Business Analytics - Hypothesis Testing
Examine practical applications of hypothesis testing.
Successful management is about making decisions to bring about desired change. A manager who has a specific objective will execute an action, hoping that the result will be the achievement of that objective. However, due to the multitude of factors that are in play in any real-life situation, it is difficult to tell if the desired outcome has been achieved or if the new numbers are only due to randomness.
For example, a sales manager implements a bonus system and sees the sales figures go up by 3% in the next month. She claims success, but was this really due to the bonus system or could the increase be the result of random fluctuations, something that would have happened anyway? Can the manager confidently claim that there has been a significant change, something that could not be the result of random fluctuations?
Another example may be where HR is worried there may be a difference between the salaries of men versus women that may need to be corrected. Is the difference between average salaries of the two genders significant, or could it be due to coincidence? Can someone claim there is discrimination at work?
Hypothesis testing answers such questions and is key to managing intelligently. Without these answers, you cannot be sure if the actions you are taking are working out or whether things really are the way you think they are. If you do not know these, how can you know what to do?
For this Discussion, describe a situation from your work experience where decisions are being made that would benefit from hypothesis testing. Follow the template below and answer all questions:
1. Describe the decision situation: What is the desired outcome or the question that needs to be answered, and what are the actions being considered? Without hypothesis testing, how are decisions being made currently?
2. How would you set up the hypothesis test? State what the null and alternative hypothesis would be, what kind of a test you would use.
3. Describe the data you would use: What do you need, how much do you need, where would it come from? Is all the data currently being recorded or will it need to be collected? If it needs to be collected, will it need to be sampled (why), or can the entire data be recorded? Please make sure to answer all these questions in your explanation.
4. How would you explain to management the significance of this test? How would you convince them that using this is necessary? What would be drawback of not using a hypothesis test in this situation?