ABSTRACT
However the proximity measures between objects are obtained, once they have been obtained we may then employ multidimensional scaling to generate spatial configurations of the objects such that the order of distances between the objects in the spatial configurations corresponds to the proximity measures be tween the objects. To put this a little more precisely,
are better able to understand and interpret the important variations among the n objects. An example might help clarify this. Jones, Sensenig, and Ashmore (1974) asked subjects to sort 36 values into categories on the basis of what they be lieved about which values “go together.” For each possible pair of values, a proximity measure was derived, based on the number of times the two values were put into the same category. There are 630 such proximity measures. By employing one of the several available versions of multidimensional scaling (Kruskal, 1964a, 1964b), these 630 measures could be represented in a twodimensional space, which is much more understandable than the 36 x 36 matrix of proximity measures. See Figure 2.3.9
ods, or multidimensional scaling in detail sufficient to allow one to use these methods; that was not our purpose. Our purpose was to call attention to a grow ing armamentarium of methods available to the researcher interested in under standing how an individual construes and interprets and organizes reality. We were led into this brief discussion of methods by pointing out the growing recognition of the need for such methods within the field of person perception.
