[en] Air surveys, whole-body counting, bioassays or combination of these measurements can be utilized for purposes or assessing internal doses to determine compliance with occupational dose equivalent limits. Air sampling with a little support provided by whole body counting and/or bioassays was often relied on in dose calculations. The utility of air sampling for internal dose reconstruction is addressed in this paper through the probabilistic analysis of environmental factors and their impact on dose estimates. In this paper we attempt to reconstruct an internal dose due to inhalation of beta + gamma emitting radionuclides for a contractual electrician, Mr. X. The data available for reconstruction of internal dose for Mr. X was found to be highly variable and uncertain. Uncertainty describes a lack of knowledge about a parameter, this lack of knowledge theoretically can be reduced, e.g., if more measurements were to be taken (for example, estimated activities for alpha-emitting radionuclides are uncertain due to the influence of naturally-occurring alpha-emitters). Variability describes the existence of different values that represent different environmental conditions (for example, the air concentrations of radionuclides may vary over time because of the different tasks performed by workers in the area). Variability can not be reduced by additional data collection because the varying values reflect the variable nature of the environment, not a lack of data. The high variability in measured air concentrations in the restricted areas of a LWR nuclear power plant where he worked do not allow adequate reconstruction of his individual internal dose using deterministic methods and therefore probabilistic methods are desirable. The guidance for probabilistic assessment developed by the United States Environmental Protection Agency as well as recommendations of the National Council of Radiation Protection provide an adequate framework for probabilistic reconstruction of Committed Effective Dose Equivalent (CEDE) for Mr. X. The total internal intake was calculated using Monte-Carlo simulations and then translated to dose using dose conversion factor. The results of the analysis reveal a broad distribution in dose estimates. The resulting dose distribution is significantly skewed to the high doses. The mean dose estimate is significantly different from the median dose. Median doses, which are often required for litigation in similar cases (to estimate if somebody, more likely than not, has received a dose above specified limit), should therefore be estimated using probabilistic techniques since using the mean for deterministic estimates may lead to erroneous conclusions for non-symmetric data distributions. Sensitivity analysis was conducted to determine which of the input parameters contributed most to the uncertainty in internal dose estimates. The spatial and temporal variations in air concentration measurements were found to be among the most significant contributors. Internal dose estimates that use individual air surveys supported by bioassay and whole-body counting data are desirable for internal dose reconstruction. (author)