Vulnerability and the Process of Recovery for Households and Companies in Phang-Nga and Phuket Provinces in Thailand
5 Methods of analysis
5.1 Structural equation modelling As the primary analysis method, the structural equation modelling approach was chosen for this study, since it offers the opportunity to assess causal relations between theoretical constructs. The advantages and disadvantages of these ap- proaches will be discussed in the following chapter. Basically, the method is a combination of different statistical methods which integrates path analysis and factor analysis into a single model. In contrast to exploratory factor analysis, factors and causal relations are specified by the researcher in advance, guided by theoretical assumptions and tested on a specific dataset (ARBUCKLE 2006:45pp). The model is designed by the user in the form of a scheme representing the hy- pothetical relations between latent variables and between the manifest variables and their latent variables (WOLD 1982). In general, a structural equation model consists of an inner and an outer model (cp. Figure 20). The outer model specifies the relations between the con- structs or latent variables and their explaining indicators. The inner model de- scribes the relations between the latent variables. These can be divided into de- pendent latent variables respectively endogenous variables, and in independent latent variables respectively exogenous variables. In the following section, the latent variables will be named � (m=1… 1) and their related indicators x(m) j=1… j. Direction and power of the relations between the latent variables are represented by the path coefficients �(ml). The connections between latent varia- bles and their connected indicators are described as factor loading coefficients �(m)...
You are not authenticated to view the full text of this chapter or article.
This site requires a subscription or purchase to access the full text of books or journals.
Do you have any questions? Contact us.Or login to access all content.