Die evaluering en verfyning van meetinstrumente en die gebruik AMONA in die modellering van latente veranderlike strukture

Abstract

In the Social Sciences, and recently also in other Sciences, constructs (latent variables) are being measured. It is however, not possible to know in advance whether the questions being asked will accurately measure the said construct. The measuring instrument that is being used to measure a construct should therefore first be refined before it is used in a model. Unrefined (raw) measuring instruments can have a major influence on the results of model fitting. The objective of the study was to refine measuring instruments and subsequently to build a model which will best fit the data and will reflect the causal relationship between constructs. Specific attention was paid to refining the measuring instruments for job satisfaction, job involvement and career orientation. a Psychometrical model was eventually built, which reflected the causal relationship between the different constructs and the biographical factors. In this study the refinement of measuring instruments (verification of reliability and validity) and model building were investigated. It was evident that the methods for the refinement of measuring instruments, that is Cronbach's Alpha and exploratory factor analysis, maximised the reliability and validity of an instrument. The measuring instruments were consequently successfully refined. It was finally also evident that RAMONA is a very successful computer package for the building of models.