and households make trade-offs among energy usage contexts, one might expect households in later life cycle stages, who spend more time at home, to use more energy for heat than households in earlier stages. This suggests that further analysis of Fritzsche's approach using only heat cost could produce a very different Pattern of results.

Beyond the issue of examining the specific energy usage context is the association between relative levels of consumption and conservation. Understanding which households have taken steps to conserve will aid in the direction of programs and products to receptive market segments. Analysis of conservation behaviors by Curtin (1976) suggest that family size demonstrates a curvilinear relationship to conservation. Households with family sizes of three to four people reported greater past conservation, less expected difficulty in future conservation, and easier adjustment of current consumption patterns to lessen energy use. Additionally, Curtin found that the elderly represented the smallest percentage reporting heat conservation and the largest percentage expecting future difficulty in conserving or adapting to energy shortages.

If the findings of Fritzsche and Curtin are compared, one might hypothesize that Fritzsche's high expenditure households are also Curtin's high conservation households. Should this be the case, distinctly different programs could be directed to segments which have already taken the first steps in reducing their demand for energy, and appear favorably disposed toward continuing such efforts, versus segments which may not be the biggest energy users but have made no attempt as yet to conserve.

This paper will attempt to clarify the relationship between energy consumption and conservation across stages of the family life cycle. In addressing the issue of energy usage context, the focus of the analysis is on home heating expenditures and heat-related conservation behaviors. In light of findings by Fritzsche and the propositions of Morrison and Gladhart, the first two hypotheses focus on consumption, relating home heating expenditures and family life cycle.

H1: The family life cycle concept explains a significant portion of the variance in home heating expenditures.

H2: A differential pattern of heating expenditures occurs over the stages of the family life cycle.

(a) Heating costs will increase with each successive stage of the family life cycle until those stages where the children have left home.

(b) Heating costs after the children have departed will decline, although they will remain higher than for those individuals in pre-child years.

In an effort to better understand the relationship between consumption and conservation, a third hypothesis drawn from Zimmerman's and Curtin's work is examined.

H3: A differential pattern of heat-related conservation behavior occurs over the stages of the family life cycle.

(a) The number of steps taken to conserve will be highest for those households in the middle stages of the family life cycle with children.

(b) The lowest level of reported conservation wilL occur during the two last stages (elderly) or the family life cycle.

METHODOLOGY

The Sample

This study was designed to examine the relationship between individuals responding to a conservation campaign sponsored by a large electric utility in the New England area and non-responding members of the general population. Self-administered mail questionnaires were used to gather the data in February 1981. A known probability sample was drawn from campaign respondents in Massachusetts and New Hampshire. A corresponding sample of the general population was drawn from matching zip code designations so that the within state geographic distribution of the samples correspond closely.

The response rate, calculated based on the number of usable responses received divided by the total mailing size less returns for bad addresses, was 75% for the conservation campaign respondents and 49% for the general population sample. Total sample size as the result of two complete mailings was N=707.

For the purpose of this analysis, where the population of interest is the entire population of the states, the two samples are weighted to reflect their relative occurrence. This adjustment is made on the basis of total response to the conservation campaign and its proportion to the total number of households in the area.

In an effort to assess the comparability of the weighted sample and the population, their relative distributions on age and income were examined. These two variables were selected on the basis of their relationship to family life cycle and possible impact on annual home heating costs. Comparison of the sample distribution on age with those of the New Hampshire and Massachusetts populations reveals almost identical distributions according to 1980 U.S. Census data. Examination of the relative income distributions using the 1979 Census projections for comparison shows the sample to be skewed somewhat higher as one might expect given the nature of the data collection process although the disparity is not severe.

Dependent Variable

Respondents were asked to review their financial records and report the total cost of their household heat for the previous twelve months. Use of wood, coal, electricity, and propane as supplemental sources of heat was also documented and included in total home heating expenditures.

Independent Variables

Family life cycle, income, and number of contributing incomes were the variables selected for examination based on research by Fritzsche. Additional demographic measures and information relating to dwelling size, heat sources, and thermostat settings were also collected as part of the overall study.

Since the Murphy and Staples conceptualization of family life cycle allowed for better categorization of the sample and showed the same predictive pattern as the Wells and Gubar treatment of family life cycle in Fritzsche's analysis, the more inclusive Murphy and Staples format is used here. Income is included because of its strong positive association with heating costs throughout the literature. Number of contributing incomes might be considered a surrogate for number of people at home during the day and hence might be expected to bear some relationship to heat expenditures, however it has no documentation to support it in the literature with the exception of Fritzsche's study which does not reference its relative predictive power. Number of contributing incomes is therefore included here for completeness to allow some comparison with its earlier usage.

