Immunoblotting is widely used in muscle physiology to determine protein regulation and abundance. However, research groups use different protocols, which may result in differential outcomes. Herein, we investigated the effect of various homogenization procedures on determination of protein abundance in human m. vastus lateralis biopsies. Furthermore, we investigated differences in abundance between young healthy males (n = 12) and type-2 diabetics (n = 4), and the effect of data normalization. Fractionated lysates had the lowest variation in total protein determination as compared to non-fractionated homogenates. Abundance of NKA_α2, NKAβ₁, FXYD1, and glycogen synthase was higher (P < 0.05) in young healthy than in type-2 diabetics determined in both fractionated and non-fractionated samples for which normalization to the stain-free signal and/or standard curve did not affect outcomes. Precision and reliability of protein abundance determination between sample types showed a moderate to good reliability for these proteins, whereas the commonly used house-keeping protein, actin, showed poor reliability. In conclusion, fractionated and non-fractionated immunoblotting samples yield similar data for several sarcolemmal and cytosolic proteins, except for actin, which, therefore appears inappropriate for data normalization in immunoblotting of human skeletal muscle. Thus, fractionation does not seem to be a major source of bias when immunoblotting for NKA subunits and GS.