Social Marketing and Consumers' Preferences For Social Consumption


James R. Bettman and Robert B. Andrews (1972) ,"Social Marketing and Consumers' Preferences For Social Consumption", in SV - Proceedings of the Third Annual Conference of the Association for Consumer Research, eds. M. Venkatesan, Chicago, IL : Association for Consumer Research, Pages: 679-688.

Proceedings of the Third Annual Conference of the Association for Consumer Research, 1972      Pages 679-688


James R. Bettman, University of California, Los Angeles

Robert B. Andrews, University of California, Los Angeles

Social consumption refers to the process of goal choice and resource use at the national level. That is, as a nation we collectively set priorities for use of our scarce resources in pursuing various areas of concern such as national defense, education, health care, pollution control, and so on. This choice process has an analogue in product choices for personal or family consumption at the individual level.

There has been much study of the marketing function as it impinges upon and influences these individual product choices. Recently there also has been rising interest in the processes of social marketing and influence, through which attempts are made to influence desired social consumption patterns of citizens. Kotler and Zaltman, in a thoughtful study of social marketing, define the concept as follows (Kotler and Zaltman, 1971, p. 5):

Social marketing is the design, implementation, and control of programs calculated to influence the acceptability of social ideas and involving considerations of product planning, pricing, communication, distribution, and marketing research.

In carrying out social marketing to facilitate planned social change, two broad issues arise. First, the marketing concept, as applied to social consumption, implies that the needs and wants of citizens for societal resource use should be ascertained, to "restore consumer sovereignty in the determination of the society's product mix and the use of national resources." (Kotler and Zaltman, 1971, p. 5). The means for implementing this concept, however, are not clear. Second, questions concerning the ethics of social marketing arise. How should social marketing be carried out so that it is not inappropriately manipulative? How should the process of planning for social change, which must precede social marketing, be designed?

The present study considers the type of information and kinds of processes that would be needed to attack these issues. Little research has been done on determining the relative importance that individual citizens hold for various national or social goals. This study shows how such social consumption preference information can be gathered and presents empirical findings for national and health care goals. These goal weighings are analyzed to determine the effects of age, race, and education on such preferences. After describing the empirical results, the potential uses of this kind of data are discussed. The discussion centers on the two issues discussed above: (l) How data on preferences for social consumption can be used to improve the effectiveness of social marketing by implementing the marketing concept and by defining specific target markets; and (2) How 'ethical' social planning and marketing processes can be designed and what role data on preferences for social consumption can play. Here many ideas proposed for planning and legislation in the area of environmental protection are useful.


[This study is based upon data collected under U.S.P.H.S. Grant 95-505 for training in Comprehensive Health Planning.]

To measure preferences for social consumption, particular societal goal areas must be designated. For the present study, data were collected for two sets of five goal areas. The first set dealt with national goals: 1) Reduce air pollution (POLL)[The abbreviations listed after each goal are used in the data summary tables.]; 2) Slow Down Inflation (INFL); 3) Improve Racial and Ethnic Relations (RACE); 4) Decrease Crime (CRME); 5) Improve Health Care Services (HLTH). The second set was goals within health care: 1) Build More Medical School Facilities to Train More Doctors (FACL); 2) Increase Medical Research (RESR); 3) Require All Voluntary Private Health Insurance Plans to Pay for All Health Services (PRVI); 4) Make It Easy to Get Medical Treatment at Night and on Weekends (NTWK); 5) Provide Government Health Insurance for All Citizens to Pay for All Health Services (GOVI). Data on age, race, and years of education of each respondent were also obtained.

The data were collected by the Survey Research Center at UCLA as part of a large survey and comprised completed interviews from 964 households in the Los Angeles metropolitan area. The sample was a probability sample with marginal constraints on median income, percent black, and geographic area. The data were collected from November 1970-January 1971. Because of refusals and other factors for particular stimulus pairs, 787 complete sets of scale values for the national goals were obtained, and 855 sets for the medical goals.

