A statistical hypothesis test is a method of making statistical decisions from and about experimental data. Hypothesis testing is a part of inferential statistics which allows us to observe the behavior of samples and subsequently predict the same for the population. We assume samples to be apt representatives of the entire population.
In this we focus on two things:
The first one being: If something has happened did it happen due to chance or something else?
The second one is deciding between the two hypotheses that are mutually exclusive on the basis of evidence from observations.
A specific vaccine is known to be only 25% effective for a period of 2 years. Now a new type of vaccine is being invented and tested on 2000 people chosen at random. If you would like to test if this new vaccine is more effective than the previous one, how will you do it?
In the previous example a decision has to be taken based on the numbers in a sample. These numbers are subject to uncertainty and you’ll have to decide if the differences that you observe are only due to chance or not.
Let’s understand Hypothesis in the plight of Analytics with a simple example:
A bill is proposed which would simplify the tax code. The proposer claims that the proposed bill is `revenue-neutral' i.e. it will not lower tax revenues. An experiment is done using 100 tax returns chosen at random and the differences between the tax paid using the old rules and new rules are noted. The average difference comes up to be -$219 with a standard deviation of $725. So the question arises can you claim that the new rules are revenue neutral?
To solve this we can put the problem in these terms: there are two hypotheses:
NULL HYPOTHESIS, H0
and ALTERNATIVE HYPOTHESIS,HA
Under the null hypothesis there is no difference in revenue and the fact that the observed value is not 0 is totally due to chance. Under the alternative hypothesis the difference is real.
So likewise if we consider the vaccine case above, H0
is the proportion 25% and HA
is the proportion higher than 25%.
Now that Null Hypothesis and Alternative Hypothesis concept is clear, so let’s understand the basic Steps in Hypothesis Testing are:
• Null Hypothesis (H0
): The hypothesis that the researcher wants to reject
• Alternate Hypothesis(HA
): The hypothesis which is concluded if there is sufficient evidence to reject null hypothesis
• Test Statistic
• Rejection/Critical Region
From the above example we can understand the importance of Hypothesis: The method in which we select samples to know more about population is called as Hypothesis Testing. It is quite a systematic way of testing claims or ideas about a population or a group. Also, called as Significance testing, we test some hypothesis by determining likelihood that a sample statistic could have been selected, if the hypothesis regarding the population parameter were true.The basic purpose of carrying out the testing is to determine whether there is enough evidence to reject a hypothesis about a process. Or in other words, it provides us a mechanism for making quantitative decisions about a process.
So technically, a result will be statistically significant if it is unlikely that it occurred by a matter of chance rather according to a pre-determined probability, also known as the significance level.
• It can justify conclusions even when no scientific explanation exists.
• It is a well-accepted statistical tool in some experimental social sciences.
• It can be used as a substitute for the traditional comparison of the predicted value and experimental results at the core of the scientific method.
Have a look at our detailed course materials which includes theories, examples and lots more to choose from to help you understand the concept with in-depth knowledge of this topic. Write to us at firstname.lastname@example.org
to gain more knowledge on this.