Answer to Question #137737 in Microeconomics for Susan Gerea

Question

2. List three ways in which we can determine a good model fit? (3 marks)

Answer 1

A good model fit is one with an ability reproduce data especially the variance-covariance matrix. A good-fitting model is one which is reasonably consistent with the data without necessarily requiring re-specification. Before the interpretation of the causal paths of a structural model, a good-fitting measurement model must be in place. A researcher must ensure that a model is good-fitting, which is the main reason for the computation of fit index. A good-fitting model is not determined by a valid model as many people have been claiming. A good-fitting model is determined when all the parameters are not significantly different from zero. However in some instances, models with statistically significant parameters might be from a poor fitting model. Models with poor discriminant validity, nonsensical results, and Heywood cases can be considered as good fitting models. A good fitting model might be obtained with the possibility of improving it through removal of specification error.