Relationships Between Consumers' Shopping and Leisure Activities and Their Attitudes Toward the Energy Crisis: a Cross Sectional Study

Citation:

Glenn S. Omura and W. Wayne Talarzyk (1975) ,"Relationships Between Consumers' Shopping and Leisure Activities and Their Attitudes Toward the Energy Crisis: a Cross Sectional Study", in NA - Advances in Consumer Research Volume 02, eds. Mary Jane Schlinger, Ann Abor, MI : Association for Consumer Research, Pages: 803-816.

Advances in Consumer Research Volume 2, 1975      Pages 803-816

RELATIONSHIPS BETWEEN CONSUMERS' SHOPPING AND LEISURE ACTIVITIES AND THEIR ATTITUDES TOWARD THE ENERGY CRISIS: A CROSS SECTIONAL STUDY

Glenn S. Omura, The Ohio State University

W. Wayne Talarzyk, The Ohio State University

[The authors gratefully acknowledge research support provided by the Fred B. and Mabel Dean Hill Fund and Consumer Mail Panels of Market Facts, Inc.]

[Glenn S. Omura is a Research Associate and Doctoral Candidate in the College of Administrative Science at The Ohio State University.]

[W. Wayne Talarzyk is Associate Professor of Marketing of the College of Administrative Science, The Ohio State University.]

Reported here is a descriptive analysis of a series of attitudinal questions relating the energy crisis to consumers' stated shopping patterns and leisure-time activities. The analysis focuses on consumer responses to key questions and those classification variables that best discriminate between those who stated their activities were affected or behavior changed and those who stated there were no effects or changes.

During the past twelve months a social issue known as the energy crisis has affected most Americans in many ways. This issue has provided an opportunity and, perhaps even more so, an obligation for consumer behavioralists to study the impact of a major change agent upon the attitudes, activities and behavioral patterns of large segments of consumers.

There are no easy answers to the multitude of changes and problems brought about by the energy shortage. In fact, even the questions that need to be explored are somewhat difficult to identify and isolate.

Therefore, while no single research project can bring all aspects of the energy crisis into perspective, hopefully such studies as the present collectively can provide a composite picture of what consumers think about the issue as well as how they have been affected by it. This paper is designed to fit into this composite framework by examining some statements related to shopping patterns, leisure activities and associated attitudinal questions.

OBJECTIVES

This paper reports on a portion of a larger data set on the energy crisis as overviewed in an earlier paper by the same authors (Talarzyk and Omura, 1974). Figure 1 illustrates the major components of the overall study.

It was reported in the earlier paper that one of the major objectives for the energy crisis research project was "to determine and evaluate what consumers think about the energy shortage and how they have been affected by lt." she data to be examined here further serve to meet the above objective by focusing on two areas of Block C in Figure 1, "Leisure Time' and

'Shopping Behavior," as they relate to "Attitudes Toward and Perceptions of the Energy Crisis" (Block B) and "Socio-Economic Variables" (Block D).

Basically the purpose here is to examine the relationships between consumers' attitudes toward the energy crisis and certain leisure-time activities and shopping patterns. It is also of interest to determine any differences between those consumers who said they were affected and those who said they were not affected by the energy shortage.

FIGURE 1

OVERVIEW OF THE ENERGY CRISIS RESEARCH PROJECT

Specifically, the objectives of this paper are to examine among what consumers and to what extent:

(1) changes in leisure time events were related to attitudes concerning the energy crisis

(2) variations in shopping patterns were related to attitudes concerning the energy crisis

The relationship among the above objectives is portrayed in Figure 2 by the explanatory variables and the statements used to measure the two research areas.

RESEARCH QUESTIONS

In the absence of a theoretical foundation to lend structure to this area of investigation, certain research questions were formulated. These questions, consistent with the prior stated objectives, involved analysis of responses to the four statements shown in Figure 2.

Agreement or disagreement with each of these four statements formed a set of dependent variables. Six areas of activities, interests, and opinions (AIO) statements served as independent variables along with selected socioeconomic characteristics. The six areas of AIO statements, developed to toe potentially related to the dependent variables included: attitudinal response to the energy shortage; expressed energy shortage affect on activities and other activities which may be affected by the shortage; responsibility for the energy shortage; rationing; economic effects; and the energy shortage as a source of harassment.

