Introduction. The system of assessment of students' performance needs to comprise certain psychometric characteristics, including in particular reliability and validity, so that it can be a highly objective source of knowledge of students' learning outcomes. Aim. Analysis of reliability and validity of the performance assessment system for Midwifery students who started a Bachelor's degree programme at Medical University of Warsaw between 2005-06 and 2012-13. Material and methods. A retrospective study enrolling a group of 922 students of eight subsequent full education cycles. The authors collected detailed data on grades for twenty courses that ended with an exam throughout the course of studies, divided into four groups according to the criteria specified in the education standards. Reliability (Cronbach's alpha coefficient), as well as theoretical (factor analysis) and criterion validity (the Pearson correlation matrix) were assessed. IBM® SPSS® Statistics version 23 was used for calculation. Results. Cronbach's alpha coefficient for each year exceeded the assumed threshold of 0.700. Total reliability of the assessment of students' performance for the period considered amounted to 0.805. The factor analysis of students' grades for 20 examination courses demonstrated a five-factor structure, which is far from the assumptions resulting from the education standards (four groups of effects). The highest level of criterion validity was observed for the D group courses ("Education in Specialist Care"), with average values of Pearson's correlation coefficient (r) amounting to 0.20. Conclusions. A good level of reliability accompanied by a low level of validity leads to a decrease in credibility of the entire system of assessment of Midwifery students' competencies. This may, in some cases, increase the risk that there would be persons with insufficient level of initial competencies among Midwifery graduates.

Key words: midwifery, educational measurement, graduate education, reproducibility of results, professional competence.