Experimental

Designs and Layout

Terms and Concepts

• Treatments
• Are the different procedures we want to
compare.
• These could be different kinds or amounts of
fertilizer in agronomy,
• Different insecticides or their doses,
• Different temperatures for insect development,
• Different food for insect growth.
• Experimental units
• Are the things to which we apply the treatments.
• These could be plots of land receiving fertilizer,
• Plots receiving different insecticides,
• Plots receiving different insect populations,
• Insects receiving different doses of chemicals
• Variable-
• Variables are quantities of interest or which serve as
the practical substitutes for the concepts of interest.
• Responses
• Are outcomes that we observe after applying a
treatment to an experimental unit.
• That is, the response is what we measure to judge
what happened in the experiment; we often have
more than one response.
• Responses for the above examples might be nitrogen
content or biomass of corn plants, or insect survival or
mortality with certain chemicals.
• Randomization
• Assignment of treatments to experimental units
by using a chance mechanism.
• Replication-
• Assignment of individual treatments to multiple
experimental units
• Blocking-
• Assignment of treatments within multiple groups
of experimental units
• Experimental Error
• Is the random variation present in all experimental
results.
• Different experimental units will give different
responses to the same treatment,
• and it is often true that applying the same treatment
over and over again to the same unit will result in
different responses in different trials.
• Experimental error does not refer to conducting the
wrong experiment or dropping test tubes.
• Measurement units (or response units)
• Are the actual objects on which the response is
measured.
• These may differ from the experimental units.
• For example, consider the effect of different fertilizers
on the nitrogen content of corn plants.
• Different field plots are the experimental units, but
the measurement units might be a subset of the corn
plants on the field plot,
• or a sample of leaves, stalks, and roots from the field
plot.
• Blinding
• Occurs when the evaluators of a response do not know
which treatment was given to which unit.
• Blinding helps prevent bias in the evaluation, even
unconscious bias from well-intentioned evaluators.
• Double blinding occurs when both the evaluators of the
response and the (human subject) experimental units do not
know the assignment of treatments to units.
• Blinding the subjects can also prevent bias, because subject
responses can change when subjects have expectations for
certain treatments.
• Control
• Has several different uses in design.
• First, an experiment is controlled because we as experimenters assign
treatments to experimental units. Otherwise, we would have an
observational study.
• Second, a control treatment is a “standard” treatment that is used as a
baseline or basis of comparison for the other treatments.
• This control treatment might be the treatment in common use, or it might
be a null treatment (no treatment at all).
• For example, a study of new chemical could use a standard chemical as a
control treatment,
• or a study on the efficacy of fertilizer could give some fields no fertilizer at
all. This would control for average soil fertility or weather conditions.
• Placebo
• Is a null treatment that is used when the act of applying a treatment
— any treatment—has an effect.
• Placebos are often used with human subjects, because people often
respond to any treatment:
• For example, reduction in headache pain when given a sugar pill.
• Blinding is important when placebos are used with human subjects.
• Placebos are also useful for nonhuman subjects.
• The apparatus for spraying a field with a pesticide may compact the
soil.
• Thus we drive the apparatus over the field, without actually spraying,
as a placebo treatment.
• Factors combinations to form treatments.
For example, the population of growth is the
result (or effected) by temperature and light.

• We vary temperature and light hours in different

combinations to observe the effect on
population growth.

• Temperature and light are the factors.

• Confounding
• Occurs when the effect of one factor or treatment cannot be
distinguished from that of another factor or treatment.
• The two factors or treatments are said to be confounded.
• Consider insecticides application on cotton in Multan and D.
G. Khan for control of whitefly.
• In this experiment, we cannot distinguish location effects
from insecticide effect—the insecticide factor and the
location factor are confounded.
• Except in very special circumstances, confounding should be
avoided.
Assignment (Practical)
• Choose a topic of interest to you. Identify an
issue that you would like to investigate with a
designed experiment. Identify and discuss the
experimental units, treatments, and response
measurement(s).
Single-factor experiments
1. Completely Randomized Design (CRD)