FIGURE 2

INTERRELATIONSHIPS BETWEEN THE THREE RESEARCH AREAS AND THE EXPLANATORY VARIABLES

With this background the research questions became:

What percentage of the sample -

- is satisfied with present leisure-time activities?

- has not had vacation plans greatly affected by the energy crisis?

- drives about the same amount of miles shopping as before the energy crisis?

- makes about the same number of shopping trips as before the energy crisis?

Which variables best discriminate between those who -

- are satisfied and those who are not satisfied with present leisure-time activities

- have had and those who have not had vacation plans greatly affected by the energy crisis?

- drive about the same and those who drive different amounts of miles shopping as before the energy crisis?

- make about the same and those who make different numbers of shopping trips as before the energy crisis?

METHODOLOGY

Data Collection Instrument

Data to evaluate the research questions were gathered through a series of AIO statements scaled on a five-point scale from strongly agree to strongly disagree. As can be seen in Table 1, presented later, both the dependent variables and the independent variables with the exception of the socio-economic variables were scaled identically. Forty-five statements comprised the total range of AIO statements. For a more elaborate description of the questionnaire and the sample base discussed in the next section see Talarzyk and Omura (1974).

Sample Base

The sample for this study involved 662 respondents, balanced to represent the United States population in terms of geographic dispersion, income, density of locale, degree of urbanization and age. The respondents completed a twelve-page questionnaire sent to 1000 members of the Consumer Mail Panel of Market Facts, Inc. The questionnaires were mailed to panel members during the first week of March, 1974.

Data Analysis

To fulfill the objectives and respond to the research questions of this paper, three basic routines were performed on the data: frequency count, chi square analysis, and step-wise multiple discriminant analyses. The frequency count provided an aggregated overview of all the AIO items taken individually and focused on the first set of research questions. The chi square analysis allowed a basic investigation of the second set of research questions through an analysis of each of the four dependent variables versus each independent variable separately.

The discriminant analyses provided a more comprehensive evaluation of the second set of research questions by developing a profile of the characteristics that best aided in distinguishing between the two groups of each dependent variable. In addition, these analyses yielded a model for the prediction of group membership given knowledge of a respondent's answers to the differentiating statements.

For the chi square and discriminant analyses, the dependent variables were separately collapsed to form an agree group (strongly agree and agree respondents) and a disagree group (disagree and strongly disagree respondents). The undecided or no opinion group was eliminated from further analysis. When a given dependent variable was being studied the other three dependent variables were allowed to enter the analysis as assumed independent variables.

Hold-out samples were used to test the efficacy of the derived discriminant functions. The individuals for each hold-out sample were randomly drawn from each respective group prior to analysis. Approximately 25% of each original group went to the hold-out samples.

TABLE 1

FREQUENCY RESPONSES TO ALL VARIABLES

Although it will be apparent from Table 1 that the sizes of the two groups belonging to each dependent variable were not exactly equal, neither were the groups greatly disparate. The greatest disparity was less than a ratio of three to one. Still, in order not to be misled by the somewhat unequal group sizes, the proportional chance criterion was calculated for each dependent variable.

FINDINGS

Frequency Counts

Table 1 shows the frequency counts for all of the AIO statements used in this analysis. For the first set of research questions the following findings may be drawn: (1) some 68% of all respondents indicated that they were satisfied with their present leisure time activities, 25% said they were not satisfied; (2) the energy crisis apparently greatly affected the vacation plans of 49% of the respondents, 38% said they were not affected; (3) a total of 56% of the sample said they drive about the same amount of miles shopping now as they did before the energy crisis, 40% said they have changed the amount of driving for shopping, and (4) some 54% of the respondents stated that they make about the same number of shopping trips as they made before the energy crisis, 42% said they have changed the number of shopping trips.

TABLE 2

RESULTS FROM CHI SQUARE AND DISCRIMINANT ANALYSES

Chi Square Pairwise Variable Tests

As a part of the analysis of the second area of research questions, chi square runs were made. The following paragraphs in connection with the information in Table 2 briefly describe the research findings in this area. [Because of space limitations the individual contingency tables are not shown here. Detailed results are available upon request from the authors.] The findings are discussed in terms of the general areas of investigation which were included in the study rather than each individual independent variable in order to conserve space.

