Metacognitive processes are important for the success in the wide range of educational activities of youth and young adults. However, the positive correlations between metacognition and academic achievements are not high enough, and the instruments used in these studies might be the reason. We explored the factor structure of the Russian version of the questionnaire “Metacognitive Awareness Inventory” developed by G. Schraw and R. Dennison and adapted by A.V. Karpov and I.M. Skityaeva into Russian. The participants of our study were 527 residents of St. Petersburg, which were studying at the university at the time. Among them there were 366 students getting their first diploma and 161students getting their second diploma (average age 23.8 ± 8.8). In this article the authors present the results of a confirmatory factor analysis of four models, which are the most frequently used in foreign and Russian literature: unidimensional model; two different two-factor models; eight-factor model. Evaluation of the model fit indices for the four models showed that none of them were a god fit. We reduced the number of items of the questionnaire and re-implemented the factor analysis of these four models. The values of indicators of a good model fit improved. In the short version of the questionnaire “Metacognitive Awareness Inventory” the authors discovered two scales – knowledge of cognition and regulation of cognition, which included 8 subscales: declarative knowledge, procedural knowledge, conditional knowledge, planning, information management strategies, comprehension monitoring, debugging strategies, evaluation.