In: Statistics and Probability
Take a moment to reflect on what you have learned in this module about hypothesis testing in general and A/B testing in particular.
- In what other scenarios or industries do you think this type of analysis would be helpful?
- What precautions should you take when designing such tests?
- How can you ensure that your results are representative of your target population?
- In what other scenarios or industries do you think this type of analysis would be helpful?
A testing hypothesis is a very important technique in statistical data analysis and we can use it for the process of decision making in various business industries, management, manufacturing industries, research, small-scale industries, service industry, government agencies, day to day life, insurance companies, etc. The technique of hypothesis testing is useful in almost all areas. It has a big scope and we can check different claims in different fields by using testing of hypothesis.
- What precautions should you take when designing such tests?
During the use of hypothesis tests, it is important to use proper test because there are so many identical or similar tests available for similar scenarios. So, selection of proper test is very important. Also, it is important to check all assumptions before using the hypothesis test. The sample size should be adequate. If the sample size is not adequate or it is small then we would get biased results.
- How can you ensure that your results are representative of your target population?
We will ensure that our results are representative of our target population by using a proper random sampling technique during the selection of sample observations. All types of biases should be avoided during the process of sampling. A selected random sample should be a good representative of the targeted population under study so that we can infer about given population. Also, sample size should be adequate as compared to population size. Approximate 10% of the population size is considered as the sample size for many research scenarios.