I’m a BB currently running a project in the bank environment. This project is about cycle time. I have done a gage R&R, collecting some data to see whether the measurement system was repetible or not. In fact, I first collected these data on an excel sheet, and I easily saw that all the measurements done by a single operator twice were absolutely the same, so the system is 100% repetible in this case.
My question : I decided to run the analysis using Minitab, so I used the gage R&R (crossed) function of minitab. I expected the software to provide me with a set of graphs, despite excel told me the system was 100% repetible. In fact, Minitab only gave me a sentence in the session window, looking like an error message, confirming there was no variation between subgroups anyway.
Is there a way to get graphs in this specific situation ?
Thanks for your help !
I am assuming its an attribute data.If that is the case you cannot use minitab gage R&R crossed. Can you send me the data file to microsoft_sucks0
You should be using the Attribute Agreement Analysis in Minitab.
The path is Stat->Quality Tools->Attribute Agreement Analysis
If you need info on how to use it, click the help button in the bottom left corner of the dialog box.
Let’s back up a second – this is about cycle time, so you should not have to use attribute measures. You should be able to measure time as a continuous variable – days/hours/minutes/seconds. If you have 100% repeatability, you probably do not have enough resolution in your measurement method. What are you measuring, and what are you using to measure?
Franz:Jim H. is right, you have a problem with resolution for your time data. It is not possible to execute a task EXACTLY the same time each time. The general rule of thumb is that the resolution of your measurement device should divide the USL and LSL into about ten intervals.
Your gage R&R has failed at 100% if it can not distinguish between different time intervals in the range you are interested.
Data gathering can be a problem when people record the time they took to perform the task the first time and ‘ditto’ the rest of them.
Thank you all for your answers.
The measurements I have taken are continuous data without a doubt. All the data collected are the same. As Jim H said, if I change a little bit any of these data, minitab provides me with all the graphs we can get using the gage R&R crossed function. But it doesn’t work if data are 100% the same. In this case, I think I should use another mean to show graphically the results of the analysis, should I ?
I’ll try to be as nice as I can.
You say you are a Black Belt, yet it seems all you are concerned about is getting your pretty graphs. You have gotten good advice from Jim and BTDT. You ought to be concerned with finding out why your gaging system exhibits NO variability (an impossibility) and fixing that rather than how you will present your results.
It’s about results, not tools.
It would be easier to give you better advice if you gave the details about what & how you are measuring, but to stay with the basics – you can’t improve what you can’t measure. If you are working on cycle time, and measuring the time in hours, and your answer is always 3 hours, you can’t see what contributes the greatest amount of time, and/or the greatest variability – therefore you don’t know what to improve. If your answer is always the same to the hour, try measuring in minutes. If it is always the same to the minute, try measuring to the second…..Your gage can’t see your process. Increase the resolution of your measurement method. If you don’t understand, or I am misstating your problem, provide more details so we can help you understand the issue.I can save you some time fooling around with Minitab – if all your measurments are continuous and exactly the same, you failed the GR&R. Don’t need any graphs to tell you that.
Very nice post. Very honest.
Jim, Dave S and Mike,
Thank you very much for your answers and comments.
As Dave S, I will try to be as nice as I can. SS is OBVIOUSLY about providing results, not showing pretty graphs. I think this is (or should be) obvious for every reader contributing this this SS website. Thank you anyway for this reminder.
My question was only a technical question about a technical output of minitab. The results are shown anyway using other tools if needed.
I thought it was clear from the beginning.
Thanks anyway for your comments.
The forum ‘General’ is closed to new topics and replies.