To obtain the preference measures, data were collected using paired comparisons with an intensity rating. For the five goal stimuli in each set, ten ratings were collected from each person; one for each distinct pair of goals. For an ordered pair of goals (i, j), the individual chose which of the pair he believed was a more important goal, and then rated how much more important that goal was to him or her on a ten point scale, where zero was indifference and nine was much more important. Denote this rating for the pair of goals (i, j) by dij. We use the sign conventions dii > 0 if i is more important than j, dij = 8 if indifference, and dij < 0 if J is more important than i. Finally, since data on only the distinct pairs of goals are collected, we set dji = -dij

The preference value for each goal held by an individual is obtained simply. For goal i, the scale or preference value Pi is given by Pi = Ej=i dij/5. Note that by their definition, the Pi sum to zero. Thus we have a measure of the relative preference of an individual for each goal.

The data on preferences for the social goals in each set were then analyzed using multivariate analysis of variance (MANOVA) [Analyses were performed using the program Multivariance, written by Professor Jeremy D. Finn, State University of New York at Buffalo.] with three factors: age (5 levels), race (3 levels), and education (3 levels). The levels for each factor are shown in Table 1. In MANOVA, the dependent variable is a vector of values rather than a single number, as in ANOVA. In this case, the values for each goal for each person are the entries in the dependent vector. However, to perform this analysis, one goal preference must be deleted from each set, since the preference scale values sum to zero. Among national goals Slow Down Inflation was deleted and among health goals Private Health Insurance was deleted. These stimuli were chosen for two reasons: relatively small variation among the factor subgroups, and a priori empirical interest of the investigators. Discriminant analyses were performed using the sums of squares for each hypothesis as an aid to interpretation.





The mean ratings for the various groups of respondents are shown in Tables 1 and 2. The MANOVA for the preferences for the national goals showed no significant interactions, and the following significant main effects: linear trends in age (p<.0001); and education (p<.0001); a quadratic trend in education (p=.052) and a main effect for race (p<.0001). For age, the results of the discriminant analysis (not presented here) show that the most important components of the linear trend effect were increasing concern with crime and with health, and decreasing choice weighting for racial relations as age increases.

For race, two independent discriminant functions were statistically significant, thus showing a two-dimensional discrimination among the racial groups. For the first dimension, health care is the most important variable for discriminating among the races, and pollution is most important for the second dimension. By examining Table 1 the decreased concern with pollution among blacks and the decreased concern among whites for health care can be seen.

Finally, for education the most important components for the linear trend were shown by discriminant analysis to be decreasing concerns for crime and health care and increasing concern for racial relations with increasing education. The quadratic trend in education was influenced largely by the U-shaped relationships for pollution as education increases.

The MANOVA for health care goals showed the following significant effects: race-education interaction (p=.028); linear trend in age (p=.043); linear trend in race (p<.0001); linear trend in education (p<.0001); and quadratic trend in education (p=.032). The major factors in the interaction, according to the discriminant analysis, involved the goals of research and of government insurance. For research, the averages (not presented here) show both whites and blacks have increasing concern as education increases, but the preferences exhibited by Mexican-Americans first increase and then decrease. For government insurance, both whites and Mexican-Americans showed decreasing concern as education increases, whereas blacks first increase and then decrease.

Since there was a significant interaction, main effects must be interpreted with caution. We examined only race and the linear trend in education. For race, only one discriminant function was significant, showing the goals of research and of night and weekend care to be most important for discrimination among the racial groups. For the linear trend in education, the discriminant analysis showed the decrease in concern for access on nights and weekends as education increases to be most important, but the other stimuli also contributed moderately.

The findings themselves are of interest as empirical results. For example, crime was the greatest concern of each racial group. Following crime, however, the differing concerns of racial groups for national goals were shown. Blacks and Mexican-Americans were concerned with health care, whereas whites were not, perhaps reflecting the barriers to care presented by lack of affluence and knowledge of the system. Surprisingly, blacks were the only group concerned with racial relations. For health care goals, one result was the increase of concern for facilities and research and the decrease of concern for access on nights and weekends and government insurance with education. As education level increased, individuals placed higher priority on system problems and less on their personal problems within the system. The overall interest in medical research is surprising. Note the concern of the elderly, Mexican-Americans and poorly educated with government health insurance.

