Attribute Control Charts

The selection depends on two questions: Are you counting defective items (pass/fail) or individual defects? And is your sample size constant or variable? Defective items with variable samples → p-chart. Defective items with constant samples → np-chart. Defect counts with constant inspection area → c-chart. Defect counts with variable inspection area → u-chart.

For p and u charts, control limits recalculate for each subgroup based on its sample size, creating the characteristic staircase pattern. SigmaResolve displays the sample size alongside each point so you can verify the data. For extremely variable sample sizes, an average-based limit option is available with flagging when individual sample sizes deviate significantly.

When your data variance significantly exceeds theoretical Poisson or binomial expectations, SigmaResolve flags over-dispersion — a signal that standard attribute chart assumptions may not hold. This commonly occurs with batched inspection data or when defect rates vary by production condition. The system suggests investigating data stratification or considering Laney P'/U' charts (on the roadmap).