Flash Player 9 (or above) is needed to view presentations.
We have detected that you do not have it on your computer. To install it, go here.

Like this document? Why not share!

The problem based tutor






Total Views
Views on SlideShare
Embed Views



0 Embeds 0

No embeds



Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
Post Comment
Edit your comment

The problem based tutor The problem based tutor Document Transcript

  • M edical Teacher, Vol. 21, No. 4, 1999The problem-based learning tutor: Teacher?Facilitator? Evaluator?ALAN J. NEVILLEMcMaster University, Hamilton, Ontario, CanadaIntroduction facilitation skills as crucial for the learning of students, in c on tradistin ctio n to tu tor s’ rele vant subjec t- m atterDespite th e structural heterogeneity of problem -b ased knowledge, a distinction which has been the subject of muchlearning (PBL) curricula, m ost PBL schools have embraced debate and considerable controversy. Barrows stated bluntly,self-directed learning, emphasizing the use of small-group ª A faculty person who is a good tutor can successfully tutordiscussion and integration of the basic medical sciences in any areaº (Barrows & Tam blyn, 1980, p. 107).with clinical problems. Self-directed learning is but one of Before dissecting the arguments for and against thethe m any terms such as discovery method or study-centred particular issue of content knowledge as a prerequisite foreducation adopted by authors since Dewey to describe an successful PBL tutoring, some background discussion iseducational approach that places the learner in control of required of the factors that relate to the role of students andhis or her learning (Knowles, 1975). The putative bene® ts tutors in the PBL small-group tutorial.of self-directed learning include enhanced opportunities to In his monograph on self-directed learning, M alcolmelaborate one’ s knowledge through active involvement and Knowles described the `fundamental and terribly difficult’verbalization, enhanced motivation through an increase inrelevance and personal control, and the practice of skills change in self-concept in moving from `teacher’ to `facilita-needed in lifelong learning (Schm idt, 1983). tor of learning’ (Knowles, 1975). He wrote: ª It required In this educational milieu, the role of the `teacher’ requires that I focus on what was happening to the students ratherrevision; new skills are required of the teaching faculty so than on what I was doing. It required that I divest myself ofthat they are willing and competent to allow students to the protective shield of an authority ® gure and expose myselftake an active role in guiding their own learning and in as meÐ an authentic human being, with feelings, hopes,teaching one another (Barrows & Tamblyn, 1980). aspirations, insecurities, worries, strengths and weaknesses. This review explores the literature that has developed It required that I extricate myself from the compulsion toaround the de® nition of the teacher or tutor role in `facilitat- pose as an expert who had mastered any given body ofing’ the learning of students in a PBL setting. Several content and, instead, join m y students honestly as acontroversies have arisen over the optimal role of the faculty continuing co-learnerº . Later, he added ª I found myself, forperson in facilitating a PBL tutorial group, including level example, functioning primarily as a procedural guide andof participation, content knowledge and involvement in only secondarily as a resource for content informationº .student evaluation. While it appears that there is probably The debate about what constitutes effective `facilitation’no completely satisfactory resolution of these controversies, of student learning in PBL m erits close scrutiny, since Ifrom a review of the frequently con¯ icting pieces of evidence, believe that the evidence cited in the literature favouring aan attempt will be made to synthesize from the literature a role dichotomy between process facilitator and contentcoherent picture of an effective tutor in the PBL setting. resource may be m ore apparent that real and that there is evidence pointing towards consideration of a balanced interaction of these functions as the optimal tutor role. ThisTh e PB L tutor: issues de® ning the debate on roles is an important concept to consider, as faculty, familiar withBarrows has claim ed that the task of the tutor in a problem- directing students in lectures or seminars, can have difficultybased tutorial group should be to facilitate the learning of adopting the role of PBL tutor, even after faculty develop-students rather than to convey knowledge (Barrows & ment training sessions (Neufeld & Barrows, 1974) and,Tamblyn, 1980). In developing this educational approach, mistakenly, m ay develop the belief that tutoring is nothingBarrows considered that tutors must allow students to more than the observation of process and tutorial dynamics.determine on their own what they need to know, and to Thus they m ay feel that they have little to contribute. Thelearn through the study of varied resources. Rather than result is a polarization of faculty attitudes to PBL, wherebytelling students what they should learn and in what sequence tutors feel constrained to act either as `wall¯ owers’ in thethey should learn, the tutor must help students determine group or, alternatively, as directive lecturers. Neither ofthis for themselves (Barrows & Tamblyn, 1980). By `facilita- these behaviours facilitates tutorial process or learning.tion of learning’ , Barrows clearly emphasized process- A second issue requiring clari® cation relates to the learning process. Andragogical self-directed learning in a PBL environm ent stresses a student-centred approach toCorrespond ence: Professor Alan J. N eville, M cMaster University, Faculty ofH ealth Sciences, Undergraduate M edical Prog ram me, H SC, Room 2E18, learning where students determine their learning objectives,1200 M ain Street West, H amilton, Ontario, L8N 3Z5, C anada. Tel: + 905 525 how to learn them and to evaluate what they have learned9140; fax: + 905 528 4727 . (Walton & M atthews , 1989). During the course of `problemISSN 0142-159X print/ISSN 1466-187X online/99/040493± 09 ½ 1999 Taylor & Francis Ltd 393
  • A.J. Nevillesolving’ students identify knowledge de® ciencies of their in the psychology literature relating to group learning andown, a process that epitomizes self-directed learning as individual achievem ent have been attributed to particularespoused by Coulson (1983). Yet it is fallacious to assume co operative lear n ing tec hniqu es, setting s, m easure s,that to be successful the entire educational process of PBL experimental designs, student characteristics and subject-must be entirely self-directed. Faculty design the overall matter. From a thorough review of the existing evidencecurriculum to a greater or lesser extent in most institutions published in 1982, Noreen Webb (1982) concluded that anand often design the vehicle of instruction, i.e. the written individual’ s role in group interaction is an important in¯ u-health-care problem . O ne can contrast highly student- ence on learningÐ and that interaction can best be predictedcentred curricula (such as the New M exico Primary Care from multiple characteristics of the individual, group andtrack, where faculty-generated objectives are not available setting. These studies from the psychology literature suggestto the students) with the relatively faculty-centred curriculum a complex interaction of student and tutor variables andof the PBL track at M ichigan State U niversity (Blumberg et bear closer scrutiny by those developing PBL curricula.al ., 1990). As an exam ple, in describing the decision-m aking proc- T he U ndergraduate M D Prog ram m e at M cM aster esses when the McMaster Undergraduate M D ProgrammeUniversity is som ewhere between these two extrem es was being planned, William