TY - JOUR
T1 - A method for ascertaining and controlling representation bias in field trials for airborne plant pathogens
AU - Deardon, R.
AU - Gilmour, S. G.
AU - Butler, N. A.
AU - Phelps, K.
AU - Kennedy, R.
PY - 2004/4
Y1 - 2004/4
N2 - The basic premise of running a field trial is that the estimates of treatment effects obtained are representative of how the different treatments will perform in the field. The disparities between the treatment effects observed experimentally, and those that would be observed were the treatments applied to the field, we term 'representation bias.' When looking at field trials testing the efficacies of treatment sprays on plant pathogens, representation bias can be caused by positive and negative inter-plot interference. The potential for such effects will be greatest when looking at pathogens that are dispersed by wind. In this paper, a computer simulation that simulates plant disease dispersal under such conditions is described. This program is used to quantify the amount of representation bias occurring in various experimental situations. Through this, the relationships between field design parameters and representation bias are explored, and the importance of plot dimension and spacing, as well as treatment to plot allocation, emphasized.
AB - The basic premise of running a field trial is that the estimates of treatment effects obtained are representative of how the different treatments will perform in the field. The disparities between the treatment effects observed experimentally, and those that would be observed were the treatments applied to the field, we term 'representation bias.' When looking at field trials testing the efficacies of treatment sprays on plant pathogens, representation bias can be caused by positive and negative inter-plot interference. The potential for such effects will be greatest when looking at pathogens that are dispersed by wind. In this paper, a computer simulation that simulates plant disease dispersal under such conditions is described. This program is used to quantify the amount of representation bias occurring in various experimental situations. Through this, the relationships between field design parameters and representation bias are explored, and the importance of plot dimension and spacing, as well as treatment to plot allocation, emphasized.
KW - Experimental design
KW - Inter-plot interference
KW - Plant pathology
KW - Simulation of plant disease dispersal
UR - http://www.scopus.com/inward/record.url?scp=1942423653&partnerID=8YFLogxK
U2 - 10.1080/0266476042000184073
DO - 10.1080/0266476042000184073
M3 - Article
AN - SCOPUS:1942423653
SN - 0266-4763
VL - 31
SP - 329
EP - 343
JO - Journal of Applied Statistics
JF - Journal of Applied Statistics
IS - 3
ER -