Peer assisted learning (PAL) has been defined as "the development of knowledge and skill through active help and support among status equals or matched companions" [1]. The same author has described the process of PAL as a situation in which "people from similar social groupings who are not professional teachers help each other to learn and learn themselves by teaching" [2]. It is this reciprocity of learning among other things that makes PAL such an attractive idea to educationalists. This is particularly true in medical education, where the restrictions on resources have forced teachers to look to creating new educational environments which can be delivered at a lower cost [3].

PAL has been embraced by medical educationalists for many years in the US [1, 4, 5]. In fact, the benefits of PAL have been extensively described in that part of the world and are reputed to positively correlate with examination performance [4, 5]. The benefits associated with PAL, however, are not confined to examination scores. Other student benefits recognised are those of lowering subjective distress and enhancing course satisfaction through the establishment of a reciprocal social support system [6]. Researchers have also shown that pairing junior and senior undergraduate students provides psychological support and aids professional and personal development [7]. In addition to knowledge and skills mentioned above, modern learning objectives will often encompass "attitudes" as part of the learning process and this is particularly relevant in PAL as peers can prove powerful role models [8].

Additionally, the advantages of PAL do not appear to be limited just to the peer learners. Peer tutors also have been shown to benefit significantly in this learning environment [9]. In fact, in some studies it has been shown that peer tutors appear to show significantly greater cognitive gains than their peer learner counterparts [10, 11].

The proposed solution was the introduction of PAL on a pilot basis for the academic year 2003-2004. We proposed that a group of fifth-year medical students would act as tutors for the small group role-play sessions. We felt it was important to fully evaluate this initiative as it was the first time that PAL had been used in our medical school. We sought to establish the views of learners and tutors towards this new method of teaching by conducting a process evaluation using qualitative methods. We were keen to evaluate the processes at work as the emphasis of PAL research to date has largely been on the consideration of student outcomes [13].

Analysis of the data resulted in the emergence of three thematic categories: Learning Environment, Educational Exchange and Communication and Modelling. A summary, descriptive account of the key findings from these categories follows. "Respondents were assigned a respondent code according to whether they were peer tutors (PT1 and PT2) or peer learners (PL), and whether the data was collected "pre" or "post" the PAL session. While every effort was made to systematically identify individual peer learners in the focus groups, poor quality of tape recording made it impossible to do so at the transcribing stage."

The "safe" learning environment is perceived as a vital factor in effective teaching and learning. It is not surprising therefore, that this aspect of the PAL process appeared as a recurring theme in the data and was highlighted by both learners and tutors as a unique and fundamental tenant of this approach. Safe learning environments are vital in order to engage students in purposeful learning experiences, encourage constructive interactions among tutors and students and enable students to control their own learning effectively. The interactions that are fostered by the 'safe' learning environment in PAL appear to encourage learners to create associative links to their existing knowledge, to evaluate the truth of their emerging understandings and to elaborate the content of the lesson – all the while avoiding being seen to contravene the rules of order of a tutor-led lesson [20]. However it is also worth noting that concern has been expressed that the learning environment in PAL can in fact lead to a decline in the quality of student learning, such that it becomes less focussed on understanding course material than succeeding through a greater awareness of assessment demands [13]. This occurs as PAL helps students come to terms with the demands of their course but depending on the context this can be at the cost of understanding the deeper material they are studying [13]. Modification in peer tutor training to reduce the emphasis on strategic learning and instead promote the intrinsic benefits of PAL has been shown to be a way of averting this trend [21].

