Abstract: Some analytical problems, which are often considered incorrect for objective reasons, are considered. The main of these reasons is an anomalously large scatter of the initial data. It can be due to either the low reproducibility of the characteristics of substances, their quantities, analytical signal intensities, process conditions, etc., or variability due to differences in the nature of the objects themselves. In the latter case, the nature of data interpretation is influenced by analytical hypotheses adopted for their consideration. The tasks considered include variations in the component composition of developers for black-and-white negative photographic materials, comparison of temperature parameters of the gas-chromatographic separation of various organic compounds, toxicity characterization (LD ₅₀) of homologues using C ₃–C ₁₂ 1-alkanols as an example, and the possibility of predicting sample preparation operations in the determination of drugs in blood plasma based on their physicochemical characteristics. The main features of data interpretation characterized by a high degree of uncertainty are revealed. It is noted that important conclusions can be drawn based on the facts of low reproducibility (one-dimensional arrays) or poor correlation of variables (two-dimensional arrays).