Hypothesis testing is defined as the formal events that statisticians use to check whether a hypothesis can be accepted or rejected. A hypothesis is a supposition about something.
Hypothesis testing is an action in statistics whereby an analyst tests a supposition concerning a population limitation. The procedure used by the analyst pivot on the nature of the figures used and the aim for the analysis. Hypothesis testing is used to conclude the result of a hypothesis completed on sample data from a larger population.
The Hypothesis is a supposition which is tested to find whether the conclusion drawn from the sample of facts stand true for the whole population or not.
Steps involved in Hypothesis Testing Procedure
The following steps are involved in hypothesis testing procedure.
- Step 1: Create a hypothesis
- Step 2: Create a suitable significant level
- Step 3: Find a suitable test statistic
- Step 4: Find a critical region
- Step 5: Perform computation
- Step 6: Conclusion
Create a hypothesis
The very initial step is to create the hypothesis to be tested. The statistical hypothesis is a supposition about the worth of some mysterious parameter, and the hypothesis delivers some numerical value or collection of values for the parameter. In this step two hypotheses about the population are made Null Hypothesis and Alternative Hypothesis.
The Null Hypothesis represented by H0 declares that there is no true change between the sample of data and the population parameter and that the change is unintentional which is produced due to the variations in sampling. Therefore, a null hypothesis states that there is no alteration between the expected and actual value of the parameter.
The alternative hypothesis symbolized by H1 is the other hypothesis about the population, which is declared true if the null hypothesis is not true. Consequently, if we reject H0 then the alternative hypothesis H1 gets accepted.
Create a suitable significant level
As the hypothesis about the population is created the researcher has to choose the level of significance, i.e. an assurance level with which the null hypothesis is accepted or rejected. The significance level is represented by ‘α’ and is usually clear before the samples are pinched such that results gained do not affect the choice. We either take 5% or 1% level of significance.
If the 5% level of significance is chose, it means that there are five probabilities out of 100 that we will castoff the null hypothesis when it should have been established, i.e. we are about 95% self-assured that we have ended the right decision.
Find a suitable test statistic
After the hypothesis is built, and the significance level is certain upon, the next step is to define a suitable test statistic and its distribution.
Find a Critical region
Before the samples are pinched it must be certain that which values to the test statistic will result in to the acceptance of H0 and which will results into its rejection. The values that result to rejection of H0 are named as the critical region.
Once the critical region is recognized, we calculate numerous values for the random sample of size ‘n.’
When all the steps are performed and the arithmetical results are concluded now the analyst can draw decisions from the arithmetical results. The decision includes either the null hypothesis is true or not. The decision that the null hypothesis is true or not, depends on whether the computed value falls in the acceptance region or the rejection region.
November 10, 2018