2. Randomized Complete Block Design (RCBD)

3. Latin Square Design (LS)

1-COMPLETELY RANDOMIZED
DESIGN (CRD)
• It is ntended for experiments where there is no
significant variation in the area or environment.
• As such, CRD is applicable for laboratory or
greenhouse experiments ONLY and should not
be used for field experiments.
• It may also be used for experiments conducted
in the open provided the conditions are the
same for the entire experiment.
• Examples are pot or seed flat experiments where the
soil is thoroughly mixed. In animal experiments, the
animals should be of the same age, weight, breed, etc.
• The main advantage of CRD is that it can be used for
experiments with equal or unequal number of
treatments or vice-versa,
• It can be used for treatments with unequal number of
replications.
• The main disadvantage of CRD is the limitation in its
use due to the restriction of providing uniform
condition in the whole experimental area.
Randomization
• All experimental units (treatment and replication
combinations) are completely randomized within the whole
experimental area.
• The steps in randomization are as follows:
1. Determine the total number of experimental units.
A) For CRD with equal number of treatments per replication, it
is the product of the number of treatments and the number of
replicates.
B) For CRD with unequal number of treatments, it is the sum
of the treatments of all replicates.
2. Assign a plot number to each experimental unit
consecutively.
3. Assign the plot numbers to the experimental
plots using a randomization scheme.
A. Using draw lots.
1. Prepare pieces of papers corresponding to the
number of experimental units.
2. Write the plot number in each of the papers.
3. Mix the papers thoroughly in a container.
4. Without looking inside the container, draw a
paper and assign the plot number on the first
experimental unit.
5. Without returning the paper previously drawn,
continue drawing individual papers and assign the
plot numbers until all the corresponding
treatments have been assigned to all
experimental unit.
Suppose T5R1 was drawn first, it will be assigned
to the first space in the experimental area.
• When the randomization is finished, the final lay-out may
look like this:
• The lay-out may also be like this:
• or any other form provided the conditions are basically the
same in the entire experimental unit.
• B. Using the table of random numbers
1. With eyes closed, point the finger to the table of random
numbers. Copy the middle 3 digits of all the consecutive digit
numbers corresponding to the number of experimental units.
2. Rank the 3-digit numbers from 1 to the highest number of
experimental units.
3. Arrange the 3-digit numbers consecutively from lowest to
highest but be sure to carry the assigned rank. The sequence
of the ranks will now serve as the randomized numbers.
4. Assign the ranks corresponding to the plot number of the
experimental units.
• CRD WITH EQUAL NUMBER OF REPLICATIONS
• Using the following lay-out as an example, the plots may be
numbered consecutively for easier data gathering:
• CRD with Unequal Number of Replications
• The reason for having unequal number of
replications may be due to limitation in the
number of available sample plants at the start of
the experiment.
• It is also possible that equal number of plants
was used at the start of the experiment but
some of the plants died due to reasons aside
from the treatment itself resulting to unequal
number of treatments.
• NOTE that if one or some of the plants died due
to the effect of the treatment,
• e.g. high concentration of a particular chemical,
and not due to other unknown factors, the data
on the dead plant should be recorded as zero
and the analysis should be done using equal
number of treatments.
• A possible lay–out for CRD with unequal number of
treatments is as follows:
• In the hypothetical lay–out,
• Treatment 1 has only 4 replications,
• Treatment 6 has 5 replications while the other 4 treatments
have 6 replications.
• Again, the lay–out may be arranged in any manner (shape or
orientation) provided the conditions for all the plants are
basically the same or equal.
Randomized Complete Block
Design (RCBD)
• Experiments in the open field is conducted using
Randomized Complete Block Design (RCBD),
• Because it is difficult to totally control the
condition in the experimental area.
• Variation in the condition of the area may be due
to the soil itself (fertility, soil type), slope or
gradient, wind direction, water direction.
• Blocking is a method of improving accuracy of an experiment
by arranging the experimental materials into groups,
• so that the units in each group are as homogeneous
(uniform) as possible.
• The basic difference between CRD and RCBD is the presence
of blocks in RCBD.
• In CRD, randomization is done in the whole experimental
area while in RCBD, randomization is done in each block and
each randomization is independent of the other.
• Since randomization is done in each block, all treatments
should appear in each block (or each treatment should
appear only once in each block).
• An ideal source of variation to use as the basis for
blocking is one that is large and highly predictable.
• Examples are:
• Soil heterogeneity or variability, in a fertilizer or variety
trial where yield data is the primary character of interest.

