Rational Subgroups

The key to successful control charts is the formation of rational subgroups. Control
charts rely upon rational subgroups to estimate the short-term variation in the process.
This short-term variation is then used to predict the longer-term variation defined by
the control limits, which differentiate between common and special causes of variation.

A rational subgroup is simply a sample in which all of the items are produced under
conditions in which only random effects are responsible for the observed variation
(Nelson, Lloyd S. "Control Charts: Rational Subgroups and Effective Applications,"
Journal of Quality Technology. Vol. 20, No. 1, January 1988). Another way to express this
is that a rational subgroup is one in which the system of causes influencing within
subgroup variation is equivalent to the system of causes influencing between subgroup
variation.

A rational subgroup has the following properties:

1. The observations within a subgroup are from a single, stable process. If subgroups
contain the elements of multiple process streams, or if other special causes occur
frequently within subgroups, then the within subgroup variation will be large relative to
the variation between subgroup averages. This large within subgroup variation forces the
control limits to be too far apart, resulting in a lack of sensitivity to process shifts.
Western Electric Run Test 7 (15 successive points within one sigma of center line) is
helpful in detecting this condition.
2. The subgroups are formed from observations taken in a time-ordered sequence. In other
words, subgroups cannot be randomly formed from a set of data (or a box of parts);
instead, the data comprising a subgroup must be a "snapshot" of the process over a
small window of time, and the order of the subgroups would show how those snapshots
vary in time (like a "movie"). The size of the "small window of time" is determined on an
individual process basis to minimize the chance of a special cause occurring in the
subgroup.
3. The observations within the subgroups are independent, implying that no observation
influences, or results from, another. If observations are dependent on one another, the
process has autocorrelation (also known as serial correlation). In many cases, the
autocorrelation causes the within subgroup variation to be unnaturally small and a poor
predictor of the between subgroup variation. The small within subgroup variation forces
the control limits to be too narrow, resulting in frequent out of control conditions, leading
to the tampering discussed in Chapter 1. Many examples of autocorrelation exist in
business processes and nature itself including:
o Chemical Processes: Samples drawn from liquids in batches are often influenced by prior
samples if the time between the samples is small. The batch is essentially homogenous
throughout, and its properties change slowly so that the measurement at one point in
time is influenced by the state of the process in an earlier state of time.
o Service Processes: The wait time of a person in a queue is influenced by the wait time of
the person in front of him/her. Recall your wait time at the supermarket check out line
when you're behind a chatty customer: You cannot be serviced until their service is
completed; likewise for the person behind you, and so on.
o Discrete part manufacturing: Feedback controls used to adjust processes based upon
past observations cause autocorrelation in data observed closely in time.
 Rational subgroup sampling

The following is an excerpt from The Six Sigma Handbook: Fourth Edition (McGraw-Hill 2014) by Paul Keller and
Thomas Pyzdek.

The basis of all control charts is the rational subgroup. Rational subgroups are composed
of items which were produced under essentially the same conditions. The statistics, for
example, the average and range, are computed for each subgroup separately, then
plotted on the control chart. When possible, rational subgroups are formed by using
consecutive units. Each subgroup's statistics are compared to the control limits, and
patterns of variation between subgroups are analyzed. Note the sharp contrast between
this approach and the random sampling approach used for enumerative statistical
methods.

To understand the concept of rational subgrouping, it's useful to understand the role of
the subgroup in the control chart. The subgroup provides a snap shot of the process at
that moment in time. On an averages (aka X-bar) chart, for example, each subgroup
provides an estimate of the process location, while the range (or sigma) chart provides
the subgroup estimate of the short-term variation. The control limits on the averages
chart are defined based on the average within-subgroup variation (estimated using the
average within-subgroup range, for example), and when the short-term within- subgroup
variation predicts the longer-term between-subgroup variation, then the process is in
statistical control. When the longer-term variation is not predicted by the short-term
within- subgroup variation, then a special cause has been identified, or more precisely,
operationally defined.

A fundamental aspect of the subgroup is thus to estimate the common cause variation
within the process, since the within-subgroup variation is used to define the width of the
control limits. Therefore, it is critical that the causes of within- subgroup variation be
representative of the causes of variation between subgroups. This is not to say that each
and every subgroup must contain an identical subset of causes of process variation.
Rather, the system that defines the causes is consistent. Consider the following cases:

1. Measure the fill volume of each of the six bottles of beer filled by a multiple-head filling
station, grouping the six measurements into a subgroup. Take data for a new subgroup
(i.e., a different set of bottles from the same six heads) every hour.
2. Measure the thickness of an extruded plate. Since the plate thickness can vary across its
width, measure at three locations to form the subgroup. Take data for a new subgroup
(i.e., at a different location along the length) every 10 minutes.

In each of these examples, the subgroup is not rational, since a different system of
causes influences the within-subgroup variation than influences the between-subgroup
variation. In the beer filling operation, the variation in fill volume is influenced by head-
to-head differences. A consistent difference in nozzle size or line pressure in one head
versus the others will cause the within-subgroup variation to be excessively large,
masking any longer-term variation in fill volume for any of the individual heads. This
produces the characteristic behavior noted in Fig. 8.23: Many subgroups hugging the
centerline of the averages chart. (The circled values are violations of Run Test 7,
discussed in the next section, which is used to detect this error in subgrouping).
SPC-PC IV Software detects irrational subgroups in X-bar /Range chart

A solution to the centerline hugging is offered with a special version of a chart sometimes
referred to as a Batch Mean chart. (Minitab refers to this chart as the Between-Within
Chart.). The range (or sigma) chart is used to monitor the process variation, as usual, but
the averages chart is replaced with an Individual-X chart, whose control limits are based
on the moving range between the subgroup averages. The data from Fig. 8.23 is plotted
with the Batch Means option as shown in Fig. 8.24.

SPC-PC IV Software displays Batch Means chart with irrational subgroups

The idea of rational subgrouping becomes a bit fuzzy when dealing with X charts, or
individuals control charts. The reader may well wonder about the meaning of the term
subgrouping when the subgroup is a single measurement. The basic idea underlying
control charts of all types is to identify the capability of the process. The mechanism by
which this is accomplished is careful formation of rational subgroups as defined above.
When possible, rational subgroups are formed by using consecutive units. The measure
of process variability, either the subgroup standard deviation or the subgroup range, is
the basis of the control limits for averages. Conceptually, this is akin to basing the control
limits on short-term variation. These control limits are used to monitor variation over
time.

See also: Rational Subgroups