Indoor radon concentration of two consecutive half-year measuring peri
ods in 3074 Finnish dwellings were analysed. The periods were the warm
est and coldest half-year periods in Finland. The mean of winter/summe
r concentration ratio in all low-rise residential buildings was 1.28,
and in houses with winter concentrations of <50, 50-100, 100-200 and >
200 Bq.m(-3), 0.97, 1.22, 1.34 and 1.55, respectively. The results are
in agreement with the results predicted by the model, which takes int
o account the varying contribution of diffusive and convective radon e
ntry as well as the contribution of stack effect and wind induced air
infiltration. The model is a useful tool for estimating the average co
rrection factor for the annual average radon concentration from the tw
o months measurements used in Finland. Typical correction factors vary
in the range of 1.0-0.7, depending on the outdoor temperature and on
radon concentration level. Model comparison with winter/summer ratios
measured in houses with slab on grade suggests an average diffusive ra
don entry rate of 7 Bq.m(-3).h(-1) and a convective entry rate of 50 B
q.m(-3).h(-1) at an average indoor/outdoor temperature difference of 1
7 K. The use of a simple averaging of winter/summer ratios for houses
with varying radon levels, creates inaccuracy in the information on se
asonal variation. Surprisingly, the behaviour of the winter/summer rat
io in blocks of flats was very similar to that in low-rise residential
buildings. The results support the interpretation that soil air leaka
ge also contributes unexpectedly to the radon concentration in hats.