Objective　To evaluate the application effect of ‘incorrect cases’ in medical statis-tics teaching. Methods　The junior students （n=307） of clinical medicine were selected as research subjects by random cluster sampling. They were divided into the teaching improving group （n=151） who accepted case teaching and control group （n=156） who accepted traditional teaching. At the end of the term， the teaching effect was surveyed by questionnaires and compared between the two groups. Measurement data was analyzed by t test， enumeration data by ?字2 test and ranked data by rank sum test. Results　Students' interests of medical statistics and understanding the importance of the course in teaching improving group were significantly higher than that in control group （Z=-2.375， P=0.018； Z=-2.971， P=0.003）. The proportion of students with learning difficulties in teaching improving group was obviously lower than that in control group （Z=-2.488， P=0.013）. The satisfaction with theory and practical teaching in teaching improving group （84.10% and 88.74%， respectively） was significantly higher that in control group （73.08% and 71.79%， respectively）（ χ2=5.530， P=0.013； χ2=13.841， P<0.001）. Although no difference was observed in the test score between the two groups（t=1.230， P=0.220）， knowledge of parameter estimation， hypothesis testing and science research design in teach-ing improving group was significantly higher than that in control group （ χ2=6.829， P=0.009； χ2=5.835， P=0.016）. Conclusions　Application of ‘incorrect cases’ in medical statistics teaching could arouse students' learning interests and improve clinical medical students' comprehensive analysis ability of medical statistics.