Every measured value has a range of uncertainty, which is commonly called accuracy. The range of uncertainty means that the true value lies within that range with a stated likelihood (typically a likelihood of \(\SI{95}{\%}\) will be applied). The uncertainty can result from the measurement itself, the calibration of the instrument and the properties of the instrument. As an example, a steel ruler is difficult to read in the sub-millimetre range and the result will slightly differ from person to person. Of course, the steel ruler was never compared to the original metre in Paris, but the scale was applied by machine, and even with the best equipment there will be a deviation. And of course, the steel ruler expands when heated, so we read different values with the same ruler at different temperatures.

The data sheets of radon monitors frequently show a mixture of different accuracy data, which the user often interprets just as it fits with his plan. Therefore, the three main uncertainty contributions in the case of radon measurement are to be explained here. These are:

• the intrinsic uncertainty of the measuring instrument,
• the calibration uncertainty,
• the statistical uncertainty of the measurement.

The total uncertainty of a measurement results from the superposition of all three individual uncertainties. Depending on the measuring instrument and the measuring conditions (especially the selected sampling interval and the present radon concentration), one of the three uncertainties may become dominant.

This specification describes the insintric accuracy of an instrument, which results from the measuring principle and the technical implementation of data acquisition. The uncertainty includes, for example, non-linearities and hysteresis in signal processing, but also the influence of environmental conditions (climatic parameters, external radiation, etc.) or the measured concentration itself. For instruments working in event counting mode, usually only the ambient conditions are of interest. The intrinsic uncertainty describes one factor of the quality of a device and should be explicitly stated in the data sheet of an instrument.

The statistical uncertainty of a measurement results from the nature of radioactive decay itself. The number of decays observed by an instrument per sampling interval is a random quantity. That means even at constant radon concentration, a different number of decays is observed in successive sampling intervals. The mean value from many (infinitely) such sampling intervals would represent the true value of the radon concentration. However, based on the absolute number of observed decays within a sampling interval, the standard uncertainty of the measured value calculated from the measured number of decays can be estimated. The following applies here: The greater the number of observed decays, the smaller the relative standard uncertainty. The absolute number of decays detected by an instrument depends on the selected sampling interval, the radon concentration and the sensitivity of the instrument. Thus, the statistical uncertainty results from the respective measurement itself. This means that there can be no general specification of the measurement uncertainty for a radon measurement. The standard uncertainty is calculated for each measured value by the device and output together with the measured value. In the data sheet, the uncertainties for typical concentrations and measurement intervals should be given to characterise an instrument.

Calibration means the comparison of the results of an instrument with a standard (also called reference). Since a direct comparison of each measuring instrument with the national standard is practically impossible, this standard is passed on by a chain of subsequent calibrations up to the standard of a calibration laboratory. In the entire calibration chain (comparison measurements), statistical uncertainties occur both in the standards and in the instruments to be calibrated. The national standard itself also has an uncertainty resulting from its derivation. The uncertainty resulting from the entire calibration chain is called calibration uncertainty. In addition to the calibration factor, the associated calibration uncertainty is an obligatory component of every calibration certificate. According to ISO, the standard uncertainty expanded by a factor of two is specified.