What is an independent correlation study?

What is an independent correlation study?

An independent correlation study checks that different measurement systems give the same answer, within a stated uncertainty. In practice that means measuring a part or reference artefact with two or more systems, then comparing results to reveal bias, systematic differences, or gaps in feature coverage. Correlation shows the measurement chain is honest and traceable.

Why do we need to check?

Measurements are the evidence used to accept parts, control processes, and prove compliance. For low-risk items a single, routine check may be fine, but for high-risk, high-precision parts, such as aerospace flight components, unverified measurements can cause scrap, rework, or worse. Correlation reduces that risk by confirming that:

• the inspection method reports the true value (within its stated uncertainty), and
• different systems and operators interpret features the same way.

Who checks the checker?

Just as manufacturing needs inspection, inspection needs verification. Correlation exercises are the practical way to “check the checker”, by comparing a primary measurement system (often a CMM) with secondary systems such as another CMM, an articulated arm, a laser tracker, optical gauging, or calibrated hand tools. The exercise should be independent of routine calibration and performed by personnel or teams that are impartial to normal inspection shifts.

What correlation exercises look like, step by step

Here is a typical correlation workflow used to verify a CMM programme has been written correctly by comparing it to another independently created CMM programme:

1. Define scope and acceptance criteria — list the features to be inspected, their tolerances, and the acceptance criteria (usually expressed as a percentage of tolerance).
2. Choose the parts — select representative production parts along with a few borderline or out-of-tolerance examples to ensure the system measures both cases accurately.
3. Run measurements — inspect each part on each CMM, recording results, operators, and any other relevant details.
4. Compile the results — use a spreadsheet or dedicated SPC software to calculate differences in average measured values for each feature across systems, expressed as a percentage of the tolerance size. If any discrepancies are found, consider rectification and repeating the measurements.
5. Decide and document — accept or reject the inspection method for each feature, noting any exemptions. Retain measurement records, calibration certificates, and the correlation report for audit purposes.

Alternatively, the same exercise can be run using calibrated artefacts such as length bars or gauge blocks to verify both machines agree to a traceable reference. This isn’t usually needed if each CMM is already being calibrated regularly with a certificate traceable to national standards, but it can be useful when you want an extra check or need to compare two systems directly.