Design and Analysis of experiments

Chapter-one
1. Introduction
1.1 Strategy of experimentation
Experiment: A test or series of tests in which purposeful changes
are made to the input variables of a process (system) so that we may
observe and identify the reasons for changes that may be observed in the
output response.
Experimentation plays an important rule in product realization activities,
which consists of new product design and formulation, manufacturing
process development and process improvement. The objective in many
cases may be to develop a process affected minimally by external source
of variability
Strategy of experimentation is the general approach to planning
,conducting experiment and analyzing the data so that valid and objective
conclusions are obtained.
1.2 Typical application of experimental design
Suppose that a metallurgical engineer is interested in studying the
effect of two different hardening process, oil quenching and salt water
quenching on an aluminum alloy. The objective is to determine which
solution produces the maximum hardness of this alloy .The engineer
decides to subject a number of alloy specimens to each quenching media
and measure the hardness of the specimens after quenching .The average
hardness of the specimens treated in each quenching solution will be used
to determine which solution is best. In this experiment, a number of
important questions come to mind,

1. Are these two solutions the only quenching media of potential

interest?
2. Are there any other factors that may affect hardness that should be
investigated or controlled in this experiment?( Temperature of the
media,…)
3. How many alloy specimens should be tested in each quenching
solution?
4. How should alloy specimens be assigned to the quenching solutions ,
and in what order should the data be collected?
5. What method of analysis should be used?
6. What difference in average observed hardness between the two
quenching media will be considered important?

1
All of these questions and perhaps many others will have to be answered
before the experiment is performed.
In any experiment, the results and the conclusions that can be drawn
depend on the large extent on the manner in which the data were
collected.
Example
Suppose in the above experiment, the engineer used specimens from one
heat In oil quench and specimens from the second heat in the salt water
quench .When the mean hardness is compared, the engineer is unable to
say how much of the observed difference is the result of the quenching
media and how much is the result of inherent differences between the
heats.
In general, experiments are used to study the performance of the process
and systems .The process is combination of machines, methods, people
and any other resource that transforms some inputs into an output that
has one or more observable response variable.
Some of the process variables and material properties x1 ,x2 ,…,xp are
controllable ,where as the other variables z1 , z2,…,zq are
uncontrollable(although they may be controllable for the purpose of a
test)
The objectives of the experiment may be to determine
1. Which variables are most influential on the response(Y)
2. Where to set the influential x’s so that the response variable Y is
almost always near the desired level.
3. Where to set the influential x’s so that the effects of the
uncontrollable variables are minimized.
General model of a process or system

Controllable factors

Process
inputs Output(Y)

Uncontrollable factors

You can understand each objective of the experiments from the forgoing
discussion with examples.

2
 1.3 Basic Principles
Statistical Design of Experiment: refers to the process of planning
the experiment so that appropriate data that can be analyzed by
statistical methods will be collected, resulting in valid and objective
conclusions.
The statistical approach to experimental design is necessary if we wish to
draw meaningful conclusions from the data. When the problem involves
data that are subject to experimental errors, statistical methodology is
the only objective approach to analysis. Thus, there are two aspects of
any experimental problem: the design of experiment and the statistical
analysis of the data. The two subjects are closely related because the
method of analysis depends directly on the design employed.
The three basic principles of experimental design are replication,
randomization and blocking.
a) Replication
It is an independent repeat of each factor combination. In
metallurgical experiment above replication will consist of treating a
specimen by oil quenching and treating a specimen by salt water
quenching. Thus, if five specimens are treated in each quenching medium
, five replicates have been obtained .
The replicates allow the experimenter to obtain an estimate of
experimental errors. This estimate of errors becomes a basic unit of
measurement for determining the observed differences in the data are
really statistically different.
If the sample mean is used to estimate the effect of a factor in the
experiment, replication permits the experimenter to obtain more precise
estimate of this effect.
b) Randomization
It is the corner stone to use statistical methods in experimental
design .By randomization we mean that both the allocation of the
experimental materials and the order in which the individual runs or
trials of the experiment are to be performed are randomly determined
.Statistical methods require that the observations or errors be
independently distributed random variables .Randomization usually
makes this assumption valid. By properly randomizing the experiment, we
also assist in averaging out the effects of extraneous factors that may be
present.
Example:
Suppose in the hardness experiment the specimens differ slightly in
thickness and the effectiveness of quenching medium may be affected by

3
specimen’s thickness. If all the specimens subject to the oil quench are
thicker than these subject to salt water quench, we may introduce
systematic bias into the experimental result. This bias handicaps one of
the media and consequently invalidates our result .Randomly assigning
the specimens to the quenching media alleviate this problem.
Randomization may be complete or restricted depending on the
types of design.
C) Blocking:
It is the design technique used to improve the precision with which
comparisons among the factors of interest are made. Often blocking is
used to reduce or eliminate the variability transmitted from nuisance
factors (factors that may influence the response but in which we are not
directly interested).For example , an experiment in a chemical process
may require two batches of raw materials to make all the required runs.
However, there could be difference between batches due to supplier to
supplier variability, if we are not specifically interested in this effect ,
the batches of raw materials are nuisance factors .
Block is a set of relatively homogeneous experimental conditions. In the
chemical process example , each batch of raw materials would form a
block, because the variability within a batch would be expected to be
smaller than the variability between bathes .Typically, as this example,
each level of the nuisance factors becomes a block. Then the
experimenter divides the observations from statistical design into groups
that are runs in each block.
1.4 Guidelines for Designing Experiments
To use the statistical approach in designing and analyzing an experiment,
it is necessary to have a clear idea in advance ,what is to be studied ,how
the data are to be collected, and at least qualitative understanding of how
these data are analyzed. An outline of the recommended procedures is
given as follows:

