Sub-grouping can be extremely insightful. Just be sure that you are doing it correctly and are also not overdoing it so that your results are impossible to interpret or compromised.
Overview: What is a sub-group?
A sub-group can be defined as a group of units that are further classified from the main group based on defined criteria and that were produced under essentially the same conditions as the main.
3 benefits of sub-groups
There are some clear benefits to sub-groups that are worth noting:
1. Balanced analysis
By breaking groups into sub-groups, you have the benefit of a more balanced and robust analysis.
With sub-groups, you are more capable of seeing relationships based on various types of criteria.
3. Better decision-making
Sub-grouping can aid in better decision-making when there is uncertainty.
Why are sub-groups important to understand?
Understanding sub-groups is important for the following reasons:
1. Helps in estimation processes
By having an understanding of sub-groups, you have a tool that can help in the estimation process of things like short-term variations. These variations can, in turn, help predict the control limits of long-term variations.
By having a working knowledge of sub-groups, you also have an understanding of their limitations. When sub-groups are misused, they can lead to misleading findings.
One concern with sub-groups is that there can be a tendency for biases. By having a clear grasp of sub-grouping, you are less likely to have biases in how you divide your group.
An industry example of sub-groups
In a machine shop, there are machines that cut pipes for large-scale farming irrigation applications. There have been some complaints from some of the end-users that their connections are leaking, and it is suspected at the machine shop that this could be due to slight variances in the standard lengths of the pipes being sold. The shop decides to run the machine long enough to get a proper sampling size and then sub-group the pipes according to what the actual measurements are. In this way, they can see how often a machine is cutting at a different measurement than the one assigned and what measurement it shifts to most often.
4 best practices when thinking about sub-groups
Keep these things in mind for the greatest likelihood of success in your analysis of sub-groups:
1. Keep it an early part of the objective
Have your sub-group analysis a key part of your objectives from the beginning.
2. Clear reasoning
Have a clear reason as to why you are sub-grouping.
3. Defining upfront
Define the sub-groups upfront as well as how many analyses will be done. Also, report on all of them, not just the ones where the findings are the most compelling.
4. Sample size
Be sure that there is a large enough sample size.
Frequently Asked Questions (FAQ) about sub-groups
Are there other words for sub-groups?
Subdivision, subset, sub-section, and sub-category are all valid interchangeable terms for sub-group.
What is the role of sub-groups in a control chart?
Sub-groups provide a snapshot of a process in a control chart. For example, on an averages chart, sub-groups provide a process location estimate, while the sigma chart provides the short-term variation’s sub-group estimation.
Should every sub-group have an identical subset of causes for process variation?
No, that is not necessary as long as there is consistency with the system that defines the causes.
Dividing into sub-groups can be an extremely useful way to gain more detailed knowledge about a sample. Just be aware that there are limitations to the kind of information that can be extracted. There can be an inclination to sub-group excessively, which can make your findings extremely difficult to interpret or appear redundant. Be sure that your dividing is not biased, and if you plan on sub-grouping, make it part of the beginning of your stated objective and not an afterthought.