B. Spaulding, in his book(Neufeld & Barrows, 1974). Whether the curriculum itself Revitalising M edical E ducation , writes: ª They [the planners]appears to be student or faculty centred need not in itself made some naive assum ptions, chief of which was that thein¯ uence the effectiveness of the tutor in facilitating the only necessary ingredients for a successful tutorial grouplearning process within tutorials; in other words, a faculty- were a sm all number of students and a faculty member.centred curriculum can be associated with a tutorial process Little or no consideration was given to group dynam ics.wherein student-centred self-directed learning can ¯ ourish. They did not foresee the difficulties that can arise when a Thus sceptical and confused faculty need some guidance group of people of diverse personalities, backgrounds, andin developing their role as PBL tutors, balancing a natural ages m ix in an intense learning situation; it was taken fordesire to be directive and `teach’ with the desired goal of granted that students and tutors would be considerate ofenhancing student-centred self-directed lear ning. O ne each other’ s learning needs and altruistic enough to helpsigni® cant area to be addressed is the following: W hat each other ® nd approaches and solutions to the topics underevidence can be adduced from the literature to enable tutors discussionº (Spaulding , 1991, p. 42). Implicit in these wordsto avoid the false role dichotomy described above? Some is that the underlying assum ptions made by programmeanswers can be found in both the cognitive psychology planners were som etimes proved false in the light of earlyliterature and in empiric studies of PBL tutorial function. experience of running PBL tutorials. Second, if there is an argument for tutors acting as a In re¯ ecting on how the prospective PBL tutor mightcontent resource in tutorials, one can ask ª how much content prepare for the role of learning facilitator, Malcolm Knowlesknowledge should the tutor process and what is the most (1975) identi® ed seven elements for an andragogical learningeffective way of expressing this knowledge? Third, is it worth process design. These can be paraphrased as follows:considering the m erits of varying the degree of tutor `direc- (1) C lim ate setting : helping the learners become acquaintedtiveness’ in learning facilitating depending on the level of with each other as persons and as mutual learningPBL training of the students? Given the emphasis on process resources, develop the skills of self-directed learningfacilitation by some authors, one could argue for the develop- and understanding the role of the tutor;ment of student or peer tutoring rather than the more (2) Planning : deciding on how tutorials will run and how`expensive’ use of faculty tutors. In assessing the literature tutorial process and function decisions are to be made;relevant to these issues, one must be cognizant of the (3) designing needs for learning : consideration of how thedifferent outcom es being measured, i.e. student satisfac- tutor can frame content objectives so that students cantion, student performance or both. Studies relating to tutor take ownership of the learning process and com pareinvolvement in tutorials, tutor content knowledge expertise their existing knowledge with the required objectives;and the role of peer tutoring will be described in the next (4) setting goals: helping the students translate the diagnosedthree sections of this review. Following this, the role of the needs into clear, feasible learning objectives;tutor in student evaluation will be explored brie¯ y, since (5) designing a lear ning plan : helping the students designthis is an area of tutor function that has been studied very their learning plans, develop strategies for accessinglittle. resources etc.; (6) engaging in learning activities : whereby the tutor considersTutor involvem ent in tutorial and sm all-group learning what part of the learning should be his/her responsibilityThe development of PBL approaches to medical education, and wh at th e students sh ould be resp onsible for,with learning taking place in small-group tutorial settings, collectively or individually;has taken place alongside increasing interest in psychological (7) evaluating learning outcomes : how to give constructiveresearch into the effects of cooperative sm all groups on feedback to the students so as to enhance the self-learning. The research into educational achievement has in, directed learning process.general, produced results favourable to the concept of Armed with Knowles’ s elements of facilitation, one cancooperative group learning (Webb, 1982), although not all begin to explore the literature describing empiric studies ofreviews have concluded that working in small groups is PBL in action.bene® cial for learning. For exam ple, M ichaels (1977) Perhaps the most important problem facing the facultyconcluded that individual competition consistently produced individual in a PBL tutorial setting is determining the degreegreater achievement than group conditions. Discrepancies of `directiveness’ to assume to achieve the necessary learning394
  • The problem -based learning tutorfacilitation without detracting from student self-directed as student directed or not, rather than the tutorial processlearning. It would be wrong to assume that the concern is itself. Wilkerson’s group would describe student-directedonly on the part of the tutors. M any students enter a PBL tutors as allowing students to initiate and sustain discus-learning situation feeling a need for the security of a clear sion, using infrequent questions to guide group process.structural plan, i.e. a course outline or syllabus, time schedule Such tutors probe for understanding, encourage students toetc. They may feel `at sea’ in the som ewhat structureless listen to one another, tolerate silence and collaborate byenvironm ent of PBL. In addition students have concerns postponing their own suggestions, acceding to students’about whether they are going to `get’ the required content selection of objectives, fostering a feeling of cooperationto pass their exams (Knowles, 1975, p. 37). Several studies rather than competition (W ilkerson et al ., 1991).reporting on students’ perceptions of PBL curricula suggest In a subsequent study, W ilkerson was able to condensethat students would prefer more `direction’ than they are this descriptive framework into two general skills that bothgetting. Blumberg & Eckenfels (1988) reported that students students and faculty agreed were important.The ® rst desiredwere dissatis® ed with the lack of structure in their PBL skill or behaviour was `guiding the work of the group’ , i.e.curriculum . In addition, M cMaster students identi® ed the essentially the roles of probing, questioning, achieving alack of de® nition of core m aterial as a weakness in their balance of clinical and basic science emphasis, and theUndergraduate M D Programm e (Woodward & Ferrier, second skill, `promoting interaction’ , described issues relating1982). to successful tutorial dynam ics (W ilkerson, 1992). These O ther g roups have repor ted som e unease on the part of two studies did not particularly address the issue of studentstudents, particularly when they ® rst approach the PBL fam iliarity with the PBL process at the tim e the tutorialtutorial m ethod. In M aastricht, m edical students at the obser vations were be ing m ad e, but one can perhapsU niversity of Lim burg were uncomfor table with th e tuto- envisage a m echanism whereby the relative em phasis onrial process at ® rst and the tutorial interactions were of the ® ve factors described by W ilkerson et al . could behighly variable quality (de Vries et al ., 1989). Unfamiliarity m odi® ed to optim ize the clim ate of learning for differentwith the PBL process and poor feedback were cited by PBL learners.students at M ichigan State