21
Bias Identification
Most samples can be analyzed for percent moisture content with a CEM drying
method, and results will compare precisely with the standard test method.
Occasionally, however, all volatiles in a particular sample cannot be removed with
microwave power. Different spread techniques can be used to eliminate the need
for a moisture bias in most samples. Sometimes browning of the sample in the
same manner as in the air oven will also produce test results comparable to the
standard test.
If microwave moisture test results are repeatable, but not comparable to the
standard test results, a bias can be identified. With most samples, the bias will
amount to only a few tenths of a percent. To determine a moisture bias, prepare a
sample for testing, ensuring that the sample is homogeneous. Test this sample by
both a standard method and the CEM microwave method. Perform at least ten tests
by each method to check repeatability, then figure an average of the ten results.
Subtract the average of the microwave test results from the average of the standard
test results. Enter the resulting number (moisture bias) into the CEM moisture/
solids instrument. Corrected moisture test results will be displayed and printed. If
the difference in the average test results is a negative number, the power level
should be reduced and the tests should be repeated.
Some meat and poultry products, whether raw or cooked, show a fat bias when
analyzed with the CEM instruments. Less fat is extracted by a short solvent
extraction than by the longer ether extraction of the traditional Soxhlet technique.
The bias will be a constant for a given product and should be considered when
generating data.
Cooked sausage (i.e. wieners, bolognas, etc. with extenders) and other cooked
products tend to bind fat in the carbohydrate/protein structure of the cooked
product and will generally yield a lower fat value when extracted in the CEM
instrument than when extracted with the Soxhlet.
For proper determination and use of biases, results from the CEM instrument
should be compared with the results from traditional “long” or standard methods
used by USDA laboratories for compliance testing. Standard methods consist of an
air oven (four hours, 125
°
C) for moisture and Soxhlet (four hour ether extraction)
for fat. USDA recommends correlation of rapid fat and moisture methods with
these long methods.
USDA Processed Products Inspection Division encourages the use of rapid
methods for fat and moisture analysis to assist in better process control and
compliance. USDA recommends periodic back-up analysis (one in every 10 - 20
analyses) by standard long methods when rapid methods are used in approved
Total and Partial QC (TQC/PQC) Programs. Backup checks can be performed less
frequently as laboratories gain confidence in the established biases. CEM
Corporation also recommends a Quality Assurance Program for establishing and
maintaining biases and to ensure USDA compliance.
Prior to establishing a fat bias for the Fat Analysis System, a laboratory should be
confident of obtaining the best possible results for moisture, since precision of the
fat results will depend on the moisture results. Moisture results should have
repeatability between duplicate analyses of the same sample and a mean or