Full factorial design versus central composite design: Statistical comparison and implications for spray droplet deposition experiments

The optimization of spraying parameters to improve pesticide efficacy and r educe environmental impacts requires a large number of experimental treatme nts. Some experimental designs can be used to reduce the required number of treatment combinations. A comparison between a 3(3) factorial design and a central composite design (CCD) regarding the model fitting ability and the ir respective capability in identifying significant coefficients and optimu m sprayer operating conditions is presented. Experiments were conducted in a spray chamber simulating an air-assisted spray boom on broccoli and potat oes. The independent variables studied were air speed, airflow and air jet angle, and the dependent variable was leaf area coverage. Small disks of Kr omekote cards were used to collect the spray dyed with the fluorescent trac er Blankophor BA(TM). Disks were placed on the upper and under side of leav es in the top, middle, and bottom portion of the crop canopy. Results showe d that both experimental designs performed equally well. Both designs lead to the same conclusions regarding the optimum settings of the operational s prayer parameters. Overall, the CCD tended to be more conservative than the factorial design. Theoretical considerations and the fact that the CCD req uires fewer experimental units, justifies the recommendation to use of CCD in spray droplet deposition experimentation.