What Is a Proficiency Testing Z-Score and How Do You Interpret It (ISO 13528)?

If you work in a testing or calibration laboratory, you have probably received a proficiency testing (PT) report full of numbers—and one of them is the z-score. But what does it actually tell you about your lab's performance? And how do you know if a z-score of 2.5 is a problem or just a warning?

This article explains the proficiency testing z-score according to ISO 13528, the international standard for statistical methods in proficiency testing. You will learn the formula, the official interpretation criteria, and how to avoid common mistakes.

What It Is

A z-score is a standardized measure of how far your laboratory's result deviates from the assigned value (the "true" or consensus value) in a proficiency testing round. It expresses that deviation in units of the standard deviation for proficiency assessment.

In simple terms, the z-score tells you: "How many standard deviations is my result away from the target value?"

Because it is unitless, the z-score allows fair comparison across different analytes, methods, and concentration levels. It is the most widely used performance statistic in proficiency testing under ISO 13528.

How It Works: Formula and Interpretation

The z-score is calculated using the following formula (ISO 13528, Section 9.4):

\[
z = \frac{x - X}{\sigma}
\]

Where:

• x = your laboratory's reported result
• X = the assigned value (e.g., the robust mean of all participants, or a certified reference value)
• σ = the standard deviation for proficiency assessment (a measure of acceptable variation)

### Official Interpretation Criteria (ISO 13528)

z-score value | Performance rating

• | z | ≤ 2.0 | Satisfactory
• 2.0 < | z | < 3.0 | Questionable (warning signal)
• | z | ≥ 3.0 | Unsatisfactory (action signal)

These thresholds are not arbitrary—they are based on the normal distribution. Assuming results are normally distributed, about 95% of acceptable results fall within |z| ≤ 2, and about 99.7% fall within |z| ≤ 3. A z-score beyond 3 is therefore statistically unlikely to occur by chance alone.

A Worked Illustrative Example

Example data (illustrative only):

Suppose your laboratory measures the concentration of lead in water as part of a PT round.

• Your reported result: x = 12.4 µg/L
• Assigned value (robust mean of participants): X = 10.0 µg/L
• Standard deviation for proficiency assessment: σ = 1.2 µg/L

Applying the formula:

\[
z = \frac{12.4 - 10.0}{1.2} = \frac{2.4}{1.2} = 2.0
\]

Your z-score is exactly 2.0. This is right at the boundary of the "satisfactory" range. According to ISO 13528, |z| ≤ 2.0 is satisfactory, so your result is acceptable—but it is a clear warning that you are at the edge of the acceptable zone. You should review your method for potential bias.

If your result had been 13.6 µg/L instead, the z-score would be:

\[
z = \frac{13.6 - 10.0}{1.2} = 3.0
\]

This would be unsatisfactory, requiring immediate investigation and corrective action.

Common Pitfalls

1. Confusing the sign. A negative z-score simply means your result is below the assigned value; a positive one means it is above. The sign matters for diagnosing bias, but the absolute value determines the performance rating.

1. Using the wrong sigma. The σ in the z-score formula is the standard deviation for proficiency assessment, not your laboratory's own repeatability or reproducibility standard deviation. Using your own SD will give misleadingly large or small z-scores.

1. Ignoring the assigned value's uncertainty. ISO 13528 notes that if the uncertainty of the assigned value is not negligible relative to σ, the z-score interpretation may need adjustment. Check your PT provider's report for this information.

1. Treating a single questionable z-score as a failure. A z-score between 2 and 3 is a warning, not a fail. Investigate the cause, but do not overreact to one borderline result.

Closing

Understanding your z-score is the first step to turning proficiency testing data into real quality improvement. Instead of manually crunching the numbers and risking errors, use our free, instant Proficiency Testing Z-Score Calculator at https://www.6sq.com/tools/zscore/ — just enter your result, the assigned value, and the standard deviation, and get your z-score and performance rating in seconds.