1.5

Pd

INTRODUCTION TO BASIC CONCEPTS IN EXPERIMENTAL DESIGN

1. Does the design permit an experimenter lo calculate a valid estimate of
— the experimental eflects and error efiects?

2. Does the data-collection procedure produce reliable results?

3. Does the design provide maximum efficiency within the constraints
imposed by the experimental situation?

4. Does the design possess suficient power to permit an adeguate test of
the statistical hypotheses?

s. Does the experimental procedure conform to accepted practices and
procedures used in the research area? Other things being egual, an
experimenter should use procedures that offer an opportunity for
comparison of his findings with the results of other investigations.

A REVIEW OF STATISTICAL INFERENCE

In the previous section an overview of experimental designs was
presented. This section is written to accomplish the same goal with respect
to procedures involved in statistical inference. It is assumed that the reader
is already familiar with basic hypothesis-testing concepts. Hays (1963)
presents an excellent introduction to this topic.

A distinguishing characteristic of the scientific method is the for-
mulating and testing of hypotheses. The testing of hypotheses reguires
the a priori formulation of decision rules to guide the decision maker.
The problem may be stated: Given two mutually exclusive hypotheses
about a population, how does one decide on the basis of sample data which
hypothesis is supported? lt will be apparent that this guestion lacks a
simple answer.

A statistical hypothesis is a statement about one or more parameters
of population distributions; and, as such, it refers to a situation that might
be true. Such a statement is always made with respect to a population
and not to a sample. Distinguishing between statistical hypotheses and
research or scientific hypotheses is important. Research hypotheses are
normally stated in general terms, at least in the initial stages of an inguiry.
In this form they are not amenable to evaluation through the use of the
procedures and theory of statistical inference. It may be possible, by means
of deductive reasoning, to transform a research hypothesis into a statistical
hypothesis that can be subjected to test. Statistical hypotheses refer to
population parameters, whereas scientific hypotheses refer to the phenom-
ena of nature and man (Clark, 1963).

In logic, the terms direct statement and indirect statement are
analogous in many ways to statistical and scientific hypotheses. A direct
statement is made in reference to limited phenomena that are directly
observable; for example, «This rat is running." The truth or falsity of
such a direct statement can be determined by observing the rat. An indirect
statement refers to phenomena that cannot be directly observed or that
are so numerous in time that it is impossible to view them all. For example,