Steps Involved in Testing Statistical Hypotheses
In testing statistical hypothesis we always beging by generating a research
question to answer. The question should be one which is testable and which
is related to some population parameter.
1. State the statistical hypothesis to be assumed true.
- a. Referred to many times as the null hypothesis.
- b. Null hypothesis means the hypothesis of no difference.
- c. Basis for all inferential statistical tests.
- d. Hypothesize a fact related to a population parameter.
2. Specify the level of significance to use.
- a. Definition: Level of significance refers to the degree of improbability
which is deemed necessary to cast sufficient doubt upon the truth of the
hypothesis to warrant its rejection.
- b. Synonymous terms --
- (1) alpha - level
- (2) Significance level
- (3) Confidence level
- c. Stated In terms of some probability level
- (1) .05
- (2) .01
- (3) .001
- (4) Sometimes .10
- d. If hypothesis is rejected means that the value of the statistic
must be sufficiently improbable to do this.
3. Specify the critical region to be used.
- a. Definition: A critical region is a portion of the scale of possible
values of the statistic so chosen that if the particular obtained value
of the statistic falls within it, rejection of the hypothesis is indicated.
- b. Criteria for choosing critical region or "region of rejection."
- (1) It must be consistent with the level of significance adopted.
- (2) Should be so located that if the hypothesis is not true the
probability of the test statistic falls within it is maximum.
4. Carry out the sampling study as planned and compute the value of
the test statistic.
5. Refer the value of the test statistic as obtained in Step #4 to the critical
region adopted. If the null hypothesis falls in the region, reject the hypothesis.
Otherwise, retain or accept the hypothesis as a tenable possibility.
Return to Statistical Methods Index Page
Return to the Advanced Statistics Index Page