Intro to Design chapter 1
Intro to Design chapter 1
Chapter – 01
Treatment
Experimental Material
The material on which treatments is to be tested. It may land, machinery etc. In the above
example the three varieties of wheat seed can be tested to assign plot of land.
Experimental units
Experimental design
A plan used to collect the data relevant to the problem understudy in such a way as to provide a
basis for valid objective inference about the stated problem.
Note: All these steps are accomplished before any experiment is performed.
2
i. Randomization
ii. Replication
iii. Local Control
Randomization
Purpose of Randomization:
Randomization is used to control bias and extraneous sources of variation present in the
experimental units.
Replication
The second principle of an experimental design is replication; which is the repetition of basic
experiment. It is a complete run for all treatments to be tested in the experiment. Some variation
is introduced because of the fact that the experimental units are not physically identical and this
type of variation can be removed by using a number of experimental units.
Purpose of replications:
The third principle of an experimental design is local control. Some sources of variation can be
minimize or removed by randomization and replication but not all. Local control is used to
choose a design in such a manner that all extraneous sources of variation are brought under
control. The term local control referring the amount of balancing block and grouping of the
experimental units.
3
A model in which k treatments or k levels of factors specifically chosen by the experimenter and
the results derived from the experiment is applicable on the treatments included in experiment
and cannot be generalize for other treatments.
𝑌𝑖𝑗 = 𝜇 + 𝜏𝑗 +∈𝑖𝑗
𝐻0 : 𝜏1 = 𝜏2 = ⋯ = 𝜏𝑘
i. ∑ 𝜏𝑗 = ∑(𝜇𝑖 − 𝜇) = 0
ii. ∈𝑖𝑗 ~𝑁𝑖𝑖𝑑 (0, 𝜎 2 )
A model in which the k treatments or k levels of a factor are chosen a random sample from a
large number of treatments (Treatment Population) and the derived results from the analysis of
sample treatments are generalize for population.
𝑌𝑖𝑗 = 𝜇 + 𝜏𝑗 +∈𝑖𝑗
In random effect model the parameters of the model are random variable.
𝐻0 : 𝜎𝑇2 = 0