U niversity who were dissatis- It is important, however, for new tutors to resist the® ed with the PBL track (Shope, 1989). Davis et al . also temptation to share their knowledge over-enthusiastically noted that students who were led by m ore directive tutors with their tutorial groups. In one study, Thomas (1992) rated their enjoym ent of PBL in sm all groups m ore highly reported that new tutors might dom inate up to 80% of (Davis et al ., 1982). tutorial time. A similar desire to share expertise was noted The discomfort that students themselves may have in the by DesM archais et al . (1992). In a contrasting vein, tutor PBL setting is not to be taken as an argument for the tutor directiveness has been related to the tim e-efficiency of the to be `directive’ th roughout th e tutorial process. The tutorial process. G ruppen et a l. (19 92) were ab le to concerns raised by Knowles earlier have been echoed by dem onstrate a potential 50% time savings in tutorial time if others. Neame & Powis (1981) have addressed the issue of the tutors were clearly directive. While this ® nding might be structure in tutorial and the role of the tutor in tutorial appealing to those in the administration of PBL curricula, discussion. They ® rmly believe that the development of the potential for detrim ental effects on students’ satisfac- stud ents as ind epen de nt lear ne rs re qu ires d elibe rate tion and self-directed learning is clearly signi® cant, albeit as curricular planning. They envisage a PBL curriculum in m entioned earlier, dependent on the familiarity of the which there is a gradual progression toward total independ- students with PBL. ence of learning ª via a g raded reduction of im posed structureº . ª In this regard, earlier course segments would Thus far, the evidence points towards steering a path be prepared with precise objectives speci® ed and with direc- between the Scylla of sti¯ ing student discussion by authorita- tive materials provided while the student becomes familiar tive intervention and the Charybdis of unconcerned detach- with the language, discipline, basic concepts and outline of m en t, in se nsitive to stud ent lear nin g ne eds. Equ ally medical studiesº (p. 889). This implies that as students’ proble m atic , ac co rd ing to C o llie r (1 9 80 ) is th e competence and knowledge increase, the tutor’s role or ª constructivelyº aggressive tutor who ª counters the collu- style should change.Thus one might be tempted to conceive sion of some groups to deny con¯ ict to arrive at a quick of the novice and expert PBL student whose needs would consensusº (p. 58). Preferable to this is the tutor ª who acts mature over curricular time and the requisite tutor role as a catalyst, clarifying and amplifying without prescribingº would need to adopt pari passu . We will return to this issue (p. 58). in the section on tutor content expertise. Obviously related to how `directive’ the tutor m ight be in It is from this perspective that one can perhaps judge the tutorial is the issue of how much the tutor knows about the ® ndings of other studies that have attempted to de® ne the content material being discussed. Frustration of faculty who processes and dynamics of the PBL tutorial.W ilkerson et al. are new to the PBL situation re¯ ects the desire to do what have de® ned ® ve qualities that they felt discrim inated self- teachers traditionally doÐ impart knowledge. Clearly the directed as opposed to faculty-directed tutors: (1) who initi- more knowledge faculty have about the matter at hand (i.e. ates topics for discussion; (2) the style and pattern of tutor content expertise) the greater the tem ptation to `direct’ the talk; (3) the use of questions; (4) the pattern of student± tutorial process. Before reviewing the evidence from studies tutor interaction; (5) silences and interruptions. These examining the controversial educational value of `expert parameters of tutorial function were de® ned from observa- versus non-expert tutors’ , two related issues need to be tions of four tutorial groups in action (Wilkerson et al ., addressed: cognitive aspects of the tutor role and peer or 1991). Thus, in this fram ework, it is the tutor who is defined student tutors. 395
  • A.J. NevilleCognitive aspect s of the tutor role content expertise of tutors, which will be discussed in full in the next section.The ® rst issue relates to the educational value in cognitive In her review of problem -based lear ning , Laenoraterms of the input of the tutor into tutorial discussion. The Berkson (1993) questions the validity of accepting thesestrategy of problem solving which best describes the activi- three studies at face value. She compares the problem-ties within the PBL small-group tutorial is the hypothetico- solving context of each of the three studies, which are thedeductive m odel.Within this context, the role of the tutor in relatively well-structured dom ains of physics, biology andthe problem-solving process can be de® ned. M any PBL logic, respectively, with what she de® nes as the ill-structuredcurricula are designed so that the students receive the domain of medicine and ® nds that direct extrapolation of`problems’ ® rst and attem pt to work through them `cold’ . the results might be inappropriate. In ill-structured domains,The rationale is that problem-oriented processing may the number of problem s to which a student is exposed orprom ote eventual processing of an analogo us problem their prototypicality m ay in¯ uence eventual success in(Szekely, 1950). The ® rst study to address this was reported analogical problem solving. In addition, the Needham &in the 1940s by Szekely, who compared the effect of two Begg study (1991) did demonstrate that memory-orientedinstructional sequences on a student’ s eventual ability to processing was a powerful way to ensure subsequent successsolve a problem in m om entum physics. Students were asked at analogical problem solving (0% success, attempting toto study a text explaining principles relevant to the eventual problem-solve cold, rising to 69% after memory-orientedtest problem. One group who had to solve an analogous processing). Despite these caveats, however, there is someproblem and were given corrective feedback prior to reading evidence from the educational and psychology literature tothe text solved the test problem m ore frequently than did support and de® ne the learning facilitator role of the tutor.the group that read the text and then had the solution of theanalogous problem presented as an example (Szekely, 1950). Peer (student) tutoringSimilar results were reported by Schmidt et al. (1989) in astudy of m edical students at the University of Limburg. In developing the `facilitator’ role theme a little further, theGroup discussion prior to the study of a text concerning faculty at the m edical school at the University of Limburgosm osis and diffusion facilitated eventual recall of the in The Netherlands considered the issue of student or peercontents of the text. tutors (De Grave et al ., 1990). They com pared staff and A third study which not only con® rms the potential student tutors with respect to tutor behaviour, group func-value of the `problem ® rst’ approach to PBL but says much tion and test achievement of students. Their rationale forabout the role of the tutor was reported by Needham & considering students as tutors rested on two theoreticalBegg (19 91). They attem pted to show that problem - behavioural constructs, Role Theory, described by Allen inoriented training promotes spontaneous analogical transfer 1976, and Cognitive C ongruence Theory, published byof information, i.e. promotes the ability to use information Cornwall in 1979 (quoted in De grave et al ., 1990, p. 124).from one problem to solve another problem without an Role theory suggests that teachers and students haveexplicit hint to use the previous inform ation. The experi- different