The Effectiveness of Peer Tutoring in Further and Higher Education: A Typology and Review of the Literature Author(s): K. J. Topping Source: Higher Education, Vol. 32, No. 3 (Oct., 1996), pp. 321-345 Published by: Springer Stable URL: http://www.jstor.org/stable/3448075 . Accessed: 14/01/2011 16:28 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at . http://www.jstor.org/action/showPublisher?publisherCode=springer. . Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. Springer is collaborating with JSTOR to digitize, preserve and extend access to Higher Education. http://www.jstor.org Higher Education 32: 321-345, 1996. ? 1996 KluwerAcademic Publishers. Printedin the Netherlands. 321 The effectiveness of peer tutoring in further and higher education: A typology and review of the literature K.J.TOPPING Centrefor PairedLearning,PsychologyDepartment,Universityof Dundee, Dundee DDI 4HN, Scotland Abstract. Quality,outcomesandcost-effectivenessof methodsof teachingand learningin colleges and universitiesare being scrutinisedmore closely. The increasinguse of peer tutoring in this context necessitates a clear definitionand typology, which are outlined. The theoretical advantagesof peer tutoringare discussed and the researchon peer tutoring in schools brieflyconsidered.The substantialexisting researchon the effectiveness of the manydifferent types and formatsof peer tutoringwithin colleges and universitiesis then reviewed. Much is already known about the effectiveness of some types of peer tutoringand this merits wider disseminationto practitioners.Directionsfor futureresearchare indicated. Researchonteachingandlearninginfurtherandhighereducationis muchless voluminousthanthaton teachingandlearningin schools.Whiletherehave beena numberof booksonthetopicof adultlearning(e.g.Rogers1977,Lovell 1980,Gibbs1981,Tight1983,Brookfield1983,EntwistleandRamsden1983, MartonHounsellandEntwistle1984,Ramsden1986,RichardsonEysenck andPiper1987,MerriamandCaffarella1991,Laurillard1993, Sutherland 1996),boththe quantityandqualityof researchin this areais surprisingly limited,consideringthe vastresourcesexpendedon thetertiarysector. of teachingandlearningin the However,thequalityandcost-effectiveness sectorare increasinglyunderthe microscope.Therehas long beenconcern thattraditional curricula,deliveredandassessedin traditional ways,promote a surfaceapproachto learningratherthana deeporevena strategicapproach (Entwistle1992).Teachingqualityassessmentexercisesconsistentlyresultin forfailingtopromotethedevelopment of transferable criticismof departments skillsin theirstudents(Barnett1992,Ellis 1993).At thesametime,increased studentnumberscoupledwithreducedresourceshaveoftenresultedin larger class sizes, thusencouraginga reversionto a traditionallecturingstyle of deliveryanda reductionin smallgroupandtutorialcontact- in short,less interactive teachingandlearing. to improveteachingqualitywhile doingmorewith Thedualrequirement less' has recentlyincreasedinterestin peer tutoringin higherand further education.However,it wouldbe unwiseto seize upon peer tutoringas a and instantpanacea.Differentformatsof peer universal,undifferentiated 322 tutoringhavebeen the subjectof researchof differingquantityandquality, withvariousoutcomes. Definitionsand typology Peertutoringis a veryoldpractice,traceablebackat leastas faras theancient Greeks.Archaicdefinitionsof peer tutoringperceivedthe peer tutoras a surrogateteacher,in a linearmodelof the transmissionof knowledge,from teacherto tutorto tutee.Later,it wasrealisedthatthepeertutoringinteraction was qualitativelydifferentfrom thatbetweena teacherand a student,and involveddifferentadvantagesanddisadvantages. At this point of development,a definitionmighthave been: 'moreable studentshelpingless able studentsto learnin co-operativeworkingpairs or small groupscarefullyorganisedby a professionalteacher'.However, as developmentandresearchin differentformatsof peertutoringproceeded apaceinmorerecentyears,itbecameclearthatpeertutoringis notnecessarily only abouttransmissionfromthe moreable andexperienced(who already have the knowledgeand skills) to the less able (who have yet to acquire them).As peertutoringhasdeveloped,definingit hasbecomemoredifficult, anda currentdefinitionseems so broadas to be ratherbland:'peoplefrom similarsocialgroupingswhoarenotprofessionalteachershelpingeachother to learnandlearningthemselvesby teaching'.However,thisdefinitiondoes includereferenceto thegainsaccruingfromthetutoringprocessto thetutor - increasingly, peertutoringprojectstargetgainsfor bothtutorsandtutees. Peertutoringis characterised by specificroletaking:at anypointsomeone has the job of tutorwhile the other(s)are in role as tutee(s).Peertutoring content.Projectsusuallyalso outline typicallyhas highfocuson curriculum forinteraction, in whichtheparticipants arelikelyto quitespecificprocedures havetrainingwhichis specificorgenericorboth.Inaddition,theirinteraction may be guidedby the provisionof structuredmaterials,amongstwhicha degreeof studentchoicemaybe available. A typologyof peertutoringcouldincludeten dimensions: 1. Curriculum Content- whichmay be knowledgeor skills orientated,or a combination.The scopeof peertutoringis verywideandprojectsare in virtuallyeveryimaginablesubject. reportedin the literature 2. ContactConstellation- some projectsoperatewith one tutorworking witha groupof tutees,butthe size of groupcan varyfromtwoto thirty or more. Sometimestwo tutorstake a groupof tuteestogether.Less andmoreintensive,is peertutoringin pairs(dyads). traditional, 3. Yearof Study- tutorsandtuteesmaybe fromthesameordifferentyears of study. 