Home / Statistical Tools / Control Charts / Levey-Jennings / Levey-Jennings How-To
Levey-Jennings How-To¶
This walkthrough builds a complete Levey-Jennings chart from twenty-five daily control results. One of them is out of control, and this chart reaches that verdict differently from every other control chart in Quantum XL, which is what makes it worth walking through.
The data is on this page rather than in a file to download. Press Copy for Excel, then paste it into a blank worksheet.
The data¶
Twenty-five consecutive days of a quality control material, measured once each day, in millimoles per litre. One value per day, so there is nothing to subgroup.
| Control Result |
|---|
| 5.19 |
| 5.22 |
| 5.01 |
| 4.89 |
| 4.84 |
| 5.00 |
| 4.85 |
| 4.78 |
| 5.03 |
| 5.02 |
| 5.08 |
| 4.86 |
| 5.00 |
| 4.99 |
| 4.77 |
| 5.08 |
| 5.05 |
| 5.36 |
| 5.03 |
| 6.13 |
| 5.18 |
| 5.03 |
| 5.14 |
| 4.95 |
| 5.03 |
Day 20 is the one to watch. Every other result sits between 4.77 and 5.36; that one is 6.13.
Steps¶
-
Put the data in Excel
Press Copy for Excel above the table. In Excel, open a blank worksheet, click cell A1, and press Ctrl+V. The header lands in A1 and the twenty-five results in A2:A26.
-
Start the chart
From the Excel ribbon: QXL Stat Tools Tab > Control Charts > Variables > Charts as Individuals > Levey-Jennings.
-
Give it the cells
In the Data Selection window set Selected Range: to A1:A26. Leave Data in Columns selected and First Row/Column is Header checked. Press Next >.
-
Check the measurement column
The Levey-Jennings Chart dialog opens on the Data tab. Check Control Result.
There is no Estimate Sigma group on this dialog
Every other individuals-form chart in Quantum XL asks you how to estimate the standard deviation. This one does not ask, because it does not offer a choice: the limits are always built from the overall standard deviation of your data. There is no Rbar, no Sbar, no pooled estimate, no moving range option, and no unbiased-constants checkbox anywhere in this dialog.
There is also no Transformation group and no Capability tab. This chart offers neither.
-
Finish
Press Finish. The chart is written to a new worksheet named Levey-Jennings, with no "Chart" on the end, and the Control Chart analysis pane opens on the right.
What you should see¶
One panel, titled Levey-Jennings Chart, with twenty-five points and straight limits. There is no second panel: no moving range, no range, no standard deviation.
| Value | |
|---|---|
| Center line (mean) | 5.06 |
| Upper control limit | 5.85 |
| Lower control limit | 4.27 |
| 2 sigma lines | 4.54 and 5.59 |
| 1 sigma lines | 4.80 and 5.32 |
One point out of control. Day 20 plots at 6.13, above the upper control limit of 5.85, and is drawn in a different color. It is the only marked point.
The limits are straight, not stepped. One standard deviation serves the whole chart and every point carries the same weight, so there is no staircase of the kind a subgrouped chart produces.
The summary table has three rows you will not see on other charts: Mean, Sigma (overall), and CV%, the standard deviation as a percentage of the mean. Here that is about 5.2%. There is no within-subgroup sigma row, because this chart never forms one.
This chart's outlier inflates its own limits
The standard deviation here is the overall one, computed from every result about the mean, and day 20 is one of those results. Including it roughly doubles the standard deviation, from about 0.14 to 0.26, and that pushes the upper control limit out from about 5.44 to 5.85.
Day 20 still signals, comfortably. But the effect is real and it is the opposite of how a chart with a within-subgroup estimate behaves, where a single wild value barely moves the limits at all. On this chart a large enough upset can widen the band until it stops flagging itself. Use Estimate Limits From to build the limits from a known-good stretch when that matters.
Which test flagged the point
This chart runs the six Westgard tests rather than the eight Nelson ones, and all six are on by default: 1-2s, 1-3s, 2-2s, R-4s, 4-1s, and 10x. Day 20 satisfies two of them at once, 1-3s for being beyond the control limits and 1-2s for being beyond 2 sigma. Quantum XL does not report which test marked a point, so the chart simply shows one point in a different color. See Out-of-Control tests.
Things to try next¶
- Build the limits without the bad day. Set Estimate Limits From to the first 19 points. The limits tighten to roughly 5.44 at the top and day 20 lands much further outside.
- Watch R-4s fire. This test is offered on no other chart. It needs a consecutive pair straddling the band: one point beyond plus 2 sigma and the next beyond minus 2 sigma. Edit two neighbouring values to opposite extremes and both get marked.
- Supply a historical mean and sigma. The Historical Values group lets you enter the values from a previous lot instead of estimating them from these twenty-five results, which is closer to how a laboratory usually runs this chart. Note that manual limits and historical values are mutually exclusive here: entering one greys out the other.