Scott and colleagues presented a study of high quality, addressing an important topic.1 The authors encourage medical faculties to increase their output of family medicine residents by selecting and targeting students based on sociodemographic and attitudinal variables of predictive value. Such an approach would require the acquisition and storage of personal information. To our mind, this is concerning, given ethical consideration and data privacy. The education level of parents, relationship status and societal attitudes should not influence admission to medical school or course and content of medical education. We see a bigger benefit in understanding the variables that influence the decision-making process in the course of studies. Recent publications show that positive family medicine experiences, longitudinally implemented in the curriculum, positively influence the attractiveness of family medicine as a career.2–5 Obligatory and longer family medicine clerkships and exposure to role models influence students committed to primary care as well as those who are undecided.3 Several studies describe the positive effect of rural training experiences.4,6 So don’t select students — convince them.