Chaughule, Thorat - Statistical Analysis/Design of Experimentsthe variables of interest cannot be controlled, but they can be recorded and analyzed.
Data collection is at the center of experimental and observational studies (Mason et al.,
2003 ; Antony, 2003; Cox & Reid, 2000).
Figure 8. 1. Important stages of statistical input in scientific investigations8 .2. STATISTICS IN EXPERIMENTATION
Statistical experimental design will be shown to be effective in eliminating known
sources of bias, guarding against unknown sources of bias, ensuring that the experiment
provides precise information about the responses of interest, and guaranteeing that ex-
cessive experimental resources are not needlessly wasted through the use of an uneco-
nomical design. Design of experiments in statistics works on assumption that, measure-
ments are subject to variation. The study of variation includes the construction of expe-
rimental designs and the development of models which describe variation. The term
measurement encompasses both qualitative responses as well as quantitative mea-
surements. In most industrial drying processes there are numerous sources of possible
variation. Frequently studies are conducted to investigate the causes of excessive varia-
tion. These studies could focus on a single source or simultaneously examine several
sources. Consider, for example, dehydration of strawberries on industrial scale that in-
volves procurement of raw materials from different sources and the dehydration which
is performed by several operators. Variation could occur because the operators syste-
matically differ in their method of operation. Variation also could occur because one or
more of the operators do not consistently adhere to the drying procedures, thereby in-
troducing uncontrolled variability to the measurement process. To investigate sources
of variability for a dehydration studies, an experiment should be statistically designed
and analyzed to ensure that relevant sources of variation could be identified and meas-
ured (Mason et al., 2003).
8. 3. ROLE OF STATISTICS IN EXPERIMENTATION
In general, there are three basic steps in the statistical design and analysis of expe-
riments as depicted in Table 8.1.
Table 8. 1. Function of Statistics in ExperimentationDefining Objectives Data^ Collection^ Conclusions^Experimental studiesObservational studies