Page 273 - Area 9 - Relevant Document

20

6. Proper choice of an experimental design
The choice of the experimental design is largely dependent on the
variability of the experimental materials of units and nature of
treatment under test. The researcher should ha a prior knowledge of
the variability of the experimental units to be used. As much as
possible, the researcher should choose experimental units that are
homogeneous so that real differences among treatments can be
observed. If the units are somewhat homogeneous then, a completely
randomized design can be employed. However, selection of
homogeneous materials is difficult to achieve but there are ways to
reduce variability among experimental units and that is by blocking
then an appropriate design applicable to heterogeneous materials may
be used such as Randomized Complete Block Design, Latin Square
Design and other suitable deign.

TESTING STATISTICAL HYPOTHESIS/HYPOTHESES

Levels of significance
Where do we draw the line that will determine as to whether or not we can
make an inference about an experiment?

The answer to this question reveals the basic nature of sciences; its
probabilistic rather than its absolutistic orientation. Most researches have
adopted the following cutoff points as the basis for inferring the operation of
non-chance factors.

When the event or one more deviant would occur one percent of the
time or less, by chance, researchers are willing to assert that the results are
due to non-chance factors. This cutoff point is known variously as the 0.05
significance level or the 5.00 percent significance level. 0.01 when the event
or one more deviant would occur one percent of the time or less, by chance,
researchers are willing to assert that the results are due to non-chance
factors. This cutoff point is known variously as the 0.01 significance level or
the 1.00 percent significance level.

The level of significance set by the experimenter for inferring the
operations of non-chance factors is known as the alpha (a) level. Thus when
employing the 0.05 level significance, a=0.05, when employing the 0.01 level
of significance, a = 0.01.

NULL HYPOTHESIS AND ALTERNATIVE HYPOTHESIS

Prior to the beginning of any experiment, the researcher sets up two mutually
exclusive hypotheses;