This letter presents a cautionary note on the assumption of Gaussian behavior for upper tropospheric humidity (UTH) derived from satellite data in climatological studies, which can introduce a wet bias in the climatology. An example study using European Centre for Medium-Range Weather Forecasts reanalysis data shows that this wet bias can reach up to 6 %RH, which is significant for climatological applications. A simple Monte Carlo approach demonstrates that these differences and their link to the variability of brightness temperatures are due to a log-normal distribution of the UTH. This problem can be solved by using robust estimators such as the median instead of the arithmetic mean.