Comparisons of relative humidity (RH) measurements between the Meisei RS-06G radiosonde and a chilled-mirror hygrometer revealed that the RS-06G radiosonde shows a stepwise change of ~3% RH at 0°C (drying when air temperature is decreasing). This is due to a discontinuous correction factor in the processing software that compensates for the temperature dependence of the RH sensor. Results from chamber experiments regarding the temperature and RH dependence of RS-06G RH sensors under steady-state conditions showed a wet bias exceeding 7% RH below ~+10°C. As this result contradicted previous in-flight intercomparisons that used the original manufacturer's correction, we investigated a possible additional dry bias caused by a thermal lag in the RH sensor. We speculated that the thermal lag of the RH sensor typically causes a dry bias during a tropospheric ascent, which largely compensates for the wet bias related to the temperature and RH dependence of the RH sensor. We observed that the experimental results of the temperature and RH dependence considering the thermal lag were in agreement with the extrapolation of the original manufacturer's correction. Consequently, we proposed to extrapolate the original manufacturer's correction, which is currently applied at temperatures between −40°C and 0°C, up to +14.5°C to resolve the artificial stepwise change at 0°C. Because the RS-06G radiosonde is a successor to the Meisei RS-01G and RS2-91 radiosondes, which have adopted the same RH sensor material installed since July 1999 and have used the same processing software, the current results should be applied to the data obtained by those radiosondes. The bias of RS-06G RH measurements using this new correction is estimated to be within 7% RH, which is within the manufacturer's specifications, being drier at +40°C and wetter between −40°C and +10°C.