7. Sources of Error

In the process of gathering soil samples for use in laboratory investigation, errors can occur at several stages:

• sampling errors

• selection errors

• measurement errors

Sampling errors. Sampling errors cannot be eliminated entirely. Since sampling means to take a part, errors can be reduced by careful selection.

Selection errors. Selection errors can occur when a sampler, eager to do a good job, over-samples the borders of a field. Sampling in rocky areas can also cause error. The errors often cancel each other. Otherwise the sampler may employ the two-step procedure of the simple random sampling which is designed to eliminate selection errors.

Measurement errors. These types of errors arise in various circumstances:

• When the measurement taken is not the true value of the unit. One example is the random error. A random error takes place when cores of soil are mistakenly assigned constant weights, even though they are variable.

• As a result of variations in analytical techniques. Fortunately, random errors of measurement tend to cancel each other as the sample size becomes larger.

• Bias errors arise either because tare weights are ignored or because of an offset calibration of the appropriate curve.