Biometrical Letters Vol. 51(2), 2014, pp. 171-179
|
AN EVALUATION OF THE EFFICIENCY OF PLANT PROTECTION PRODUCTS VIA NONLINEAR STATISTICAL METHODS - A SIMULATION STUDY Ewa Skotarczak, Ewa Bakinowska, Kamila Tomaszyk Department of Mathematical and Statistical Methods, Poznan University of Life Sciences, Wojska Polskiego 28, PL-60-637 Poznań, Poland, efalsa@up.poznan.pl |
A nonlinear statistical approach was used to evaluate the efficiency of plant protection products. The methodology presented can be implemented when the observations in an experiment are recorded as success or failure. This occurs, for example, when following the application of a herbicide or pesticide, a single weed or insect is classified as alive (failure) or dead (success). Then a higher probability of success means a higher efficiency of the tested product. Using simulated data sets, a comparison was made of three methods based on the logit, probit and threshold models, with special attention to the effect of sample size and number of replications on the accuracy of the estimation of probabilities.
threshold model, logistic model, plant protection, simulated data