stereotypical roles with different expectations,ments, which were carried out using ® rst-year psychology responsibilities and status. Theoretically, the role similaritystudents as subjects, basically showed that, if subjects tried of student tutors and students should have bene® cial effectsto solve a training problem before hearing its solution, or on m otivation owing to active, com m itted enthusiastictried to explain a training story’s solution before hearing the student tutors.correct explanation, spontaneous transfer was more likely Cognitive congruence theory states that experts differthan it was if subjects had studied the same training passage from novices in their cognitive structures. Thus there existsfor memory before hearing its solution or explanation. O f cognitive incongruence between students and staff tutors,more signi® cance for the tutor role in PBL was the ® nding but cognitive congruence between students and studentthat corrective feedback during the preliminary problem tutors. The result of cognitive congruence would be bettersolving was important. W ithout feedback, problem solution assessm ent of prior knowledge of th e students whenrates dropped to about 66% for the `problem -oriented explaining som ething, and greater fam iliarity with theprocessors’ (from 90%) and to about 57% for the `memory- language, concepts and exam ples used by the students (Deoriented processors’ (from ~ 70%). In other wo rds, if Grave et al ., 1990). These two theoretical constructs werestudents are in a problem-solving situation, working through exam ined by analysis of a questionnaire administered toa clinical problem, and are allowed to proceed with incor- medical students in a PBL curriculum . The students wererect conceptual or procedural understanding of the relevant also given a true/false (76-item) factual knowledge achieve-issues, the bene® t of the problem-solving approach to explain ment test. From the 26-item questionnaire, a series offuture analogous problems is lost.The tutor becomes central, discriminant analyses was performed to discriminate betweentherefore, in `facilitating’ learning by stepping in to correct student-led and faculty-led tutorial groups.basic misconceptions that might be leading individuals (or The results indicated that if one used the test achieve-the group as whole) astray. This is not to assert that the m e nt sco re as an o utc om e, tu tor ial pro cess-train edtutor should become the self-promoted expositor of all the undergraduate students perform the role of tutor as well asbasic conceptual issues in tutorial but that he/she should faculty, since the scores were similar in each group. Thus,remain alert to the discussion, able to step in and steer the for this test of `factual recall’ , the cognitive congruencediscussion appropriately. This epitomizes the role of facili- theory was not supported. In terms of tutorial process vari-tator. Clearly the tutor needs to know when to step inÐ and ables, as assessed by the questionnaire, students fromcan only do so if he/she understands som ething of the faculty-led g roups experienced the tutorials as `m orecontent under discussion. Again, this brings up the issue of pleasant’ than the student-led tutorials, which tends to396
  • The problem -based learning tutorcontradict the role theory assumptions described earlier appeared that, indeed, expert tutors showed a trend towards(However, student-tutors were felt to be more concerned using behaviours such as lecturing or directing the discus-with the emotional climate of the tutorial group, thereby sion, leaving students less time to introduce their own ideas.supporting role theory). Irrespective of these ® ndings and Conversely, students in groups led by expert tutors not onlytheir putative consonance or dissonance with the two theories scored higher on microbiology item s in the ® nal examina-laid out in the premise of this study, one signi® cant result tion but also rated their PBL experiences higher in term s ofwas that subject-m atter expertise of either type of tutor was being time well spent and enjoyable. They gave themselvesa crucial factor in the functioning of tutorial groups on higher ratings in self-assessment and ability to apply relevantmost aspects of tutorial process addressed by the question- information to the problem s being studied (Davis et al .,naire. Although those identi® ed as subject-expert student 1992).tutors and subject-expert faculty tutors led student tutorials An earlier study was carried out in 1988 at Harvard andwhich were overall more `satisfying’ , this was not re¯ ected published by Silver & Wilkerson (1991). The tutorials ofin the achievement test scores. The results from this Dutch four randomly selected ® rst-time tutors (out of a pool ofstudy suggest that if student satisfaction is the outcome of 15) were audiotaped. The research was carried out duringinterest, content-expert student tutors may perform as well an 1 1 -w ee k in te rd isc iplinar y co ur se in patho log y,as faculty tutors but the ® ndings have not been replicated immunology and microbiology.Two sessions on two separateelsewhere. A similar study of law students at the University cases were taped for each of the four tutors, for a total ofof Lim burg purported to show that student-tutored groups eight tutorial sessions. Content expertise of the tutors wasscored higher on essay-type questions of `higher order’ assessed by asking the tutors to rate their expertise on eachconcepts compared with students in faculty-led groups of the topics discussed using a ® ve-point scale. Silver &(Moust et al ., 1989). Q uite frankly, the data from this study W ilkerson showed that tutors with expertise played a moreand a similar study of medical students by Moust & Schmidt directive role in tutorials, spoke more often and for longer(1992) show m inimal differences between study groups. periods and provided more direct answers to the students’The evidence for using students as peer tutors therefore questions. Expert tutors also suggested m ore of the topicsrem ains unconvincing. for discussion. The authors noted that tutor-to-student exchanges predom inated, with less student-to-studentTh e tutor as content exper t discussion. They concluded that these behaviours of expertThere is a clear divergence of opinion in the literature as to tutors would endanger the development of students’ skillsthe bene® t or disruptiveness of tutor content-knowledge in active, self-directed learning, although their results didexpertise on the facilitation of student learning and tutorial not explicitly demonstrate this.function in PBL. The issue is clearly linked to but remains Using quite different methodology, de Volder (1982)distinct from directiveness addressed earlier in this review. studied the relationship between tutors’ self-perceivedThis is important, because, in some reports, the concern is subject-matter expertise and students’ evaluations of thenot only that content expert tutors detract from students’ tutors’ functioning. `Functioning’ was assessed by m eans ofself-directed learning by `teaching’ or `lecturing’ the students a 13-item questionnaire, consisting of items such as `Thewith their expertise, but also that they dominate the group tutor intervened when the discussion became incoherent’ .dynamic, resulting in less collaborative learning (Zeitz & Tutor content expertise was positively correlated with itemsPaul, 1993). One can examine the topic from the point of re¯ ecting somewhat directive behaviour, a ® nding that deview of learning outcomes, student satisfaction or both Volder interestingly interpreted as re¯ ecting characteristicsparameters. of more effective tutor functioning. An important aspect of the expert tutor debate relates to Somewhat akin to the Davis study is the work of Eagle etthe de® nition of expert (Camp & Anderson, 1993; Zeitz & al. (1992), who demonstrated that content