Choosing the right statistical tool is like selecting the right car. Howsoever tempting it may look, we would always have to think and make the right decision for a long term benefit. The Lean Six Sigma tools help us take such decisions and make the changes. These tools we select must also indicate false alarms and must predict the probability of a risk.
A Control Chart is one of the Lean Six Sigma tools that help see the trends and the plausible causes of variation. But from the numerous types available, which is the right one for your project?
Standard X-bar-S chart is good for rational subgroups. Individual X chart works for individual data, but then again depends on the distribution. To get the data from a stable system, we need to have a probability or distribution model. The probability model and the data free from variation is always interconnected.
Don Wheeler is of the understanding that if the observations are not within the three-sigma control limits it is not a stable process. A random dataset, which has no concrete data associated with it, leaves a myriad of unanswered queries.
As and when an explainable context is attached, the dataset makes sense, or at least the variations can be clarified. Lean Six Sigma Tools such as the Box-Cox transformation can then be used to explain the datasets in which the lower boundary is zero.
The Control Chart can measure the false alarms and do them in terms of the ARL (Average Run Length). The theory states that 1 point out of 370 will be outside either of the control limits, even when there is no variation.
Although Wheeler claims that normal-based individual X chart allows for a greater risk undertaking, but champions disagree on this front, mostly because it fails to sustain the reliability.
With the advancement is technical process, measurement tools such as the Control Charts have also transformed. Therefore, it is always important to choose the most appropriate tools, based on the events and on the experience, to be able to attain maximum success.