The atmospheric water vapor is an important indicator of the Earth's climate state and evolution. We therefore aimed at calculating the content and long-term variation of the precipitable water vapor at five coastal Antarctic stations, i.e., Casey, Davis, Mawson, McMurdo, and Mario Zucchelli. To do that, we processed the 12-year time series of GPS and radiosounding (RS) observations acquired at those stations, with the purpose of ensuring the utmost accuracy of the results adopting homogeneous, consistent, and up-to-date processing strategies for both data sets. Using the two fully independent techniques, rather consistent contents and seasonal variations of precipitable water were detected, mainly ranging from 1 (Austral winter) to 10 mm (Austral summer). At each site, correlation coefficients varying from 0.86 to 0.91 were found between the GPS and RS time series, with mean discrepancies $\leq$0.75 mm. There is no clear indication regarding the possible dry or wet biases of one technique with respect to the other, with only a notable GPS wet bias identified at Mawson and a dry bias at Casey that, nevertheless, correspond to an average difference of $<$ 1 mm on the two series; the biases at the other sites are much smaller. Although extremely small, i.e., ranging from −0.03 to 0.04 mm/year, the linear trends of the series are not always consistent in sign. In accordance with the major climate models, the RS linear trends are mostly positive, whereas depending on the site, GPS exhibits a (very small) decrease or increase in water vapor. [ABSTRACT FROM AUTHOR]