Sitonce
Country: US
Show exams for United States Hong Kong
Sign in

NRSB RMS Quality Assurance: Duplicates, Blanks and Spikes

Updated 11 min read
Key takeaway

Duplicates, blanks and spikes test different parts of a radon measurement quality system.

  • Duplicates reveal precision and reproducibility; blanks can reveal contamination or background contribution; spikes or known challenges help assess recovery and analytical performance.
  • The exact design, frequency and acceptance limits come from the current quality plan and ANSI/AARST standard.
On this page12 sections
  1. QA is a system, not a form
  2. Duplicates assess agreement
  3. Blanks look for unwanted contribution
  4. Spikes test recovery and response
  5. Calibration and control samples answer related questions
  6. A QA plan assigns people and timing
  7. What a failed control should trigger
  8. Worked example: duplicate disagreement
  9. Worked example: a blank is unexpectedly positive
  10. Keep field QA separate from lab QA
  11. Study QA through diagnostic questions
  12. Use current primary references

The April 2026 NRSB RMS guide lists laboratory relationships, quality-assurance responsibilities, blanks, duplicates, spikes, calibrations, a QA plan and control charts. These are connected controls. A reliable result depends on the device, deployment, records, lab analysis, calibration, data review and corrective action. One completed checklist cannot compensate for a detector that was misplaced or a result that fails a required control.

Exam scope
NRSB Radon Measurement Specialist (RMS)
Primary concept
Quality Assurance: Duplicates, Blanks and Spikes
Use
Apply current method and quality plan; verify standards and jurisdictional rules

QA is a system, not a form

A quality assurance plan defines how an organization will demonstrate that measurements are fit for their intended use. It assigns responsibilities, sets review steps and specifies what happens when a control is outside limits. The current ANSI/AARST MS-QA standard describes requirements for measurement quality systems. The detailed schedule and acceptance criteria are method-specific and should not be invented from a general article.

For exam purposes, focus on what each control is designed to reveal. A duplicate asks whether results agree when measurements are repeated under comparable conditions. A blank asks whether unwanted signal or contamination appears when the target exposure should be absent or controlled. A spike or known challenge asks whether the system recovers an introduced known amount appropriately. Those labels point to different diagnostic clues.

Duplicates assess agreement

A duplicate is a second measurement intended to compare results under defined similar conditions. It may use paired devices deployed side by side or another design specified by the protocol. The comparison speaks to precision and reproducibility, not necessarily absolute accuracy. Two devices can agree closely and still share a systematic bias if both have the same calibration or method problem.

The expected agreement depends on the standard, concentration range, method and calculation. Near low concentrations, a percentage difference can become unstable because the denominator is small; some quality programs therefore use concentration-dependent criteria. Do not apply one memorized percent threshold to every dataset. Use the approved formula and acceptance rule supplied by the current plan or exam question.

When duplicate results disagree, first verify that the pair is truly comparable: same interval, location, method, lab process and exposure conditions. Check device identifiers, start and stop times, handling, placement and potential disturbance. Then apply the specified statistic or tolerance. Do not discard the inconvenient result simply because it is higher.

Blanks look for unwanted contribution

A blank is designed to show whether a target signal may be present when the device or sample should not have received the intended radon exposure, or whether handling and analysis introduce a background contribution. The exact blank type and procedure depend on the device and QA plan. A blank is not a second ordinary room sample and should not be interpreted as one.

An elevated blank may point toward contamination, leakage during storage or shipping, laboratory background, improper preparation or another process failure. The blank alone may not identify the cause. Review its chain of custody and handling alongside other controls, and consult the lab’s documented procedure before attributing fault.

The most important response is traceability. Identify affected batches or measurements according to the QA plan, assess whether client results remain valid and document corrective action. A blank result cannot be “subtracted” from a client result unless the validated method explicitly specifies that correction. Ad hoc correction risks replacing evidence with guesswork.

