Structured subgroup sampling using Minitab

QUESTION: I want to produce subgroup samples for statistical process control (SPC) purposes. I have a database from our call centre that holds information on the following: time call received, type of call (response graded 1-5 in order of severity – 1=emergency call, 5=call taker resolved with no police attendance required), time resource allocated to incident, time of arrival, time resource left incident, time incident closed, etc.

From calls received, we create approx. 30,000 incidents per month. If I take one month’s worth of data and import it into Minitab, how can I filter it to create a subgroup sample? Based on my (somewhat limited) knowledge of statistical process control, what I think I want to do is create a subgroup sample that takes n measurements every x minutes (e.g. say 10 measurements every hour), over the course of a day and then over the days in the month. I’d like to use this sample to determine what is currently ‘normal’ in our response process so that I can show how the time taken between call receipt and police attendance differs by type of call, hour of day, day of week, etc.

Is there a simple way in Minitab to achieve this type of sample?

RESPONSE: Control charts are set up to check for process stability, not assess inputs that can impact a response. You should be able in Minitab pull a sample and put it into a control chart. However, for what you want to do I suggest decrease your frequency of subgroup (30,000-foot-level control chart format) so that potential common-cause input variability occurs between subgroups; e.g., might subgroup weekly. Process capability can also be assessed as described in: x-bar and R charts and alternative

You should then pull data from regions of stability to statistically test hypotheses that you mentioned above using tests such as Analysis of Means, which can help gain insight to what you might be done to improve.