experts werePaul, 1993). W hen th e M cM aster Undergraduate M D superior to non-experts for both process and outcomeProgramme was being planned, the term `expert tutor’ was measures. Eagle et al. studied 70 students at the Universityapplied to ª someone combining clinical and physiological of Calgary during an eight-week integrative course, whichexpertise whose research and care of patients focused on consisted of 24 ill-de® ned clinical cases portrayed byone body systemº (Spaulding, 1991, p. 43). At the tim e, the simulated patients and supported with added written mate-M cMaster Education C om mittee advocated and supported rial. The 17 tutors were all clinicians, and expert tutors werea role for th e `non-e xp er t tutor’ , reasoning th at th e de® ned as clinicians who would be likely to see the particularnon-expert would be less inclined to emphasize detail and type of clinical problem in their practice. Thus tutors weremore inclined to see things from the view point of students. experts for som e problems and non-experts for others. TheLater, Spaulding averred in his history of the early days at results demonstrated that students guided by content expertsM cMaster that ª the notion of an expert tutor was a falla- produced more then twice as many learning issues for self-cious simpli® cationº (Spaulding, 1991, p. 43). directed learning and spent almost twice the amount of Others have de® ned the term `expert’ very narrowly. tim e in self-study. Expert-led students also generated two toDavis et al. (1992) studied 201 students in 27 groups in a three times more learning issues congruent with the learningsecond-year problem -based course in microbiology. They objectives of the problems.applied an extrem ely strict de® nition of what constitutes an Contrary to the positive outcome m easures in the Eagleexpert tutor. Only those who had an active research interest study, investigators at Sherbrooke found little to supportin the speci® c topic being covered by the students were the need for expert tutors (DesM archais & Black, 1991).considered content experts, i.e. research m icrobiologists. In They analysed achievement data of two consecutive cohortscomparing groups led by expert and non-expert tutors, it of students, 200 in total, and 170 tutors, half of whom were 397
  • A.J. Nevilleconsidered to be content experts.The latter term was defined DesM archais & Black (1991). This wide-ranging studiedin term s of the tutor belonging to the appropriate clinical covered 1120 medical students in four curriculum yearsdiscipline. Three elements of achievement were evaluated: a and included 152 tutors participating in 336 tutorials.multiple-choice test, short essay questions and a tutor judge- Achievement was measured by 100± 150 item true± falsement regarding the individual student’ s performance. W hile tests for ® rst-year students and by short essay questions forsome differences emerged in one of the classes on the essay subsequent years. An 11-item Likert-typ e rating scale wastests, overall a composite score of the three measures revealed used to assess tutor functioning (addressing a variety ofno signi® cant differences for groups led by expert tutors. tutor behaviours). Students’ estimates of self-study timeSimilar results were published by Calvin & Wetzel (1989) in was also measured in light of the ® ndings of Eagle et al.a study of the New Pathway at Harvard. (1992) described earlier. W hile assessing objective learning outcom es or observ- Schmidt et al. found that students tutored by expertsable tutorial processes, some studies have also canvassed spent more time on self-directed study and had higherstudent opinions about their level of satisfaction. Davis et al. achievement scores. These ® ndings were most noticeable(1992) included student ratings in their study and an for ® rst-year students, suggesting that novice students wereAustralian study from the University of Newcastle (Feletti more dependent on their tutors’ expertise than advancedet al ., 1982) speci® cally addressed medical students’ evalu- students. In addition to the tutors’ knowledge-relatedation of tutors in a PBL curriculum. Feletti et al. constructed be haviour s, process-facilitation skills affected stud enta 19-item questionnaire and administered it to 50 ® rst-year achievement. Indeed, these two sets of behaviours weremedical students at the regular end-of-term program me correlated, suggesting that both were necessary conditionsevaluation sessions. Twenty-seven tutors were evaluated. A for effective tutoring (Schmidt et al ., 1993).varimax rotated factor analysis wa s perfor m ed on th e Re¯ ecting on the diverse results from studies at his ownquestionnaire results and resulted in the identi® cation of institution and work elsewhere, Schm idt sought to resolvefour major factors in tutors’ behaviour which were important some of the inconsistencies in tutor expertise research byin the rating process. These factors were (a) ability to care re-analysing the data from the U niversity of Limburgfor students, (b) a knowledge of course structure and (Schmidt, 1994). From his survey of the literature, twoteac hing staff philoso phy, (c ) ability to e nc ou rag e issues emerged, over and above the de® nition of the termindependent thinking in students and (d) knowledge of the expert tutor discussed earlier.speci® c m edical problems being studied. Using discriminant The ® rst observation is that, depending on the tutor-analysis, the authors found that a thorough up-to-date training provided, expert tutors may not always demonstrateknowledge of the particular problem being studied was the their content expertise in tutorials to assist students’ problemitem that discrim inated best between tutors rated in the solving. Such was the case in the Sherbrooke study, whereupper and lower half of the g roup according to their the ratings of various facilitative behaviours were equivalentperceived effectiveness. Clinician experts were rated highly for expert and non-expert tutors for six out of seven criteriain their willingness to allow students to develop and explore evaluated (DesM archais & Black, 1991). Schmidt also pointsproblem solving, a ® nding similar to Wilkerson’ s (1992) out that some studies that do show differences in behaviourstudy which showed that students preferred physician tutors between expert and non-expert tutors do not report studentto PhD tutors in balancing basic science and clinical applica- achievement (e.g. see earlier, the studies of Wilkerson, 1992tions, prom oting critical appraisal and synthesising multiple and Feletti et al ., 1982).perspectives. The second issue relates m ore to the students them selves In an attempt to resolve some of the discrepancies am ongthe studies debating the value of tutor content knowledge, I in the PBL setting. Schmidt declares that it is often observedhave deliberately left until last a review of the bulk of the that students who have little or no experience with PBL relyresearch from the University of Limburg and a more recent more heavily on their tutors as sources of guidance andstudy from the University of M ichigan (Davis et al ., 1994) inform ation. Consequently, if these tutors are familiar withbecause this body of work brings into stark relief the issue the subject-matter, this may have an effect on studentof `oversimpli® cation’ alluded to so presciently by W illiam achievement. This hypothesis could explain the ® ndingsSpaulding (1991). from Lim burg that positive effects on achievement of expert The early studies assessing the impact of tutor content tutors were most obvious for ® rst-year students. In a sim ilarexpertise at Maastricht de® ned expertise in terms of whether vein, the 1992 ® ndings of Davis et al. at the University ofthe tutor came from biom edicine, clinical medicine or the M ichigan could re¯ ect that the PBL microbiology coursesocial sciences (Schm idt, 1977; Swanson