Spikes test recovery and response

A spike or known challenge introduces a known quantity or exposure so the system’s response can be compared with the expected result. It can help evaluate whether the measurement or analysis process recovers the challenge appropriately. The design varies among methods; do not assume that every radon detector can be spiked in the same way or that the operator personally prepares the challenge.

The relationship between observed and expected values can be expressed as recovery, subject to the defined procedure and uncertainty. A result that is too low may signal loss, under-response or an analytical issue; one that is too high may indicate contamination, calibration or calculation trouble. These are hypotheses to investigate, not conclusions from one number alone.

A known challenge also tests the whole chain. Correct preparation, identification, timing and lab analysis matter. If the challenge is mislabeled or the assigned value is wrong, the apparent recovery failure could be a record problem. The measurement professional should preserve the control data and follow the lab’s investigation process.

Calibration establishes a relationship between instrument indication and a reference under defined conditions. A calibration record does not replace routine QA because performance can drift, handling can fail and deployment can introduce new error. NRSB maintains policies and information for device evaluation and calibration; the current requirements depend on the instrument and use.

Controls may be scheduled at defined intervals, included in each batch or triggered by an event. An equipment check before deployment can catch a fault; a blind duplicate can test field and analysis components; a spike can challenge recovery. A layered program is useful because no single check covers every source of error.

Know the vocabulary but do not collapse it. Calibration, verification, duplicate, blank and spike are not interchangeable. Calibration adjusts or characterizes response against a reference; verification checks performance under stated criteria; duplicates examine agreement; blanks look for unwanted contribution; spikes assess response to a known challenge. Specific program definitions govern.

A QA plan assigns people and timing

A laboratory and a field professional may divide QA responsibilities. The plan should say who selects devices, tracks calibration, prepares or submits controls, reviews results, communicates failures and authorizes corrective action. NRSB’s syllabus asks candidates to understand levels of lab service and division of responsibility because a result can fail at the handoff between field and laboratory.

Before a job, confirm that the lab is appropriate for the method and that required controls are included. Verify sample identifiers, forms, deadlines and packaging. At receipt, reconcile the lab’s report with device IDs and requested analysis. A missing control or mismatch is a quality signal, not a clerical nuisance to ignore.

If a client hires a measurement specialist and the lab performs analysis, neither party should assume the other owns every control. Put the division of responsibilities in the quality system and service agreement. Clear accountability makes it possible to locate a failure and determine which results may be affected.

What a failed control should trigger

A failed control should trigger a documented response under the QA plan. Preserve the original data, identify the failure, determine the scope of affected measurements and investigate plausible causes. The plan may require reanalysis, a repeat measurement, equipment removal, client notification or a hold on reporting. The correct action depends on control type and acceptance criteria.

Do not wait for a client complaint to investigate a known out-of-control condition. If a report has already been issued, follow the organization’s correction and notification procedure. Communicate facts plainly: which result is affected, what is known, what remains uncertain and what corrective step is planned. Avoid minimizing a failure because the number seems convenient.

A failed blank, duplicate or spike does not automatically invalidate every result produced by a lab or device. Assess scope using the plan and evidence. Conversely, passing one control does not prove that all field conditions were valid. QA is a set of evidence streams, each with a limited purpose.

Worked example: duplicate disagreement

Two collocated passive devices are deployed for the same interval. Their results differ more than the acceptance rule allows. First confirm identifiers and that both devices returned from the same location and interval. Check whether either was moved, damaged, exposed to a draft or handled differently. Confirm the lab used the right analysis and units, then calculate agreement using the prescribed method.

If the comparison remains outside limits, document the discrepancy and follow the plan’s escalation. Depending on the rule, the result may need qualification or a repeat under controlled conditions. Do not choose one result based on which supports the client’s preference. The quality issue is precisely that the pair did not meet the stated agreement criterion.

