Predicting winter wheat yields by comparing regression equations

• Predicting of regression equations as a tool for determining yield reduction.
• Three stages to compare regression equations are needed.
• The first step is predicting adequate regression equations.
• In the second step regression coefficients are compared.
• The last step: comparison of the intercepts - distances between hyperplanes.

The aim of this study was to determine a statistical model described by regression equations as a tool for predicting the reduction in winter wheat yields as a function of the degree of pathogenic damage to the flag leaf and first leaf under the flag leaf at the stage of full heading, at growing sites with varied conditions. The study was carried out in the years 2007–2009, at two locations in the Wielkopolska region of Poland. The research was based on one-factor experiment on winter wheat, variety Tonacja. The method is based on the adaptation of the procedure of comparing of parallelism and separation for regression equations. These results indicated that expressing the yield as a function of the degree of infection by pathogens by means of a single regression equation is a process that requires several stages, entailing comparison of the regression coefficients and, independently of that, comparison of the intercepts (distances between hyperplanes). Determination of regression equations may be used as a tool for determining yield reduction, but it requires the consideration of factors including habitat conditions, leaf type and location. The program package R (3.0.2) was used for the calculations.