• Direction of insect migration, in an insecticide trial where

insect infestation is the primary character of interest.
• With these things in mind, blocking can now be done so
that the variability of the area of a particular block is
made as small as possible.
• The guidelines for blocking are as follows:
• When the gradient is unidirectional, use long
and narrow blocks so that their length is
perpendicular to the gradient.
• When the fertility gradient occurs in two
directions with one gradient much stronger than
the other, ignore the weaker gradient and follow
the preceding guideline for the case of
unidirectional gradient.
• When the fertility gradient occurs in two
directions with both gradients equally
strong and perpendicular to each other,
choose one of the alternatives:
• Use blocks that are as square if possible.
• Use long and narrow blocks with their
lengths perpendicular to the direction of
one gradient.
• Use the Latin Square Design with two-way
blockings, one for each gradient.
• Randomization
• The steps in randomization for RCBD are as follows:

1. Divide the experimental area into the desired

number of blocks.
2. In the example (next slide), there are 6 blocks
with 6 treatments per block.
3. The shape and orientation of blocks will depend
on the factors that will reduce or minimize
intrablock error while maximizing interblock error.
• Suppose the blocks will be arranged like this:
4. Assign a plot number to each treatment consecutively
in each block.
5. Assign the treatments to the experimental plots using
a randomization scheme.
A. Using the table of random numbers
(1) With eyes closed, point the finger to the table of
random numbers.
(2) Copy the middle 3 digits of all the consecutive 5--
digit numbers corresponding to the number of
treatments in a block.
(3) Rank the 3-digit numbers from 1 to the highest
number of treatments.
(4) Arrange the 3-digit numbers consecutively from
lowest to highest but be sure to carry the assigned
rank. The sequence of the ranks will now serve as the
randomized numbers.

(5) Assign the ranks corresponding to the number of

treatments of the first block.
(6) Repeat steps 1-5 to the remaining blocks. Once
finished, the lay-out may look like this:
• If the fertility of the area is not known, the blocks and
plots may be arranged this way:
• B. Using draw lots

(1)Prepare pieces of papers corresponding to the number of

treatments in each block.

(2) Write the treatment name in each of the paper.

(3) Mix the papers thoroughly in a container.

(4) Without looking in the container, draw a paper and

assign the treatment name on the first experimental unit of
the first block.
(5) Without returning the paper previously drawn,
continue drawing individual papers and assign the
corresponding treatments until all the treatments
have been assigned to the first block.

(6) Repeat steps 1-5 to the remaining blocks.

• Do the same for the remaining blocks.
LATIN SQUARE DESIGN
• This design is most useful in areas where the direction
of soil fertility/heterogeneity is bidirectional.
• With this type of soil fertility, it is not advisable to use
RCBD because the blocking will take care of only one
gradient while the other gradient will be confounded
(or added) to the treatment effect.
• As a result, the variation observed among treatments
cannot be attributed to the effect of the treatment
alone because part of the variation may be due to
differences in fertility gradient.
• Latin Square is the more appropriate design because
the two-directional blocking,
• commonly referred to as row-blocking and column-
blocking, is accomplished by ensuring that every
treatment occurs only once in each row-block and
once in each column-block.
• As such, LS design is considered more powerful than
RCBD in that aside from detecting differences due to
treatments, it also detects differences due to rows
and columns and not due to blocks alone.
• Some examples of cases where LS design can be
appropriately used are:
• Field trials in which the experimental area has two
fertility gradient running perpendicular to each other,
• or has a unidirectional fertility gradient but also has a
residual effects from previous trials.
• Insecticide field trials where the insect migration has
a predictable migration which is perpendicular to the
dominant fertility gradient of the experimental area.
• Greenhouse trials where the experimental pots are
arranged in straight line which is perpendicular to the
glass or screen walls,
• such that the differences among rows of pots and the
distance from the glass wall (or screen wall) are
expected to be the two major sources of variability
among the experimental pots.
• Laboratory trials with replication over time,
• such that the difference among experimental units
conducted at the same time and among those
conducted over time constitute the two known
sources of variability.
• One important feature of the design is that the
number of replications is always equal to the
number of treatments.
• As such, the LS design can only be used when
the number of treatments is a perfect square (9,
16, 25, 36, 49, etc.).
• Because of this requirement, the main
disadvantage of the design is that it is not
advisable for large number of treatments.
• Thus, in practice, the LS design is applicable only for
experiments in which the number of treatments is not
less than four but not more than eight.

