Studies on the Relative Efficiency of Different Experimental Designs for Sunflower (Helianthus annuus L.)
Nishu Lohmor
Department of Mathematics, Statistics and Physics, Chaudhary Charan Singh Haryana Agricultural University, Hisar, Haryana, India.
Mujahid Khan *
Department of Mathematics, Statistics and Physics, Chaudhary Charan Singh Haryana Agricultural University, Hisar, Haryana, India.
Kiran Kapoor
Department of Mathematics, Statistics and Physics, Chaudhary Charan Singh Haryana Agricultural University, Hisar, Haryana, India.
Ramesh Kumar Tripathi
Department of Agricultural Statistics, Narender Deva University of Agriculture and Technology, Kumarganj, Faizabad, India.
*Author to whom correspondence should be addressed.
Abstract
The best experimental design to use in any given condition is the one which estimates the desired effects and contrasts with maximum precision or efficiency. In uniformity trial data, the treatments being considered as dummy, the relative efficiencies of various experimental designs were determined using the yield data taken from the uniformly raised sunflower crop during February 2014 to June 2014 at CCSHAU research farm, Hisar. Randomized Block Design (RBD) was found to be more effective in reducing error variation over Completely Randomized Design (CRD). Latin Square Design (LSD) was found to be more advantageous over CRD as well as RBD only when columns were considered as blocks. Split plot design was more efficient than RBD on comparison with sub plots (10.82% gain in precision) while less efficient with main plots in comparison to RBD. In case of factorial experiments, the gain in efficiency of the confounded design of order 25 increases upto 68 per cent for the 4-plots block.
Keywords: Experimental design, relative efficiency, uniformity trial, completely randomized design, randomized block design, Latin square design