The analysis of these preferences for social consumption is interesting in its own right. However, the major thrust of this paper is the inputs such measurements can make to the social marketing process, and in particular to the two major issues discussed in the introduction. Now that it has been shown that such measures are feasible and relatively easy to obtain, the uses of such data for social marketing will be discussed.


The Marketing Concept in Social Marketing

To apply the marketing concept to social goals, preference ratings of citizens for alternative goals are necessary. As in private markets, the design of the product or service requires knowledge of consumer desires and goals. The measures described above attempt to ascertain such needs and desires for social goals. These means can serve to direct the social marketer to use societal resources toward the achievement of those goals most valued by the citizen consumers. For example, if any social marketing effort for health care programs were to be mounted, the health care priority measures given above could be quite useful. In particular, the somewhat surprising results on the relative ranking of national health insurance and of allocation of funds to research would be of great importance if consumer sovereignty is to be maintained.

There is a further use of social consumPtion preference data that aids making social marketing more effective that is related to the idea of implementing the marketing concept. If the social marketer determines what citizens' desires are, and if he designs a 'social product' or program that is consistent with these desires, it will be easier to market that program. As Kotler and Zaltman (1971) point out, this is the essence of Lazarsfeld and Merton's (1949) idea of canalization. There is an existing base of preferences and attitudes upon which the social marketer can build. Since social marketing often deals with 'products' that affect areas where deep-rooted attitudes are found (e.g., environmental issues), it is even more important for the social marketer to have some knowledge of his target market's preference structure. Wiebe's (1952) idea of force, involving the person's predisposition toward a goal, also implies that knowledge of preferences for social consumption can lead to more effective social marketing.

Finally, analyses of the differences in the preference patterns of different subgroupings within the population of interest, such as those performed above for age, race, and education, can help the social marketer gain insights into the value structures of different target market segments. For example, the data show that blacks and Mexican-Americans value certain health care goals differently (e.g., national health insurance and research). Thus, to be effective, social marketing programs dealing with minority health care problems would have to take these differences into account before implementing any programs. However, note that the definition of target segments is not this simple. As the race-education interaction for health care goals shows, the simplistic one-variable characterizations of interest groups often used by social planners is not sufficient. Groups must be defined by more complex patterns of variables rather than by one variable alone (e.g., by race, income, age, and education rather than race alone). [This type of result has also been found in a different context of attempting to relate personality variables to product use behavior (Sparks and Tucker, 1971).]

An Ethic of Social Marketing

Data on preferences for societal consumption can also be useful in designing an "ethical" social marketing process. The problem that arises in social marketing is the possibility that citizens may be 'sold' social programs that they do not really want. Resources may not be allocated according to their desires. Kotler and Zaltman express this as a possibility of manipulation (Kotler and Zaltman, 1971, p. 11). How can a social planning and marketing process be designed that avoids this problem? An attempt will be made to sketch elements that such a process should contain, and then these elements will be integrated into a process design. Finally, current progress toward this type of design will be assessed.

Elements necessary for an 'ethical' social planning and marketing process include: 1) open planning; 2) impact statements and evaluation of alternatives; 3) citizens' preferences for social consumption; 4) an agency independent of the change agency or social marketer that serves as a system entry point and information system. These elements will each be discussed in more detail.

The concept of open planning has been discussed by Borelli et. al. (1971) in the context of power plant siting decisions. Open planning entails a) involvement of citizen groups before major decisions are made, not after the point of no return for a project has been passed; b) open discussion and even solicitation of views, including extreme views; c) full disclosure of relevant information; d) compromise. Of course, in many social marketing arenas, the heated controversy developed in power plant siting decisions will not arise. However, the principles involved are important: (1) for social consumption patterns, public views must be heard, and (2) the social marketer must be open about his plans. Social marketing of family planning and birth control programs might be an instance where open planning would be particularly relevant.

