# Outliers in WPC testing: investigate before excluding

Canonical: https://woodplasticcomposite.net/guides/outliers-in-wpc-testing-investigate-before-excluding/
Language: en
Updated: 2026-10-15
Version: 1
Review: AI-assisted English editorial draft. Human technical review has not been recorded.

## Summary

An unusual result should prompt an investigation, not automatic deletion. It may reveal a preparation error, a measurement problem or genuine material variation relevant to production.

## Technical context

Removing a low result because it looks inconvenient biases the reported performance. Conversely, a documented equipment malfunction may justify excluding a measurement under the agreed method. The decision needs evidence and a consistent rule.

https://woodplasticcomposite.net/guides/outliers-in-wpc-testing-investigate-before-excluding/#technical-context

## What to measure and document

Retain raw results and record the reason for any exclusion. Review specimen identity, conditioning, test setup and failure observations. Where a retest is justified, state whether the original result remains part of the report and why.

https://woodplasticcomposite.net/guides/outliers-in-wpc-testing-investigate-before-excluding/#verification-plan

## Implications for development and specification

A credible technical file shows how uncertainty was handled. For buyers, unexplained perfect consistency can be less informative than a transparent account of variation. The objective is an honest estimate of performance, not the most flattering average.

https://woodplasticcomposite.net/guides/outliers-in-wpc-testing-investigate-before-excluding/#practical-decision

## Related reading and evidence limits

General engineering-statistics guidance; not a WPC product standard. Use the linked work as related reading. The practical checks above are editorial recommendations, not reported test results or a substitute for product qualification.

https://woodplasticcomposite.net/guides/outliers-in-wpc-testing-investigate-before-excluding/#reading-scope

## Key points

Removing a low result because it looks inconvenient biases the reported performance.
Retain raw results and record the reason for any exclusion.
A credible technical file shows how uncertainty was handled.

## Comparison

Review step | Evidence to retain
Material and geometry | Exact specimen or product identity
Verification | Method, conditions and individual observations
Decision | Approved scope and documented limitations

## What evidence is needed for outliers in WPC testing: investigate before excluding?

Retain raw results and record the reason for any exclusion. Review specimen identity, conditioning, test setup and failure observations. Where a retest is justified, state whether the original result remains part of the report and why.

## Sources

1. NIST engineering statistics handbook — https://www.itl.nist.gov/div898/handbook/