A multi-dimensional study of differences in perceptions and preferences of social responsibilities of marketers

This study was undertaken in an effort to illustrate a potential method for examining the attitudes of different groups of people with respect to selected social and consumer issues facing marketers. Such a technique, if successful and communicable, could aid the understanding of differences in attitudes of marketers, critics, customers, and other members of society. Marketers would then be better able to respond to the demands of vocal critics and to the needs and desires of customers and other members of society. Fifteen issues were studied. In general, they were selected to represent the criticisms discussed in current literature. The subjects represented three groups: 19 housewives, 28 business (25 industrial and 3 retail), and 23 students. No group was selected to be representative of a larger population. Attitudes were studied with respect to perception and with respect to preferences. Techniques of multidimensional scaling were used for this examination. To provide the necessary data for the multidimensional scaling programs, the subjects were asked to complete two questionnaires. The first questionnaire contained ten pairs of semantic differentials. Subjects rated each of the fifteen issues on each pair of semantic differentials. Euclidean distances were calculated between each pair of issues for each individual and used as input for the perceptual scaling program. The second questionnaire asked the subjects to rank the fifteen issues in order of preference. This data was used as input to the preference scaling program. The perceptual scaling algorithm that was used was designed by J. D. Carroll and J. J. Chang of Bell Laboratories. This program, INDSCAL, developed an overall perceptual scaling (called a group stimulus space) for all subjects. Each subject, however, was assigned idosyncratic weights for each dimension. These weights, or saliences, indicate the degree that the dimension was used by the individual in developing his perceptual differences between the issues. The best scaling of the fifteen issues appeared to be in four dimensions. These were interpreted as follows: Dimension 1. This dimension is dominated by one issue, layoffs of workers in times of sales decline. Dimension 2, The second dimension ranges from product oriented issues to price and promotionally oriented issues. Dimension 3. This dimension ranges from promotionally oriented issues to product and price oriented issues. Dimension 4. The fourth dimension includes societal oriented issues on the negative side and consumer oriented issues on the positive side. When the scalings were recomputed in three dimensions for fourteen issues (without the layoff issue), the dimensions matched dimensions 2, 3, and 4 of the four dimensional scaling of fifteen issues. When the perceptual saliences were tested for overall differences between the groups (using multivariate methods) the hypothesis of significant differences between the groups was rejected at the .05 level. In other words, no overall significant difference between the groups was proved with respect to perception. However, analysis of the individual dimensions showed some apparent differences. In these "univariate" analyses, the following interpretations were made: —Housewives tend to lump product and price oriented issues together and compare these with promotional issues. —Students tend to perceive layoffs more closely to the other issues than do the other groups. —Students tend to attach more salience to the difference between promotionally oriented issues and the issues of price and product than do the others. Further, students tend to perceive the issues as price, product, and promotion issues more distinctly than the other groups. —Students tend to see less difference between issues on the basis of being societal or consumer issues than do the other groups. The first of these is significant at the .05 level using univariate t-tests (i.e., housewives do not perceive a difference between product and price oriented issues which is significantly different from other groups). Preferences were analyzed using Carroll and Chang's PREFMAP algorithm. This program allows four models of analysis. The simplest model, a vector representation of preferences, provided adequate representations. Using this model, a preference vector was fitted for each individual on the group stimulus space provided by the INSCAL model. The relative preferences for the dimensions can be interpreted by analyzing the projections of the vectors on the dimensional axes. The following interpretations were made from the preference vectors: —Housewives have less preference for action on product oriented issues than the other groups. —Businessmen have less preference for action on price oriented issues than the other subjects. —Students prefer actions on societal oriented issues while businessmen and housewives prefer action on consumer oriented issues. The last of these observations is significant at the .05 level using univariate t-tests (i. e., students prefer societal oriented issues significantly more than housewives and businessmen). When tested for overall differences in preference between the three groups (using multivariate methods) the hypothesis of overall significant differences between the groups was accepted at the .05 level. In other words, there were overall significant differences between the groups on the basis of preference. While the findings listed above are interesting and may point to true differences between the populations of students, housewives, and businessmen, the true value of this study is in the method which it presents. If some improvements are made in data collection procedures and in the computer programs, the method shown here can be used by marketers to evaluate the opinions of the various populations regarding issues relevant to their business. These evaluations should aid the marketers in formulating policy regarding social and consumer issues.