Traditionally, the EFA approach has been used in organizational behaviour and marketing research. Basically, EFA is designed for the situation where the relationships between the observed and latent (factors) variables are unascertained or uncertain and the approach proceeds in an exploratory mode to discover the underlying factors, thereby illustrating the relationships between the factors and the observed variables. However this approach has certain limitations. The foremost limitation of this approach lies in the fact that in EFA, items are assigned to those factors on which they load most substantially. Therefore, it is possible for an item to load to a significant extent on more than one factor and hence the distinctiveness/identity of the factors is affected. Furthermore, in pure EFA items are loaded on to a factor only statistically and not on any theoretical basis, thereby affecting the valid identity of the factors. And, finally the concept of unidimensionality (i.e. extent to which items on a factor constitute or govern one single construct) has not been taken care of in EFA approach (Ahire et al., 1996). In essence, EFA is particularly useful only in the absence of a sufficiently detailed theory about the relationships of the observed variables to the latent constructs.In contrast, the CFA approach overcomes the above mentioned limitations and addresses the situation wherein the researcher specifies a model a priori,and tests the hypothesis that a relationship between the observed and the latentvariables does in fact exist. In other words, the hypotheses that form the constraints are an integral part of the CFA technique. This is because theresearcher is aware of the number of factors that are required to explain the inter-correlations among the measured variables. Furthermore, he/she knows which observed variables are presumably reliable indicators of each of the factors, and which variables are not related to a factor. The postulated model draws its logic from research outputs and other theoretical perspectives, and if the researcher has a reasonably good idea about the likely number of factors to be found, and the variables that are expected to be highly influenced by a particular factor, it is more appropriate to use CFA rather than EFA (Bentler,1995). As TQS is at an advanced stage of research and in view of the increasing acceptance of the CFA approach in both marketing and organizationalbehaviour literatures, the present work chose to adopt the factor analysis in aconfirmatory fashion.Once a scale has been developed, its construct validity must be ensured sothat one can have confidence that explanations based on the proposed model reflect reality. Construct validity is broadly defined as the extent to which an operationalization measures the concept it is presumed to measure. In order to check for the goodness of the overall model fit, the following hypothesis has been formulated.
đang được dịch, vui lòng đợi..