962892_en-US

liliconcom

Levey Jennings charts with subgroups gives wrong limits

Hi,

As indicated in other threads, there are several issues with applying Levey Jennings control charts. But when your data set has subgroups JMP just seems to miscalculate the control limits! The chart generated is for the average within subgroup but the calculated limits are for individual measurements as seen below, giving too wide limits.  It appears that division by the square root of subgroup size is missing from the formula.  

 

Best regards,

Stefan

Quality and Process EngineeringRe: Levey Jennings charts with subgroups gives wrong limits

Quality and Process Methods > Control Chart Builder > Control Chart Builder Options > Options Panel and Right-Click Chart Options 

Statistical Details for Levey-Jennings Charts 

Re: Levey Jennings charts with subgroups gives wrong limits

I would prefer not to use the Levey Jennings chart. If the process is expected to have between group variation, I would probably use the Three way chart. If not, I would use the X bar and R chart.

Best regards,

Stefan

Re: Levey Jennings charts with subgroups gives wrong limits

Then don't use it? And if there is a bug in JMP, please contact JMP support.

Re: Levey Jennings charts with subgroups gives wrong limits

I wouldn't use Levy Jennings charts either, they were design for specific purpose. In QC labs, some standards for very defined Stddev. In a LJ chart you can specify that sigma to see if you method is in control.   Everyone else that uses to hide OOC events that they should be responding to, since frequent recalculation of the limits just keeps fitting the common and special cause in the process.