One Outlier or a Warning Sign?
In a batch of eight parts, seven show a consistent result while one part differs noticeably. The key question is whether the variation is normal for a small run or a signal of a process problem.
Is the difference isolated?
Confirm whether only one sample is affected or if the pattern repeats.
Does it affect function?
Verify fit, load behavior, and intended use performance.
Has the same condition appeared before?
Check previous runs for the same condition and location.
Is another sample needed?
Decide if extra evidence is required before final acceptance.
Should the batch decision change?
Assess whether the outlier changes pass/hold outcome.
Does the outlier require separate documentation?
Define if the outlier needs its own exception record.
Sample Sufficiency
Is an n=8 sample size sufficient to confirm batch consistency, or is additional sampling required to validate the outlier status?
Decision Logic
Should the 'one-off' defect trigger a full batch rejection or can the remaining seven parts be cleared for deployment?
Exception Protocol
Determining when a documented exception is acceptable versus when a process recalibration is mandatory.
