Biostatistics in Perennial crops
Perennial crops live longer than annual crops, and since they are likely to be larger, experiments with them face special problems. The experiments often taken several years, so the plants are at risk longer, increasing the likelihood of mishap and thus unbalanced data, so greater caution is needed in making plants about them. The experimental design must be able to tolerate the inevitable missing data. Whatever design is used, it must be throughly practicable, for it is no use producing a scheme on paper that will give endless trouble in the field. Experimental objetives may change over time, requiring modification of the original study plan. It is better simple experimental design than complex ones, these are inflexible and are seriously undermined by missing data in both computation and interpretation. Since perennials are in general larger than annuals are of greater interest as individuals, and will probably use single.plant or severalplant plots, and environmental variation in rarely of sole importance. With perennials the experimenter should not assume de error variation as positional, Le envirenmental,some of it may well derive from the plants themselves.Attempting to reduce error variance by merely increasing replicationwithin a standar design may not be satisfactory, The analysis of covariance has proven to be very effectivealternative in many cases, after reducing de error variance by 25.50%. The data will probably include several observations collected over time on the same plant. The statistical analysis must take this into account, and several possible methodsare suggested.
| Main Author: | |
|---|---|
| Format: | Digital revista |
| Language: | spa |
| Published: |
Universidad Nacional de Colombia - Sede Bogotá - Facultad de Ciencias Agrarias
1989
|
| Online Access: | https://revistas.unal.edu.co/index.php/agrocol/article/view/20990 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|