• Because of such limitation, the LS design has not been

widely used in agricultural experiments despite its
great potential for controlling experimental error.
• The process of randomization and lay-out for a LS design is
shown below for an experiment with six treatments, A, B, C,
D, E and F.
• 1. Consider a sample LS plan with six treatments. For our
example, the 6 x 6 Latin Square preliminary plan is:
2. Randomize the row arrangement of the plan selected
in step 1, following one of the randomization schemes
described in RCBD.
For this experiment, the table-of-random numbers
method is applied.
• With eyes closed, point the finger to the table of
random numbers.
• Copy the middle 3 digits of all five consecutive 5-digit
numbers corresponding to the number of treatments.
For our example, 717, 569, 223, 478, 036 and 132 were
taken.
• Rank the 3-digit numbers from lowest to highest:
• Use the rank to represent the existing row number of
the selected plan and the sequence to represent the
row number of the new plan.
• For our example, the fifth row of the selected plan
(rank = 1) becomes the first row (sequence = 1) of the
new plan;
• the sixth row of the selected plan becomes the
second row of the new plan, and so on.
• The new row plan, after the row randomization, is
given on the next slide:
• 3. Randomize the column arrangement, using the
same procedure used for row arrangement in step 2.
• For our example, the six random numbers selected
and their ranks are:
• The rank will now be used to represent the column
number of the plan obtained in step 2 (i.e. with re-
arranged rows) and the sequence will be used to
represent the column number of the new plan.
• For our example, the first column obtained in step 2
remains the first column of the final plan,
• the sixth column of the plan of step 2 becomes the
second column of the final plan, and so on.
• The final plan, which becomes the lay-out of the
experiment is on next slide:
• Consequently, the field lay-out is shown below. Note
that the dimension of the lay-out cannot be changed
unlike in CRD or RCBD.
SPLIT–PLOT IN RANDOMIZED COMPLETE
BLOCK DESIGN
• Application
• Split–plot in Randomized Complete Block Design is used for
field experiment involving two factors tested at the same
time,
• with one factor (and its interaction) considered more
important that the other factor.
• t is also recommended if application of treatment of a factor
requires a large area to minimize border effect, or to
facilitate better management.
Randomization
• One of the basic differences between factorial and
split–plot in RCBD is in randomization.
• In factorial in RCBD, randomization of all factor
combinations is done in each replication.
• In split– plot in RCBD, only the main plot is
independently randomized in each replication.
• After randomizing all the main plots, the subplot is
randomized in each main plot in each replication.
• With this randomization scheme, the size of the main
plot is much larger than that of the subplot,
• so it is expected that the degree of precision is less in
the main plot than in the subplot.
• Degree of precision is more in subplot due to small
size.
• The procedure in randomization is as follows:
1. Determine the number of treatments in the main
plot.
In the example, there are 3 levels in the main plot.
2. In replication 1, assign a number to each main plot
treatment consecutively.
3. Assign the main plot treatment to the experimental
plots using a randomization scheme as discussed in
RCBD:
• A. Table of random numbers B. Draw lots.
• 4. Determine the number of
subplots.
• In the example, there are 4 subplot
levels.
• 5. Randomize the subplots in each
main plot treatment of independent
of the other main plot treatments
using
a randomization scheme discussed
in RCBD.
• 6. After finishing randomization of the main plots
within Replication 1 and all subplots within each
main plots in Replication 1,
• proceed to Replication 2 and follow Steps 2 to 4.
• 7. Continue until all replications are completed.
• Using the randomization discussed above, a possible lay–out of a 3 x 4
split–plot with 4 replications is as follows.
• 1. If the plots are long and narrow or if there is a one–way fertility
gradient in the area, the following lay–out may be used:
• 2. If the fertility of the area is not known, the
following lay–out may be used:
• 3. The lay-out may also be
like this depending on the
dimension and availability
of the area:
STRIP-PLOT (SPLIT-BLOCK) IN
RANDOMIZED COMPLETE BLOCK
DESIGN
• Application Strip–plot in RCBD is most appropriate in
experiments dealing with two factors which are more
or less equally important while the interaction is more
important.
• This is also applicable if the two factors can be applied
easily using large plots than using small plots.
• A typical example is a combination of irrigation system
and frequency of spraying of pesticide.
• In this design, one treatment is applied in horizontal
position while the other treatment in vertical position.
• As such, both Factor A and Factor B cover large plots
while the interaction is much smaller than any of the
two factors.
• The degree of precision is more or less equal for both
Factors A and B but that of the interaction is more
precise than any of the two factors.
Randomization
• Randomization of Factor A is done independent of the
randomization of Factor B but both will be randomized
within each block.
• The procedure is as follows:
• 1. Determine the number of treatments in Factor A.
• 2. For Block 1, randomize the treatments in factor A in one
direction (vertical or horizontal) using a randomization
scheme described in RCBD.
• Note that all treatments must appear together as strips in
one block, that is, use of two sub–blocks is not possible.
• 3. Determine the number of treatments in Factor B.
• 4. Randomize the number of treatments in Factor B by the
same randomization scheme used in Factor A but is the other
direction.
• When each cell is completed, the combination of A
and B will be:
• 5. Follow Steps 1 to 4 for the succeeding blocks.
• Following the randomization scheme above, a
possible lay–out for strip–plot in RCBD with
three levels of Factor A and four levels of Factor
B with four replications is shown if the total area
is very long and the fertility gradient is irregular
or not known.
• The lay-out may also be like this

