Chaughule, Thorat - Statistical Analysis/Design of Experimentsone another in any systematic fashion and are as alike as possible on all characteristics
that might affect the response. While there is inherent random variation in all experi-
mental units, the ability to detect the effects of important factor and to estimate these
effects with satisfactory precision depends on the degree of homogeneity among the ex-
perimental units.
If all the responses for one level of a factor are taken from experimental units that
are produced by one manufacturer and all the responses for another level of the factor
are taken from experimental units produced by a second manufacturer, any differences
noted in the responses could be due to the different levels of the factor, due to the differ-
ences in design from two manufacturers, or to both. In this situation the effect of the fac-
tor is said to be confounded with the effect due to the manufacturers. Confounding vari-
ation can occur when using dryers from two different fabricators for same response.
When a satisfactory number of homogeneous experimental units cannot be obtained,
statistically designed experiments are often blocked so that homogeneous experimental
units receive each level of the factor(s). Blocking divides the total number of experimen-
tal units into two or more groups or blocks (e.g., raw materials) of homogeneous expe-
rimental units so that the units in each block are more homogeneous than the units in
different blocks. Factor levels are then assigned to the experimental units in each block.
The natural variation that occurs in a process, even when all conditions are maintained
at the same level, is termed as noise. When the effect of a particular factor on a process
is studied it becomes extremely important to distinguish the changes in the process
caused by the factor from noise.
The terms design and layout often are used interchangeably when referring to expe-
rimental designs. The layout or design of the experiment includes the choice of the fac-
tor-level combinations to be examined, the number of repeat tests or replications (if
any), blocking (if any), the assignment of the factor–level combinations to the experi-
mental units, and the sequencing of the test runs (Mason et al., 2003 ; Antony, 2003 &
Montgomery, 2003).
8.5. SELECTING A STATISTICAL DESIGN
To avoid various potential pitfalls during experimentation, several key factors
should be considered. Important design considerations for selecting a statistical design
are as follows.
8.5.1. Consideration of Objectives
Defining conceptual terms is prerequisite to understand and quantify outcome of
experiment. Concept definition and determination of observable variables influence
both the experimental design and the collection of information on uncontrollable factors.
Consider an example, in case of drying of carrots, it is prerequisite for experimenter to
be clear about response outcome such as colour retention, β carotene retention (prin-
cipal ingredient in Carrots) etc. To design such an experiment it is important to know
different physical properties of β carotene in order to retain maximum content. So con-
cept definition is important in order to select the experimental conditions for quantifia-
ble outcome. Consideration of the nature of anticipated conclusions can prevent unex-
pected complications when the experiment is finished and the result is being written.