부산대학교 도서관

PURPOSE This study was conducted to refine and verify the argument by Berger & Pope (2011). Based on expectancy theory, we expected that the existence or type of regression discontinuity could vary depending on the situation and time in Korean Professional Basketball League. METHODS We conducted a regression discontinuity design using the R 4.3.1 package on the data of 4,531 games played during the 17 regular seasons (2006–2007 and 2022–2023) of the Korean Basketball League. We divided the sample into top-tier teams (1st and 2nd place) and mid to lower-tier teams (3rd to 10th place). Furthermore, we segregated the sample into the early phase of the regular season (1st and 2nd round) and mid to final phase of the regular season (3rd to 6th round). RESULTS First, a discontinuity of the regression was found after the third quarter. Second, the regression discontinuity appeared differently depending on the performance level. More specifically, discontinuity effects did not appear among the top teams whereas the regression discontinuity effects were identified among mid to lower-tier team samples. Third, the regression discontinuity appeared differently depending on the time point of the regular season (round). A discontinuity of regression line appears from the 3rd to 6th rounds in the Korean Professional Basketball League. CONCLUSIONS This research provides theoretical and practical contributions by systematically exploring that the probability of winning can vary even with the same score difference depending on the time point and situations perceived by the players.