Page 274 - Area 9 - Relevant Document

21

1. The null hypothesis (Ho) which specifies hypothesized values for one
or more of the population parameters; and

2. The alternative hypothesis (Ha), which asserts that the population
parameter is some other value other than one hypothesized.

The alternative hypothesis may be either directional or non-directional.
When (Ha) asserts only that the population parameter is different from the one
hypothesized, it is referred to as a non-directional or two tailed hypothesis.

Occasionally Ha is directional or one tailed. In this instance, in addition to
asserting that the population parameter is different from the one
hypothesized, we assert the direction of that difference. In evaluating the
outcome of the experiment, one tailed probability values should be employed
whenever our alternative hypothesis is directional. Moreover, when the
alternative hypothesis is directional so also is the null hypothesis.

THE INDIRECT PROOF

The null can neither be proved nor can the alternative hypothesis be
directly proven. However, if we can reject the null hypothesis, we can assert
its alternative, namely that the population parameter is some value other than
the one hypothesized.

The support of the alternative hypothesis is always indirect. We support
it by rejecting the null hypothesis. On the other hand, since the alternative
hypothesis can neither be proved nor disproved directly, rejecting the
alternative hypothesis can never prove the null hypothesis. The strongest
statement that can be made in this respect is that we failed to reject the null
hypothesis.

From the above, the conditions for rejecting the null hypothesis are:

1. When employing the 0.05 level of significance, the null hypothesis is
rejected when a given result occurs, by chance, 5 percent of the time or less.

2. When employing the 0.01 level of significance, the null hypothesis is
rejected when a given result occurs, by chance, 1 percent of the time or less.

Under these circumstances, the alternative hypothesis is affirmed. In
other words, one rejects the null hypothesis when the results occur, by
chance, 5 percent of the time or less (or one percent of the time or less),
assuming that the null hypothesis is the true distribution. That is one
assumes that the null hypothesis is true, calculates the probability of this
assumption and if the probability is small, rejects the assumption.