Background
Learning is the basis for undergraduate study and continuing professional development in a broad range of medical specialties. A range of learning methods is used by both undergraduate and postgraduate students during their career development: previous studies indicated that certain methods, such as problem based learning, are favoured over traditional lecture based courses. However, little is known about whether these preferences differ between members of different specialties or whether they become more focused as postgraduate careers develop. Consequently it is also not known whether current educational practices favour groups of students with particular career aspirations.

Methods
We surveyed 65 medical students at various stages of their undergraduate course, as well as 18 general practice registrars and 20 surgical specialist registrars and asked them to rank their preferred learning methods. Medical students were also asked to indicate a preferred specialty.

Results
Medical students and registrars all like practical and small group interactive learning and dislike reflective methods. These preferences are more pronounced among surgical registrars, who also rate didactic and self-study methods lower than their GP colleagues. This difference is not apparent between the medical students who are considering a career as GPs and those who are planning a future in general surgery, except for a trend among potential surgeons to rate practical methods more highly. In general medical students are less comfortable with problem-based learning approaches than qualified doctors, though this difference is not apparent in graduate entry medical students who routinely employ this mode of study. Similarly, as students progress through the course they become more appreciative of bedside teaching and experiential methods but become more entrenched in their dislike of reflective learning.

Implications
This small but novel study highlights similarities and differences between learning preferences of very diverse medical specialties. Medical students and doctors in general prefer practical and interactive methods, and are less comfortable with reflective learning. This difference is more pronounced in surgeons. Unselected undergraduates are less homogenous in their stated preferences, less comfortable with problem-based learning approaches but show some patterns akin to the preferred learning methods of doctors in their envisaged specialty.

The described learning preferences are supportive of current practices in medical education, which seek to increase practical and small group teaching at all levels, with reliance on lectures for a significant part of the undergraduate curriculum. However the increasing emphasis on reflective exercises at a postgraduate level appears to run against the inherent preferences of doctors.