Abstract
The use of nearest neighbors and spatial models (SPAT) to analyze field trial data has become commonplace in recent years. These two types of analyses improve precision compared to ANOVA when trials are poorly blocked, but results are less clear in well-blocked trials. We examined data from wheat trials containing 60 cultivars, conducted at five locations, where each location was set up as an alpha lattice design. We compared the relative efficiency of detecting cultivar differences for spatial models and nearest neighbors analyses (NNA) to ANOVA, fit of the models, and correlations of ranked cultivars. Though the SPAT and NNA generally outperformed the ANOVA, the selection of desirable cultivars remained relatively unchanged when using a well-blocked design analyzed with an ANOVA.
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.
Recommended Citation
Landes, Reid D.; Eskridge, Kent M.; Baenziger, P. Stephen; and Marx, David B.
(2001).
"ARE SPATIAL MODELS NEEDED WITH ADEQUATELY BLOCKED FIELD TRIALS?,"
Conference on Applied Statistics in Agriculture.
https://doi.org/10.4148/2475-7772.1227
ARE SPATIAL MODELS NEEDED WITH ADEQUATELY BLOCKED FIELD TRIALS?
The use of nearest neighbors and spatial models (SPAT) to analyze field trial data has become commonplace in recent years. These two types of analyses improve precision compared to ANOVA when trials are poorly blocked, but results are less clear in well-blocked trials. We examined data from wheat trials containing 60 cultivars, conducted at five locations, where each location was set up as an alpha lattice design. We compared the relative efficiency of detecting cultivar differences for spatial models and nearest neighbors analyses (NNA) to ANOVA, fit of the models, and correlations of ranked cultivars. Though the SPAT and NNA generally outperformed the ANOVA, the selection of desirable cultivars remained relatively unchanged when using a well-blocked design analyzed with an ANOVA.