323 4. Ability- while manyprojectsoperateon a cross-abilitybasis (even if they aresame-year),thereis increasinginterestin same-abilitytutoring (wherethetutorhassuperiormasteryof only a verysmallportionof the curriculum,or a pairareof equalabilitybutworkingtowardsa shared, deeperandhopefullycorrectunderstanding). 5. Role Continuity- especiallyin same-abilitytutoring,the roles of tutor Structured andtuteeneednotbepermanent. switchingof rolesatstrategic canhavetheadvantage of involvinggreater moments(reciprocal tutoring) noveltyanda widerboostto self-esteem,in thatall participants get to be tutors. 6. Place- Peertutoringmayvaryenormouslyin locationof operation. 7. Time - peer tutoringmay be scheduledin regularclass contacttime, outsideof this, or in a combinationof both,dependingon the extentto or supplementary. whichit is substitutional 8. TuteeCharacteristics projectsmay be for all studentsor a targeted subgroup,e.g. the especiallyable or gifted,thoseconsideredat risk of failureor dropout,andthosefromethnic,religious under-achievement, andotherminorities. - thetraditional 9. TutorCharacteristics assumptionwas thattutorsshould be the 'best students'(i.e. those most like the professionalteachers). However,verylargedifferentialsin abilitycan proveunder-stimulating forthetutor.If tutorsarestudentswhoaremerelyaverage(orevenless), bothtutorandtuteeshouldfindsome cognitivechallengein theirjoint activities(e.g. Fantuzzo,DimeffandFox 1989).Althoughtuteegainmay notbe so great,theaggregategainof bothcombinedmaybegreater.Many projectsin schoolshavedeployedstudentswithlearningandbehaviour difficultiesas tutors,to thebenefitof the tutorsthemselves(Scruggsand Osguthorpe1986,AshmanandElkins1990). 10. Objectives- projectsmay targetintellectualgains, formalacademic achievement,affectiveandattitudinal gains,socialandemotionalgains, self imageand self conceptgains,or any combination.Organisational objectivesmightincludereducingdropout,increasingaccess,etc. Theoreticaladvantagesof peer tutoring The cognitiveprocessesinvolvedin peer tutoringhave been exploredby variouswritersoverthe years,manyof whomemphasisedthe valueof the andquestioning(e.g. Gartner,Kohlerand Riessman inherentverbalisation 1971, DurlingandSchick 1976,BarghandSchul 1980, Webb1982, Foot, Shute,MorganandBarron1990,Forman1994).A neo-Piagetian interpretation of individualdevelopment throughthe cognitiveconflictandchallenge 324 involvedin manyformsof peer assistedlearningis offeredby Doise and Mugny(1984).However,peertutoringis morefullyunderstoodthroughthe view of cogsocial interactionist (or socio-culturalor socialconstructivist) nitivedevelopment.Supported(or 'scaffolded')explorationthroughsocial andcognitiveinteractionwith a moreexperiencedpeerin relationto a task of a level of difficultywithinthe tutee's 'zone of proximaldevelopment' remainsa theoreticalcornerstoneof peerassistedlearning(Vygotsky1978). This themehas beenfurtherdevelopedby BarbaraRogoff(1990)underthe in thinking'. labelof 'apprenticeship Peer tutoringis often promotedon the groundsthat,for the tutors,it is 'Learningby Teaching'.This view is expandedin the old saying 'to teach is to learntwice'. Sternberg's iden(1985)theoryof intelligentperformance tifies componentswhichmightbe enhancedduringpeertutoring(Hartman 1990):the meta-cognitiveskills of planning,monitoringandevaluatingand the associateduse of declarative,proceduralandcontextualknowledge;and thecognitiveprocessesof perceiving,differentiating, selecting,storing,inferand ring,applying,combining,justifying responding.Justpreparingto be a peertutorhasbeenproposedto enhancecognitiveprocessingin thetutor- by increasingattentionto andmotivationfor the task,andnecessitatingreview of existingknowledgeandskills.Consequently, existingknowledgeis transnew associations anda newintegration. formedby re-organisation, involving The act of tutoringitself involvesfurthercognitivechallenge,particularly clarification andexemplification. withrespectto simplification, An excellentstudyby Annis (1983) comparedthreerandomlyallocated groupsof students:one whichmerelyreadthe materialto be studied,one which readthe materialin the expectationof havingto teach it to a peer, anda thirdwhichreadthe materialwith the expectationof teachingit to a peer and then actuallycarriedthis out. On a 48 item test of both specific andgeneralcompetence,the 'readonly' groupgainedless thanthe 'readto teach'groupwhichin turngainedless thanthe 'readandteach'group.The tutorsgainedmorethanthe tutees.A similarstudyby Benwareand Deci (1984) comparedthe relativeeffectivenessof readingto learnfor a test and readingfor learningto teach a peer.Subjectswere randomlyassignedto conditionsandthe outcomemeasurewas a 24 itemtest of bothrotememory While both groupsperformedequallywell and conceptualunderstanding. on rotelearning,the 'learnto teach'groupperformedbetteron higherorder andon a questionnaire conceptualunderstanding, regardingmotivationand their as more active and learningperceived experience interesting. Manyotheradvantageshave been claimedfor peer tutoringand related formsof peer assistedlearning(e.g. Greenwood,Cartaand Kamps1990). Pedagogicaladvantagesfor the tutee includemore active, interactiveand 325 participativelearning,immediatefeedback,swift prompting,loweredanxiety with correspondinglyhigherself-disclosure,and greaterstudentownershipof the learningprocess. The 'pupil/teacher'ratio is much reducedand engaged time on task increased. Opportunitiesto respond are high, and opportunities to make errorsand be corrected similarly high. In addition to immediate cognitive gains, improvedretention,greatermeta-cognitiveawareness and better