Association and agreement for method comparison: which method to choose?

When comparing two measurement methods in clinical research, two statistical approaches are frequently used: association and agreement.

Association (correlation) measures the strength of the linear relationship between two variables. However, a strong correlation does not mean that the two methods give interchangeable values.

Agreement assesses how closely two methods produce the same results. The Bland-Altman method is particularly suited to this analysis.

In conclusion, to validate a new measurement method against a gold standard, agreement is generally preferable to simple association.