Shewhart variables control charts

Variables control charts plot quality characteristics that are numerical (for example, weight, the diameter of a bearing, or temperature of the furnace).

There are two types of variables control charts: charts for data collected in subgroups, and charts for individual measurements. For subgrouped data, the points represent a statistic of subgroups such as the mean, range, or standard deviation. For individual measurements, the points represent the individual observations or a statistic such as the moving range between successive observations.

An important concept for data collected in subgroups is that of a rational subgroup. A rational subgroup is a number of units produced under the same set of conditions. The measurements in each subgroup should be taken within a short period and should minimize the chances of special-cause variation occurring between the observations. That is, it should represent a snapshot of the process and the inherent common-cause variation.

Subgroups are often preferable when learning about a process as they can better estimate the process variability or detect smaller shifts in the process average. However, when monitoring a process, if the amount of time taken to collect enough observations to form a subgroup is too long, or the cost of sampling is expensive or destructive then the benefits may be negated.