Identification of Efficient Experimental Design(s) by Comparing Different Commonly Used Designs for Large Data Sets
Presentation Type
Event
Abstract
Following basic principles of experimental design and using appropriate filed design usually play a significant role in successful plant breeding program. Breeders usually seek suitable field designs and use them for their experiments to minimize the experimental error. Finding out the most appropriate design for the experiment may greatly improve the data analysis and help breeders to take right decision. For the current study, 64 corn hybrids (genotypes) were evaluated in six counties in North Carolina. The main objective of our study is to analyze and compare results using different models or experimental designs (CR, RCBD, Sub-block, and Rectangular) and finally determine which design(s) is the most suitable for each county. Linear mixed model (LMM) with jackknife resampling technique was used for data analysis. Among all the tested designs, we found that Rectangular design is more suitable for handling large data sets and CR design was least effective. Rectangular design was well over 2 times and close to 2 times better than CR and RCB design respectively in terms of Relative Efficiency (RE) for most of the counties.
Start Date
2-12-2018 12:00 PM
Identification of Efficient Experimental Design(s) by Comparing Different Commonly Used Designs for Large Data Sets
University Student Union: Volstorff A
Following basic principles of experimental design and using appropriate filed design usually play a significant role in successful plant breeding program. Breeders usually seek suitable field designs and use them for their experiments to minimize the experimental error. Finding out the most appropriate design for the experiment may greatly improve the data analysis and help breeders to take right decision. For the current study, 64 corn hybrids (genotypes) were evaluated in six counties in North Carolina. The main objective of our study is to analyze and compare results using different models or experimental designs (CR, RCBD, Sub-block, and Rectangular) and finally determine which design(s) is the most suitable for each county. Linear mixed model (LMM) with jackknife resampling technique was used for data analysis. Among all the tested designs, we found that Rectangular design is more suitable for handling large data sets and CR design was least effective. Rectangular design was well over 2 times and close to 2 times better than CR and RCB design respectively in terms of Relative Efficiency (RE) for most of the counties.