Do Your Two Methods Actually Agree?

The method, in full

The practical guide behind this tool: when it applies, how to read the output, and the traps that make it say the wrong thing.

Read the guide →

How it works

The analysis works on each row's difference between the two methods, plotted against the pair's mean (the Bland-Altman plot). It computes the bias (mean difference) with a t-based 95% confidence interval, the 95% limits of agreement (bias ± 1.96·SD of the differences) each with its own confidence interval via the classical Bland-Altman standard error, a regression of the difference on the mean to detect proportional bias, and the observed share of points inside the limits. The correlation between the methods is reported alongside and explicitly contrasted with agreement.

Use it whenever two methods measure the same quantity on the same units and you need to know whether they can be used interchangeably: a new device against a gold standard, two instruments, two labs, two assays.

Not for two measurements of different quantities or different units (the difference is meaningless), not for asking whether a measurement changed over time (use the paired comparison tool), and not for categorical ratings (use an agreement statistic for categories).

What data do you need?

Two numeric measurements of the same quantity on the same samples, one pair per row. For example, a portable meter against the lab reference:

Minimum 10 rows · Best with 40-5,000 paired rows

What's in the report?

Standard-library analysis: do two measurement methods agree well enough to be used interchangeably? Map two numeric columns measuring the same thing on the same units: a new device against a reference, two instruments, two assays, two raters scoring a continuous quantity, and get the Bland-Altman plot with bias and limit lines, the bias (mean difference) with a 95% confidence interval, the 95% limits of agreement each with its own confidence interval, a proportional-bias check, the observed share of points inside the limits, and the correlation coefficient explicitly contrasted with agreement.