application of knowledge and skills to new situations have been claimed. Motivationaland attitudinalgains can include greatercommitment, self-esteem, self-confidence and empathy with others. Much of this links with work on self-efficacy and motivatedlearning(Schunk 1987), leading to the self-regulation of learningand performance(Schunk and Zimmermann 1994). Modelling and attributionalfeedback are importanthere - perhaps peer tutoring can go some way towardscombating the dependency culture associated with superficiallearning.From a social psychological viewpoint, social isolation might be reduced,the functionalityof the subject modelled, and aspirationsraised,while combatingany excess of individualisticcompetition between students.Moust and Schmidt (1994a) found that studentsfelt peer tutorswere betterthanstaff tutorsat understandingtheirproblems,were more interested in their lives and personalities,and were less authoritarian, yet more focused on assessment. Economic advantages might include the possibility of teaching more studentsmore effectively, freeing staff time for otherpurposes.Politically, peer tutoringdelegates the managementof learning to the learners in a democraticway, seeks to empower students rather thande-skill them by dependencyon imitationof a masterculture,and might reduce studentdissatisfactionand unrest. Peer tutoringcan have disadvantages,however (Greenwood et al. 1990). Establishing it does consume organisationaltime in designing and effecting appropriatepeer selection and matching, and it may also necessitate some adaptationto curriculummaterials.Certainlythe requirementsfor training studentsin teaching and learningskills are greater,althoughit can be argued that peer tutoringmerely serves to bringto the surface needs that traditional teaching tends to overlook. All these may involve increased costs in the short term, with a view to reducedcosts and/or greatereffectiveness in the medium and long term. The quality of tutoringfrom a peer tutor may be a good deal inferior to that from a professionalteacher (although this should not be assumed), and the need for monitoringand quality control cannot be overstated.This also significantlyconsumes time and resources. Likewise, the tutor'smasteryof the contentof tutoringis likely to be less than that of a professionalteacher,so curriculumcontentcoverage in peer tutoringmay be much morevariable.Forthese reasons,projectco-ordinatorsmay experiment 326 andpractice,ratherthanthefirst initiallywithpeertutoringforconsolidation learningof new material,utilisingit on a smallscalewithsuitabletopics. Researchon peer tutoringin schools A recentreview(Topping1992) identified28 previousreviewsand metaanalysesof researchon peertutoring,mostlyin schools.SharpleyandSharpley (1981) conducteda meta-analysisof 82 studiesin schools, reporting substantialcognitive gains for both tutees and tutors.Same-agetutoring appearedas effectiveas cross-agetutoring,andtrainingof tutorssignificantly improvedeventualoutcomes.Cohen,KulikandKulik(1982)discovered500 titlesrelatingto tutoring.In 65 studieswithcontrolgroups,tutoredstudents controlsin 45. Therewas againevidencethattutortraining out-performed effect.Highlystructured tutoringwas producedlargersizes of experimental that also associatedwithlargereffectsizes.Therewasevidence peertutoring improvedtuteeattitudesin class,as well as tuteeself-concept.In 38 control tutorsout-performed controlsin groupstudiesmeasuringtutorachievement, 33. Improvedtutorattitudesandself-conceptwerealso reported. evidencethatpeertutoringis effectivein schools. Thereis thussubstantial relative cost-effectiveness this, mayalsobe considered.Levin,Glass Beyond and Meister(1987) conducteda cost-effectivenessanalysisof four differin primary ent interventions designedto improvereadingandmathematics schools(elementaryschools)in theUSA:computerassistedlearning,reducing class size, lengtheningthe schoolday,andcross-agepeertutoring.The most cost-effectiveintervention(peertutoring)was four times morecosteffectivethantheleast.Theleastcost-effectivewasreducingclasssize.While evidenceconcerningpeertutoringin schoolscancertainlynot be automatifood callygeneralisedintohigherandfurthereducation,thereis considerable for thoughtin thesefindings. Peer tutoringin higher education- previousreviews Previousreviewsand surveysof peertutoringin higherandfurthereducation includethoseof GoldschmidandGoldschmid(1976),Cornwall(1979), Whitman(1988), Lee (1988),Lawson(1989), Maxwell(1990) andMooreWest,Hennessy,Meilman,andO'Donnell(1990).All of theseareinteresting, butthe earlierpaperswerecompletedat a timewhenmostof the literature wasdescriptivein nature.TheGoldschmids'ownempiricalwork(1976)was well beforeits time in this respect.Cornwall(1979)offereda wideranging andproblemsolving. overviewof the field,includingadviceon organisation 327 In a surveyof 93 colleges,Lee (1988)madea comparative analysisof seven differentkindsof programmes targetedon increasingretentionandreducing studentdropout.Programmes involvingpeersas resourcesshowedupparticwerenot moreeffectivethan most well. The ularly expensiveprogrammes cheaperonesandsize of institutionwasnota factorin retentionanddropout rates.Peertutoringandpeercounsellingbothshowedgoodcost-effectiveness, Lawson remedialprogrammes whiletraditional provedverycost-ineffective. 