Passive tracer gas measurements of ventilation were performed in a single-family house and in a flat in Stockholm during four consecutive winter seasons. Measured averages of ACH data are reported for 47 two-week periods for the two naturally ventilated dwellings. The variation of ventilation is discussed in terms of variation in the ventilation driving forces induced by inside-outside temperature differences. The detached house shows a slightly better correlation between ACH and indoor-outside temperature difference than the town flat. It is concluded that

the correlations are not good enough for predictive use for either dwelling. Therefore it does not seem possible to “normalise” ventilation measurement data, obtained during one time period with a specific weather condition, to what would be expected during a period with “normal” weather conditions. A possible reason for the lack of good correlation between ACH and natural driving forces is a highly variable influence from occupant behaviour.