I'm using JMP 11. Using a script, I build a (generic) control chart to plot our lab test results. The problem is that I sometimes don't have a upper or lower control limit (e.g.: trace metals: no UCL, Assay: No LCL). When there is a one-sided limit, I don't want to see the missing limit on the chart for obvious reasons. How can I remove one limit from the control chart?
As far as I know this is not currently supported.
'Control Chart Builder' has an option to suppress the display of the control limits. If you do this 'by hand', you can then reinstate only the required limit with an 'HLine()' graphics script. If doing this routinely, these steps could be scripted.
I've tried the Control Chart Builder.
This was not the result I was hoping for. The problem is that, you end up with a bunch of lines and no way to identify what line is what. You can do this visually, but this has absolutely no scientific value. Next to it, you get a small table next to chart, with average values and limits. Since we need to print this on paper, this means that the width of the chart is seriously reduced.
Thanx for the suggestion, but we need a different solution.
You might be better off using a run chart instead of a control chart anyway. Just a thought - I have always been a big fan of control charts, but recently, have come to appreciate run charts as well.
I've tried this chart, but the limit won't appear on them. I can put a line on it (I can even define how the line looks like). But in the end, it's just a line. As stated above, scientifically useless unless named.
Run charts do not have limits. The analysis is based on number of runs and length of runs above/below the median. I recently wrote and article for Quality Digest about run charts which might be useful to you:On Alleviating Tortured Data | Quality Digest
In the end, JMP doesn't support one sided limits. For reporting purposes we wanted to chart to show that the limits are one-sided. (After all, you cannot have a negative concentration of metals in a solution.)
I divide K_Sigma by 2 in case of a one sided limit. I manually set the value for the limit to 0. And I get the control charts with 2 limits :-(
I hope that in the future, we get a JMP version that will support scientific statistics. But for now I guess we'll communicate to our customers that this is the limitation of the JMP-software we use.
Thank you all for your suggestions.
Could you just set the Y axis to zero, this would hide the negative control limit.