PROC SHEWHART: Classification of Shewhart Charts

Classification of Shewhart Charts

Shewhart charts are broadly classified according to the type of data analyzed.

Shewhart charts for variables are used when the quality characteristic of a process is measured on a continuous scale.
Shewhart charts for attributes are used when the quality characteristic of a process is measured by counting the number of nonconformities (defects) in an item or the number of nonconforming (defective) items in a sample.

Shewhart charts for variables are further classified according to the subgroup summary statistic plotted on the chart.

$\text{[math]}$ and $\text{[math]}$ charts display subgroup means (averages) and ranges. Typically the two charts are presented on the same page, with the $\text{[math]}$ chart aligned above the $\text{[math]}$ chart to facilitate the simultaneous analysis of the central tendency and variability of the process.
$\text{[math]}$ and $\text{[math]}$ charts display subgroup means (averages) and standard deviations. Typically the two charts are presented on the same page, with the $\text{[math]}$ chart aligned above the $\text{[math]}$ chart.
Median and range charts display subgroup medians and ranges. Typically the two charts are presented on the same page, with the median chart aligned above the $\text{[math]}$ chart.
Charts for individual measurements and moving ranges display individual measurements and moving ranges of two of more successive measurements. In this case the subgroup sample consists of a single observation.

Likewise, Shewhart charts for attributes are classified according to the subgroup summary statistic plotted on the chart:

A $\text{[math]}$ chart displays the proportion of nonconforming (defective) items in a subgroup sample.
An $\text{[math]}$ chart displays the number of nonconforming (defective) items in a subgroup sample.
A $\text{[math]}$ chart displays the number of nonconformities (defects) per unit in a subgroup sample consisting of an arbitrary number of units.
A $\text{[math]}$ chart displays the number of nonconformities (defects) in a unit (here, a subgroup sample typically consists of one unit).

You can create all of the preceding types of Shewhart charts with the SHEWHART procedure. In addition, you can create a wide variety of nonstandard Shewhart charts, including

a trend chart displaying a time trend plot and an $\text{[math]}$ chart (or median chart) that has been created removing the time trend from the data. The trend chart and $\text{[math]}$ chart are presented on the same page, with the $\text{[math]}$ aligned above the trend chart, to facilitate the detection of special causes after accounting for the time trend effect. Trend charts are applicable when a time trend (for instance, due to tool wear) is observed in a preliminary $\text{[math]}$ chart of the original data.
a box chart displaying a box plot (box-and-whisker plot) for each subgroup and control limits for the subgroup means. This chart facilitates detailed analysis of the subgroup distributions and is applicable with large subgroup sample sizes (ten or more).