What needs to be clarified
Check the sample–population relationship, reproduce the nonconformity proportion, and assess whether extrapolation to goods not directly examined is justified by the report.
Materials to prepare
- Report with original test results and calculation formula
- Batch register, sample identifiers and selection records
- Complaint log, inclusion criteria and substituted-sample information
What to compare
Trace each sample’s selection, repeated complaints and exclusions. Check the denominator and coverage of delivery periods, models and usage conditions. Distinguish targeted problem samples from random sampling; a confidence interval does not repair selection bias. List unsupported generalization assumptions explicitly.
Possible outcome of the analysis
A review may verify the selected-sample calculation and generalization limits; the batch-wide conclusion may need qualification or further testing under a justified sampling scheme.
Limits of the conclusions
Sample size alone does not remove selection bias. Without batch access, a review establishes no actual defect prevalence and does not replace repeated product testing.