Variable 1 (Satisfied with leisure activities) Eleven variables were determined to be significantly related to satisfaction with leisure activities. As seen in Table 2, each of the areas except responsibility for the energy shortage is found to be related to Variable 1 (read across the columns as Independent Variable 1, with the agree group as X1 and disagree group as Y1) through at least one of the area's subcomponents.

Variable 2 (Energy crisis not affecting vacation plans): Ten variables were found related to Variable 2 but not all of the dependent areas were represented. None of the attitudinal response variables were determined to be related to Variable 2, nor were any socioeconomic variables related.

Variable 3 (Miles shopping unchanged in past 6 months): Although seventeen variables were discovered to be related to Variable 3, the entire area of anticipation of economic conditions was unrelated.

Variable 4 (Number of shopping trips unchanged in past 6 months): Thirteen variables were found to be related to Variable 4 with the area of anticipation of economic conditions not related.

While the chi square analyses indicated that relationships existed between the two groups of variables in the sample under study, only a gross association could be stated. Information useful to a consumer behavioralist concerned with the energy shortage was obtained more concisely through discriminant analyses.

Profiles through the Multiple Discriminant Analyses

Before profiling the groups under each of the five dependent variables, it should be noted that while significant discriminating variables were uncovered, few of the coefficients were to any great extent different relative to the corresponding variable in the opposing group. Thus, while the written descriptions below tend to draw out relative differences, the actual coefficients of each of the two groups should be observed also in Table 2. The number in parentheses following each of the characteristics below refer to the corresponding variable number from which the descriptions were drawn.

Variable 1 (Satisfied with leisure activities)

Those more likely to be satisfied:

- reported their lifestyles to be more affected by the energy crisis   (12)

- reported less often that everyone should vacation away from home   (23)

- worked more on home projects  (24)

- held greater belief that the energy crisis would disappear with the lifting of the embargo  (28)

- were more optimistic about their personal future  (41)

- held less belief that the purchase limit was a hassling device   (44)

- older  (49)

The attitudinal response to the energy crisis and the rationing areas were not included as sources of discriminating characteristics under Variable 1.

Variable 2 (Energy crisis not affecting vacation plans)

Those less likely to be affected:

- more satisfied with their leisure-time activities  ( 1)

- drive same number of miles shopping presently as before the crisis   ( 3)

- held greater belief that responding to energy-saving pleas was patriotic   ( 5)

- reported their lifestyles to be less affected by the energy crisis   (12)

- worked less on home projects  (24)

- held greater belief that they would have more money to spend next year   (38)

- held less belief that the purchase limit was a hassling device   (44)

- less educated  (46)

- had higher income  (48)

The areas of responsibility for the energy shortage and rationing were not represented in terms of contributing significant discriminating characteristics to Variable 2.

Variable 3 (Miles shopping unchanged in 6 months)

Those who have not altered the number of miles shopping:

- less likely to have their vacation plans affected by the crisis  ( 2)

- reported taking the same number of shopping trips presently as before the crisis  ( 4)

- held less belief that responding to energy-saving pleas was patriotic   ( 5)

- reported themselves to be ignoring the energy crisis less  ( 8)

- held more belief that the country was recession-bound  (36)

- held less belief that prices would be higher next year  (37)

- less educated  (46)

The only areas which led to discriminating between those who have not altered the number of miles shopping and those who have, include attitudinal response to the energy shortage, economic effect, and socio-economic characteristics.

Variable 4 (Number of shopping trips unchanged in 6 months)

Those who have not altered the number of shopping trips:

- drive same number of miles shopping presently as before the crisis   ( 3)

- reported their lifestyles to be less affected by the energy crisis   (12)

- tried less to include many stores in one shopping trip  (13)

- held less belief that the country was recession-bound  (36)

- more educated  (46)

- older  (49)

Three areas contributed to discriminating between the two groups of this dependent variable: energy shortage affect on activities, economic effect, and socio-economic characteristics.