A sense of cooperation and closeness appears to permeate PAL as a whole and probably explains some of the reasons behind it's success. This is reported to be due to the concept of "promotive interaction". Promotive interaction describes how individuals encourage and facilitate each other's efforts in order to reach the groups goals [22]. This can be done by exchanging resources and information; giving and receiving feedback; and mutually influencing each other's reasoning and behaviour which is evident in the communication taking place and the 'modelling' effect noted in the data. Therefore, perhaps what PAL attempts to achieve is to make "the implicit nature of social learning explicit by encouraging active learning within social settings" [23]. In PAL the interaction between tutor and learner takes place on a number of different levels simultaneously. In this study, the PT uses the subject matter and techniques of the communication skills session to facilitate the teaching process. This not only provides a useful illustration of these techniques for the PL but also gives the PT an opportunity to practice the techniques during the session itself. So simultaneously, the PT communicates, tutors, learns and illustrates and in so doing provides, and is provided with, a very rich learning experience. It would seem that this correlates with the assertion of Parr and Townsend that there are peer effects in learning that originate in peer interactions and association [8]. It would appear that these "peer" effects in learning are multiple, complex and often occur simultaneously and in a reciprocal fashion [8].

There is also the suggestion that those involved in PAL learn not just about the material in question but also appear to learn about learning [24]. The transferable skills gained in this way enable heightened performance in areas of study other than those targeted by PAL such as the development of critical thinking skills [25]. Contrary to what one might think it is not just stronger students that can or should work as tutors. The benefits of working as a tutor in the PAL setting has been recognised in children [26] as well as adults [11, 27]. In fact results in studies among children, suggest that serving as a tutor may be a particularly useful method for enhancing the academic performance of low-achieving children [26]. Annis et al compared the learning taking place between peer groups tutoring and being tutored and found that tutoring resulted in significantly greater content-specific and generalized cognitive gains than being tutored [11]. Peer tutoring thus appears to be a potentially powerful technique for increasing all levels of student learning. So why is that when we learn to teach we appear to learn better than when we learn to be tested. Benware [28] suggests when we learn to teach, we learn with a more active orientation. He described how students who learn to teach, are more intrinsically motivated, have higher conceptual learning scores, and perceive themselves to be more actively engaged with the environment than students who learn in order to be tested [28].

One of the chief concerns raised in this data by the PL was about the quality of the teaching. Indeed, this is probably one of the crucial distinctions that must be made in relation to PAL, in that the role of the PT should be one of facilitation rather than of teaching [29]. An interesting distinction was made between the large lecture setting in which learners felt the PT would lack credibility and knowledge; and small group work, which it was felt would ideally suit PAL. Indeed, there are obvious perils in making students teachers in that they are not qualified to be teachers, it is unfair to place such demands on them, they can give incorrect or misleading information and it detracts from what PAL is supposed to be about. Other concerns that have been noted in previous studies besides peer competence, are those of informed consent and accountability [30]. Additionally, it would appear that PAL does not suit all professional groups. For example, in a study using student nurses as teachers, it became apparent that the student nurses were uncomfortable with being used as tutors. They repeatedly questioned the intrinsic worth of this approach as a developmental tool, and considered the responsibility for teaching the content of parallel resource sessions to lie with nurse educators [31].

The small group sessions described in this study have traditionally been facilitated by medical teachers from different disciplines but were limited by lack of faculty resources, increased class sizes and difficulties in recruiting medical teachers. Peer tutoring can provide an innovative way of overcoming such difficulties. The cost of this approach is the provision of training for peer tutors. Although the issue of cost was not addressed in this study per se, the three hour preparatory workshop for prospective peer tutors, who can then go on to facilitate multiple role-play sessions in the place of medical teachers has the potential to be extremely cost-effective.

By and large, healthcare and academic staff are enthusiastic about teaching, but are often limited by resources and staff numbers [33]. PAL is one way of answering these issues. This study has demonstrated a concordance of the views between peer tutors and learners on barriers and levers of this approach as well as a heightened awareness of the learning environment and the flow of communication therein. This is significant as it not only demonstrates a high level of acceptability among tutors and learners but also indicates the reciprocity of educational exchange that appears to occur within the PAL setting. Additionally, some of the learning experiences of tutor and learner appear to be unique to the PAL setting and so strengthen the argument for the formal inclusion of PAL in the curriculum. This study has highlighted some of the unique characteristics of peer assisted learning and we recommend the development of further qualitative studies around peer learners and tutors views of this process.

This study was funded by an Educational Innovations Fund grant awarded by the Centre for Excellence in Learning and Teaching, National University of Ireland, Galway, 2003/4.