We examine the impact of post-collection sample handling on the protein composition of human tear samples. In particular, we characterize diffusion-based protein extraction from Schirmer strips. These strips of filter paper membrane are the de facto standard for tear fluid collection and storage, with diffusion-based protein elution off the strip being the most widely reported protein extraction strategy. Nevertheless, the diffusion-based protein elution strategy remains uncharacterized regarding downstream functional protein assays. Here, the time-dependence, concentration-dependence, and repeatability of the diffusion-based protein recovery protocol are characterized. Levels of protein irrecoverable from the Schirmer strip and lost during sample handling are isolated and compared for several major tear proteins. Further, the impact of the Schirmer strip and sample handling on the downstream concentration of proteins ranging in molecular weight, surface charge, and surface hydropathicity is quantified. Diffusion-based protein extraction from Schirmer strips was observed to be protein-dependent. Schirmer strips retained tear proteins to varying extents: 14.2% of lysozyme, 9.5% of human serum albumin, 27.7% of secretory IgA, and 30.9% of mucin 4. Tear protein loss during sample handling ranged from 2% (lysozyme) to 41.2% (mucin 4). Strip retention of protein was observed to be associated with protein molecular weight and hydrophobic surface area. Greater sample handling loss was associated with increased hydrophobic surface area of model proteins. Surface charge or surface hydrophilicity was not significantly associated with protein loss. We therefore conclude that, although diffusion-based processing of Schirmer strip-collected tear samples is widely used, these protocols may result in total post-collection protein loss which is considerable, consistent, and protein-dependent. This loss alters the relative and absolute protein concentrations in the sample. A priori prediction of strip-losses for individual proteins does not appear to be facile, based on cursory knowledge of protein surface properties. Thus, we emphasize "spike and recover" control experiments to determine expected elution profiles for target proteins when using diffusion-based protein sample preparation for Schirmer strip-collected tear fluid.