What are the data requirements for a capability analysis?
- Distribution fit
- Your data must follow the assumed distribution for the analysis, such as a normal distribution (or nonnormal data transformed to fit a normal distribution) for Capability Analysis–Normal. The validity of the capability indices critically depends on the validity of the fit of the distribution used for the analysis. If your data do not follow the assumed distribution, the results will be inaccurate. If you are unsure which distribution best fits your data, use Individual Distribution Identification to identify an appropriate distribution or transformation.
- Process stability
- Your process must be stable and in control. A process is stable if it contains only variation from common causes, but no variation due to special causes.
Tools for determining whether a process meets the requirements for capability analysis
You can use the following tools in Minitab to assess whether your process satisfies the requirements for the analysis:
- Xbar or Individuals charts to determine whether the process is in control. If the process is out of control, the capability indices are invalid.
- R, S, or MR chart to track the variation in the data and assesses whether the process variation is acceptable.
- A run chart to look for evidence of patterns in your data.
- Probability plot to verify that a chosen distribution fits your data.
- Capability histogram and capability plot to visually compare the distribution of data from your process to the specification spread. It also includes capability statistics to assess capability of your process quantitatively.
You can use Minitab’s Capability Sixpack to evaluate the main requirements for capability analysis. This analysis includes control charts, a histogram, a probability plot, and a capability plot, as well as major capability indices. Chooseand select the normal, nonnormal, or between/within analysis.
Example of using Capability Sixpack to check assumptions
Source: Minitab Support files.