• It is not advisable to use

very long plots because
• it will result to one of
the factors having very
long area coverage
which might result to
high variability within
the block.
• As such, square or
nearly square plots are
recommended.
MULTIPLE FACTOR EXPERIMENTS
• The basic designs can be used when dealing with a
single factor or treatment.
• However, some experiments require two or more
factors or treatments to be analyzed together.
• As a general rule, it is always advisable to deal with
simple experiments rather than complex ones, however,
• experiments involving two or more factors are
conducted when the response to the factor of interest
is expected to differ under different levels of the other
factor,
• in short, an interaction occurs between the two factors.
• In interaction, the effect between two factors can be
measured only if the two factors are tested together in
the same experiment.
• When interaction is absent, the simple effect of a
factor is the same for all levels of the other factor(s)
and equals the main effect.
• When interaction is present, the simple effect of a
factor changes as the level of the other factor changes.
Consequently, the main effect is different from the
simple effects.
• Based on the above points when dealing with
interaction effects, the following points must be
observed when interaction effect between two factors
is present:
• The simple effects and not the main effects should be
examined.
• The result from a single–factor experiment is
applicable only to the particular level in which the
other factors were maintained in the experiment and
there can be no generalization of the result to cover
any other levels.
• FACTORIAL
• Factorial, whether in CRD or RCBD, is applicable when an
experiment is dealing with two factors to be evaluated at the
same time.
• The other assumption is that both factors (A and B) and the
interaction between the two (A x B) are of equal importance.
• Factorial can also be considered as simple CRD or RCBD if the
factor combination will be considered as single treatment.
• After analysis, the treatment effect is partitioned into three
meaningful components:
• due to A, due to B and due to AxB.
• FACTORIAL IN COMPLETELY RANDOMIZED DESIGN
• Randomization
• The randomization procedure in factorial in CRD is very similar
to CRD if the factor combination will be considered as single
treatment.
• The steps in randomization are as follows:
1. Determine the number of A factor x B factor x Replication
combinations in the whole experiment. In the example,
there are 48 combinations (3 levels of A, 4 levels of B, and 4
replications).
2. Assign a plot number to each combination consecutively.
3. Assign the factor combinations to the experimental plots
using a randomization scheme.
• A. Using the table of random numbers
• 1. With eyes closed, point the finger to the table of random
numbers. 2. Copy the middle 3 digits of all the consecutive
digit numbers corresponding to the number of factor
combinations.
• 3. Rank the 3–digit numbers from 1 to the highest number of
combinations.
• 4. Arrange the 3–digit numbers consecutively from lowest to
highest but be sure to carry the assigned rank. The sequence
of the ranks will now serve as the randomized numbers.
• 5. Assign the ranks corresponding to the number of
combinations of the whole experiment.
• B. Using draw lots
• 1. Prepare pieces of papers corresponding to the number of
A factor x B factor x replication combinations of the whole
experiment.
• 2. Write the combinations in each of the paper.
• 3. Mix the papers thoroughly in a container.
4. Without looking in the box, draw a paper and assign the
combination on the first experimental unit of the whole
experiment.
5. Without returning the paper previously drawn, continue
drawing individual papers and assign the corresponding factor
combination name until all the factor combinations have been
assigned to the whole experiment.
• Using the randomization discussed
above, a possible lay–out of a 3 x 4
factorial with 4 replications is:
• Note that in CRD, the lay–out could
be
anything provided the conditions are
similar in the whole area.
However, it is recommended that the
factor combinations are placed side by
side and not too far from one another
to be sure that the conditions are
similar.
FACTORIAL IN RANDOMIZED
COMPLETE BLOCK DESIGN
• Factorial in RCBD is normally used in field experiment
involving two factors that are equally important
together with their interaction (A x B).

