Errors of Measurement
To ensure the quality of a product, errors of measurement is very important, as test cases are executed at various levels defects if found any in the tested area. The defects are logged and data is collected. The defects are then fixed. The testing team verifies if the defect was really fixed and closes the defects. This information is very beneficial. Proper tracking ensures that all defects were fixed. This information also helps in future projects.
Variation in a process has two sources:
- Variation of the process that makes the product.
- Variation of the measurement process.
Even when rightly used, a measuring instrument may not give a true reading of a characteristic. The distinction between the true value and the measured value can be due to one or more five sources of variation.
Bias is the difference between the observed average of measurements and the reference value. The reference value is also termed as accepted reference value or master value. It is a value that serves as an agreed-upon reference for the measured values.
Any statement of bias and repeatability must be preceded by three conditions:
- Definition of the test method: this definition includes the step-by-step procedure, equipment to be used, preparation of test specimens, test conditions and so on.
- Definition of the system of causes of variability: these include materials, analysts, apparatus, and laboratories.
- Existence of a statistically controlled measurement process: the measurement process must have stability for the statements on bias and precision to be valid.
Errors of measurement can cause incorrect decisions on:
- Individual units of product.
- Lots submitted to sample the plans.
In the classification of a product, two types of errors can occur, they are:
- A non conforming unit can be accepted (the consumer's risk).
- A conforming unit can be rejected (the producer's risk).