Positioning Educational Programs with Nonmetric Data; Perceptual versus Preference Maps

Most of practical methods for constructing positioning maps use as dependent variables data which are measured either on interval scales with normal distribution or on ordinal scales. Furthermore, ordinal data considered for this purpose are scores obtained either with Likert type items or with pair-comparisons scales. However, there are practical situations when data that describe customers' perceptions do not fit into these frames. One of these situations takes place when stimuli are considered simultaneously and ranked on a single ordinal scale. In educational marketing settings, this is the case when students express their options for specializations within a college by using a one-dimensional scale, such as a hierarchy of preferences list. These rank type scales have at least two major sources for drawbacks. The first source acts during data collecting stage, when respondents do not have options to fully reflect the magnitude of differences in their preferences or to consider several dimensions when making stimuli comparisons. The second source is located at the statistical processing stage; rank type data will not satisfy the requirement of normality and may produce biased results that threaten the validity of final conclusions. Despite the lack of conformation to statistical requirements for MDS procedure, rank-type data can be successfully used for constructing positioning maps. This article is an analysis of methodological and managerial peculiarities that arise when using this type of data. The empirical research is based on observations over a real choice situation, and managerial considerations are made under an internal competition perspective.