Abstract: To investigate the influence of data processing on the expanded uncertainty accuracy in the chemical determination of nitrite in pickles, the uncertainty components of calibration curve were fitted with a general linear model and a more rigorous nonlinear model, respectively. Meanwhile, the uncertainty components of solution volume were assessed with a general body expansion coefficient method and a more strict density correction method, respectively. The uncertainty components of spike recovery were evaluated using an approximate formula and a more rigorous formula. The results showed that the expanded uncertainty accuracy obtained from the strict method was higher than that from the general method. For those test results close to the limited boundary, the strict methods, but not the general methods could be used to evaluate the quality of nitrite determination. When the correct testing process is followed, a higher and reasonable evaluation accuracy can be achieved using real data.

Key words: uncertainty, nitrite, pickles, ion chromatography, nonlinear fitting, volume change, recovery rate, conformity assessment