Structural Equation Modeling Analysis

1588 Words7 Pages
Structural Equation Modeling (SEM): Hoyle (2014) states that SEM is a comprehensive statistical technique of testing hypotheses about the relations among observed and latent variables . SEM is a multivariate technique and is an advance method of general linear model. It helps the researcher in testing multiple regression equations. It makes a conceptual model, path diagram and regression to measure complex relatioships.The first step normally consists of developing a theoretical model. The collected data acts as an input and the method fits the data to the developed model. This is further evaluated in terms of model fit statistics. This study employs Confirmatory Factor Analysis (CFA). It is a type of SEM that deals specifically with measurement…show more content…
PLS SEM comprises of two models, namely the measurement model and the causal model. The following chart shows the measurement model. A measurment model is the chart, which specifies the constructs, the indicators and the relationship between constructs: FIGURE 31 MEASUREMENT MODEL From the diagram it can be seen that the model consists of three latent variables and forty manifest variables. The relationship between latent variables is represented in the chart with the help of arrows. Each arrow signifies the hypothesis, which the model tests. In the above diagram one can see three blue circles, which represent latent variables. The forty yellow rectangels represent the manifest variables. Latent Variables: The variables of interest in the model, which cannot be observed directly. These variables are also known as constructs/ factors/ hypothetical variables. Manifest Variables: These are the observed variables, which can be directly measured. These variables help in indirect measurement of the latent variables. These variables are known as indicators/ items. The latent and manifest variables in the study are categorized in the following table: TABLE 10 VARIABLES OF THE…show more content…
There is significant impact of attitude towards advertisement on purchase intention. 3. There is significant impact of attitude towards brand on purchase intention. Causal Model: The causal model given below, provides the graphical presentation of the relationship between the varibles: attitude towards advertisement, attitude towards brand and purchase intention: FIGURE 32 CAUSAL MODEL Evaluation of the Model: The causal model was evaluated based on specific parameters. The first parameter was to take a look at the factor loadings of all the manifest variables. Factor loading can be defined as regression slopes for predicting the indicators from the latent variables. They are shown on the arrows between the indicators and constructs. Heir and Sheshkin opine that the acceptable limit of factor loading is greater than 0.5. From the causal model it can be seen that the factor loading of some indicators is very low. The four indicators (namely ATTB 5.1, ATTB 5.2, ATTB 5.3 and PI 4 ) with factor loading less than 0.5 were removed. After removing these indicators the model was run again.Thus the new model after removing the select four indicators was as follows: FIGURE 33 REVISED

More about Structural Equation Modeling Analysis

Open Document