About this lesson
Effective hypothesis testing is a disciplined process. From writing the process, to designing the study or experiments, and finally analyzing the data, there are proven best practices that should be applied. This lesson presents and explains the hypothesis testing process as used in Lean Six Sigma.
Hypothesis Test Process
While there are many different statistical hypothesis tests, the process of hypothesis testing is simple and straightforward. This lesson covers the steps of the hypothesis testing process.
When to use
Hypothesis testing is used in Lean Six Sigma projects during the Analyze phase and the Improve phase. During the analyze phase it is used to identify root causes and understand relationships between process variables and other factors. During the Improve phase it is used to demonstrate the effectiveness of a proposed solution to the problem.
The Hypothesis Testing process is a five-step process:
- Define the problem or issue being investigated and state the goals. This clarifies the issue you are investigating.
- Form the Null Hypothesis (Ho) and the Alternative Hypothesis (Ha). These two are always created in pairs and are essentially opposite statements. The Null hypothesis always states that the variable or factor being investigated is not impacting the process in a statistically significant manner. The Alternative hypothesis states the variable or factor does impact the process. The Null and Alternative hypothesis should be written so that it is clear the population being considered and the specific variables to be analyzed.
- Collect data for analysis. Based upon the elements of the hypothesis. It should be clear what data sets are required. Often the data exists within the data collected during the Measure phase, but there may be additional factors needed for the analysis.
- Calculate the Hypothesis Test statics. Depending upon the nature of the data and the goal of the question, different hypothesis tests will be used. All will determine if there is a statistically significant impact on the process based upon the condition being studied.
- Based upon what the evidence suggests, decide:
- Reject Null Hypothesis Ho
- Fail to reject Null Hypothesis Ho
Based upon the conclusion, the Hypothesis Testing process may need to be repeated with a different set of variables or factors.
Hints & tips
- Be sure your purpose is clear. This will make writing the hypotheses easier.
- If you need to collect more data, be sure your measurement system is adequate. You may need to do a measurement systems analysis to demonstrate that you trust the data.
- The decision is not to accept the Alternative hypothesis, it is a decision to reject the Null hypothesis. A poorly written Null and Alternative hypothesis may lead to the rejection of the Null, but since there are multiple other options, you cannot say with confidence that rejection of the Null leads to acceptance of the Alternative. They may both be false.
Lesson notes are only available for subscribers.