19 in identified as (1989) surveyed collegesanduniversities Canada having comto be more was found Peer assisted learning programmes. tutoring peer of goals,selection,training, monthanpeercounselling.Detaileddescriptions aregiven,butlittlehard logisticsandmethodsforevaluationof programmes Peerassistedlearneffectivenessandcost-effectiveness. dataon comparative medical schools were in United States surveyed by Mooreing programmes Westet al. (1990).Of 127collegesin anassociation,62 replied,andof these and while40 had'advisingprogrammes' 47 hadpeertutoringprogrammes, 13 had 'peerassessmentprogrammes'. Cross-year small-group tutoring In this review of the more recent substantiveliteratureon differentforms of peer tutoring,the format most like surrogateprofessional teaching will be consideredfirst.This is whereupperyear undergraduates(or post-graduates) act as tutors to lower year undergraduates,each tutor dealing with a small group of tutees simultaneously.The literaturesearch revealed 18 studies of note (Bobko 1984, MeredithandSchmitz 1986, Cone 1988, Moust, De Voider and Nuy 1989, Button, Sims and White 1990, House and Wohlt 1990, Lidren, Meier and Brigham 1991, Longuevanand Shoemaker 1991, Moust and Schmidt 1992, 1994b, Johansen,Martensonand Bircher 1992, Ameman and Prosser 1993, Johnston1993, AmericanRiverCollege 1993, McDonnell 1994, Moody and McCrae 1994, Mallatrat1994 and Schmidt, Arend, Kokx and Boon 1994.) Many of these gatheredonly subjectivefeedback outcome data. Of eleven studies doing this, nine reportedvery positiye outcomes, one noted outcomes as good as those from teaching by professionalfaculty, and one reportedless good outcomes than for professional faculty. Three studies reportedreduceddropoutin association with such tutoring.Five studies reportedimprovedacademicachievement,anotherfour reportedacademic achievement as good as that from professional teaching and one reported achievementslightlybutsignificantlyworsethanthat.Muchof the research is not of the highest quality, but good quality studies (e.g. Lidren 1991 and American River College 1993) do clearly demonstrateimprovedacademic achievement. 328 InBobko's(1984)study,thepeertutorshadgroupsof 25 tuteesfor12hours overprevious perweek.Coursegradesdidnotshowa significantimprovement have been not with Interviews but comparable. years, previousgroupsmay tuteesyieldedmanyreportsof increasedconfidenceandless anxiety,while in theirknowledgeandabilityto communicate. tutorsreportedimprovements Meredithand Schmitz(1986) reporteda studyinvolvingmanysubjective ratings,andalthoughsomefavouredpeertutoringcomparedto facultytutoring, othersindicatedthe opposite,and a greatmanywere not significantly different.A mixedmethodprojectreportedby Cone(1988)involvedrotating recitationandtestingbetweensame-yearpeerswithcoachingandtestingby cross-yearpeerteachingassistants.Tutoringobjectivesand materialswere Outcomesontestweremarkedly higherthannormalexpechighlystructured. tations,butthe lackof propercontrolgroupsandtheabsenceof information aboutassignationto groupslimitsthe conclusionsthatmightbe drawn. A comparativestudyby Moustet al. (1989)in law includedprocessmeasures which indicatedthat studenttutorbehaviourswere very similarto those of professionalfaculty.Nevertheless,on outcometest scoresthe faculty tutoredstudentsscoredhigherthanthose tutoredby peers.Buttonet al. (1990) reportedcross-yeartutoring(whichtheytermed'proctoring')in mechanicalengineeringandcomputingin relationto specificdesignprojects. The subjectivefeedbackfromthevastmajorityof tutorsandtuteeswas very outcomeson Grade positive.HouseandWohlt(1990)comparedachievement PointAveragesfor peertutoredandnon-tutored students.Malepeertutored butfemaletuteesdid not. studentsachievedhigherGPA'sthannon-tutored, Thesubjectswereself selectedintogroupsandtheoutcomemeasurewasvery effects.Student generalandprobablyinsensitiveto smallscaleintervention A also better Lidren et al. (1991) used qualitystudyby drop-out improved. randomizedcontrolgroupsandcomparedoutcomesfor peertutoredgroups of six with groupsof twenty.Both groupsperformedbetteracademically in termsof examinationresultsand positivesubjectivefeedbackthannontutoredstudents.The smallerpeer tutoredgroupsyieldedbetteroutcomes thanthe largerones. LonguevanandShoemaker(1991)deployedupperyearstudentsandclerical staffas volunteertutors.Thetutorswererequiredto attendthesamelecturesas thetuteespriorto givingtutorialassistance.Thistutoringprogramme in the institutionparchargeda fee to tuteesand 10-15%of undergraduates ticipated.Therewas someevidencethatlargeramountsof tutoringresulted in higherGradePointAverages,althoughthe size of differencewas small and its significancenot easy to establish.Johansenet al. (1992) reported subjectivefeedback,with tuteesmostlysatisfiedbut tutorsratheranxious. AmemanandProsser(1993) studiedpeertutoringin dentistryin Australia. 329 Subjectivefeedback indicatedconfidencegains in tutorsand tutees.Johnston (1993) deployed traineeteachersas tutorsfor economics studentsin 'microlearning groups' of four. Although subjective feedback was very positive, the examinationand test resultsof participantsand non-participantswere not very different. American River College (1993) deployed twenty-four paid 'learning assistants' for three hours per week with groupsof two to six tutees. Tutees' subjective feedback was very positive, and tutors felt their own knowledge of their subject improved.Most strikinglyhowever,althoughtutees had lower general Grade Point Averages than non-tutoredstudents, they scored as well or betterthan them in tutoredsubjects. In the areaof computerscience, McDonnell (1994) researchedtutoringby thirdyear studentsof small groups of up to four second year students,and reportedvery positive subjectivefeedback.Moody and McCrae(1994) reportedon cross-yeartutoringin groupsof six to fourteenin law. Subjectivefeedbackfromtutorswas positive.Mallatratt (1994) targetedreduceddrop out rate for a peer tutoringproject in computing. Half the studentsutilisedthe scheme, a quarterregularly.Tuteesreported findingthe experiencesupportiveand achievedimprovedgradescomparedto previous cohorts of students.Seven studentsreportedthat peer tutoringhad been the critical factorin preventingthem from leaving the course, and other subjectivefeedback was positive. Moust and Schmidt (1992, 1994b) found studenttutoredand staff tutored groups gained equally in achievementduring an eight-week problem-based law course. Schmidt et al. (1994) compared the achievement of 334 peer tutoredand 400 faculty tutoredgroups in a problem-basedhealth sciences course. Overall, the latterachieved slightly but significantlybetter,but peer tutoringwas equally beneficial in the firstyear of the course. The Personalised System of Instruction Fred Keller is credited with the 'invention' of the Personalised System of Instruction,which is also called the 'Keller system'. In 1968 he described the procedure,which is based upon programmedlearningmaterial,through which each studentproceedsattheirown pace with the goal of masteringeach step. The peer tutor'sinvolvementis largely as a checker,testerand recorder, to ensure tutee mastery.In 1977 Robin and Heselton comparedtrainingPSI tutors interactively with training by a written handbook only. The direct trainingproducedhigherqualitytutoringbehaviour,butno differencein tutee outcomes. Davis (1978) discussed the components of the tutoring role in PSI, and queriedwhetherthe tutorsbenefitedmore thanthe tutees. The most substantialreview of the effectiveness of PSI was producedby Kulik, Kulik 330 75 controlledstudies.Of 61 studies andCohen(1979), who meta-analysed evaluatingin termsof class marks,48 foundPSIto give superiorresults.Of in the targetgroup,18found 20 studiesscrutinisingvariationin achievement PSIwasassociatedwithreducedvariability. final Sixtyonestudiesconsidered and57 of thesefoundPSItuteessuperior. examinationperformance Elevenstudiesalsoconsideredstudentsubjectiveratingof teachingquality, ten of these findingthatPSI studentsgave morefavourableratings.Eight studiesalso measureddelayedretentionof the materiallearnt,andall found PSI studentssuperior.PSI was foundto be effectiveacrossthe wholeability range.It raisedthe finalexaminationscoreof a typicalstudentin a typical class fromthe fiftiethto the seventiethpercentile.Effectswereeven more andthesedifferencesweremorepronounced strikingon delayedexamination on essay than on multiplechoice examinations.PSI effects were evident in studieswith bothgood andless goodresearchdesigns.Despitethis very convincingevidence,Sherman(1992)notedthatPSIusereacheda plateauand speculatedthatcomputeraidedlearningmaybe currentlymorefashionable to teachers. becauseit is less threatening SupplementalInstruction Instruction aimsto reduce Anotherwell known'brandname',Supplemental risk courses rather than rate and highriskstuusuallytargetshigh drop-out dents.It is often used in courseswithnew anddifficultcontent,a predominanceof lecturesandlow ratesof interactive teaching,andwhereassessment It operateson a cross-agebasiswith andmonitoringarerelativelyinfrequent. attheUniversityof Misone 'leader'workingwithseveraltutees.Originated souriat KansasCity (UMKC)in 1975, it has come to be offeredto almost Over300 institutions have halfof thefirstyearstudentsin its hostinstitution. beentrainedto use SI in the USA andmorethan15 institutionsnowuse SI or somevariantthereofin the UK.Leadersaretrainedto 'model,adviseand content.Theyhavealways facilitate'ratherthandirectlyaddresscurriculum as the same course the tutee,andusuallyagainattend previouslycompleted thetutees'lectures. MartinandArendale(1990)reporta controlledstudyof SI at UMKC.The drop-outrate halved,the averagecoursegradewas 0.5 to 1.0 higherand graduationoutcomeswere 12.4%higher.The NationalCentrefor SupplementalInstruction (1994) reviewedevidencefortheeffectivenessof SI from UMKCandotheruniversitiesin theUSA.In UMKCdatafrom14 successive academicyears,involving295 coursesand 11,855SI participants, indicated statisticallysignificantdifferencesin gradesfor participants comparedto initial even when academic wascon(pre-SI) performance non-participants, 331 trolled.Therewas widespreadevidenceof effectivenessacrossthe whole was also associatedwithhigherre-enrolment abilityrange.SI participation rates.Similardatawerereportedfrom 146 otherinstiratesandgraduation (see also Martin, tutions,involving2875coursesand298,629SI participants BlancandDeBuhr1983,MartinandArendale1992). coursKenneyandKallison(1994)reporttwostudiesof SI inMathematics and One es, usingcomparable participant non-participant groups. studyfound significantdifferencesfavouringthe SI group,theotherfoundno difference. In bothstudiestherewasevidenceof low abilitystudentsresponding disproa and used to well SI. (1992) regression Bridgham Scarborough portionately modelto predictmedicalstudents'expectedfinaloutcomesfromtheirentry for SI level, findinga subsequentstatisticallysignificant'over-achievement' one half of a between one third SI effect and size was Average participants. deviationin finaltestscores. standard Researchin the UK was reportedby Rye, WallaceandBidgood(1993), Wallace(1993), Rust(1993), Rustand Wallace(1994), Healy(1994) and Bidgood(1994). Wallace(1993) reportedthat levels of attendanceat SI sessions werecorrelatedwith finalcoursemarks.However,furtherdetails werelacking.Rust(1993) reportedthatthe courseworkmarksof SI tutees wereon average5%higherif theyhadattended2 or moresessions,although was the SI tutees were far frombeing modelstudents.This improvement