How sample preparation affects test results

Why this decision needs structure

An instrument measures only the portion of sample that reaches the measurement system. Errors introduced while splitting, mixing, drying, dissolving, filtering, or transferring material can create bias that a more precise instrument cannot correct.

This guide does not replace an official method, risk assessment, manufacturer instruction, or applicable regulation. Use the cited sources as a starting point, then adapt decisions to the application, laboratory capability, and local requirements. The intended output is a requirements record that can be reviewed, tested, and updated.

A practical decision framework

Representativeness

Record relationship between the lot, laboratory sample, subsample, and the test portion actually analysed. Treat this as a decision that can be verified, not a preference. Record the starting condition, acceptable value or limit, the person who confirms it, and the evidence required. Connect the decision to the method, safety, workload, and post-implementation work. Mark missing information as an open assumption so users, procurement, facilities, and suppliers do not interpret it differently.

Homogenization

Record mixing or size-reduction technique, sequence, duration, and evidence that material is sufficiently uniform. Begin with evidence from real work, then separate mandatory needs from features that merely add convenience. Examine the effect on results, time, user competence, space, utilities, and recurring cost. Each conclusion should have a source, an owner, and a verification method so that later changes can be reviewed without reopening the entire discussion.

Contamination

Record tool cleanliness, blanks, container material, dust, carryover, and preparation sequence. Consider normal operation, peak demand, credible failures, and conditions after future change. An agreement that works in only one scenario is fragile. Use numbers where available, state units and tolerances, and retain the decision rationale. This helps the team distinguish risks that require control from features that do not create practical value.

Loss and transformation

Record evaporation, adsorption, degradation, reaction, temperature, light, holding time, and analyte recovery. Treat this as a decision that can be verified, not a preference. Record the starting condition, acceptable value or limit, the person who confirms it, and the evidence required. Connect the decision to the method, safety, workload, and post-implementation work. Mark missing information as an open assumption so users, procurement, facilities, and suppliers do not interpret it differently.

Dilution factors

Record mass and volume at every stage, moisture correction, material purity, and calculation propagation. Begin with evidence from real work, then separate mandatory needs from features that merely add convenience. Examine the effect on results, time, user competence, space, utilities, and recurring cost. Each conclusion should have a source, an owner, and a verification method so that later changes can be reviewed without reopening the entire discussion.

Preparation quality control

Record method blanks, duplicates, spikes, reference materials, acceptance limits, and action after control failure. Consider normal operation, peak demand, credible failures, and conditions after future change. An agreement that works in only one scenario is fragile. Use numbers where available, state units and tolerances, and retain the decision rationale. This helps the team distinguish risks that require control from features that do not create practical value.

Applying the framework to real work

A solid sample that looks uniform may still contain different particle-size and composition distributions. If the test portion is removed before adequate homogenization, replicate readings can agree closely while failing to represent the original material. Use a case like this in a short working session with users, method owners, facilities, safety, procurement, and technical support. Put known data beside open questions. This allows different expectations to be resolved before they become quotation revisions, rework, or downtime.

Maintain one decision table with requirement, rationale, evidence, acceptance limit, owner, and status. Do not combine facts, assumptions, and preferences in one sentence. When new information appears, update the relevant row and note its effect elsewhere. This simple discipline makes evaluation transparent and reduces dependence on meeting memory.

Mistakes to avoid

• Starting with a product name or feature list before agreeing the purpose and performance boundary.
• Using words such as good, complete, fast, or suitable without an acceptance measure.
• Ignoring user work, site conditions, recurring materials, documentation, and support after installation.
• Leaving decisions in conversation without a controlled document, owner, and review date.

Action checklist

• The chain from sample to test portion is documented
• Homogenization has been shown suitable for the matrix
• Containers and tools do not add contaminants
• Analyte loss or change is controlled
• Every dilution factor is traceable
• Blanks and control samples are prepared with the batch
• Preparation deviations are recorded before reporting

Bringing the decision together

A strong decision record need not be long, but it should connect the requirement, risk, evidence, and acceptance method. Start with the laboratory work, involve the right roles, and use data to narrow the options. When conditions change, revisit the documented rationale and limits rather than repeating an earlier choice by habit.

Implementation note

Before approval, arrange a cross-review by someone who did not prepare the first draft. Ask the reviewer to find missing units, limits that cannot be tested, terms with more than one interpretation, and dependencies without owners. Confirm that the requirement still matches the current method, workload, building conditions, user capability, and safety policy. Record the review outcome and next review date so the document remains active. If a change affects result quality or risk, reassess it before work continues.

Sources

1. Sample Processing and AnalysisUnited States Environmental Protection Agency
2. ISO/IEC 17025:2017 - General requirements for the competence of testing and calibration laboratoriesInternational Organization for Standardization
3. Learn about Drinking Water Analytical MethodsUnited States Environmental Protection Agency