Variability in water vapor plays an important role in determining how much a signal gets delayed during the propagation through the atmosphere. This delay through the neutral atmosphere can be split up into a hydrostatic and a wet part. The latter can also be converted into precipitable water and its variation is an indicator for climate change. In this thesis, a long time series of tropospheric wet delays in zenith direction is computed from three different sources and compared to each other. The time frame chosen for the analysis is from 1985 - 2014, including data from a number of stations with different measuring activity and therefore varying data availability. Eighteen Very Long Baseline Interferometry (VLBI) stations and eight Global Positioning System (GPS) stations, co-located with the VLBI stations, are used for the analysis. Data from ray-tracing through numerical weather models (NWM) with operational data, available with the Vienna Mapping Function 1 (VMF1), are used for comparison as well. The data processed with the Vienna VLBI Software (VieVS) show a satisfying correlation with different pressure inputs, whereas the results obtained from VMF1 and GPS are contradicting. The data obtained from VMF1, using operational NWM data, shows that this kind of data is not ideal for determining a trend over many years. For a trend analysis, reanalysis data would be better suitable but is not always available. The GPS data have their limit in the small observation time span of only twelve years. The VLBI data using pressure values from in-situ pressure records, show quite uncertain results, especially at the station Zelenchukskaya. Outliers and missing values have been corrected for, but could still influence the quality of the results, as well as random errors. The results with all methods show a worldwide increase in the zenith wet delay (ZWD) of 0.1 mm/year (single solutions ranging from -0.5 to 0.4 mm/year) averaged over all stations or a decrease of -0.02 mm/year (single solutions ranging from -0.5 to 0.2 mm/year) when excluding the station Zelenchukskaya. The ZWD trends for single stations (averaged over all methods) range from -0.9 to 1.7 mm/year or -0.9 to 0.3 when excluding the station Zelenchukskaya, yielding mostly realistic results with a few exceptions. In general, there is a common pattern visible, even with all the possible errors in the data, but more research is needed to get conclusive and congruent results across all methods. Also the detection of potential remaining offsets is important to avoid a negative influence on the result. For future studies, the usage of longer time series should be considered in order to further improve the reliability of the trends.