Total Grain-Size Distribution (TGSD) of tephra deposits is key to the characterization of explosive volcanism, plume-dispersal modeling, and magmatic fragmentation studies. Nonetheless, various aspects that includes deposit exposure and data fitting make its determination extremely complex and affect its representativeness. In order to shed some lights on the reliability of derived TGSDs, we examine a large TGSD dataset in combination with a sensitivity analysis of sampling strategies. These analyses are based both on a well-studied tephra deposit and on synthetic deposits associated with a variety of initial eruptive and atmospheric conditions. Results demonstrate that TGSDs can be satisfactorily fitted by four distributions (lognormal, Rosin-Rammler, and power-law based either on the absolute or cumulative number of particles) that capture different distribution features. In particular, the Rosin-Rammler distribution best reproduces both the median and the tails of the TGSDs. The accuracy of reconstructed TGSDs is strongly controlled by the number and distribution of the sampling points. We conclude that TGSDs should be critically assessed based on dedicated sampling strategies and should be fitted by one of the mentioned theoretical distributions depending on the specific study objective (e.g., tephra-deposit characterization, physical description of explosive eruptions, tephra-dispersal modeling).