Abstract
Plant scientists are interested in measuring plant response to quantitative treatment factors, e.g. amount of nutrient applied. Response surface methods are often used for experiments with multiple quantitative factors. However, in many plant nutrition studies, second-order response surface models result in unacceptable lack of fit. This paper explores multi-factor nonlinear models as an alternative. We have developed multi-factor extensions of Mitscherlich and Gompertz models, and fit them to data from experiments conducted at the University of Nebraska-Lincoln Horticulture department. These data are typical of experiments for which conventional response surface models perform poorly. We propose design selection strategies to facilitate economical multi-factor experiments when second-order response surface models are unlikely to fit.
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.
Recommended Citation
Landes, R. D.; Stroup, W. W.; Paparozzi, E. T.; and Conley, M. E.
(1999).
"NONLINEAR MODELS FOR MULTI-FACTOR PLANT NUTRITION EXPERIMENTS,"
Conference on Applied Statistics in Agriculture.
https://doi.org/10.4148/2475-7772.1263
NONLINEAR MODELS FOR MULTI-FACTOR PLANT NUTRITION EXPERIMENTS
Plant scientists are interested in measuring plant response to quantitative treatment factors, e.g. amount of nutrient applied. Response surface methods are often used for experiments with multiple quantitative factors. However, in many plant nutrition studies, second-order response surface models result in unacceptable lack of fit. This paper explores multi-factor nonlinear models as an alternative. We have developed multi-factor extensions of Mitscherlich and Gompertz models, and fit them to data from experiments conducted at the University of Nebraska-Lincoln Horticulture department. These data are typical of experiments for which conventional response surface models perform poorly. We propose design selection strategies to facilitate economical multi-factor experiments when second-order response surface models are unlikely to fit.
