Generalized Row - Column D esigns for Single and Multi - Factor Experiments
Loading...
Date
2015
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
ICAR - INDIAN AGRICULTURAL STATISTICS RESEARCH INSTITUT E LIBRARY AVENUE, PUSA
Abstract
In
field and animal experiments
,
whe
re
there are two sources of variation in experimental
units
that may influence the response variable
,
row
-
column designs are
used
.
Most of the
row
-
column designs developed in the literature have only one uni
t
corresponding to the
intersection of row and column
. However,
for the
instances
when the number of
treatments is large with limited
experimental resources
,
Generalized Row
-
Column (GRC)
designs
are used
where there is more than one unit in each row
-
column
intersection.
The
GRC designs developed in the literature are to study all possible pair
-
wise treatment
comparisons.
There
may arise experimental situations where it is desired to compare
treatments belonging to two disjoint sets
and
t
he interest is to es
timate the contrasts
pertaining to treatments
from different sets
with as high precision as possible.
Balanced
Bipartite Generalized Row
-
Column (BBP
-
GRC) designs have been defined and
series of
BBP
-
GRC
have been developed
in which
the contrast
o
f first set
versus second set of
treatments is estimated more precisely.
The presence of missing observations, outliers in
the data, etc. are some of the disturbances that may occur during experimentation. These
disturbances may lead to less precise comparisons among
treatments.
Robustness of
different classes of GRC designs against missing of one or more observations
has
been
investigated.
It is found that
the
efficiency is quite high
(more than 90%)
for most of the
designs and the designs are robust
and
there
is a d
ecreasing trend in efficiency with
increase in
number of missing observations.
The
GRC designs developed in the literature
are
mostly
for single factor experiments
.
Situations may arise wherein the experiment
consist of more than one factor with each facto
r having more than one levels
.
Generalized
confounded row
-
column (GCRC) designs, generalized partially confounded row
-
column
(GPCRC) designs and fractional GCRC designs have been developed
which
ensure that
all lower order interactions including main effec
ts are estimable
.
For easy accessibility
of
GRC designs
, a
web solution
named
WebGRC has been developed that provides the
online generation of
randomized layout of the
se
designs
along with an online catalogue
within a permissible range
.
Description
t-9931
Keywords
null