LEAN SIX SIGMA WORLD CONFERENCE ABSTRACT
Assessing Process Capability When the Data are Non-Normal
Presenter:
|
Dr. Scott Kowalski, Senior Technical Trainer, Minitab, Inc., Sanford, FL, USA
|
Keywords:
|
Process Capability, Nonnormal Data, Case Study
|
Industry:
|
Manufacturing
|
Level:
|
Intermediate
|
Process capability measures, such as Cpk, Ppk and tolerance intervals, typically assume that your data follow a normal distribution. But what should you do when your data fail the test of normality? There are many reasons for non-normal data. However, one of the most common reasons is that the data are naturally skewed due to a physical bound on the data or when extreme values tend to occur in only one tail of the data distribution. Case studies will be used to facilitate strategies for assessing process capability that you can successfully apply to your own data that are not normally distributed.
Following the presentation, you will be empowered to implement successful strategies for assessing process capability when the normality test fails:
- Transform data to use capability statistics that assume a normal distribution.
- Choose an appropriate distribution to model data that are non-normal.
