Regression models in planning of experiments

The paper deals with the results of an experiment investigating the influence of nitrogen, phosphorus and potassium on the sugar beet yield. Relationships between yield and the major factors of yield were based on the analysis of the multiple regression model with three independent variables: the used quantities of nitrogen, phosphorus and potassium. The adequacy of the model was established by selecting the variables to be included in its final form. The potassium, the third independent variable, was insignificant statistically and therefore, this variable and linear interactions of initial inputs were eliminated from the final model. There was a positive influence of nitrogen and phosphorus on the growth of sugar beet yield in the observed period of 25 years (1966-1990). This was shown through the positive values of linear and the negative values of quadratic expressions of independent variables. These relationships are illustrated graphically by response surface of increasing yield with the diminishing amount of growth.