Case 3.2 The Demographic Discovery of the New Millennium 1.
HMOs HMOs must must fir first st und under erst stan and d the the want wantss and and need needss of the the matu mature re seg segme ment nt.. Thus, Thus, usi using ng marketing research to uncover these wants is required. Exploratory research can be used to gain insights into the differences between the segments, which can then be tested by descriptive research methods. HMOs should obtain information from the mature market on their lifestyles, demographics, attitudes towards medical care and preventative medicine, past usage of medical services (especially in the recent past), knowledge of HMOs, and psychographics. Differences in the health care needs of the segments can be found using a one-way ANOVA design. Each segment can be considered a separate treatment and differences across illnesses, accidents, size of bill, etc., can be measured. Discriminant analysis can be used to identify the variables that best discriminate the four segments.
2.
Firs Firstt it shou should ld be be note noted d that that the the stu study dy has has a limi limita tati tion. on. The The sam sampl plee cons consis iste ted d only only of of women who did not work. Thus, there were many other elderly consumers who were not considered and the results should not be extended to all elderly consumers. Given this limitation, the research strategy is appropriate. Clustering the respondents on advertising items results in relatively homogeneous groups that can be interpreted as segments of the sample. From these segments the differences in psychographics can be assessed. Also, the factor analysis results allow us to determine a psychographic profile of each cluster.
3.
Engaged —Relatively negative attitude towards advertising, conceiving it to be biased
and untruthful at times, nevertheless, the information provided by ads is helpful. They rely on friends more than other groups. Autonomous —A somewhat negative attitude towards advertising, conceiving it to be
biased and untruthful at times and relatively unhelpful in providing information affecting
105
their purchases. They tend to be very independent in their consumption behavior, not consulting friends. Receptive —A relatively positive attitude towards advertising, believing that it helps them
in gathering information and basing most of their decision to buy on advertisements, not friend’s opinions. Like the other two groups, they are also skeptical of comparative advertising, but less so. 4.
Factor 1—Fashionable; Factor 2—Traditional Mother; Factor 3—Concerned Mother; Factor 4—Skeptics of Big Business Note that students may label the factors differently, but they should reflect the same general theme.
5.
Discriminant analysis would be helpful in determining the factors that most influence each of the segments and whether a significant difference exists among the groups in terms of the factors. The factor analysis results provide no insight into these questions. Discriminant analysis can be conducted by using the factors as the predictor variables and the clusters as the criterion variables. This will yield discriminant functions that will indicate which factors best discriminate between the clusters.
6.
Instead of using cluster analysis, factor analysis could have been used. Factor analysis is commonly used to elicit underlying psychological constructs based on the correlations in the data, whereas cluster analysis groups subjects on less abstract factors based on distance heuristics. A factor analysis of the 200 AIO items would produce X factors representing the underlying lifestyle dimensions of the elderly ladies. Factor scores for each subject could then be calculated and cluster analysis of the factor scores run. This would produce Y clusters of elderly lifestyles. To relate these clusters to attitude towards advertising, analysis of variance is used. Each of the four attitude items is used as the dependent variable, and the Y lifestyle clusters are used as independent variables. Differences in mean values on each attitude statement can then be assessed. Multiple discriminant analysis can be conducted by using the factors as the predictor variables and the clusters as the criterion variables. A similar analysis can also be run on the overall attitude towards advertising by summing the scores for each of the four items and repeating the analysis.
106
