Sample processing procedure should result in analytical portions that are representative of the analytical sample, even if 1-5 g portion are withdrawn from a large-size crops. Therefore, the efficiency of sample processing should be regularly tested during method validation and internal quality control studies. In the scope of this study, sample processing uncertainty was investigated by surface treatment of the cucumber, papaya, and jackfruit representing large-size crops with radio-labelled chlorpyrifos methyl. After homogenization in a chopper, five small and five large test portions were withdrawn from "statistically" well-mixed materials and their residue content was determined with liquid scintillation counter, which provided a quick and well reproducible mode of quantitative determination of residues. The efficiency of sample processing characterized with the sampling constant, changed between 0.78 and 3.01 kg with typical value of 1.61 kg for processing at ambient temperature. Once the sampling constant was established, it was used to predict the uncertainty of sample processing for different test portion sizes. Uncertainty of sample processing at 30 g analytical portion varied between 4.70 and 10.55 %. Furthermore, addition of dry ice to papaya resulted in well-mixed samples as well as 50 % reduction in sampling constant value.