Why calculate LCL in P-chart?

[khumbutrekker]

Hi to all,
Why is there a need to calculate the LCL in P chart? Since its a proportion of non-conforming units, it should be better if it is lower than p-bar?
Thank you,

 

[Miner]

This will alert you to a shift in the process for the better. That would allow you to determine what has changed and potentially sustain the improvement.

 

[khumbutrekker]

Thank you Miner.

 

[Bev D]

I can’t tell you how many times someone wou claim ‘victory’ when the proportion defective went below the mean or even less than the previous point. Of course they use this claim to stop working on the Problem. The lower control limit serves to demonstrate that if the control rules for the lower results are not violated then the process hasn’t gotten ‘better’.

 

[David-D]

Understanding when it drops below the LCL provides two indications in my experience. First, as mentioned, it might indicate that things have improved; in those cases its good to be able to identify what has changed to drive down the defect rate so it can be permanent captured as part of the process improvement. On the other hand, it is also possible that the number of defects identified is also driven by the inspection system effectiveness; in those cases the drop is actually a bad thing because your inspection is failing to identify the nonconformances. In those situations, you want to perform root cause and correctice action to the inspection system.

Either way, the LCL provides an indication that something has changed - root cause analysis will tell you if it is a good change or a bad change and you can then take corrective action accordingly (to fix it or make it permanent).

David