Control-chart selection

Learning objectives

• Explain the purpose of control-chart selection in a garment-factory improvement project.
• Apply the described method to a representative shop-floor situation.
• Recognize the common mistakes and how to avoid them.

Concept

Chart selection depends on data type, subgroup structure, sample size, event opportunity, and monitoring purpose.

Continuous subgroups → Xbar-R/S; individuals → I-MR; defectives → p/np; defects → c/u.

Control limits describe process behavior; specifications describe customer requirements — never conflate them.

Variables, units, and assumptions

Data type (continuous/attribute), subgroup size n, opportunity constant/variable.

Stable measurement, rational subgroups, independent samples.

Free-spreadsheet workflow

Free spreadsheet: enter values in time order, compute mean and MR/R/S, calculate 3-sigma limits from constants, plot.

Interpretation and limitations

Points outside limits or nonrandom patterns (runs, trends, shifts) indicate special causes.

Small baselines produce wide limits; violated independence inflates false signals.

Garment-factory example

Use p chart for varying inspected quantities and I-MR for one daily shrinkage result on a swatch.

Method

1. Operationally define the measure.
2. Confirm gauge.
3. Determine rational subgroup.
4. Choose chart.
5. Establish baseline.
6. Investigate special causes before freezing limits.

Common mistakes

• Using specification limits as control limits.
• Selecting chart by software default without examining data type.