The article has not been subjected to CARB review and does not necessarily reflect their views. Additional funding from NIEHS Superfund Program under Grant 5 P42 ES04699-05 is gratefully acknowledged.Previous studies have reported a large and persistent discrepancy between field measurements and model predictions of pressure-driven entry of soil gas into housesthe phenomenon that causes high concentrations of radon indoors. The discrepancy is often attributed to poor understanding of inherently complex field sites. This paper compares measurements of soil-gas entry made at a fullscale test basement located in natural solid with predictions of a three-dimensional finite difference model. The results corroborate the earlier findings, with the model underpredicting the soil-gas entry rate by a factor of 7. The effect of seasonal changes in soil conditions on soil-gas entry is also examined. Despite large seasonal changes in near-surface soil moisture content and air permeability, there is no observable effect on soil-gas entry, apparently because critical soil conditions near the soil-gas entry location in the structure floor remain relatively constant.