11.4. The p-Chart
The p-chart is used to study classification type attribute (nominal scaled) data; for example, the proportion of nonconforming items by month. Subgroup sizes in a p-chart may remain constant or may vary. The p-chart assumes that:
There are only two possible outcomes for an event. An item must be classified as either conforming or nonconforming.
The probability, p, of a nonconforming item is constant over time.
Successive items are independent over time.
Subgroups are of sufficient size to detect an out-of-control event. A general rule for subgroup size for a p-chart is that the subgroup size should be large enough to detect a special cause of variation if it exists. Frequently, subgroup sizes are between 50 and 500 per subgroup, ...