et al ., 1990). Both was undertaken in a curriculum that was otherwise relativelyof these studies demonstrated, in end-of-unit m ultiple- conventional and thus the students were not particularlychoice tests, achievem ent levels u nrelated to content familiar with PBL (Davis et al ., 1992). In addition, theexpertise of the tutor. Clearly this de® nition of expert differs Michigan study, like several others, addressed a single course,from that of several of the other authors quoted in this rather than an entire curriculum or large block of curricularreview. Thus, as the authors at the University of Limburg units.and authors elsewhere have pointed out, the variability in From th is bo dy of research, Sc hm id t de velops ade® nition of tutor content expertise may partly explain the hypothesis that the novice student arrives to begin a PBLdiscrepant ® ndings. Schmidt et al. , in Limburg, took this cu rr icu lum with little prio r con te nt knowle dg e andissue a step further in a study published in 1993 (Schmidt unfamiliar with the PBL process. Thrust into a self-directedet al ., 1993). In this study, Schmidt et al. chose to adopt the environment which is unstructured, the novice student withde® nition of expert employed in the Sherbrooke study of little prior knowledge on which to build a scaffolding for398
  • The problem -based learning tutornew knowledge needs guidance and relies on the content- which encourages these attributes. Assessing one’ s ow nexpert tutor. If the curriculum itself is heavily structured, knowledge de® cits as a springboard to self-directed learningsuch as at Sherbrooke, the effect of tutor content expertise would also clearly represent the same educational philosophy.m ight be lost. C urricular structure and students’ prior Application of some basic psychometric principles toknowledge are thus viewed by Schmidt as interrelated factors tutorial evaluation, however, unearths some signi® cantin determ ining both students’ learning strategies and the problems with the use of tutorial ratings. In addition, atextent to which the content expertise of the tutor might least one PBL medical school has refused to utilize tutorimpact on the learning. Schmidt argues, quite reasonably, and peer ratings in summative evaluation because of a `belief ’that curricular and student heterogeneity across medical that the tutor (teacher) and evaluator roles are incom patibleschools studied could account for the differences in observed (van der Vleuten & Verwijnen, 1990). This has not reallyeffects of tutor content expertise on student satisfaction and been formally tested.achievement. He concludes that students need a minimum W hat has been tested extensively is the relationshiplevel of structure in order to pro® t from problem-based between self-assessment and objective (written or facultyinstruction. Structure can be provided either internally by observed) performance (Gordon, 1991). In all but a fewprior knowledge or, if prior knowledge is lacking and the instances where the criteria for self-assessment were verycurriculum lacks structure, by falling back on the tutor for explicit (Geissler, 1993), the validity and reliability have`content’ suppor t. These conclusions of Schm idt have been low (van der Vleuten & Verwijnen, 1990). Inter-itemimplications not only for tutors and tutor training, but also correlations related to process issues are frequently high,for curriculum design in PBL. however, suggesting that raters are providing only a `global’ O ne last comm ent on curriculum structure and its impression, rather than identifying individual attributesrelationship to tutor content expertise can be left to Wayne (Gordon, 1991). Blumberg et al. at McMaster, however,Davis’s group at the University of M ichigan. Having reported have challenged the notion that tutorial assessm ent providesin their 1992 study on the positive effects of achievement only global impressions by performing a cluster analysis ofand student satisfaction of `very expert’ tutors, this group six dimensions or domains evaluated in tutorials (i.e. problemrepeated their examination of this issue (Davis et al ., 1994). solving, knowledge, critical appraisal, clinical skills, learningIn this m ore rec ent iteration, all tu tors (e xp er t and skills and personal qualities) (Blum berg et al ., 1994). Theynon-expert) were given highly structured and comprehensive found, surprisingly, that the six domains appeared to beinformation about a particular case the students were about evaluated relatively independently with no single domainto study, i.e. the situation was highly focused. Results using particularly in¯ uencing the outcome of others.multiple-choice test achievement as an outcome showed no Cohen et al. (1993) from the same group have alsosigni® cant effect of tutor content expertise, which was reported troubling evidence of faculty reluctance to awardm irrored in the assessm ent of student satisfaction. As students unsatisfactory tutorial evaluations.While there maySchm idt would probably concur, if enough highly focused be several reasons for this, it would appear that agreedcurricular structure is achieved, the signi® cance of tutor `valid’ criteria for levels of perform ance need to be statedcontent knowledge expertise may be diminished. explicitly and understood by tutors and students alike From th is review o f th e literat ure o n th e debate (Cohen et al., 1993). Even though much of what is evalu-surrounding the issue of PBL tutor content expertise, one ated in tutorials appears to relate to process issues, studentscan conclude that if the conditions of study are varied with repeated signi® cant de® cits can potentially be identi-enough, either bene® cial or detrim ental effects on PBL self- ® ed (Blum berg et al., 1995), especially if previous evalua-directed, student-oriented learning can be inferred. W hen tion s are forwarde d to subseque nt tutors ( C ohen &one combines the results of the empiric studies with informa- Blumberg , 1991). In contrast, Blake et al. (1996) found thattion gleaned from the cognitive psychology literature, one written tutorial evaluations (accum ulated over m ultiplegets the impression that problem solving in a knowledge curricular units or blocks) did not predict perform ance on avacuum , without corrective feedback, is frustrating for the practice licensing examination.learner and inefficient or ineffective in term s of educational Given that the Association of Am erican M edical Collegesachievem ent. The deg ree of tutor conte nt knowledge (AAM C) has recom mended that subjective faculty assess-required for effective learning facilitation in PBL is not an ments should be the main focus of performance evaluationabsolute quantity but needs to be tailored to the particular and that psychometric `solutions’ should serve as supple-student groups’ level of prior knowledge and familiarity ments to, not substitutes for, faculty judgement, it wouldwith PBL. seem from the evidence presented here that m uch work needs to be done to increase the `credibility’ of tutorialTutor as student evaluator (tutor and peer) assessment (DaRoza, 1993).The debate surrounding the use of the tutor in PBL as astudent evaluator rem ains as unresolved as that of the issue C onclusionsof knowledge content expertise. A signi® cant difference isthe lack of studies to address the former question. In 1981, Johnson et al. published a meta-analysis of the Tutorial evaluation is central to the entire evaluation effects of cooperative, competitive and individualistic goalprocess at M cMaster (Neufeld &Barrows, 1974). Tutor, structures on achievement (Johnson et al ., 1981). In a verypeer and self-ratings are used to evaluate many domains wide-ranging review, not restricted to studies of learners insuch as knowledge, effort, self-directed learning, group skills the health professions, Johnson et al. dem onstrated theand communication skills. This is entirely consistent with superior ity of cooperative learning over interpersonalthe philosophy of problem-based learning in small groups competition in terms of achievement, knowledge retention, 399
