Uniformity, Reliability, and Accuracy

We do not measure things. Rather, we measure properties of things. For example, we cannot measure a box, but we can measure its properties such as its mass, length, and temperature. 

To properly compare and analyze the things in our universe, we need to compare and analyze their properties. Because people all around the world are making measurements, we must ensure there is agreement on all the various types of measurements used. Difficulties arise because measurements have both quantitative and qualitative aspects. The fact that the two lines on the international prototype platinum-iridium bar in France are one meter apart is quantitative how you measure other objects with that bar is qualitative. The ability to match up two lines may seem simple, but, depending the desired accuracy, such simple operations are in fact difficult. This difficulty is why using measuring equipment is a qualitative art. 

Uniformity requires that all people use the same measurement system (i.e., metric vs. English) and that all users intend that a given unit of measurement represents the same amount and is based on the same measurement standard used 

Reliability requires the ability to consistently read a given measurement device and also requires a given measurement device to perform equally well, test after test. It is the user s responsibility to know how to achieve repeatable data from the equipment being used. 

Accuracy refers to how well a measurement device is calibrated and how many significant figures one can reliably expect. It is the user s responsibility to know how to read his equipment and not interpolate data to be any more accurate (i.e., significant figures) than they really are. 

---

Laboratory research is dependent on reliable quality measurements and the use of uniformity, reliability, and accuracy to achieve this goal and this cannot be emphasized enough. Poor or inaccurate measurements can only lead to poor or inaccurate conclusions. A good theory can be lost if the experimental data are misread. 

---