Validation of the Discriminant Functions

In order to determine how well the variables used in the discriminant functions discriminate, the proportional chance criterion Cpro was calculated. This criterion serves as an index of how well both groups are correctly identified (Morrison, 1969). The criterion Cpro. indicates the probability an individual may be correctly classified by chance. If the percentage of the total number correctly classified exceeds Cpro., then the model derived is better than what might have occurred through chance. Table 3 provides the critical figures.

TABLE 3

VALIDATION OF DISCRIMINANT FUNCTIONS

For both the original sample and the holdout sample, the derived discriminant functions showed high levels of correct classifications. The 65.16 correct prediction of those satisfied/unsatisfied with their leisure time activities and the 57.3% correct prediction of those reporting the energy shortage not affecting/affecting their vacation plans both exceeded their respective proportional chance criterion reasonably well. [For a comparison of how well discriminant models in another context predicted, see Ostlund (1974).] The last two groups, those reporting no change/change in miles shopping and shopping trips since the energy shortage were predicted correctly exceptionally well, 83.46 and 79.7%, respectively, far exceeding their respective Cpro.

DISCUSSION

The findings demonstrated that relationships indeed existed between consumer attitudes toward the energy shortage and certain leisure time activities and shopping patterns. The relationships were evaluated through the frequency analysis, chi square analysis, and discriminant analysis. In the frequency count, there was by no means unanimity in the direction of responses to the dependent variable statements, suggesting that consumers were differentially affected by or differentially responded to the energy shortage. The chi square analysis indicated certain general associations between the selected variables, while the discriminant analysis pinpointed characteristics of those reporting basic satisfaction with leisure time pursuits and the reported affect of the energy shortage on leisure activities and shopping patterns. The discriminant analysis indicated that the dichotomous groups could be differentiated, and even better, that membership in the groups could be successfully predicted on the basis of the discriminating variables found to be significant in the respective models.

In examining the profiles reported above, a number of interesting observations may be made. Those more likely to be satisfied with their leisure time activities reported their lifestyles to be more affected by the energy shortage, but apparently the satisfaction arose or remained because their leisure activities did not include a great deal of energy consumption. Also, a general degree of optimism was present.

The respondents who indicated that the energy shortage was not going to affect their vacation plans apparently did so because of financial well-being. This conclusion is suggested by the importance of these financial variables (38 and 48 ) in the discriminant model seen in Table 2. At the same time it is curious that these respondents held greater belief that responding to energy-saving pleas was patriotic.

Perhaps the explanation is that the life styles of this group do not include extensive traveling on vacations, or, rather, that they simply ignored the energy-saving pleas despite recognizance of the obvious patriotic response. If the latter explanation is more correct, and the complete profile tends to indicate this, patriotic or patriotic-types of pleas through the media apparently will not induce a large segment of people, 38% of this sample, to reduce energy consumption. Should the country face another energy crisis or suffer further deterioration in energy supply, further research appears necessary to determine the type of appeal that would be positively received and acted upon other than the past choice of patriotic-oriented pleas.

The group that indicated no change in the numbers of shopping miles driven and the group with no change in the number of shopping trips made exhibited the expected similar responses to the classification variables. As groups, in general, they reported less influence on their life styles by the energy shortages. There were some variations in certain independent variables (belief that the country was recession-bound) and socio-economic variables (age and education). But the differences were slight.

While this investigation has provided only an abbreviated examination of the relationships between consumer attitudes toward the energy shortage and certain activities, it will hopefully serve as part of a composite research effort, Such a larger effort is required for consumer behavioralists to fully understand the spectrum of interrelationships involving this contemporary social issue.

As part of on-going research, additional analyses of the larger data set from which this paper is based will aid in a further understanding of the impact of major social issues on consumers. The issues of the energy shortage and the presidential impeachment will be examined through socio-psychological theories of opinion leadership, innovativeness, social character, gregariousness, and social desirability.

REFERENCES

Morrison, D. G. On the interpretation of discriminant analysis. Journal of Marketing Research, 1969, 6, 156-163.

Ostlund, L. E. Perceived innovation attributes as predictors of innovativeness. Journal of Consumer Research, 1974, 1, 23-29.

Talarzyk, W. W. & Omura, G. S. Consumer attitudes toward and perceptions of the energy crisis. Proceedings, Fall Conference, American Marketing Association, 1974.

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