In 2006 a new type of tutorial, called Map Meeting, was successfully trialled with novice first year physics students at the University of Sydney, Australia. Subsequently, in first semester 2007 a large-scale experiment was carried out with 262 students who were allocated either to the strongly scaffolding Map Meetings or to the less scaffolding Workshop Tutorials, which have been run at the University of Sydney since 1995. In this paper we describe what makes Map Meetings more scaffolding than Workshop Tutorials--where the level of scaffolding represents the main difference between the two tutorial types. Using a mixed methods approach to triangulate results, we compare the success of the two with respect to both student tutorial preference and examination performance. In summary, Map Meetings had a higher retention rate and received more positive feedback from students--students liked the strongly scaffolding environment and felt that it better helped them understand physics. A comparison of final examination performances of students who had attended at least 10 out of 12 tutorials revealed that only 11% of Map Meeting students received less than 30 out of 90 marks compared to 21% of Workshop Tutorial students, whereas there were no differences amongst high-achieving students. Map Meetings was therefore particularly successful in helping low-achieving novices learn physics.