Chaughule, Thorat - Statistical Analysis/Design of Experimentsfrom many sources. Two general sources of importance to the statistical design of expe-
riments are experimental error and measurement error. Experimental error is intro-
duced whenever test conditions are changed. For example, dryer settings are not always
exact enough to be fixed at precisely the same value or in case of microwave vacuum
dryer, vacuum chamber pressure changes with respect to atmospheric temperature &
humidity if the chamber pressure is not set. Dehydration batches of same Sapota fruit do
not always show exactly the same drying nature and quality due to different initial com-
position. Measurement errors arise from the inability to obtain exactly the same mea-
surement on two successive test runs when all experimental conditions are unchanged.
8.2.3. Analysis Phase
A statistical analysis of the experimental results should allow inferences to be drawn
on the relationships between the design factors and the measurements. This analysis
should be based on both the statistical design and the model used to relate the mea-
surements to the sources of variation. If additional experimentation is necessary or de-
sirable, the analysis should guide the experimenter to an appropriate design and, if
needed, a more appropriate model of the measurement process (Mason et al., 2003).
Thus, the role of statistics in food engineering and scientific experimentation can be
described using three basic categories: project planning, experimental design, and data
analysis.
8.4. DIFFERENT TERMINOLOGIES USED IN EXPERIMENTAL DE-
SIGN
The words and phrases used in experimental design are not uniform across discip-
lines or even, in textbooks. For this reason statistical experimental design with a brief
definition of terms is compiled as below. Table 8.2 contains definitions of many terms
which are in common use.
A response variable is an outcome of an experiment. It may be a quantitative mea-
surement such as the percentage of moisture removed in batch drying process, the dry-
ing time or it may be a qualitative result such as extent of color changes in drying of cer-
tain food product for example. A factor is an experimental variable that is being investi-
gated to determine its effect on a response. It is important to understand that a factor is
controllable by the experimenter; that is, the values, or levels, of the factor can be de-
termined prior to the beginning of the test runs and can be executed as planned in the
experimental design. To give an example, in hot air drying of Garlic slices experiments, if
a response is retention of Allicin (the principle constituent of garlic), the factors would
be drying time, drying temperature and slice thickness then the possible levels for a fac-
tor drying temperature would be 40, 50 and 60°C. However, one will try to avoid higher
drying temperature as it may result in unacceptable product quality. An experimental
region, or factor space, consists of all possible levels of the factors that are considered
for inclusion in the design. For quantitative factors, the factor space is often defined by
lower and upper limits for the levels of each factor. Additional variables that may affect
the response but cannot be controlled in an experiment are called covariates. Covariates
are not additional responses; that is, their values are not affected by the factors in the
experiment. Rather, covariates and the experimental factors jointly influence the re-