A Critique of the Proposed National Education Policy Reform
Experiment by design.pptx
1. EXPERIMENT BY DESIGN
PRESENTED TO:
PROF. S. K. JAIN
DR. DHARMENDRA JAIN
PRESENTED BY:
ROHIT KUMAR
M. PHARMA, I SEM
Dr. Harisingh Gaur University, Sagar, Madhya Pradesh
Department of Pharmaceutical Sciences
3. OVERVIEW:
• Experimental by design means creating a set of procedures to systematically test a hypothesis.
• DOE, or Design of Experiments, is a method of designed experimentation where you manipulate
the controllable factors (independent variables or inputs) in your process at different levels to see
their effect on some response variable (dependent variable or output).
• Experimental by design is the process to organize, conduct, and interpret results of experiments in
an efficient way, making sure that as much useful information as possible is obtained by
performing a small number of trials.
• In experimental design, the researcher combines different variables, in a full factorial design or
fractional pattern to save time and money, measures the response to these combinations.
4. • There are three basic principles behind any experimental design:
• Randomisation: the random allocation of treatments to the experimental units.
• Randomize to avoid confounding between treatment effects and other unknown effects.
• Replication: the repetition of a treatment within an experiment allows:
• To quantify the natural variation between experimental units.
• To increase accuracy of estimated effects.
• Reduce noise: by controlling as much as possible the conditions in the experiment. A classical
example is the grouping of similar experimental units in blocks.
MAIN PRINCIPLES OF EXPERIMENTAL DESIGN: 3 “R’s”
5. Response variable: the property of prime interest that is measured and that you want to study (e.g.
grain yield, plant height, etc).
One more explanatory variables: the property or properties that you think will affect the response
variable and that you want to investigate. There are two major types of explanatory variables:
1. Factor: categorical explanatory variable classifying each observation as belonging to a specific
group. Each group of a factor is called a factor level. A distinction can be made between:
• Treatment factors: factors with levels that are of direct interest and that we want to compare, for
example Variety with levels variety A, B, and C.
• Blocking factors: factors with levels that are not of direct interest, but that are important to control
variation in the experiment, so are part of the experimental design, for example blocks.
2. Covariate: continuous explanatory variable that quantifies some property of each observation on a
continuous scale. Again, we can make the distinction between covariates that are of interest (for
example the amount of nitrogen in kg/ha when you want to investigate the effect of nitrogen on say,
grain yield) or covariates that are used to correct for differences between experimental units that are
not related to the treatments
TERMINOLOGY OF EXPERIMENTAL DESIGN:
6. Treatment: is what we want to compare in the experiment. It can consist of the levels of a single factor,
a combination of levels of more than one factor, or of different quantities of an explanatory variable.
Experimental unit: is the physical unit that receives a particular treatment, for example, a plot in the
field. It is essential that the allocation of a treatment to a particular experimental unit is at random.
Measurement unit: the level at which observations are made, usually one measurement per
experimental unit, but it is possible to make more than one observation within a single experimental
unit (repeated measures).
Replication: is the number of independent instances of a treatment that occur within an experiment i.e.
several experimental units receive the same treatment.
Block: is a group of experimental units that show some similarity/homogeneity between each other.
Random allocation of treatments to units within blocks reduces experimental error.
TERMINOLOGY OF EXPERIMENTAL DESIGN:
8. SCREENING:
• The term 'Screening Design' refers to an experimental plan that is intended to find the few
significant factors from a list of many potential ones.
• The primary purpose of screening design is to identify significant main effects, rather than
interaction effects.
• Types of design used for screening:
1. Full factorial design
2. Fractional factorial design
3. Plackett-Burman Design
9. SELECTION OF DESIGN FOR OPTIMIZATION:
Selection of design depends on the objective of experiment and number of factors to be investigated.
Types of objectives:
1. Comparative
2. Screening
3. Response Surface Method
4. Optimizing responses when factors are proportions of a mixture objective
5. Optimal fitting of a regression model objective
11. REFERENCES:
1. Antony j. , “Screening design”, sciencedirect, published by Elsevier second
edition, 2014, PP: 1-2.
2. Jankovic A. & et.al., Designing the design of experiments (DOE) – An
investigation on the influence of different factorial designs on the
characterization of complex systems, 2021, published by Elsevier, PP: 1-
10.
3. Plant and soil e science library, “Intro to statistics”, PP:1-7