Repeatability describes variation in repeated observations under the stated measurement conditions.
Why this matters in the industry
Research teams need an evidence-based view of variability rather than selecting one favorable reading.
The technical reasoning
Repeatability depends on what is repeated: instrument readings, connection cycles, full setups or entire sessions. Each procedure samples a different range of variation. Research records should identify the repetition level so that a narrow within-connection spread is not presented as evidence of complete experiment reproducibility.
From instrument readings to defensible results
Calibration, correction and verification have different roles. Calibration establishes a relationship under stated conditions; correction uses a model to adjust an indication; verification checks selected behavior against a defined criterion. A calibrated instrument does not automatically characterize the cables, adapters, fixtures and software around it. Repeated readings can estimate some random variation, but they do not expose every systematic error. The method must identify the measured quantity and the route through which its value is inferred.
How to structure the investigation
Define what is repeated, including whether connections are remade. Keep operating conditions documented and collect enough observations for the intended analysis. Report individual readings or appropriate statistics and distinguish repeatability from changes caused by different setups.
Define the plane, frequency range and operating state. Preserve raw readings, correction files and reference identities, and distinguish measurements made without reconnecting from repetitions of the full setup. Use an independent reference check where practical. When comparing two routes or stations, collect repeated observations and look for frequency-dependent offsets and spread. Investigate unexplained differences before treating a software correction as a solution.
Worked example or engineering scenario
Twenty readings from one unchanged connection can have low spread, while five independent remountings show a much larger shift. These datasets describe different sources of variation.
Evidence to collect
| Record | Purpose |
|---|---|
| Repeated action | Defines the tested state and scope of the comparison. |
| Operating conditions | Makes the stimulus or route condition reproducible. |
| Observation count | Supports interpretation of variation and possible confounding effects. |
| Reported variation | Connects the observation with the stated engineering decision. |
Trade-offs and common interpretation errors
Agreement between two systems can conceal a shared error. A stable reference can also drift or be damaged. State the scope of the comparison and the evidence supporting the reference's stability. A small observed difference should be interpreted alongside uncertainty and repeatability, rather than assumed to be a meaningful device improvement.
What the result can support
Design repetitions to represent the intended use and report the repetition procedure with its statistics.
Repeatability alone does not establish the absence of systematic error.
Further technical reading
Related industry knowledge
- How to Check an RF Lab Setup Before a Measurement Session
- How to Characterize RF Components Across a Frequency Sweep
Numerical scenarios are illustrative assumptions, not reported measurements of a supplied product or installation.