ABSTRACT
Analytical errors and undesirable performance are discovered as the result of a laboratory's own quality assessment efforts, from its participation in collaborative test programs, or from external performance audits. The first objective should be to positively identify the kind of problem that is present, whether due to imprecision or bias. In the event of both possibilities, problems of precision must be addressed first. The identification of measurement problems is facilitated by using the approach described by Ishikawa based on the use of a cause-effect diagram. However, one should remember that improvement of a measurement process results only from the elimination or reduction of chance causes and their corresponding effects. Performance evaluation is often based on the results of a collaborative test. The meticulous analyst can become unduly concerned with the quality of data to the point that every value is critically screened for its credibility.
