Correlation Examples and Practice Datasets

Try the same dataset with multiple methods to see how linear, monotonic, and binary-group assumptions change the result.

Examples are also useful for checking whether a reported coefficient matches what the raw data actually shows.

Start by loading a dataset that resembles your real use case, then inspect the coefficient, p-value, confidence intervals, and chart together. Looking at only one metric usually hides important structure.

After that, switch methods on the same data. If Pearson and Spearman disagree materially, that is often a signal that the relationship is monotonic but not well described by a straight line.

The example library is intentionally mixed. Some datasets are clean and strongly aligned with one method, while others are ambiguous on purpose so you can see where method choice becomes a judgment call.

That makes the page useful for teaching, self-study, and quick calculator verification when you want known benchmark-style inputs.

Example datasets are excellent for learning, but they are simplified. Do not assume that a neat teaching example captures all the complications of production data such as missingness, measurement error, clustering, or confounding.

Treat the examples as practice for method selection and interpretation, then verify your own data with the calculator and, when stakes are high, with established statistical software.