Bias

If the difference in the results of an analysis with an outcome that is considered true (reference) is always greater or smaller, then the event is called bias.

Bias can occur at any stage of sampling, either manual or mechanical sampling, and the time distribution of samples in the sample preparation stage, or other conditions. Although this bias occurs in both stages of sampling and sample preparation stage, but very difficult to be able to measure the level of bias.

Processes that are most likely to do the measurement bias in the sampling process is the conveyor belt.

The most common and many people do is measurement bias (bias test) at sampling mechanically (mechanical sampling).

Bias test is done by comparing the results of analysis (usually using ash as a comparison parameter) of the stopped belt samples (as a reference) with the results of analysis of samples taken with the mechanical sampling.

Analysis results obtained from the reference sample (taken with a stopped belt sampling) is considered as the result of the closest to the actual values.

The true value of a parameter no one ever knew. The value of a parameter that is considered closest to the true value is the average value of the results of analyzes performed by the laboratory as much as possible.

Reference :
AS 4264 : Part 4 - Determination of precision and bias
AS 2646 : Part 8 - Determination of precision and bias
ASTM D6518-03a - Practice for Bias Testing a Mechanical Coal Sampling System