For the exam, the best answer typically investigates the process and applies the written acceptance limit. Distractors may propose averaging automatically, deleting the high value, or declaring the entire laboratory incompetent from one pair. Those responses overreach the evidence.

Worked example: a blank is unexpectedly positive

A blank reports a measurable signal above its expected range. Review how it was stored, transported and analyzed; inspect the batch, control history and other blanks. Determine whether the blank traveled with field samples or was prepared in the lab. These details help locate possible contamination or background contribution.

The plan may require holding a related batch or evaluating the effect on client results. Do not subtract the blank value by hand unless the validated method requires it. A positive blank is evidence to investigate. It does not, by itself, reveal the correction factor or prove which client sample is wrong.

Keep a complete record of the response: control ID, result, acceptance range, investigation, affected data, decision and corrective action. Another qualified reviewer should be able to follow the logic without guessing. This documentation is part of the result’s defensibility.

Keep field QA separate from lab QA

Field QA includes location, deployment, duration, device protection, environmental notes and chain of custody. Laboratory QA includes analytical controls, calibration checks, blanks, spikes and data review as applicable. Some controls cover both sides. A good RMS candidate asks where in the workflow the signal entered rather than assuming all error belongs to the lab.

For example, a duplicate pair may test several sources of variation if devices share a location and lab, while a reference challenge may focus more closely on analytical response. Interpret the design before interpreting the result. The same label can be implemented differently by a method, so the written QA plan remains authoritative.

When the lab reports an exception, contact its technical staff and preserve their explanation. A rushed field professional who paraphrases a lab note incorrectly can create a second quality problem. The report should accurately identify which measurements are affected and what the client should do next.

Study QA through diagnostic questions

Make a table with control, question answered, likely failure signal and first follow-up. Fill it from the current standards and quality plan. Then practice cases where controls conflict: duplicates agree but spikes fail; blank is clean but a device was moved; calibration is current but a continuous trace shows data gaps. This tests whether you know each control’s limits.

Avoid memorizing universal thresholds from an old study aid. Acceptance limits can depend on method and concentration range, and standards change. The April 2026 NRSB guide points toward MS-QA and current ANSI/AARST sources. Use those sources for exact criteria and understand their rationale well enough to apply a rule when the numbers are supplied.

When studying QA, ask what uncertainty each check is designed to reduce. Connecting each control to the uncertainty it addresses makes it easier to explain why a documented quality system matters.

Use current primary references

The NRSB RMS guide identifies ANSI/AARST MS-QA as a source and its standards portal describes that standard as setting minimum requirements for quality systems for qualified professionals and laboratories. NRSB also maintains accredited laboratory and device resources. Check those current pages before citing a lab, approval or acceptance criterion.

This article explains the purpose of common controls; it does not reproduce proprietary standards language or specify a universal control frequency. For professional work, obtain and follow the current standard, the lab’s method and your organization’s QA plan. For exam preparation, note how the guide maps the concepts to laboratory relations and professional responsibility.

A valid quality system has a path from observation to decision: define the control, review the evidence, apply the criterion, investigate exceptions, correct the process and communicate affected results. Knowing that sequence is more useful than simply recognizing the words blank, duplicate and spike.

Common questions

What does a duplicate radon test tell you?

A duplicate compares results from repeated measurements under defined comparable conditions. It primarily evaluates agreement or precision. Close agreement does not prove the measurements are absolutely accurate because both could share the same systematic bias.

What is a blank sample used for in radon QA?

A blank helps identify unwanted signal or contamination introduced by handling, storage, transport or analysis, depending on the method. The QA plan defines the blank procedure and acceptance criteria.

What does a radon spike test?

A spike or known challenge checks how the system responds to a known amount or exposure. It can provide evidence about recovery and analytical performance, but the exact design and limits are method-specific.

Should a failed blank simply be subtracted from each result?

No. Do not apply an informal correction. Follow the validated method and QA plan to investigate the blank, assess affected measurements and determine the appropriate corrective action.