Selecting medical student is a huge task and great challenge. Data gathered as selection criteria have also been shown to correlate with students' performance. So the selection of student should be based on criteria that can predict the outcome in the end.A cross 3Rs study was conducted at the School of Medical Sciences, Universiti Sains Malaysia to explore the predictive power of students' profile and entry streaming on their academic performance in MD programme. The study identified that Gender, Ethnicity, Entry streaming and having Biology in pre entry course have strong predictive power on subsequent academic performance. The study concluded that female graduates' mean marks were significantly higher than male counterparts. Chinese graduate performed significantly better than Indian and Malay graduates. STPM stream and having Biology appeared as strong predictors of good performance in medical school compared to Matriculation stream. The study found no correlation between household income of the graduates and their parents' occupation with their performance in medical school. The study findings provided important evidences for future student selection in Medical school.