Impact statements are assessments of the impact of a proposed program on various societal groups, the environment, health levels, etc., to the extent to which these impacts can be predicted. For any adverse impacts, alternative programs or subprograms should be described that would avoid the impact, if possible. Further, if these alternatives are not chosen, reasons for this must be given. The social marketer would be responsible for providing a detailed statement of this information for any major 'social product'. In essence, this information is an attempt to make the decision process of choice among alternatives an open decision process rather than an implicit or private one.

Preference ratings for social consumption represent a complement to, and not a substitute for, open planning. Open planning should generate some data on the views of the public. However, by its very nature, open planning is a forum of limited size, and hence gives a potentially biased view. A survey methodology such as that used in this study provides a mechanism for sampling the views of various interest groups in much more depth. Of course, such information sacrifices the implications that can possibly be derived by deeper probing in open planning. There is a tradeoff between breadth and depth. Finally, there is always the thorny question of the extent to which the public's priorities should govern in programs. That is, it may not always be appropriate to follow 'public wisdom'.

Kotler and Zaltman (1971, p. 10) described the change agency or social marketer as having a research unit that collects information to feed into a planning unit. It is the contention of the present study that there should be instead an agency independent of the change agency. This is for two reasons: (1) to avoid conflicts of interest with public interest groups and insure full disclosure of relevant information and (2) to serve as a highly visible entry point for interest groups into the system. Consumer groups are often frustrated, particularly in social marketing situations, by overlapping jurisdictions and complex governmental structure, so that the group does not know how to enter the system (this is often particularly true of health care and environmental problems). Having such an independent agency would serve Wiebe's functions of direction and mechanism (Wiebe, 1952). The independent agency would be responsible for gathering the preference data discussed above and also technical data needed to assess impacts.

The social planning and marketing process would then function as follows: the change agency would be continually apprised of social consumption preferences by the information agency. If the need for a new 'social product' to be socially marketed was discerned, open planning for this product would be instituted. Normally this would be initiated by the change agency. However, it is possible that the information agency, in its function as an entry point, could ask the change agency to open plan on a social product desired by an interest group that had entered the system in this manner. As plans progress, impact statements and alternatives would be proposed. At each step of the process, monitoring of the larger public's social consumption preferences would be used in conjunction with open planning inputs to guide further action.

Such a process certainly takes time and resources. However, we have entered a stage of societal awareness and concern where such processes for 'social products' may be necessary. It is "decision-making in the face of paradigm changes, when new fundamental assumptions enter the picture." (Borelli et. al., 1971, p. 19). Efficiency, even efficient social marketing, can no longer be the sole criterion. Pragmatically, however, resources are limited. Every social program probably cannot go through this extensive process. Only major programs should be accorded selectively this extensive consideration. Other social marketing endeavors could be dealt with in a more abbreviated fashion by using only the social consumption preferences supplied by the information agency. Finally, it would seem that there should be a single information agency or possibly one such agency for each of a few maj or problem areas.

What actual progress has been made in implementing such a system design in practice? Most progress in this area has been made in the area of environmental quality decisions. Let us take each of the system elements in turn. Open planning has been attempted in a few cases. Northern States Power Company in Minnesota developed a Citizens Advisory Task Force that essentially participated in open planning for a power plant siting decision (Borelli et. al., 1971, pp. 36-39). Impact statements are required for major Federal actions affecting environmental quality by the National Environmental Policy Act of 1969. There is to be "a detailed statement by the responsible official on:

(i) the environmental impact of the proposed action,




(iii) alternatives to the proposed action. . ."

[Pub. L. No. 91-190, Title I, ' 102 (C), 83 Stat. 853 (1970) codified at 42 U.S.C. ''4332 (1970)].

