A hypothesis is an assumption or a claim about some
characteristic of a population, which we should be able to support or reject on the basis of empirical evidence. For example, an electrical bulb manufacturing company may claim that the average life of its bulb is at least 1000 hours.
Hypothesis testing is the process of testing between
different alternatives. The alternatives have to be mutually exclusive and exhaustive. Being mutually exhaustive means that there should not be any possibility of any other relationship between the parameters. In the example of the of the electric bulb manufacturer, the following two options will have to be considered to verify the manufacturer’s claim: Average life of the bulb is greater than or equal to 1000 Average life of the bulb is less than 1000 hours. We can see that these options are mutually exhaustive as well exclusive. Typically in hypothesis testing we have two options to choose from. These are term as null hypothesis And alternative hypothesis
Null hypothesis ( H0 ) is the presumption that is
accepted as correct unless there is a strong evidence against it. It represents the status quo Alternative hypothesis ( H1 ) is accepted when null hypothesis is rejected .it is the negation of the status qou situation. The negation of the null hypothesis also means that some corrective action is needed to ensure that e.g the average life of bulbs is at least 1000 hours. Significance Level ( P-value) P value is a statistical measure that helps scientists determine whether or not their hypotheses are correct. P values are used to determine whether the results of their experiment are within the normal range of values for the events being observed. Usually, if the P value of a data set is below a certain pre-determined amount (like, for instance, 0.05), scientists will reject the "null hypothesis" of their experiment - in other words, they'll rule out the hypothesis that the variables of their experiment had no meaningful effect on the results. There is always aprobabilistic component involved in the accept-reject decision in testing hypothesis. The criterion that is used in accepting or rejecting a null hypothesis is called the significance or the P-value P-value represents the probability of concluding (incorrectly)that there is a difference in your samples no true difference exists. It is a stastitics calculated by comparing the distribution of given samples data and an expected distribution ( normal,t ,F,etc) and is dependent upon the stastistical test being pperformed. e.g if two samples are being compared in a t- test ,a P-value of 0.05 means that there is only 5% chance that you would be wrong in concluding that the population are different or 95% confident of making a right decision.
One-tailed and Two- tailed Tests
A directional hypothesis is tested with One-tailed test
whereas a non directional hypothesis is tested with Two- tailed test. The following three relationships are only possible between any two parameters,
a) u1=u2
b) u1<u2
c) u1>u2
To be able to formulate mutually exclusive and exhaustive null and
alternative hypotheses from these relations we can choose either
b) or c) as alternative hypothesis and combine one of these two
with a) to formulate null hypothesis. Thus we have H0 and H1 as:
H0: u1,>=u2 OR u1,<=u2
H1: u1,<u2 OR u1,>u2 The above hypotheses are called directional hypotheses and one tailed tests are done for their analysis. If the null hypothesis is given by a) only and b) and c) are combined to formulate alternative hypothesis we have the following H0 and H1 as:
H0: u1, =u2
H1: u1, < u2 OR u1, > u2
The above hypotheses called non directional as we are only
concerned about the equality or non directional equality of a
relationship. non directional hypothesis non directional hypothesis non directional hypothesis directional hypothesis