• It is not an ideal design for treatments with extensive

border effect, e.g. spray, watering or chemicals that
may be carried easily by wind.
 Randomization
• The randomization procedure in factorial in RCBD is very similar
to RCBD if the factor combination will be considered as single
treatment.
• The steps in randomization are as follows:
• 1. Determine the number of factor combinations per replication.
• This is the product of the number of Factor A and number of
Factor B
• 2. Assign a plot number to each factor combination
consecutively.
• 3. Assign the factor combinations to the experimental plots using
a randomization scheme.
• A. Using the table of random numbers
• 1. Before randomization, decide which way will the numbers
be assigned to the plots: horizontal or vertical.
• With horizontal assignment, will it be left to right or right to
left?
• With vertical assignment, will it be top to bottom or bottom
to top?
• With eyes closed, point the finger to the table of random
numbers.
• 2. Copy the middle 3 digits of all the consecutive 5–digit
numbers corresponding to the number of factor
combinations per block.
• 3. Rank the 3–digit numbers from 1 to the highest
number of factor combinations.
• 4. Arrange the 3–digit numbers consecutively from
lowest to highest but be sure to carry the assigned
rank. The sequence of the ranks will now serve as the
randomized numbers.
• 5. Assign the ranks corresponding to the number of
factor combinations of the first block.
• The first block may look like this:

• 6. Repeat steps 1–5 for the other blocks.

• B. Using draw lots.
• 1. Prepare pieces of papers corresponding to the
number of factor combinations.
• 2. Write the factor combinations per block in each of
the paper.
• 3. Mix the papers thoroughly in a container.
• 4. Without looking in the box, draw a paper and
assign the factor combination name on the first
experimental unit of the first block.
• 5. Without returning the paper previously drawn,
continue drawing individual papers and assign the
corresponding factor combination name until all the
factor combinations have been assigned to the first
block.
• 6. Repeat steps 1–5 for the remaining blocks.
• As a general rule, try to make each block as square as
possible to reduce the distance between the two
farthest plots.
• In a very long block, higher soil variability may occur
between two plots on both ends of the area.
Using the randomization discussed, a possible lay–out of a 3 x 4 factorial
with 4 replications is as follows if the plots are short and wide, or the soil
fertility gradient is not known:
• Or it could be this way:
• If the plots are long and narrow,
• or there is one–directional soil fertility gradient,
• or a the wind direction will have an effect of the
treatment,
• the following lay–out may be applied with the
length of the plot perpendicular to the strong
gradient:
SPLIT–PLOT IN COMPLETELY
RANDOMIZED DESIGN
• Randomization
• The basic differences between factorial and split–plot in CRD
are in lay–out and randomization.
• In factorial in CRD, all factor combinations are assigned and
randomized in the whole experimental area (since there are
no blocks).
• In split–plot, only the main plots are assigned and
randomized in the experimental area.
• After randomizing all the main plots, the subplot is
randomized in each main plot.
• With this randomization scheme, the size of the main
plot is much larger than that of the subplot,
• so it is expected that the degree of precision is less in
the main plot than in the subplot.
• Moreover, the size of the subplot is the same as that
of the interaction so the degree of precision of the
subplot is equal to that of the interaction
• The procedure in randomization is as follows:
• 1. Determine the total number of Main Plot x Replication
combinations in the experiment. In the example, there are 3
x 4 combinations of Main Plot x Replication.
• 2. Assign a plot number to each main plot treatment
consecutively.
• 3. Assign the main plot treatments to the experimental plots
using a randomization scheme as discussed in CRD:
• A. Table of random numbers
• B. Draw lots
A possible randomization of Main Plot x Replication is
shown below:
• 4. Determine the number of subplots.
• 5. In each Main Plot x Replication combination, randomize
the subplot levels independent of the other Main Plot x
Replication combinations.
• 6. Repeat the procedure in the other combinations. When
completed, a possible lay–out of a 3 x 4 split–plot in CRD
with 4 replications is as follows:
• The lay-out may also be
like these depending on
available area: