Process capability (Cp, Cpk, Pp, Ppk, Cpm)
Compares the spread and centring of a process with its specification limits. Capability indices (Cp, Cpk, Cpm) use the within-subgroup sigma as SPSS SPCHART does (R-bar/d2, S-bar/c4, or moving range for individuals); performance indices (Pp, Ppk) use the overall standard deviation. The histogram shows both normal curves and the limits.
Menu: Graphs → More charts → Process capability…
When should I use it?
- Showing whether a stable process meets its tolerance.
- Reporting Cpk to a customer or auditor.
What are the assumptions?
- The process is in statistical control (check a control chart first).
- Measurements roughly normal.
How do I run it in StatLab?
- Choose the measurement and (if any) the subgroup variable.
- Type the lower and/or upper specification limit and, optionally, a target.
How do I read the output?
| Output | Meaning |
|---|---|
| Cp / Pp | Tolerance width ÷ 6σ: potential capability. |
| Cpk / Ppk | Distance from the mean to the nearer limit ÷ 3σ: actual capability (≥ 1.33 is a common requirement). |
| Cpm | Penalises a mean away from the target. |
What are common mistakes?
- Computing capability on a process that is out of control.
- Confusing Cpk (within sigma) with Ppk (overall sigma).