When it Comes to Measurement, Defining Uncertainty = Increased Confidence in Results

By Dr. Anna Zhenchuk, published on 3rd June 2016

In recent years, while working in the field of micronutrient measurement in food and biological samples, I’ve often been asked: “How reliable is a measurement result?” or “How accurate is the method used to get the result?” 

According to the FAO guidelines, understanding and addressing the uncertainty inherent in all measurements is as important as ensuring accuracy. This ensures that results are both valid and comparable. In this article, I’ll break down the concept of measurement uncertainty—a key indicator of measurement reliability—and explain how embracing it can increase confidence in results. 

To reject or not to reject

Imagine you’re deciding whether a batch of fortified food meets quality standards. This decision might depend on a single test result. Without understanding the uncertainty in that result, you could reject a batch that is actually acceptable—or worse, approve one that falls short of safety standards. 

This is where measurement uncertainty becomes critical. It provides a numerical estimate of how much trust you can place in a measured value. As required by the International Standards Organization (ISO), all measured results should be reported alongside their measurement uncertainty (ISO/IEC Guide 98-3:2008). 

No result is perfect

Let’s start with a simple truth: no measurement is perfect. Every result comes with some level of imperfection, no matter how carefully the measurement is performed.  

Measurement uncertainty is the term used to quantify this imperfection.  What causes uncertainty? Several factors, such as: 

• Sampling methods 
• Sample preparation 
• Equipment calibration 
• Environmental conditions 
• Operator’s technique 

In simpler terms, measurement uncertainty includes: 

1. The possibility of measuring unintended substances or missing some of the target analytes. 
2. Variation between repeated measurements of the same sample. 
3. Differences between a sample and the portion analyzed. 

By understanding these factors, we can better evaluate the reliability of any result. 

How is measurement uncertainty calculated?

Measurement uncertainty is typically evaluated during the validation of a method. There are detailed guidelines and protocols provided by both the Association of Analytical Chemists (AOAC) as well as ISO on how to evaluate the measurement uncertainty. Validation involves testing a sample with a known concentration of the analyte under various conditions. This helps determine two key components: 

1. Bias: The systematic error, or the difference between the average of measured values and the true value. 
2. Precision: The random error, or the variability in repeated measurements, expressed as a standard deviation (SD). 

To calculate uncertainty: 

Add the bias and 2x the standard deviation (±2xSD) to create a range. This range gives a 95% confidence level, meaning there’s a 95% chance that the true value falls within it.  

Example: 

If the bias is 1 mg/kg and the SD is 0.5 mg/kg, the uncertainty is: 1 ± 2(0.5) = ± 2 mg/kg. This calculation ensures a practical estimate of uncertainty for decision-making in food testing. 

When we measure something, measurement uncertainty reflects the range around the result where the true value is likely to lie. This range often corresponds to a confidence interval tied to a certain level of confidence: for example, a 95% confidence interval typically spans ±2 standard deviations, meaning we are 95% sure the true value falls within this range. (Figure 1) Similarly, at ±1 standard deviation, the confidence is about 66%, and at ±3 standard deviation, it’s 99%, showing how uncertainty is directly linked to both the spread of measurements and our confidence in the result. 

Understanding Results with Uncertainty

Let’s apply this concept to real-world food testing. 

Without Measurement Uncertainty: 

Imagine you test four batches of fortified food. Results are compared to tolerance limits, which define the minimum and maximum acceptable nutrient levels. If you report results as single values, two out of four might fall outside the tolerance limits, even though they’re very close as seen in Figure 2.  

With Measurement Uncertainty: 

When the uncertainty is included, the same results might all fall within the acceptable range. Reporting the uncertainty provides confidence that these results are valid, preventing unnecessary rejection of good batches, see Figure 3. 

Knowing and understanding measurement uncertainty allows: 

• Comparison of results from different labs or devices. 
• Confident decision-making in quality control and regulatory assessments. 
• Improved reliability of results. 

Key Takeaways

1. No result is perfect: Every measurement has some degree of uncertainty. 
2. Every method should be validated: Understanding bias and precision is key to ensuring accuracy. 
3. Knowing your uncertainty leads to confidence: It helps make informed and fair decisions. 

From Lab Precision to Field Efficiency

Whether you’re fortifying food products, conducting quality control, or ensuring compliance with nutritional standards, BioAnalyt’s iCheck tools offer dependable solutions. These tools are rigorously validated and include detailed performance guides, empowering professionals to make confident, data-driven decisions. 

 Each tool is backed by validation and detailed performance guides. For product support, and training, or to place an order, contact us at support@bioanalyt.com  

Together, let’s ensure your measurements are accurate, reliable, and impactful. 

References

1. Uncertainty of measurement — Part 3: Guide to the expression of uncertainty in measurement (GUM:1995). ISO/IEC Guide 98-3:2008(en). Available from www.iso.org. Reviewed and confirmed in 2023 
2. Bell S. A Beginner’s Guide to Uncertainty of Measurement. Measurement Good Practice Guide No. 11 (Issue 2). National Physical Laboratory, UK, 2001.
3. De Bièvre, P. and Günzler, H. Measurement Uncertainty in Chemical Analysis. Paul Springer Science & Business Media, 29 Jun 2013.
4. Ellison, S.L.R and Williams, A. Eurachem/CITAC guide: Quantifying Uncertainty in Analytical Measurement, Third edition, (2012) ISBN 978-0-948926-30-3. Available from www.eurachem.org.
5. Guidelines for Laboratories Performing Microbiological and Chemical Analyses of Food, Dietary Supplements, and Pharmaceuticals – An Aid to the Interpretation of ISO/IEC 17025:2005. (2015) (ALACC Guidelines). Available from www.aoac.org.
6. White, G.H. Basics of Estimating Measurement Uncertainty. Clin Biochem Rev. 2008 Aug; 29 (Suppl.1): S53–S60.
7. FAO Guidelines on Measurement Uncertainty, 2021.

ABOUT THE AUTHOR

Dr. Anna Zhenchuk is the Managing Director at BioAnalyt, where she has overseen new product development and provided hands-on support to customers in over 30 countries. Her background is in biotechnology, immunology and nutrition as well as business administration.