ABSTRACT
This chapter proposes several research questions and discusses a variety of methodological issues. These include the overall conceptual framework of the study, the conceptualization of the dependent and predictor variables, measurement and data development issues, strategies for the control of unwanted influences, and issues related to violation of assumptions inherent in ordinary least square (OLS) regression analysis. For the purpose of model development, the author has assumed that the growth of the service sector is stimulative to the formation of new firms in the manufacturing sector. One of the important assumptions of the classical linear regression (CLR) model is that the variance of the disturbance term conditional on particular values of the independent variables is a constant. One of the problems that may arise in regression analysis is multicollinearity. Multicollinearity exists between two (or more) independent variables when the relationship between (among) them is perfectly (or almost perfectly) linear.
