This article analyzes the relevance and theoretical foundations of typical errors (misconceptions) of 7th-9th grade students in mathematics education and the modern psychometric method of their detection - the Item Response Theory (IRT) methodology. Based on data from international studies (TIMSS, PISA), systemic gaps in the mathematical training of Uzbek schoolchildren have been identified, including typical errors related to fractions, percentages, and linear equations. The article highlights the advantages of IRT (1PL, 2PL, 3PL models), its superiority over classical test theory, and the possibilities of cognitive diagnostic assessment (CDA). The presented analysis serves as a methodological basis for correcting gaps in mathematical training and developing intervention programs.