1. Recognition of and Statement of the problem

2. Choice of factors , levels and ranges. Pre-experimental planning
3. Selection of the response variable.
4. Choice of experimental design.
5. Performing the experiment.
6. Statistical analysis of data.
7. Conclusion and Recommendation
__________________________________________________________________
In practice, step 2 and 3 are done simultaneously or in reverse order

4
 1. Recognition of and Statement of the problem
A clear statement of the problem contributes substantially to better
understanding of the phenomena being studied and the final solution of
the problem .It is also important to keep the overall objective in mind:
There are many possible objectives of an experiment including
confirmation (is the system performing the same way that it did in the
past?), discovery (what happens if we explore new materials, variables,
operating condition, etc..) , and stability (under what conditions the
response variables of interest seriously degrade?) Obviously, the specific
questions to be addressed in the experiment relate directly to the overall
objectives.
2. Choice of factors, levels and ranges
When considering the factors that may influence the performance of
processes or system, the experimenter usually discovers that these
factors can be classified as either potential design factors or nuisance
factors.
 The potential design factors are these factors that the
experimenter may wish to vary in the experiment. Often we find
that there are a lot of potential design factors, Some useful
classifications are
 Design factors
 Held-constant factors
 Allow to vary factors
The design factors are the factors actually selected for the study in
the experiment.
Held-constant factors are variables that may exert some effect on
the response, but for the purpose of the present experiment these factors
are not of interest so they will be held at specific level.
Allowed –to-vary factors: They are factors which are allowed to vary
assuming the effects are relatively small.
As example, the experimental units or the materials to which the design
factors are applied are usually non homogeneous, yet we often ignore this
unit to unit variability and rely on randomization to balance out any
material or experimental unit effect.
 Nuisance factors: are factors that may have large effect that must
be accounted for yet we may not be interested in them in the
context of the present experiment. Nuisance factors are classified as
 Controllable
 Uncontrollable
 Noise factors

5
A controllable nuisance factor is one whose levels may be set by the
experimenter. For example, the experiment can select different batches of
raw materials or different days of the week when conducting the
experiment. The blocking principle is useful in dealing with controllable
nuisance factors.
If the nuisance factor is uncontrollable in the experiment, but it can be
measured, analysis of covariance can often be used to compensate for its
effect. For example , the relative humidity in the process environment
may affect process performance, and if humidity cannot be controlled ,it
probably can be measured and treated as a covariate .
When a factor that varies naturally and uncontrollably in the process can
be controlled for the purpose of an experiment we call a noise factor. In
such situations, our objective is usually to find the settings of the
controllable design factors ,that minimize the variability transmitted
from the noise factors.
3. Selection of the response variable.
In selecting the response variable, the experimenter should be certain
that this variable really provides useful information about the process
under study. Most, often the average or standard deviation (or both) of the
measured characteristics will be the response variable.
It is usually critically important to identify issues related to defining the
response of interest and how they are to be measured before conducting
the experiment.
4. Choice of experimental design.
Choice of design involves the consideration of sample size ( no.of
replicates), the selection of a suitable run order for the experimental
trials ,and the determination of whether or not blocking or other
randomization restrictions are involved. There are also software packages
that support this phase of experimental design. The experimenter can
enter the information about the number of factors, levels , and ranges
and these programs either present a selection of designs or recommend a
particular design.
5. Performing the experiment
When running the experiment , it is vital to monitor the process
carefully to ensure that everything is done according to plan. Prior to
conducting the experiment a few trial runs or pilot runs are often helpful.
These runs provide information about consistency of experimental
material, a check on measurement system, a rough idea of experimental
error, and a chance to practice the overall experimental techniques.

6
 6. Statistical analysis of data.
Statistical methods should be used to analyze the data so that the results
and conclusions are objective rather than judgment in nature. If the
experiment has been designed correctly and if it has been performed
according to the design, the statistical methods required are not
complicated. Because of the questions that the experimenter wants to
answer can be cast into an hypothesis-testing frame work, hypothesis
testing and confidence interval estimation procedure are very useful in
analyzing data from a designed experiment.
It is also helpful to fit the empirical model, an equation derived from the
data that express the relationship between the response and the
important design factors.
Residual analysis and model adequacy checking are also important
analysis techniques.
Statistical methods cannot prove that a factor has a particular effect.
They only provide guidelines as to the reliability and validity of the
results. Properly applied .statistical methods do not allow anything to be
proved experimentally, but they do allow us to measure the likely error in
a conclusion or to attach a level of confidence to a statement.
7. Conclusion and Recommendation
Once the data have been analyzed, the experimenter must draw practical
conclusion about the results and recommend a course of action.