  • A.J. Nevillesatisfaction and development of interpersonal communica- have described `context-free’ generalizations that assumetion skills. In this domain of cooperative or collaborative constant tutor behaviour in different situations, i.e. tutors inlearning, students teach one another; however, PBL medical different tutorial groups, courses, subject-m atter, etc. Aschool curricula have modi® ed this domain by adding a recent study by Gijselaers from M aastricht (1997) raisesfaculty tutor . some concerns about interpreting previous work in the field. From all that has been reviewed to this point, one m ay Using a previously validated evaluation instrument comprisingperhaps need to take issue with Barrow’s comment, quoted a 12-item questionnaire of tutor behaviour administered toearlier: ª A faculty person who is a good tutor can success- medical students, Gijselaers assessed (a) stability of tutorfully tutor in any areaº (Barrows & Tamblyn, 1980). behaviour, i.e. the extent to which a tutor’s behaviour as The evidence presented suggests that the presence of a measured in one course correlated with the same measure intutor facilitates both tutorial process and learning achieve- a different course, and (b) generalizabilty of tutor behaviour,ment. The matter is not completely closed, however, since i.e. the extent to which measures of tutor behaviour werethere are no studies directly comparing the learning of tuto- stable across courses. He also considered the effect ofrial groups with or without tutors. departmental affiliation of PBL tutors. At UCLA School of Medicine, a tutorless format of PBL This study showed that the stability and generalizabiltyhas been introduced to reduce the demand on faculty time of tutor behaviour were low and, interestingly, moderately(Duek et al ., 1996). From a study conducted in that institu- related to departmental affiliation. This study in no waytion of learning issues identi® ed by students in tutorless invalidates previous research in the area of effective PBLtutorials, it was found that the m ean core overlap of student tutoring but emphasizes that different situations requireand faculty-derived objectives was only 25%, compared with different tutor behaviour to facilitate student learning,41% in a study from Limburg using faculty-tutored tutorials. consistent with the thesis proposed earlier by SchmidtThese ® ndings from UCLA suggest that, at least in terms of (1994). The caveat, however, is that future studies of effec-de® ning the learning issues, tutors do play a role, especially tive PBL tutoring should probably consider context-specificif other sources of structure are absent from the curriculum. character istics an d, pe rhaps, d epar tm e ntal and /or Em ploying advanced students as tutors puts th ese organizational background of PBL tutors.students in the same role as faculty tutors and should not be Finally, as observers of tutorial process, tutors may be ableconfused with peer collaborative learning. Parenthetically, to evaluate certain aspects of professional behaviours andthe literature on the educational value of student tutors interpersonal communication skills, but there is really verycom es mostly from one institution and I think the data are little evidence, from the sparse literature available, that confirmsconfusing and inconclusive. the tutor as an effective evaluator of student performance, M ost authors have therefore agreed that faculty should particularly knowledge. It would appear that assessment ofmaintain a facilitative role, but the word facilitation is open student achievement is best left outside tutorials.to interpretation and clearly has elements of both tutorial The extent to which faculty in¯ uence learning in PBLprocess and tutorial learning. The degree of directiveness of remains the subject of debate and further research is requiredtutors in tutorials and the content knowledge they are to elucidate the effects of the tutor on the extent or breadth ofexpected to display are widely debated in the literature, and learning and the development of self-directed learning skills.perhaps the best conclusion that can be drawn from the However, one can discern from the literature a picture ofliterature is that the tutor’s leadership behaviour should not the ideal tutor who is ¯ exible and sensitive to the student’sremain the same in all PBL situations, but needs to be learning needs and knowledgeable about the curriculum .varied according to student level or curriculum as describedby Schmidt (1994). There is som e congruence in this Note on contributordevelopm ental approach to teaching between Schmidt’ s P ROFESSOR A LAN J. N EVILLE is Associate Professor of M edicine andproposal and Neame & Powis’s (1981) conclusions described Chair of the Undergraduate M D Program at McMaster University,earlier, which I think are germane to tutor function. The Ham ilton, Ontario, Canada.common link is structure. Novice students, with little experi-ence of PBL or prior knowledge, probably bene® t from Referen cesdirective and knowledge expert tutors to provide the neces- B AR ROW S , H.S. & TAM BLYN , R. (1980) Problem-B ased Learning: A nsary structure or foundation upon which to build their Approach to Medica l Education (N ew York, Springer).learning. Davis et al. (1994) would add that a highly focused B ERKSON , L. (1993 ) Problem-based learning: have the expectationscurriculum m ight compensate to some extent for the level been m et? A cademic M edicine, 68(10 ) Suppl, pp. 579± 588.of knowledge expertise on the part of the tutor. As students B LAKE , J.M . et al. (1996 ) Validity of Progress Tests and Tutorial Assessm ent in Predicting Licensing Exam ination Perform ance,mature, in knowledge as well as familiarity with PBL, the unpublished paper.tutor should becom e m ore participatory or delegatory, B LU M B ERG , P. & E C KENF ELS , E.A. (1988 ) Comparison of studentallowing the students more leeway in deciding what and satisfaction with their preclinical environm ent in a traditional andhow they will learn. Sim ply put, novice students in PBL a problem -based curriculum , in: Research in Medical Education:need some direction or structure, or they may ¯ ounder. Proceedings of the Twenty-Seventh A nnual Conference, pp. 60± 65Mature students can ¯ ourish in a far less structured learning (Washington, DC, Association of American M edical Colleges). B LU M BERG , P. et al. (1990 ) Roles of student-generated learning issuesenvironment as they becom e m ore self-sufficient. Indeed, in problem -based learning, Teaching and Learning in Medicine, 2(3),directive tutoring for the latter may frustrate and antagonize pp. 149± 154 .such students. B L U M B E R G , P. et a l. (1 99 4 ) A n alys is of acad em ic p rob lem s As an interesting postscript to this issue, it should be encountered by m edical students, Teachin g and Learning in Medicine,pointed out that the majority of studies cited in this review 6(2), pp. 96± 191.400
