“A statistical hypothesis is an assumption about a population parameter which may or may not be true.”
Dissertation research usually begins with hunches, guesses and questions which are to be tested. The research hypothesis states one’s expectation in positive manner. The best way to test whether statistical hypothesis is true or not would be by examining the entire population, since it is impractical researchers simply examine it through random sample population.
The main steps involved in hypothesis testing are:
- The first step is to state the Null and Alternative hypothesis clearly, where Null hypothesis (H0) signifies that sample observations results are obtained purely by chance an Alternative hypothesis (H1) signifies that the sample observations results are obtained by non-random causes.
Symbolically Hypothesis can be expressed by following equations:
H0: P = 0.5
H1: P ≠ 0.5
(Where, H0 is Null hypothesis and H1 is alternative hypothesis, P is the probability, at 95% confidence level the value of α should be less than equal to 0.05 for the alternative hypothesis to be true)
- The second step is to determine the test size. In this step the researcher would decide whether the test should be one-tailed or two-tailed to determine right critical value and rejection region.
- One Tailed Test: A test where, the region of rejection is only on one side of sampling distribution
- Two Tailed Test: A test where region of rejection is on both sides of sampling distribution.
- Significance level: Researchers can choose any value between 0 and 1 but significance levels equal to 0.01, 0.05 and 0.10 are most commonly chosen.
- Test Method: Selection of test method depends on test statistic and sampling distribution. Methods most commonly used are; mean score, difference between means, ANOVA, Chi-Square, Cross tabulations, z-score, t-score, etc.
- Type I Error: When one rejects a true null hypothesis it is called type I error.
- Type II Error: When one accepts incorrect null hypothesis it is called type II error.