Measurement Error

Here I am in silence
Looking 'round without a clue
I find myself alone again
All alone with you
--Information Society

All measurement systems have error. All of them. Error is the difference between a observed (measured) value and the actual (true) value.

error = observed - actual

Measurement error can be of two types. There can be a systematic error, or bias in the measurement system that causes a consistent offset from true value. For example, a scale that has not been 'zero-ed' might consistently read two pounds light. 

Systematic error affects accuracy, or the ability of the measurement system to capture the actual value. The higher the systematic error, the lower the measurement system's accuracy.

Measurement error can also be random in nature. Random error causes offsets from the true value, but they are neither consistent in magnitude or direction. A forehead thermometer might measure a different temperature depending on where precisely it is placed on the head.

Random error affects precision, or the ability of the measurement system to reproduce the same result repeatedly under similar conditions. The higher the random error, the lower the measurement system's precision.

The below picture summarizes the two types of measurement error well.

Because all measurement systems have error, the big challenge for decision-makers is to determine just how much error is baked into information systems that they want to use. Too much error renders a measurement system worthless.