Home / Statistical Tools / Control Charts / Laney P' Chart / Options
Options¶
The Laney P' chart shares its dialog with the p chart and the np chart. One ribbon entry, P, NP, and Laney's P', opens a single window, and the Variant: radio row at the top selects which chart you get. Every control is documented once, on p Chart Options. This page lists only what is different when the variant is Laney P'.
What changes when you select Laney P'¶
| Window title | Laney P' Chart |
| Worksheet name | Laney P Chart, without the apostrophe, or Laney P Chart - {GroupName} under GroupBy |
| Y axis | Proportion Defective, the same as the p chart |
| Plotted statistic | defectives over sample size, the same as the p chart |
| Limits | every subgroup's standard deviation multiplied by \(\sigma_Z\) before the limits are placed |
Nothing is added to the dialog and nothing is removed from it. The variant changes what the engine does with the numbers, not what the window offers.
The controls whose meaning shifts on this variant¶
Diagnostics: sigma_Z stops being advice¶
On a p or np chart, Compute and report sigma_Z statistic produces a number for the summary table and nothing else. On this variant the same number is the limit calculation. It is on by default.
That has a consequence worth stating: with the box on, this chart cannot be built without \(\sigma_Z\), because \(\sigma_Z\) is not optional to it. If the estimate cannot be formed at all, from too few usable subgroups, the engine falls back to a factor of 1, and the chart you get is a p chart wearing a Laney P' title. Check the reported figure rather than assuming a correction was applied.
Force Straight Limits: inert here¶
| Variant | What an assumed n changes |
|---|---|
| P | nothing |
| NP | the center line and the limits |
| Laney P' | nothing |
The engine never substitutes an assumed sample size on this variant, so the box and its Assumed n: can be set, will be stored, and will change nothing. The limits keep stepping with the real sample sizes, just multiplied by \(\sigma_Z\).
Manual and None limit types: they bypass the correction¶
Under Manual, all three lines are the numbers you typed, so \(\sigma_Z\) plays no part and the chart is a Laney P' chart in name only. Under None, no limits are drawn at all and \(\sigma_Z\) is not computed. Both are the same on the p chart, but here the loss is larger, because the correction is the reason to be on this variant.
Split Control Limits: a sigma_Z per phase¶
Splitting re-estimates the correction inside each phase, so each phase carries its own \(\sigma_Z\) and its own band width, and the Per-Split Breakdown block reports them separately. If the overdispersion came from a step change partway through the record, a split can return two phases whose \(\sigma_Z\) values are both near 1.