Water Measurement Uncertainty Builds Confidence in Tests

That's where water measurement uncertainty comes in. Water measurement uncertainty accounts for variables that could impact the reliability of a given test. Rather than undermining test results, this acknowledgment of uncertainty boosts transparency by providing context and strengthening confidence. By presenting the whole picture, labs, utilities, regulators, and customers know how much trust they can place in a number.

"Measurement uncertainty doesn't mean your result is wrong," Patsy Root, senior manager of government affairs at IDEXX, said during the IDEXX Water Academy webinar "Basics of Measurement Uncertainty." "It means you understand the possible variation in your result—and that makes your data stronger."

Below, we'll take a closer look at measurement uncertainty, including its definition, significance, influencing factors, calculation process, and validation and verification.

Defining Measurement Uncertainty

The International Standards Organization (ISO) defines measurement uncertainty as the lack of absolute accuracy in a result. This measurement describes the range that encompasses the true value of a measurement with variable factors.

In practice, labs account for measurement uncertainty in situations where variation is inevitable, such as when they conduct microbiological tests that count living organisms capable of moving and interacting. Whether lab managers are conducting these tests using Most Probable Number (MPN) or Colony Forming Units (CFU), no two counts will be identical. In this case, measurement uncertainty quantifies this variability.

Understanding Measurement Uncertainty's Significance

The implications of measurement certainty for public health are significant, ensuring that test results are defensible, transparent, and actionable.

Take, for example, EPA beach monitoring, which requires advisories when enterococci hit a limit of 35 per 100 mL. If a water test lab reports a result of 34, a beach can theoretically safely stay open. However, unaccounted-for uncertainty could bring the actual number to 35 or higher, leaving decision-makers in a gray area, questioning the best course of action.

"There is almost nothing worse than having to retest," Root said. "It takes more time, more money, creates a lot of angst, and undermines the confidence of the customer."

When labs include uncertainty in their results, a report of 34 becomes a reading of 34.6–38. This range provides public health officials with the information needed to act decisively. To continue with our example, the reported measurement uncertainty may mean closing a beach to protect swimmers or avoiding unnecessary closures that frustrate beachgoers.

Influencing Factors for Measurement Uncertainty

Every lab process can introduce variability. During the IDEXX Webinar, Root highlighted several of the most common contributors that influence measurement uncertainty:

• Method used: Standardized, validated methods (ASTM, ISO, Standard Methods) reduce uncertainty.
• Media supply variation: Different suppliers or lots may affect organism growth.
• Sample collection: Poor technique or chain-of-custody issues introduce error.
• Incubation temperature: Uncalibrated incubators or water baths cause variability.
• Preprocessing: Labs must perform dilutions, filtration, or acid treatments aseptically.
• Standard operating procedure (SOP) adherence: Shortcuts or lack of attention paid to SOPs increases risk through inconsistency.
• Media storage: Not following the manufacturer's instructions affects test reliability.

"You'll never be able to get rid of measurement uncertainty, but you can certainly minimize it," Root said. Mitigation strategies include calibration with certified providers, technician training, internal audits, control charts for media performance, SOP training, and ongoing demonstrations of capability.

Calculating Measurement Uncertainty

"The data you generate is used to make decisions—whether to open or close a beach, if water is safe to drink, or if a building has risk for Legionella," Root said. By reporting uncertainty, labs ensure those decisions rely on an accurate picture, not a false sense of precision.

Labs can quantify uncertainty in two primary ways: standard deviation of replicate measurements and 95% confidence intervals.

Standard Deviation of Replicate Measurements

Standard deviation is a statistical measure of data spread that shows how much individual data points deviate from the average.

For example, a plate reading of 16 colonies with a standard deviation of ±2.1 means the true result lies between 13.9 and 18.1. If the regulatory limit is 17, that range signals the need for action.

95% Confidence Intervals

Labs widely use this method in microbiology for both MPN and CFU analyses. Tables from Quanti-Tray systems or Standard Methods provide ready references.

For example, five positive wells in a 51-well tray equate to 5.3 MPN/100 mL, with a 95% confidence interval of 2.3–12.3.

Teasing Apart Validation and Verification

Reducing uncertainty also depends on starting with sound methods that have been both validated and verified.

"Starting with a validated method that you've verified in your laboratory is the very first step to mitigate as much measurement uncertainty as possible," Root said.

With that in mind, the distinction between validation and verification is critical.

Method developers or third parties validate a method to demonstrate its performance characteristics, including sensitivity, specificity, reproducibility, and false positive/negative rates. In the U.S., this is typically handled by developers and reviewed by the EPA. In Europe, validation often follows ISO 13843.

Labs everywhere verify that they can achieve similar results under their own conditions before adding a method to their accreditation scope.

Building Confidence Through Transparency

Many view measurement uncertainty as complex, but it serves a straightforward purpose: building confidence through transparency. Every measurement carries uncertainty, and quantifying this variance ensures decisions rely on reliable, transparent data. By acknowledging the variability inherent in microbiological testing, labs protect themselves, reassure clients, and safeguard public health.

Root said, "If you can explain the reliability of your results, you're protecting both your laboratory and the community you serve."