In light of the hypothesis which suggests that conservation steps related to heat consumption should bear a relationship to heating expenditures, a second predictive model including a simple additive index of ten self-reported conservation measures is included as an independent variable in addition to family life cycle and income. It is recognized that the conservation behaviors examined in this study have different perceived relative economic costs and personal costs to consumers which might suggest the use of differential weighting in compiling the index. Lack of data on consumer utilities for-these measures, however, suggests that more detailed research is necessary in order to explore a more sophisticated index of this type.

Method of Analysis

In order to reconsider Fritzsche's analysis, an MCA using family life cycle, income, and number of contributing incomes -to predict household heating expenditures was run. In addition to evaluating relative predictor importance, family life cycle category means on heating costs, adjusted for income and number of contributing incomes, are provided. A second predictive model incorporating the conservation index is also examined using MCA.

FINDINGS

Influence of Family Life Cycle on Heat Expenditures

Results of the MCA for the model including family life cycle, income, and number of contributing incomes as explanators of heat expenditures reveal an adjusted R2=.15. As such, this value is not particularly noteworthy, but the model provides an important basis for comparison in evaluating the relative influence of the independent variables on consumer heat expenditure patterns.

The beta coefficients of the MCA indicate the relative predictive power of each independent variable when the remaining predictors are held constant. As shown in Table 1, family life cycle has a beta of .34 compared to .32 for income and .07 for number of contributing incomes. Both family life cycle and income are statistically significant at the .01 level while number of contributing incomes fails to be significant at the .05 level.

TABLE 1

MEASURES OF RELATIVE PREDICTOR IMPORTANCE - MULTIPLE CLASSIFICATION ANALYSIS

Beyond the relative strength of predictors, a measure of marginal predictive power is also of interest in evaluating the importance of specific explanators. The squared part correlation indicates the variance in the dependent variable (household heat costs) that can be marginally explained by each predictor relative to the total variance in household heat costs. As shown in Table 1, family life cycle constitutes 8 percent of the total variance; income, 7 percent. In contrast, the number of contributing incomes, with less than one-half of one percent marginal explanatory power, explains little of the variance in household heat costs.

In summary, both the beta coefficients and the squared part correlations illustrate that family life cycle is an important predictor of consumers' heat expenditures in support of H1. Income is of somewhat less importance in predicting these costs. Despite its inclusion in previous research by Fritzsche, the number of contributing incomes in a household adds little to the understanding of heat expenditure patterns.

Heating Expenditure Patterns Across Family Life Cycle

The average household expenditure for heat across the entire sample was $875.64 for the year 1980. Deviations from the grand mean for each family life cycle group appear in Table 2. In contrast to Fritzsche's conclusion that household energy costs peak during the full nest or middle stages of the family life cycle, this study of heating expenditures reveals a different pattern.

The deviations across family life cycle stages-shown in Table 2, as generated by the MCA, are adjusted for the effects of income and number of contributing incomes as in the Fritzsche study. The prevailing pattern indicates a direct positive relationship between age of head of household and heating expenditures rather than the curvilinear pattern noted by Fritzsche which seemed dominated by the influence of family size. The categories which Fritzsche found to have the highest total household energy costs were young and middle-age married couples with children. In this case these two groups appear to be well below the grand mean when considering heating costs (-$87.13 and -$57.83 respectively). Older couples (+$238.23), older singles (+$232.29), and households with children headed by middle-age non-married adults (+780.43) report the highest expenditures above the mean.

TABLE 2

HEATING EXPENDITURES AND RELATED CONSERVATION EFFORTS ACROSS FAMILY LIFE CYCLE

The reader should be cautioned that this latter group, which anchors the high-end of the distribution with heating costs $780.43 above the grand mean, may exhibit unstable estimates because of a small cell size. Similarly, the low-end of the distribution at $444.27 below the grand mean for household; with no children headed by young, widowed-separated-divorced adults also exhibits a critically small cell size as does the group of young, widowed-separated-divorced individuals with children (+$85.75). It should be noted that while unstable estimates for these groups may result. the inclusion of these groups in the MCA does not effect the stability of the other category estimates and the reported trend of increasing costs associated with increasing age still holds. This is not to say that households will necessarily use more energy as family life cycle advances, which would be making a time series conclusion from cross-sectional data, but rather that these older consumers appear to have very different heating needs from other households surveyed.

The results of the MCA indicate that the differential pattern of heat costs across family life cycle as suggested in H2 appears to be present. The predicted magnitude of expenditures as stated in H2a and H2b based on Fritzsche's work, however, cannot be supported by these data.

