The use of volunteer radon measurements for radon mapping purposes: an examination of sampling bias issues

Abstract

National and regional radon surveys are used in many nations to produce maps detailing the spatial variation of indoor radon concentrations. National surveys which are designed to be representative use either a geographically-weighted or a population-weighted sampling scheme. Additionally, many countries collect a large number of data on indoor radon concentrations from volunteers who have chosen to have the indoor radon concentration measured in their own dwellings. This work examines the representativeness of volunteer-based samples in radon measurement and explores the effect of potential volunteer bias on radon mapping results. We also investigate the influence that media attention has on volunteer sampling of indoor radon concentrations. The result of our work indicates that volunteer measurements are biased due to over-sampling of high radon areas. Consequently such volunteer radon measurements should not be used for radon mapping purposes.

Results of the first national indoor radon survey performed in Serbia

“…that the number of collected detectors was 5300. Since distribution of detectors was on a voluntary basis, this could create bias in the survey as it was reported in some other national surveys (Burke and Murphy 2011). Since participants in the projects were mainly high-school students and volunteers, there could be a tendency for participation in such studies to be biased towards people who are more highly educated, which can further be correlated with their level of wealth and consequently their housing.…”

Establishing a regional reference indoor radon level on the basis of radon survey data

“…In practice the random method of dwellings sampling for radon survey is impractical due to the often poor cooperation of citizens and homeowners, selected at random. The alternative approach based on inviting volunteers is criticised due to over-sampling from high radon areas and introducing large biases in the survey results [12][13][14]. In our work to create a representative sample, as best as practically achievable, we involved groups of citizens in Ekaterinburg who were not organised on a territorial basis but as employees of different institutions and companies, university students, school teachers etc, who are more accessible and more cooperative than randomly selected citizens.…”

Qualitative overview of indoor radon surveys in Europe

“…Any deviation from pure random sampling can cause biases (IAEA, 2013). It was shown that volunteer measurements could be biased due to the over-sampling in radon priority areas (Burke and Murphy, 2011).…”