ABSTRACT
Linear regression in a special case of the general multiple regression model. Statistical techniques replace thinking. Regression analysis cannot identify the cause and effect relationship or identify the appropriate equation form. Regression analysis merely offers an arithmetic procedure for quantifying a relationship that has been defined by the researcher. A good way to start the search for an appropriate equation form is to plot the data. Admittedly, if there are more than 5 or 6 explanatory variables this can be both time consuming and misleading. In practice, it is common to use a standardized computer program to calculate the regression equation. The concept of interaction was presented in within the framework of an analysis of variance model. The regression model can also handle data where interaction is a consideration. The estimated the relationship between corn yield and applications of nitrogen fertilizer.