Influence of Conservation Efforts on Heat Expenditures

Rather than trying to draw policy conclusions or make marketing strategy recommendations to suppliers of energy conservation products based solely on a study of household expenditures, it is important to consider what action consumers may already be taking to conserve or restrict heat usage. Using the index of self-reported conservation behaviors as a measure of consumer action, the relationship with heat cost is examined using MCA.

The marginal explanatory power of the conservation index, when included as a predictor in the MCA, was .05 resulting in an improved adjusted R2=.20. The relative betas for each independent variable were found to be similar to those in the first model, as shown in Table 1. The conservation index, which was not included in the first model, has a beta value of .26. As such, conservation races third in importance after family life cycle and income in its ability to explain household heating costs. All three are significant at the .01 level. Further refinement of index components and index construction should lead to improved predictive power as well as to greater understanding of both consumption and conservation.

Conservation-Related Behavior Across Family Life Cycle

The mean on the ten-item conservation index for the entire sample is 5.5. Deviations from the grand mean for each family life cycle group were reported earlier in Table 2 in conjunction with heating expenditures. Comparison of relative deviations shows an interesting pattern. Somewhat surprisingly, the differences among groups in reported conservation behaviors are not large when compared with the range of reported heat costs. As shown in Table 2, four groups (I,III,VI,X), however, stand out as being below the grand mean. Group I, young, single people with no children, exhibits a conservation score that averages 1.1 activities below the grand mean. Group III, young, widowed-separated-divorced people with no children, also averages 1.1 activities below the grand mean on the conservation index. Middle-age, non-married people with no children (Group VI) score 0.3 below the grand mean whereas the average for older, non-married people with no children (Group X) is 1.0 below the grand mean. Interestingly, the common denominator among these four groups is that they represent single person households.

The data offer general support for H3 that differences in levels of conservation behavior occur across family life cycle stages. While there is some support for H3a that number of conservation steps taken will be highest for households in middle stages of the life cycle with children, results are far from conclusive. Households that conserve the least appear to be best characterized by their single person status rather than their age in contradiction of H3b.

Table 3 presents the percentage participation on each of the ten conservation index components across the total sample. Deviations from the grand mean are shown for each of the family life cycle groups. Inspection of the individual conservation items for each family life cycle group reveals that chose living alone are below the sample grand mean on almost every item. Of these four groups (I,III,VI,X), an exception of interest is the percentage of middle-age and older singles who have had energy audits for heating efficiency performed by utilities or specialized service companies. Compared with a very low grand mean of 6.7%, positive deviations of 4.9% and 4.6% respectively seem curiously high. This may indicate a propensity for future investment in energy saving products or changing behavior patterns.

TABLE 3

PERCENTAGE DEVIATIONS FOR INDIVIDUAL CONSERVATION BEHAVIORS RELATED TO HEAT USAGE

Examination of heat conservation patterns across the entire sample indicates that people are responding to appeals for conservation on a selective basis. Six of the ten examples of conservation behavior or investment in products to save energy show greater than 50% of the sample reporting compliance. The steps most likely to be taken to conserve appear to be those which require minimal investment such as turning the thermostat down when no one is at home, closing drapes and shades at sunset, adding weatherstripping around windows and doors, adding heating bills for an annual analysis of expenditures, cleaning the furnace, and closing-off unused rooms.

The findings reported here are similar to those of a Department of Energy survey (1980). In that survey, more than two-thirds of the households adding conservation-related equipment or insulation in 1977 and 1978 added inexpensive materials such as caulking, plastic covering, and weatherstripping. Only about 10% of the households surveyed by DOE made relatively expensive additions such as attic insulation, heat pumps, and new furnaces in each of those years. Similarly, far and away the most widespread method of reducing energy consumption is simply turning down the thermostat (Forbes 1982). The data reported here reflected that national pattern with approximately 80: of the sample indicating that they lower the thermostat when away from home.

IMPLICATIONS

The purpose of this section is to examine the relation ship between heating expenditures and conservation behavior in order to better understand implications for marketers and policy makers. Segmentation opportunities for conservation products and programs are discussed.

The relative deviations of heating expenditures and conservation behaviors are shown in the Figure. Households in the middle stages of the family life cycle where one might expect family size and dwelling size to peak show conservation behaviors above the mean and heat cost very close to the mean. Early stages of the life cycle which exhibit little interest in conservation also seem to have low expenditures. This would be consistent with small size dwellings, rental units, and multiple family dwellings which would improve heating efficiency and reduce conservation options related to insulation and furnace modifications.

The deviations from heat cost and conservation index means is cause for concern in life cycle Groups IX,X,XI. Results for Group IX may be unstable due to small size as mentioned earlier, however, Groups X and XI exhibit a consistently high consumption level. Heating costs nearly 25% greater than the mean are reported yet the number of conservation steps taken is not radically different from other life cycle groups.