modestandagaindetailswerelacking. inannualexamination results performance Healy(1994)reportedimproved of SI studentsas well as reductionsin dropoutrates,coupledwithenhanced of the skillsanda deeperunderstanding andothertransferable communication as area in the of curriculum the However, (engineering). question principles no control was and were self doubtful, selected, group comparability groups wasused.Healy(1994)notedtheneedforlongertermfollowupof SIeffects. and thatend-of-yearcoursework Morepersuasively, Bidgood(1994)reported examinationmarksin two successiveyearsof a computersciencecourseat KingstonUniversitywerestatisticallysignificantlybetterfor SI participants andstart-of-year withequivalent thanfornon-participants entryqualifications marks.SI studentsdidnotfigurein failureor resitlists. It has been claimedthatSI in the UK has also demonstrated improved as well as gainsin self gradesfor SI leaderscomparedto non-participants, confidenceand communication skills, but detailsof the data are difficult to find. In the USA SI leadersare usuallypaid,whereasthis is muchless frequent in the UK. A relateddevelopment is the establishmentof faculty- widecross-yearsmall-group'StudentSupported Learning',withmanyof the featuresof SI butmuchmorefocuson gainsforthetutors,whoareunpaidbut 332 transferschemefortheirparticipation receivecreditsin a courseaccreditation Hill and 1996). Thompson, Simpson, (Topping, Same-yeardyadicfixed-roletutoring Moreinnovative(andperhapseasierto organise)is tutoringbetweenpairs (dyads)in the sameyearof study,i.e. at the samepointin the course,where Sevenstudies,someof conone memberretainsthe roleof tutorthroughout. siderableage,havefocusedon achievement gainsresultingfromthispractice. The classic studiesby Annis(1983) andBenwareandDeci (1984) referred to earlierwereexamplesof this format.Rosen,PowellandSchubot(1977) workedwithsamegenderpairsin whichthe tutorswereeithermore,less or roleswere equallycompetentthanthetutees.Also,forhalfof theparticipants, halfwaythroughtheproject.Subjectsreceivedonly 20 minutes reciprocated of trainingand48 out 90 pairsdid not supplyfull data.Outcomemeasures included20 item pre- andpost-testsand satisfactionquestionnaires. There was some evidencethe changingrole from tutee to tutorwas associated withan improvement in achievement. Therewasalsoan indicationthatpairof someone or with greater equalabilitywas associatedwith a greater ing achievement. FremouwandFeindler(1978) studiedthe effectivenessof dyadicsameyear tutoringin contrastwith thatof tutorialsin groupsof nine led by a professionalfacultymember.The peer tutorswere given some additional contenttraining.Two controlgroupswere used, one given equalattention of a differentsortandanothera non-participant waitinglist group.Thepeer tutoredgroupachievedoutcomesas goodas theprofessionally tutoredgroup. A studyin Esperantoteachingwasreportedby McKellar(1986).Tutorswere trainedin newmaterialandstudyguideswereprovidedto supportthetutoring. High accountabilitywas inbuilt,since post-testtutorandtuteescoreswere combinedas a performanceindicator.The researchers foundthatthe more tutorsgave information,the higherwas the tutorscoreandcombinedtutor andtutee score.High scoreswerealso associatedwith the tuteeaskingfor clarificationandaskingfor the mainpointsto recall.However,wheretutors gave wronginformation,this was associatedwith reducedscoresfor both tutorand tutee. The tutorsimply askingif the tuteeunderstoodwas also associatedwithpoorerscores. Two studiesin Edinburghare reportedby Falchikov(1990). One study allocatedparticipants andstudyaloneconditions,but randomlyto tutor/tutee foundno significantdifferencesin achievementbetweenthese conditions. Althoughsome tutorsreportedsubjectiveperceptionsthatthey hadgained morefromtutoringthantheywouldhavedonefromindependent study,some 333 tutees reported lacking confidence in their tutors. As in the Rosen (1977) study, it appearsthat randomallocationcan create its own problems. In the second study,following tutoringsome participantsbecametutees again while some became tutors.Althoughthere was less global satisfactionat role repetition, some tutees expressedmoreconfidence in theirtutors.No significant differences in achievement were found as a function of role repetition or non-repetition,but attritionat post-testwas high. In summary,most of the studiesof dyadicsame-yearfixed-rolepeer tutoring have not comparedthe procedureto an alternativeprocedure,but considered organisationalvariationswithin the procedureand their relationshipto outcomes. However,one study(FremouwandFeindler 1978) showed this format of peer tutoringto be as effective as small group tutoringby a professional, two studies that it was more effective than independentstudy, but one study found no difference. The literaturedemonstratesthe side-effects of random allocation to conditions and the potential problem of 'the blind leading the blind.' Same-yeardyadicreciprocalpeer tutoring Although this format might be consideredeven more innovativethan sameyear dyadic fixed-role tutoring,the first relevant study dates back to 1976. Although there is relatively little work in the area, some is of high quality. Goldschmid and Goldschmid(1976) used dyadic reciprocalpeer tutoringin an undergraduatepsychology course of 250 students. They compared outcomes for three groups:one involvedin a seminarwith faculty,one pursuing independentstudy,and the thirdinvolved in peer tutoring.The peer tutoring groupdid the best of the threeon an unexpectedpost-testand they ratedtheir learningexperiences more positively. More recently,John Fantuzzoand his colleagues have