1. Statistical Reference Guide
  2. Method comparison / Agreement

Testing commutability of other materialsTest the commutability of processed samples, such as EQA, PT, QC or reference materials, with unprocessed samples, using a prediction interval or a preset criterion around the regression line.

A processed sample is commutable if its result lies within the scatter of the unprocessed samples about the regression line. There are two ways to judge it:

  • Prediction interval. The processed sample is commutable if it falls within the prediction interval of the unprocessed samples. Use it when a clinically significant difference has not been established, or when the measurement procedures are imprecise or nonspecific. You must fit a Deming regression to use a prediction interval.
  • Preset criterion. The processed sample is commutable if its difference from the regression line is within a clinically acceptable difference. Use it when such a difference has been established. It avoids rejecting a sample for a clinically insignificant difference when the unprocessed samples lie close to the line, and detects a difference the prediction interval would miss when their scatter is wide.
  1. On the Analyse-it ribbon tab, in the Method Comparison group, click Restrict to Group.
  2. In the Group / Color / Symbol drop-down list, select the variable indicating the type of sample.
  3. In the Restrict fit to group drop-down list, select the group of unprocessed samples.
  4. Select the criterion:
    • To use a prediction interval, on the Fit Y on X task pane, in the Prediction band edit box, type 95%, and if required select the Familywise coverage check box. If each sample is measured with the same number of replicates by both methods, in the Method drop-down list, select CLSI to compute the interval as CLSI EP14.
    • To use a preset criterion, on the Descriptives task pane, select the Scatter plot and Allowable difference band check boxes, and then in the Absolute edit box, type the criterion in measurement units, and/or in the Relative edit box, type the criterion as a percentage (suffix with % symbol). In the Around drop-down list, select Fit line.
  5. Click Recalculate.

The scatter plot shows the points in different colors depending on the type of sample. The fit is restricted to the unprocessed samples, and the prediction band or allowable difference band is drawn around it.

A table lists each additional sample with its predicted Y. For a prediction interval, it shows the interval and marks Y as not commutable if it lies outside it. For a preset criterion, it shows the difference from the predicted Y and the allowable difference, in measurement units or as a percentage of the predicted Y when the relative criterion applies, and marks the difference if it exceeds the allowable difference.

Available in
Method Validation Edition Ultimate Edition

From the Statistical Reference Guide for Analyse-it 6.24.0: https://analyse-it.com/docs/user-guide/method-comparison/testing-commutability

Statistical Reference Guide v6.24.0

Administrator's Guide

User's Guide

Tutorials