Adaptive testing is used to improve efficiency and enhance the test taker experience. One type of adaptive testing is multi-stage testing (MST), in which the next modules are selected based on the test-taker's performance on previous modules. MST has been used in the Programme for International Student Assessment (PISA) since 2018, and will be used in the National Assessment of Educational Achievement (NAEA) in Korea. Adaptive testing can also benefit adaptive learning by reducing the number of items and testing time. This allows teachers and students to have more time for teaching and learning. However, it is necessary to examine whether conventional statistical methods are still applicable to data from adaptive tests. In this study, we show that the proportion correct may be biased when data are collected under MST designs. To adjust the proportion correct, we propose using item response theory to generate pseudo-complete data. Using the simulation study and empirical analyses of PISA 2022 mathematics data, we show that the proposed method is promising for adjusting the proportion correct.