No measurement is perfect. All have some degree of uncertainty, even when performed correctly. Measurement uncertainty is the range of outcomes expected to surround the true value of the measurement. The range is defined by the limitations of the technology, method, technique, equipment, environment, and the operator. It does not represent a mistake by the scientist, but rather a quantified expression of confidence in the result. =
Importance
Understanding measurement uncertainty is important for the lab so it can apply the information obtained and communicate how best to use its results to the lab’s key stakeholders. Defining measurement uncertainty enables the lab to make defensible decisions, such as pass/fail criteria, trend changes, and release specifications. It allows the lab to understand differences in results produced by different instruments, analysts, or labs. Having a clear understanding also reduces rework, conflicts, disputes, and data misuse. Communicating uncertainty is part of strong data integrity and helps build trust in the lab and its results.
Components
Measurement uncertainty is driven by a statistical analysis of measured results. A single measurement cannot have a quantified measurement uncertainty. Often, these calculations are driven by metrics such as the mean, standard deviation, and confidence level. These values aim to establish the accuracy and precision of the measurement. Accuracy indicates how close a measurement is to the true value, and precision indicates the repeatability of results.
Results are often reported as value + uncertainty, where uncertainty reflects a defined confidence level. This approach helps lab staff and stakeholders understand how much to trust the value and how to use the results appropriately.
Sources of uncertainty
Measurement uncertainty comes from two primary sources—random effects and systematic effects. Random effects are driven by uncontrollable, unpredictable variation in measurements due to chance and fluctuations. Random effects can be estimated from repeated measurements. Systemic effects are driven by consistent biases and are not eliminated by averaging. Systemic effects arise from issues in instrument calibration, the resolution of a measurement, the environment in the lab, and issues with reference standards.
Other contributors to measurement uncertainty include an incomplete definition of what needs to be measured, the scientist’s skills and attention applied to the measurement, and flaws in the measurement procedure or setup.
Fit for purpose
- Not all measurements require the greatest possible accuracy and precision. Lab managers can help their scientists and stakeholders balance the effort invested in a measurement with the importance of confidence in the result. The level of effort invested in the work should be proportional to the risk, cost, or regulatory impact of the decision. Measurements with higher value require greater confidence. However, applying high-confidence level approaches to low-risk measurements wastes time, resources, and materials.
Management
Lab managers can make decisions to help improve the lab’s control of measurement uncertainty. Here are some important ways to manage measurement uncertainty:
- Develop and use validated or well-understood methods. These can include detailed studies of the measurement uncertainty.
- Commit to completing routine calibration and preventive maintenance. Ensure the instruments are operating properly and calibrations are done properly and as needed.
- Document consistent standard operating procedures (SOPs) to ensure lab staff are properly trained and proficient in making the needed measurements.
- Chart key measurements to track any trends that could lead to increased uncertainty. Investigate and address trends that increase doubt about the results.
- Clearly communicate lab results with other staff who may need them or with stakeholders who request them. Ensure anyone using the data can make good decisions around its use.
Responsibilities
Staff across the lab have responsibilities around measurement uncertainty:
- Scientists need to be attentive to their measurements and use the knowledge available to do the right tests, apply them correctly, and keep their equipment in working condition.
- Lab managers need to develop a culture of quality that aims to deliver the right results, first time, for their stakeholders. They need to drive higher confidence in results that are high-importance or high-risk decisions for the lab. They also need to treat measurement uncertainty as a performance driver for the lab, not as a statistical afterthought.
- Lab quality professionals need to ensure that methods, procedures, and workflows are well-defined and include the best practices to reduce measurement uncertainty.
Training
Part of a successful culture of lab quality includes ensuring that all staff understand the importance of measurement uncertainty, why it matters for certain decisions, and how their actions affect variability and measurement uncertainty in their results. Training staff to standardize procedures for critical tests helps reduce measurement uncertainty, especially when multiple people are responsible for running the same types of tests. Cross-training helps to share best practices across the lab, enabling everyone to benefit from the lessons learned by each staff member.
Lab managers can reduce the uncertainty driven by staff actions by:
- Investing time and budget into targeted training that aims at the critical issues, rather than blanket retraining that wastes time and resources.
- Actively talking with staff about the importance of attention to detail, so that staff understand this is an expectation and a priority for the lab
- Participating in discussions where data are used in investigations, analyses, or problem-solving to show how the data can be used appropriately.
Stakeholders
Communicating measurement uncertainty to key stakeholders can be challenging, especially if they are not technically trained or if they’ve been out of the lab for an extended period. It is important that people using the data to make decisions have a clear understanding of how the data can be used appropriately and of the key errors that can arise from its uncertainty. Educating stakeholders helps them use data effectively and make informed, data-driven decisions. Helping them use the data properly also builds trust in the lab.
Lab Quality Management Certificate
The Lab Quality Management certificate is more than training—it’s a professional advantage.
Gain critical skills and IACET-approved CEUs that make a measurable difference.
An example of a key misunderstanding of data is knowing when values are sufficiently different to make a decision about them. Using measurement uncertainty properly helps identify the differences in data that are significant. Sometimes, non-technical stakeholders try to make decisions about datasets they perceive as different, even though they all fall within the same confidence level.
Lab managers can ensure that a measurement-uncertainty SOP is documented in the quality program and that all staff are trained on its contents. They can also make uncertainty part of routine conversations in the lab. Talking about data includes the uncertainty of the tests. This normalizes questions and discussions about uncertainty and helps people confidently report issues and surface observations that may impact data accuracy. Once measurement uncertainty is part of the lab’s culture, it can also align with the lab’s key performance indicators, objectives, and results.
By making small, informed decisions, lab managers can significantly improve measurement reliability, lab performance, and build stronger relationships with key stakeholders.