  • The problem -based learning tutorB LU M BERG , P. et al. (1995 ) The prog ress of students who encounter K NOW LES , M. (1975) Self-directed Learning. G uide for Learners and repeated difficulties in a problem -based m edical school, Teaching Teachers (Toronto, Prentice Hall, Canada). and Learning in M edicine, 7(1), pp. 23± 28. M IC HAELS , J.W. (1977 ) Classroom reward structures and academ icC ALVIN , R.B. & WETZEL , M.S. (1989 ) Pathology in the new pathway perform ance, R eview of Educational Research, 47, pp. 87± 98. of medical education at Harvard Medical School, A merican Jour nal M OU ST , J.H .C. & S CHM IDT , H .G. (1992 ) underg raduate students as of Clinical Pathology, 92(Supp l., O ctober), pp. S23± S30. tutors: are they as effective as faculty in conducting sm all-groupC A M P , M .G . & A N D ER S O N , A .S. (19 93) Expe rt tutors versus tutorials?, paper presented at the Annual M eeting of the American non-expert tutors in PBL (letter), Academic M edicine, 68, p. 353. Educational Research Association, San Francisco, California, April.C OH EN , G. & B LUM BERG , P. (1991 ) Investigating whether teachers M OU ST , J.H.C . et al. (1989 ) Peer teaching and higher level cognitive should be g iven assessm ents of students m ade by previous teachers, outcom es in problem -based learning, H igher Education, 18, pp. Academic Medicine, 66, pp. 288± 289. 737± 742.C OH EN , G.S. et al. (1993 ) Do ® nal grades re¯ ect written qualitative N EAM E , R.L .B. & P OW IS , D.A. (1981 ) Toward independent learning: evaluations of student perform ance, Teaching and Learning in curricular design for assisting students to learn how to learn, Journal M edicine, 5(1), pp. 10± 15. of Medical Education, 56, pp. 886 ± 893.C OLLIER , K.G. (1980 ) Peer-group learning in higher education: the N EEDHAM , D. & B EG G , I. (1991 ) Problem -oriented training promotes developm ent of higher order skills, Studies in Higher Education, 5, spontaneous analogical transfer: mem ory-oriented training promotes pp. 55± 62. m emory for training, M emor y and Cognition, 19, pp. 543± 557.C OU LSON , R.L. (1983 ) Problem -based student-centered learning of N EU FELD , V.R. & B AR ROW S , H.S. (1974 ) The `McM aster Philosophy’ : the cardiovascular system using th e Problem -Based Learning an approach to m edical education, Jour nal of M edical Education, M odule (P.B.L.M.), Physiologist, 26, pp. 220 ± 224. 49, pp. 1040± 1050.D A R OZA , D.A. (1993 ) Commentary on `Do ® nal grades re¯ ect written S CH M IDT , H. et al. (1989 ) Explanatory models in the processing of qualitative evaluations of student perform ance?’ , Teachin g and science text: the role of prior know ledge activation through sm all- Learning in Medicine, 5(1), pp. 16± 17. group discussion, Journal of Educational Psychology, 81, pp. 610 ± 619.D AVIS , W.K . et al. (1992 ) Effects of expert and non-expert facilitators S CH M IDT , H .G. et al. (1993 ) In¯ uence of tutors’ subject-m atter on the small-group process and on student perform ance, A cadem ic expertise on student effort and achievem ent in problem -based M edicine, 67, pp. 407 ± 474 . learning, Academic M edicine, 68, pp. 784 ± 791.D AVIS , W.K . et al. (1994 ) In¯ uence of a highly focussed case on the S CH M IDT , H .G . (1977 ) Niet-medicials tutor: m aakt het een vershil? effect of small-group facilitator’s content expertise on students’ (N on-m edical staff as tutors: does it m ake a difference?), internal learning and satisfaction, Academic M edicine, 69, pp. 663± 669 . report, U niversity of Lim burg, Maastricht, The Netherlands.D E G RAVE et al. (1990 ) Peer teaching and problem -based learning: S CHM IDT , H.G. (1983 ) Problem -based learning: rationale and descrip- tutor characteristics, tutor functioning, group functioning and tion, M edical Education, 17, pp. 11± 16. student achievem ent, in: Z.M . N OOM AN , H.G. S C HM IDT & E.S. S CH M IDT , H.G. (1989 ) Resolving inconsistencies in tutor expertise E ZZAT (Eds) Innovation in Medica l Education: A n Evaluation of its research: does lack of structure cause students to seek tutor guid- Present Status, pp. 135 ± 137 (N ew York, Springer). ance, Academic Medicine, 69, pp. 656± 662.DE VOLDER , M .L. (1982 ) Discussion groups and their tutors: relation- S HOPE , T.R. (1989 ) Student-initiated analysis and change of a medical ships between tutor characteristics and tutor functioning. Higher school curriculum, Academic Medicine, 64, pp. 330± 301. Education, 11, pp. 269 ± 271. S ILVER , M . & W ILKERS ON , L. (1991 ) Effects of tutors with subjectD E V R IE S , M . et al. (1989 ) D utch com parisons: cognitive and expertise on the problem ± based tutorial process, Academic M edicine, m otivational effects of problem -based learning on m edical students, 66, pp. 298± 300 . in: H.G. S C HM IDT , J.R. L IPKIN , M . D E V R IES & J.M . G REEP (Ed s) S PAULDI NG , W.B. (1991 ) Revitalising M edical Education. M cM aster New Directions for M edical Education : Problem-B ased Learning and M edical School.The EarlyYears 1965± 1974 (Ham ilton, B.C. Becker). Community-Oriented Medical Education, pp. 230± 238 (N ew York, S W AN SO N , D.B. et al. (1990 ) Effect of tutor characteristics on test Springer-Verlag). perform ance of students in a problem -based curriculum, in: W.D ES M ARC HAIS , J.E . & B LACK , R. (1991 ) Effect of tutor content B ENDER , R.J. H IEM STRA , A. J. J. A. S CHER PBIER & R. P. Z W IER STRA expertise on student academ ic achievem ent in the Sherbrooke (E ds) Teachin g and A ssessing Clinical Competence ., pp. 129 ± 134 Problem -Based Curriculum , unpublished manuscript, U niversite  (G roningen, T he Netherlands, BoekWerk Publications). de Sherbrooke, Quebec, Canada. S ZEKELY , L. (1950 ) Productive processes in learning and thinking,D ES M ARC HAIS , J.E . et al. (1992 ) From traditional to problem -based A cta Psychiatrica, 7, pp. 388± 407 . learning: a case report of com plete curriculum reform, M edical T H OM AS , R.E. (1992 ) Teaching m edicine w ith cases: student and Education, 26, pp. 190 ± 199. teacher opinion, M edical Education, 26, pp. 220± 297 .D U EK , J.E . et al. (1996 ) Learning issues identi® ed by students in VAN DER V LEU TEN , C. & VERW IJNEN , M. (1990 ) A system for student tutorless problem-based tutorials, Advances in H ealth Sciences Educa- assessm ent, in: C. VAN DER V LEUTEN & M . VERW IJNEN (Eds) Problem- tion, 1, pp. 29± 40. B as ed L ear n ing: Persp ective s from th e M a astr icht A p p roa chE AG LE , C.J. et al. (1992 ) Effects of tutors with case expertise on (Am sterdam , Thesis-publ.). problem-based learning issues, Academic Medicine, 67, pp. 465± 469. WALTON , H.J. & M ATTH EWS , M .B. (1989 ) Essentials of problem -F ELETTI , G.I. et al. (1982 ) M edical students’ evaluation of tutors in based learning, M edical Education, 23, pp. 542± 558 . a group-learning curriculum , M edical Education, 16, pp. 319± 325. WEBB , N.M . (1982 ) Student interaction and learning in sm all groups,G EISSLER , P.R. (1973 ) Student self-assessm ent in dental technology, R eview of Educational Research, 52(3) , pp. 421± 445 . Journal of Denta l Technology, 37(9) , pp. 19± 21. W ILKERS ON , L. et al. (1991 ) A case study of student-directed discus-G IJSELA E RS , W.H . (1997 ) Effects of contextual factors on tutor sion in four problem ± based tutorial groups. In research in m edical behaviour, Teachin g and Lear ning in M edicine, 9(2), pp. 116 ± 124. education, Academic M edicine, 66(Suppl., Septem ber), pp. S79± S81 .G ORDON , M.J. (1991 ) A review of the validity and accuracy of self- W ILKERS ON , L. (1992 ) Identi® cation of skills for the problem -based assessm ents in health profession training, Academic Medicine, 66, tutor: student and faculty perspectives, paper presented at the pp. 762 ± 769. Annual M eeting of the American Educational Research Associa-G RU PPEN , L.D. et al. (1992 ) Tutor-led and student-led sm all groups: tion. San Francisco, California, April. no differences in learning?, paper presented at the Annual M eeting WOODW ARD , C.A. & F ERR IER , B.M . (1982 ) Perspectives of graduates of the Am erican Educational Research Association, San Francisco, two or ® ve years after graduation from a three-year m edical school, California, April. Jour nal of M edical Education, 57, pp. 294± 302.JOH NSON , D.W. et al. (1981 ) Effects of cooperative, com petitive and Z EITZ , H.J. & P AU L , H . (1993 ) Facilitator expertise and problem - individualistic goal structures on achievem ent: a meta-analysis, based learning in pbl and traditional curricula (letter), A cadem ic Psychological B ulletin, 89(1) , pp. 47± 62. M edicine, 68, pp. 203± 204 . 401