Contemplating Precision and Bias

Form and Style Can Help You Along

The first things you'll notice when you begin to read A21 are references to ASTM standards that will aid you in the pursuit of precision and bias:

  • E177, Practice for Use of the Terms Precision and Bias in ASTM Test Methods;
  • E456, Terminology Relating to Quality and Statistics, and
  • E691, Practice for Conducting an Interlaboratory Study to Determine the Precision of a Test Method.

Form and Style states that precision is "the closeness of agreement between test results obtained under prescribed conditions." A statement on precision provides potential test method users the means to assess how useful the test method would be for proposed applications.

As A21.2.2 notes: "Every test method shall contain a statement 1) regarding the precision of test results obtained in the same laboratory under specifically defined conditions of within-laboratory variability (repeatability conditions), and 2) regarding the precision of test results obtained in different laboratories (reproducibility conditions)." An example precision statement is provided.

A temporary statement must be included if the responsible committee decides that an interlaboratory study should be delayed. This statement addresses repeatability and notes that reproducibility is being determined.

Bias is defined in A21.3.1 as "a systematic error that contributes to the difference between the mean of a large number of test results and an accepted reference value."

The bias statement must describe the bias and the methods employed to provide corrected test results. A21.3.2 notes that if the bias is not known but the direction or the bounds on the bias, or both, can be estimated, these must be reported in the bias statement.

Subsections A21.4.1 through A21.4.2 provide general considerations to keep in mind when preparing a precision and bias statement. These include providing a brief descriptive summary of the interlaboratory study that will permit test method users to judge the reliability of data.

Finally, some test methods may involve exceptions to the directions given in A21, and instructions on handling these cases are given in A21.5 along with sample precision and bias statements to address these issues. Examples of exceptions include, but are not limited to:

  • When a test method specifies that a test result is a non-numerical report of success or failure or other categorization or classification based on criteria specified in the procedure (A21.5.1), and
  • When a test method specifies that the procedure in another ASTM test method is to be used without modification, no statements of precision and bias are necessary if those in the other test method are applicable (A21.5.2).

While this column covers some of the basics, committee members are urged to study A21 further when considering the precision and bias statements for their test methods.

Industry Sectors

Issue Month
May/June
Issue Year
2011