SPSS Analysis
Budget: $30 – $250 USD
-Generalised Method of Moments (GMM)
-Data coding will be used for the transformation of nominal data into statistical data. Descriptive statistics for the variables will be tabulated into means for the sample, standard deviation, mode, minimum and maximum values
-Diagnostic tests were applied to the above model before it was estimated. The Hausman (1978) test will be performed to decide on the most suitable estimation technique between the Fixed-Effect model and the Random-Effects model.Several estimation techniques will be used because the General Method of Moments uses all possible instruments that include Least Squares Dummy Variable (LSDV) corrected for Kiviet bias (Kiviet, 1995), the Generalized Least Squares (GLS) primarily as a means for rigorous testing (robustness). The data will be tested for heteroskedasticity, serial correlation and multicollinearity to prevent spurious results of the regression analysis. Heteroskedasticity is tested using the Breusch-Pagan test. To identify any multicollinearity amongst the variables through the use of a correlation matrix. To establish the nature of the correlation between the dependent and independent variables through the application of the pooled Ordinary of Least Squares (OLS), it will be applied to multiple regression. To investigate the validity of the instruments, Arellano & Bond (1991) have suggested the Sargan Test of over-identifying restrictions. This is to test the null hypothesis of the overall validity of the instruments used. It will be used to estimate the parameters and value of the GMM objective function. The Arellano–Bond test will be used to test autocorrelation, where there is a null hypothesis then there is no autocorrelation. Tests of AR(1) and AR(2) will be performed. The basic model will be transformed into the first differenced form, and then, the first differenced lag dependent variable is instrumented with its past levels
-Data coding will be used for the transformation of nominal data into statistical data. Descriptive statistics for the variables will be tabulated into means for the sample, standard deviation, mode, minimum and maximum values
-Diagnostic tests were applied to the above model before it was estimated. The Hausman (1978) test will be performed to decide on the most suitable estimation technique between the Fixed-Effect model and the Random-Effects model.Several estimation techniques will be used because the General Method of Moments uses all possible instruments that include Least Squares Dummy Variable (LSDV) corrected for Kiviet bias (Kiviet, 1995), the Generalized Least Squares (GLS) primarily as a means for rigorous testing (robustness). The data will be tested for heteroskedasticity, serial correlation and multicollinearity to prevent spurious results of the regression analysis. Heteroskedasticity is tested using the Breusch-Pagan test. To identify any multicollinearity amongst the variables through the use of a correlation matrix. To establish the nature of the correlation between the dependent and independent variables through the application of the pooled Ordinary of Least Squares (OLS), it will be applied to multiple regression. To investigate the validity of the instruments, Arellano & Bond (1991) have suggested the Sargan Test of over-identifying restrictions. This is to test the null hypothesis of the overall validity of the instruments used. It will be used to estimate the parameters and value of the GMM objective function. The Arellano–Bond test will be used to test autocorrelation, where there is a null hypothesis then there is no autocorrelation. Tests of AR(1) and AR(2) will be performed. The basic model will be transformed into the first differenced form, and then, the first differenced lag dependent variable is instrumented with its past levels