REMINDER: This article is not a tutorial on applying and using statistics, but simply is a functional instruction on using the CESP documents themselves.
There is also Attribute 2 Sample Capability Analysis available (Before vs After capability comparison) as a separate document. It works very much the same way as this guide, so it is recommended you also use this guide when completing the 2 Sample statistics document. Note: 2 Sample does not contain Capability Sixpack but does contain all other tabs.
The Attribute Capability Analysis contains both Binomial and Poisson analysis:
- Binomial = Defectives (% Defective & PPM)
- Poisson = Defects (DPU & DPMO)
Which one is finally used is determined by a drop down box. Usage of the document asides this in either case is the same.
Figure 18 - Data Entry
The data entry tab includes all setup parameters as well as data considerations to be taken into account before using the capability analysis for robust inferences.
- Type: Binomial or Poisson. Select the relevant capability analysis to run based on your needs. When you change this, the instructions for all settings will change to guide you more effectively. For example, when entering a target, the instruction next to this box will change depending on whether you are running Binomial or Poisson. All graphical analysis will update to suit your selected type.
- Target: Select a target value for your process. This is % Defective for Binomial, or DPU for Poisson
- Title of Data: Enter a title to appear at the top of your capability analysis for easy identification.
- Data: On the right this represents the defectives/defectives in the process being counted. This is your data to be analysed.
- Sample Size: This represents the batch size or sample size that the defectives or defects were recorded in. This is your data to be analysed.
In the default example above, all sample sizes are varied, indicating that the data was NOT taken from a fixed size batch.
Up to 500 rows of data can be added if required. This can be typed in manually or copy/pasted from another source.
Once the data has been entry, select the Capability Analysis tab at the bottom of the page.
Figure 19 - Capability Pack
The Capability Pack tab will display the relevant control chart (P, Np, C, U) in the top left, a summary Histogram on the top right, a cumulative % Defective or DPU chart in the lower left, and statistical information on the lower right, including Mean, Standard Deviation, n value, PPM/DPMO value, and Total Defective or DPU values. There is also standard text, which will not change regardless of the results, that prompts the user to make deductions and ask questions of the results.
Select the Process Capability tab at the lower edge of the sheet.
Figure 20 - Process Capability Histogram
The histogram displayed will not have a “bell shaped curve” but will display the results in a basic histogram format. To the right is the statistical data that is obtained from the histogram itself. This is due to the data typically being non-normally distributed (Binomial for example).
CESP selects an optimal bin size and number of bins from multiple methods (Sturges, Freedman, Sqrt). You can override the number of Histogram bins if you so wish by entering a value in the top right yellow box. Deleting this value will revert back to optimal selection.
Finally, select the Control Charts tab at the bottom of the sheet.
Figure 21 - Control Charts
The tab will display a larger scale version of the relevant control chart first, but will also provide a summary of the Nelson’s Rules that have been applied to the data, and the overall result. If the data is shown to be out of control, refer to the table on the bottom of the tab as shown, to identify which data points have breached Nelson’s Rules, and which rule they have breached.
Figure 22 - Binomial Analysis - Nelson's Rules Breaches
This will indicate a FAIL next to each data point and what Nelson rules it infringes.
Full training regarding Nelson’s Rules will be provided by your Capella trainer in formal training sessions.