No abstract

No abstract

88 officers in the U. S. Navy were asked to state how much they would like each of 14 political personalities elected President of the United States. These respondents were also asked to rate each pairing of the political candidates for how similar-dissimilar they were in general appeal to them. Multidimensional scalings conducted with the similarity data yielded two dimensions. The cross-validated multiple correlation between the set of Presidential preferences and the two sets of coordinates from the multidimensional scaling was .808.

In the comparison of military units or systems, many attributes, including performance, might be measured.One general approach to determining which system is best involves forming a composite measure of the differentially weighted component measures.It is a fairly common practice to control the component means and variances; it is less common, and more difficult, to control component covariances and, thus, to control the contribution of a component to the variance of the composite measure (variable).This paper presents a method leading to a computer procedure for assigning equal or differential influences on the composite by the use of component means, standard deviations, and covariances. Two numerical examples are given herein, one illustrating equal and the other unequal weighting. The technique appears to be both fast and precise.

(Continue on reverae aide If necaaaary and Identify by block number)Multidimensional scaling, cluster analysis and factor scoring techniques were used to study personnel data collected at the Naval Supply Center, San Diego.The results indicate that these techniques can be used to gain insight into the management process of an organization.