reporteda series of high quality studies of reciprocalpeer tutoring(RPT), consistently showing that it results in greater achievement, greater satisfaction and less feeling of stress in comparison to other treatmentand control groups. Fantuzzo, Dimeff and Fox (1989) allocated psychology students randomly to three conditions: reciprocalpeer tutoring,questioning only, and placebo control. The RPT group reciprocatedroles within each session, creating tests for each other before the session, administeringthem to each other, scoring them, discussing the outcome and coaching their partneras necessary. The questions only group created the tests alone but never administeredthem they studiedto give the test. Thisgroupalso saw the questionsgeneratedby the RPT pairs. In the placebo condition, studentsmet and watched instructional videos with the same curricularcontent and answered the questions on the 334 videos.Onexamination scores,allthreegroupsgained,buttheRPTgroupdid significantlybetterthanthe othertwo groups,whichwerenot significantly differentfromeachother.Studentsatisfactionwassignificantly improvedand distressindicatorssignificantlyreducedfortheRPTbutnottheothergroups. Subsequently,Fantuzzo,Riggio,ConnellyandDimeff (1989) conducted a componentanalysisto attemptto determinewhatelementsof RPTwere implicatedin its effectiveness.One hundredandtwentyfive studentswere allocatedto five conditions:a dyadicpeertutoringgroupwith a structured contactgroupinvolvinggeneral interactionprocess,a dyadicunstructured discussionrelatedto upcomingexam topics, an independentunstructured conditionin which individualshad to submita shortessay on up-coming structuredlearningconditionsimilarto examinationtopics,an independent the 'questionsonly condition'in the previousstudy,and a no treatment foundthatdyadicinteractionwas associated controlgroup.The researchers withgainsin achievementon pre-posttests,anda higherdegreeof structure was also associatedwith betteroutcomes.They also foundthatstructured methodswere associatedwith betterscores on studentstress inventories. Theirconclusionwas thatit was not merelypairingbutstructured exchange whichwaseffective. Riggio, Fantuzzo,Connellyand Dimeff (1991) soughtto replicatethe studybut with morediversestudentsin a differentsetting.The RPTgroup showedsignificantlyhigherachievementscoresthanthe othergroups,and therewas generallya significantmaineffect for dyadicconditions,butnot for structure.However,structuredid yield betterscoreson two out of three stressinventories.Satisfactionratingsfor the RPTgroupweresignificantly higherthanthoseof the othergroups.Thuscomparedto thepreviousstudy, factorsless impact dyadicfactorsshowedless impacton stressandstructure on achievement.Riggio et al. (1991) note that the subjectswere from a 'commuter'college whowerenotalreadywell socialisedwitheachother. In the UK, all 45 studentsin a year-longundergraduate calculusclass were involvedin same-yeardyadicpeer tutoring(Topping,Watson,Jarvis andHill 1996),the 12 one-hourpeersessionssubstituting fortraditional lectures.Degreeexamination resultsin calculusweresignificantly betterforthe experimentalgroupthanfor the previous(comparison)year,especiallyfor studentswhowerenotmathsmajors,buttheyearcohortswerenon-equivalent in some respects.Structured subjectivefeedbackfromthe studentssuggested thatpeer tutoringhad improvedtheirtransferable skills in a numberof areas.Similarly,a projectwith 125 undergraduates in a year-longclass in mathematical economicswas reportedby Topping,Hill, McKaig,Rogers, RushiandYoung(1996).Finaldegreeassessmentresultsfortheexperimental groupwere in generalnot statisticallysignificantlydifferentfromthose 335 of the previous (comparison)year. However, subjective feedback from the students indicated that peer tutoringhad improved their transferableskills in a numberof areas. Furthermore,studentswho regularlyattendedthe peer tutoring sessions obtained significantlybetterdegree assessment outcomes, and gave significantly better feedback about improved transferableskills, thanthose who did not. Additionally,studentdrop-outrateswere lower in the experimentalthan in the comparisonyear. Dyadiccross-yearfixed-rolepeer tutoring This formatis reportedin fourstudies,threefromAustralia.Schaffer,Wile and Griggs (1990) analysed the exam results of a cohort of students, some of whom had participatedin a peer tutoringprogramme.There was a positive relationshipbetween degree of participationin tutoringand examination results. However, no control groups were used and no demonstration of causality is evident. A study by Black (1993) focuses on ethnic minority group tutees in nursing and midwifery, and claims 'higher than expected' pass rates, but lacks sufficientdetail to enable this to be verified.Loh (1993) deployed paid peer tutorsin a course for Anatomyfor Nurses with a previous high failure rate. Subsequentlythe peer tutoringparticipantfailure rate was less than the non-participantrate,but no informationwas given aboutassignment to groups. Subjective feedback was positive however,tutees reporting feeling more confident.Quintrelland Westwood(1994) pairednewly arrived internationalstudents with host national students,expecting twice monthly contact during the year. Tutees showed more positive attitudesthan a comparison group matched for course of enrolment,but not significantlybetter academic performance.Manyof these studiesappearto sufferfromproblems of self-selection to groupsand consequentnon-comparability. Same-year group tutoring Fourstudies have consideredsame-yeargrouptutoring,often in the formatof rotatingpresentationsby individualstudentsto