Troubleshooting / Common Issues
Observed SD or Within-Laboratory CV% exceeds Goal/Claim at one or more levels
Possible causes
- –Operator technique variability (pipetting, sample handling)
- –Instrument instability — check maintenance logs and calibration records
- –Control material instability — verify storage conditions and expiration
- –Environmental factors (temperature, humidity fluctuations)
Recommended action
Investigate the Levey-Jennings chart for trends or shifts that coincide with specific days or operators. Repeat the study after addressing identified root causes.
Between-day or between-run variance is much larger than within-run variance (long-term design)
Possible causes
- –Calibration drift between days or runs
- –Reagent lot changes mid-study
- –Different operators with different technique
- –Instrument requiring preventive maintenance
Recommended action
Confirm within-run performance is acceptable first, then investigate between-run and between-day factors systematically. A large between-day component relative to within-run often points to a calibration or reagent stability issue.
Outlier data points on the Levey-Jennings chart
Possible causes
- –Transcription or data entry errors
- –Sample handling issues (incorrect volume, wrong vial)
- –Instrument error during that run
Recommended action
Apply statistical outlier criteria appropriate for repeatability data, then review the raw data for that point. If exclusion criteria are met and a valid, documented reason exists (instrument error, sample handling issue), exclude the point in the platform with justification and re-run analysis. Do not exclude points based solely on visual appearance.
95% CI for CV% is very wide
Possible causes
- –Insufficient sample size (too few replicates or days)
- –High variability in the data
Recommended action
A wide confidence interval indicates that the precision estimate itself is imprecise. Consider increasing the number of days or replicates. For short-term Complex designs, ensure you have met the EP15-A3 minimum of 5 days and 5 replicates per run (dfR ≥ 20).