Sorbent tubes (STs) are highly efficient and versatile tools for collecting gaseous samples of volatile organic compounds (VOCs). In order to assess the experimental bias in their application, the relative recovery (RR) of 19 VOCs (acetaldehyde, propionaldehyde, butyraldehyde, isovaleraldehyde, valeraldehyde, benzene, toluene, m-xylene, p-xylene, o-xylene, styrene, methyl ethyl ketone, methyl isobutyl ketone, isobutyl alcohol, butyl acetate, propionic acid, butyric acid, isovaleric acid, and n-valeric acid) was evaluated in accordance to three different sorbent tube types: (1) Tenax (Te), (2) Carbopack X (X), and (3) Tenax + Carbopack B + Carboxen 1000 (TBC). Calibration was initially made by direct injection of the liquid working standard (L-WS) into the link tube (LT) of the thermal desorption (TD) system. The LT was modified to optimize the transfer of all the VOCs to the focusing trap of the TD. Three different types of ST were then loaded with the same L-WS using (1) direct injection (DI) to tube and (2) injection to tube via vaporization (VAP). The results of the two ST/TD methods (DI and VAP) were normalized to those of the LT/TD to derive RR. The mean RR values of all target VOCs were 62.6 (X), 83.9 (Te), and 82.8% (TBC) for the DI and 76.5, 75.8, and 81.0% for the VAP. Although the results using TBC and Te tubes exhibited better recovery in DI than in VAP, those of X tubes had the reverse trend. As such, the relative affinity of each sorbent material can be affected by the mode of introduction of the standard. The practicality of the LT/TD, especially as a reference for the ST/TD, was also validated from some environmental samples in an ancillary experiment. As such, we were able to demonstrate the usefulness of the LT/TD method to evaluate analyte recovery from spiked sorbent tubes.