Create a null hypothesis. The null hypothesis is best described as the results that is not covered by the mechanism you are testing or coincidental results. It is denoted by H0 (H sub 0). Usually the null hypothesis will contain equality or a lack of evident change. For example, let's say you had a car in your front yard. The car had been sitting there for over a year and you wanted to know if you should try to sell it. You don't know if it works, but since it sat idle for a long time, it was accepted that it did not run. However, the key to if you will sell it depends on if it is operational, since the market for dead cars is small.
The null hypothesis is:
H0: The car does not work.
Find an alternative hypothesis. The alternative hypothesis is what you sought to test. In our example, the alternative hypothesis is that the car is not operational. This is expressed as
H1: The car is operational.
After your hypotheses have been constructed, perform statistical analysis to find the proper decision. The decision is either that you reject the null hypothesis, or you fail to reject the null hypothesis, since conclusions are drawn always in reference to the null hypothesis.