Further exploration of the data reveals that Groups X and XI where the head of household is over 65 show no significant difference in dwelling size compared to other childless households but do report significantly higher thermostat settings. While this may not explain all of the increase in heat cost, it is a distinction with important implications.

With an apparent greater physical need for warmer temperatures exhibited by Groups X and XI, one might suggest efforts to increase heating efficiency through investment in insulation or a more economical heating system. Since these two groups reported the lowest median incomes of any of the family life cycle stages, however, such investment is not likely to be freely made unless extensive personal savings can be relied upon.

FIGURE

RELATIVE HEATING EXPENDITURES AND CONSERVATION BEHAVIORS ACROSS FAMILY LIFE CYCLE

Operating under such constraints these individuals may be unable to break their present consumption pattern without some form of public assistance which is not likely to be forthcoming at this time. As a result, while marketers can identify a target with substantial need, its potential buying power is limited.

Single person households under age 65 represent a market with relatively high disposable income, yet these individuals have taken few conservation steps. They may represent a market segment receptive to products which require low-involvement or little personal commitment. Opportunities may- exist for marketing products such as automatic thermostat controls which would have little impact on individual life-style but have a positive economic impact.

Those who are currently conserving at rates above the mean, Groups IV,V,VIII, and IX, still represent a viable opportunity for further energy consumption reductions. As Curtin noted, past conservation of energy was associated with a more favorable predisposition toward further conservation efforts. As long as such steps have not adversely affected people's sense of well-being, Curtin hypothesizes they should remain open to other conservation options. This suggests that satisfaction, not self-sacrifice, should be an effective theme for marketing conservation programs and products.

CONCLUSIONS

On the basis of the evidence presented, the understanding of consumer expenditures for home heating is enhanced by the inclusion of the family life cycle construct. While only a very limited model is examined here, one of the purposes of this paper is to compare energy consumption and conservation patterns. Greater emphasis on prediction of heat consumption would require a more expansive set of variables including characteristics of the residential structure. Family life cycle stages, however, reflect different household needs, priorities, and activities which can help to explain household heat expenditures.

The data analyzed in this study reveal a heat expenditure pattern across family life cycle unlike total energy consumption previously noted in the literature. The strong association between increases in heat cost and age suggests different directions for policy makers and marketers than does the curvilinear pattern where total energy costs peak with family size (Fritzsche 1981). Analyzing energy consumption within the specific context of its usage, such as home heating, clarifies the nature of the consumption process and aids in the understanding of consumer motivations and expected benefits from consumption. By expanding the analysis of energy consumption patterns to include corresponding levels of energy conservation behavior, researchers gain additional insight into opportunities for promoting behavioral change.

REFERENCES

Business Week (1981), "Energy Conservation: Spawning a Billion Dollar Business." 6 April. 58-61.

Curtin, Richard T. (1976), "Consumer Adaptation to Energy Shortages," Journal of Energy and Development, 2, 38-59.

Duvall, Evelyn M. (1971), Family Development, 4th. ed., Philadelphia: J.B. Lippincott Company, pp. 106-32.

Forbes (1982), "The Great Conservation Fallacy," 10 May, 156-7.

Fritzsche, David J. (1981), "An Analysis of Energy Consumption Patterns by Stage of Family Life Cycle,'; Journal of Marketing Research, 18, 227-32.

McDougall, Gordon, H.G., Claxton, John D., and Ritchie, J.R. Brent (1980), "Social Marketing: The Case of Energy Conservation in Canada," Waterloo, Ontario: Wilfrid Laurier University, Research Paper Series No. 1980.

Morrison, Bonnie Maas, and Gladhart, Peter Michael (1976), "Energy and Families: The Crisis and the Response," Journal of Home Economics, 68, 15-18.

Murphy, Patrick E., and Staples, William A. (1979), "A Modernized Family Life Cycle," Journal of Consumer Research, 6, 12-22.

Norton, Arthur J. (1974), "The Family-Life Cycle Updated: Components and Uses," in Selected Studies in Marriage and the FamilY, eds. Robert F. Winch and Graham B. Spanier, New York: Holt, Rinehart and Winston, pp. 162-7.

U.S. Department of Energy, Residential Energy Consumption Survey: Conservation, DOE/EIA-0207/3 (1980).

Wells, William C., and Gubar, George (1966), "Life Cycle Concept in Marketing Research," Journal of Marketing Research, 3, 355-63.

Zimmerman, Carol A. (1981), "Household Travel Patterns by Life Cycle Stage," in Consumers and Energy Conservation: International Perspectives on Research and Policy Options, eds. John D. Claxton, C. Dennis Anderson, J.R. Brent Ritchie, and Gordon H.G. McDougall, New York: Praeger, pp. 81-95.

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