THE RELIABILITY OF THE SINGLE POPULATION APPLICATION OF KYST
Abstract
The present study was designed to demonstrate the reliability of KYST, a multidimensional scaling technique appropriate with psychological data. Multidimensional scaling techniques, used in a variety of disciplines employing multivariate data, have advantages over other techniques in that they do not require an underlying assumption of linearity of data and that they can visually display relationships among the data parsimoniously. A sample of employees in a large insurance company were asked to respond to a Likert-type questionnaire on opinions toward various aspects of work. Thirty subsets were constructed randomly from the 4,773 completed questionnaires. Each subset was independently analyzed using KYST. Results revealed a three-dimensional solution for each subset. Twenty-one, or 70% of the subsets, contained dimensions identified as Anxiety, Satisfaction, and Fairness. An additional seven, or 23.3% of the subsets, contained some of these dimensions. It was concluded that KYST is a reliable technique for use with these type of data, particularly in identifying the overall dimensions necessary to explain the data accurately. Recommendations were made for further research as well as for improving the available computer routine for KYST.
Subject Area
Educational psychology
Recommended Citation
DEL GAIZO, EDWARD RICHARD, "THE RELIABILITY OF THE SINGLE POPULATION APPLICATION OF KYST" (1983). ETD Collection for Fordham University. AAI8326167.
https://research.library.fordham.edu/dissertations/AAI8326167
