Experimental Designs under Three-way Blocking Structure
KrishiKosh
View Archive InfoField | Value | |
Title |
Experimental Designs under Three-way Blocking Structure
M.Sc. |
|
Creator |
ARVIND KUMAR
|
|
Contributor |
Cini Varghese
|
|
Subject |
research methods, replication, experimental design, sets, land resources, biological phenomena, economic resources, yields, manpower, area
|
|
Description |
When heterogeneity present in the experimental material is from three sources, then three-way blocking of the experimental units is recommended for control/ reduction of experimental error. For these situations, experimental designs under three-way blocking structure can be advantageously used. These designs are capable of marking out stony patches and other features that tend to clump in compact areas or similarity that may arise due to same group of housing of animals. Gerechte designs and Sudoku square designs are some popular and useful classes of such designs. To have a greater application potential, more classes of such designs are constructed to fit into various experimental situations. Here, eight series of designs under three-way blocking structure have been developed for regular regions. Out of these, two series of designs are structurally incomplete one having one empty node and the other having two empty nodes, in each region. Further, two more series of designs are structurally incomplete having either one or ( 1 2 v ) empty regions. Again, there are many trials where it is unrealistic to replicate every treatment the same number of times in a three-dimensional blocking setting. Three series of designs have been obtained for comparing two sets of treatments, one set consisting of test treatments and the other control treatment(s). Out of these three series, two series have one control treatment and the third one has two control treatments. Many times the regions may be irregular, having no uniform shape and size. So, four methods of constructing designs under three way blocking structure for irregular regions have been developed in which two methods give structurally incomplete designs. Each method of construction has been illustrated with appropriate examples. Considering a four-way classified model, the information matrix for the estimation of treatment effects has been derived for designs under three-way blocking set up. Properties of these classes of designs have been studied and canonical efficiency factor of the designs constructed in comparison to an orthogonal design having same number of treatments has also been worked out. SAS codes have been developed using PROC IML to compute the information matrix, variances pertaining to different groups of treatment comparisons and efficiency factor. |
|
Date |
2016-03-21T18:50:26Z
2016-03-21T18:50:26Z 2013 |
|
Type |
Thesis
|
|
Identifier |
http://krishikosh.egranth.ac.in/handle/1/65351
|
|
Language |
en_US
|
|
Format |
application/pdf
|
|
Publisher |
IARI INDIAN AGRICULTURAL STATISTICS RESEARCH INSTITUTE INDIAN AGRICULTURAL RESEARCH INSTITUTE PUSA, NEW DELHI
|
|