Please enable JavaScript.
Coggle requires JavaScript to display documents.
Chapter 21: Hypothesis Testing - Coggle Diagram
Chapter 21: Hypothesis Testing
Hypothesis Test Basics
Testing whether the data you have fits a data model. In Chapter 18, we conducted a hypothesis test to determine whether data fit the normal curve. We used the Chi-Squared Goodness-of-Fit Test, but ultimately, it was a hypothesis test.
Comparing a statistic to a hypothesis about the data or population
Answering the question whether something changed within the data, often after a team has modified an input or other part of the process. In the case of most Six Sigma projects, the team probably wants to find out whether the process or outcome is improved.
Selecting the Right Hypothesis Test
To determine which hypothesis test to run, you must know:
• Which type of data you have (continuous/variable or discrete/attribute)
• The number of levels of interest for the input in question (1, 2, or more than two)
• Distribution of data (normal or non-normal)
• What you are testing (means, medians, variance, count, or proportions)
Hypothesis Tests for Discrete Data
Hypothesis Tests for Continuous Normal Data
Why Run Hypothesis Tests
Six Sigma teams can’t just answer the question “Is this number different?” They must answer the question “Is this number so statistically different that we can take action on this information?”
Running Hypothesis Tests
State the null and alternative hypotheses. The null hypothesis is always written from the perspective that no change or difference occurs or is present. The alternative hypothesis is always written from the perspective that a change or difference is present – either not equal, greater than, or less than.
Set the confidence level for alpha. Usually, the confidence level is set at 95 percent for alpha, but other common settings are 99 and 99.9 percent. A confidence level of 95 percent means an alpha value of 0.05; a confidence interval level of 99 percent means an alpha value of 0.01; alpha value of 0.05; a confidence interval level of 99 percent means an alpha value of 0.01; a confidence interval of 99.9 percent means an alpha value of 0.001.
Decide which hypothesis test you are going to use. Use the information in the above section to find the appropriate test based on:
a. The type of data you have
b. The statistic you are dealing with (mean, variation, etc.)
c. How many sets of data you have (level of interest in x)
d. Remember, testing more than two sets of data is covered in Unit 7 on Design of Experiments and ANOVA
Decide whether your sample size is fixed or whether you can select a sample sized based on your beta setting. Setting sample size is covered in Chapter 22.
Run the test in Minitab. Specific steps for each hypothesis test – as well as other Minitab functions – are covered in Unit 6.
Interpret the p-value against your alpha setting, which tells you whether or not to reject the null hypothesis.
Translate the statistical analysis into real-world, business-relevant language.
A Review of Hypothesis Testing Terms