In addition, recent court cases [In particular, the 'Wildlife Preserves' case, Texas East. Trans. Corp. versus Wildlife Preserves, Inc., 48 N.J. 261, 225A. 2d 130, 137 (1966).] have changed burden-of-proof rules to essentially demand open decision making in environmental cases. The principle espoused has been well stated by Krier (1970, p. 115):

Once a reasonable showing is made that a proposed course of action poses a probable threat of significant environmental damage, the body desiring to initiate that action should be required to come forward with evidence on the likelihood of such damage, the unavailability or unfeasibility of alternatives. and the justification for its activities.

Flack and Summers report the use of priority data for water quality goals in resolution of conflict for water resources planning (Flack and Summers, 1971). They made ingenious use of interactive man-computer systems and computer graphics to resolve conflicts among decision makers. In particular, conflicts between those desiring a social marketing product and citizen representatives could be attacked.

With regard to the idea of an information agency and entry point, Krier reports that the California Environmental Quality Study Council performed many of the same functions during its first year of operation as those proposed for the information agency (Krier, 1971). The Council actually went even further and became nearly a public ombudsman and advocate for environmental affairs. Finally, the most extensive total system design that involves many of the ideas noted above is the proposed design of a system for the government of Puerto Rico involving citizen feedback, open planning, and an information agency (Chandler, 1970).

In conclusion, gathering data on citizens' preferences for social consumption has been shown to be both feasible, as evidenced by the empirical study of this paper, and highly central to social marketing. Implementing the marketing concept and designing an 'ethical', non-manipulative social marketing process needs such inputs. Since the proposed social planning and marketing process has used examples mainly drawn from the environmental quality arena, it may seem somewhat overdrawn for less heated social marketing areas. However, the principles espoused are just as relevant for such areas. Openness in decision making and acknowledging citizen preferences are necessities in the present and foreseeable social climate of public awareness and concern.


Borelli, Peter, Mahlon Easterling, Burton H. Klein, Lester Lees, Guy Pauker, & Robert Poppe. People, Power, and Pollution. Report Number 1, Environmental Quality Laboratory, California Institute of Technology, September 1971.

Flack, J. Ernest & David A. Summers. Computer Aided Conflict Resolution in Water Resource Planning: An Illustration. Water Resources Research, 1971, 7, 1410-1414.

Kotler, Philip & Gerald Zaltman. Social Marketing: An Approach to Planned Social Change. Journal of Marketing, 1971, 35, 3-12.

Krier, James E. Environmental Litigation and the Burden of Proof. Law and the Environment, 1970, 105-122.

Krier, James E. Environmental Watchdogs: Some Lessons from a 'Study' Council. Stanford Law Review, 1971, 23, 623-675.

Lazarsfeld, Paul F. & Robert K. Merton. Mass Communication, Popular Taste, and Organized Social Action. In Wilbur Schramm, (ed.), Mass Communications, second edition Urbana, Illinois: University of Illinois Press, 1960), 492-512.

Sparks, David L. & W. T. Tucker. A Multivariate Analysis of Personality and Product Use. Journal of Marketing Research, 1971, 8, 67-70.

Stevens, Chandler H. Science, Government, and Citizen Feedback. Operations Research, 1970, 18, 577-591.

Wiebe, G. D. Merchandising Commodities and Citizenship on Television. Public Opinion Quarterly, 1951-52, 15, 679-691.



James R. Bettman, University of California, Los Angeles
Robert B. Andrews, University of California, Los Angeles


SV - Proceedings of the Third Annual Conference of the Association for Consumer Research | 1972

Share Proceeding

Featured papers

See More


Situation Neglect Underlies Both Psychological Myopia and Psychological Hyperopia

Sarah Wei, University of Warwick
Christopher Hsee, University of Chicago, USA

Read More


C6. How Does Unsatisfied Curiosity Stir Our Craving For Food?

Chen Wang, Drexel University, USA

Read More


Exploring the Intersection of Digital Virtual Consumption and Family Rituals

Linda Tuncay Zayer, Loyola University Chicago, USA
Jenna Drenten, Loyola University Chicago, 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.