the peergroup.Unfortunately, only one of these reportedachievementoutcomes. Autonomousstudentstudy groups were establishedby Beach (1960), who measuredachievementgains with pre- and post-tests. Results indicatedthat extrovertsdid better in peer tutoringthan did introverts,the introvertsgaining equally in traditionallectures.The study raisedquestionsregardinginteractionsbetween teachingand learning methodologies and studentpersonalityor learning style. Fineman (1981) reportedon rotationalpresentationsto the peer group by members 336 behaviour.Peerassessment of a groupof twelve studentsof organisational criteriawas included.Thesubjectiveevaluationby the on peerbrainstormed was participants positive. Similarly,Hendelmanand Boss (1986) found rotatingpresentationsto groupsto yield positivesubjectivefeedbackfromthe students.The tutees reportedthatpeertutoringwas as effectiveas facultytutoring,andthetutors thatpeertutoringwasmoreeffectivethanfacultytutoring.A coursein ComputerAided EngineeringDesign was the focus of a studyby Maginand Churches(1993), occasionedin partby a lack of sufficientaccess to hardware.Thosestudentswhohadhadaccessto machinestutoredthosewhohad not hadsuchaccess,overa fourweekperiod.Subjectivefeedbackindicated the tuteesfoundthetutoringas or moreeffectivethantutoringby faculty. Peer assistedwriting Withinthe traditionalhighereducationsystem,writtenoutputis oftenused as a vehiclefor assessmentof the individual,andcollaborativewritingcan be problematicto assess. However,in recentyearstherehas been greater interestin writingas a devicefor improvinglearningandthinking,coupled with the advocacyof 'writingacrossthe curriculum','writingcentres'and 'collaborative writing'(Olson1984,Gere1987).Rizzolo(1982)describedthe useof peertutorsin a writingcentre,alsostaffedbyEnglishfaculty.Thetutors It wasnotedthattutoringin writing werepaidandtrainedthroughinternship. Thetuteesratedtheirpeertutors hadto be morethanmerelyproof-reading. feedback. on therole Similarly,Bell (1983)emphasised veryhighly subjective of peertutorsin a writingcentrein promotingconfidenceandencouraging new studentsto view writingmoreas a processandless as a product.More substantialdatawereofferedby O'Donnell,Dansereau,Rocklin,Lambiotte, HytheckerandLarson(1985),whocomparedrandomlyassignedco-operative writingand writingalone conditions.The writingof the 36 studentswas assessedfor communicative quality.The co-operativewritersdid betteron to a furtherindividualwritingtask. the initialpost-testandon transference on Holladay(1989, 1990) reported the use of peer tutorsin a 'writing at MonroeCommunityCollege.Seventy acrossthe curriculum'programme six percentof tuteesfoundtheirtutorshelpfulor veryhelpful,facultyfeltthe classes,and qualityof papersimprovedin tutoredclassesversusnon-tutored all the tutorsfelt theirownwritinghadimprovedas a resultof tutoring.This programmecontinuedin subsequentyearswitheven betterresults.A study by Levine(1990) also yieldedverypositivesubjectivefeedback.Theexperimentalclass improvedin meetingdeadlinesandthefailureratereducedfrom 35%to 3%.However,gradesandexamresultswereverysimilarfor exper- 337 is unclear.Students imentalandcomparisongroups,althoughcomparability who hadtutoringin writingfromfacultyandpeerswerecomparedby Oley (1992) with thosewho hadtutoringfrompeersonly or facultyonly.Many had been identifiedas weakwriters,andsome received of the participants helpvoluntarilyandsomeon a compulsorybasis.Assignationto conditions was random.Thosewho receivedpeertutoringsubsequently attainedhigher did who not. those than grades Louthand MacAllister(1990) assignedfreshmancompositionstudents randomlyto threeconditions:somestudentswroteina traditional independent manner,otherswrote(partially)interactively althoughproducingindividual writtenproducts,whilea thirdgroupwrotewhollyinteractively producinga writinggroup,whichscoredhigherthanthe joint product.The independent othertwo groupsat pre-test,did not improveduringthe project,while both conditionsimprovedtheirperformance, collaborative althoughthe statistical significanceof this was debatable.The use of mixedabilitywritinggroups of fourstudentsin geographywasreportedby Hay(1993),who emphasised skillwhichis vocationallyvalued. of writingas a transferable theimportance their the students reviewed In groups, essay assignments,readeach others' writingandmadewrittenreviewsof eachothers'work,witha rotatingchair person.Haynotedthatit waspossibleto do thereadingactuallyin thegroup sessionsto avoidanypossibilityof plagiarism.Twogroupsgave subjective feedback:in one65%werepositiveandin theother80%.Problemsincluded criticalandthaterrorswerenotalwaysdetected. thatpeerswereinsufficiently the felt that writinggroupshouldcontinue.Theco-operative Ninetypercent writingdid not necessarilysavefacultytimeon marking,as monitoringthe groupprocessoccupiedsometime. In summary,of nine studieson peerassistedwriting,five give only subjectivefeedback,butthisis generallyverypositive.Fourstudiesgivedataon gainsin writingcompetenceandof these,twogoodqualitystudiesshowtutee gains,oneshowsno statisticallysignificantdifferenceanda thirdshowssome includeraiseddeadline tuteegainsof equivocalstatus.Otherimprovements attainment rates,reducedfailurerates,andself reportof improvedwritingin thetutors. Peer assisted distance learning In distancelearningfeedbackandsupportfromany peer groupis problematic. Attemptsto build this in by way of occasionalsummer-schoolsare little more than a token gesture, and the loneliness of the long-distance learner is a widespreadphenomenon.Distance learningis also fundamentallydifficultto research,and the quantityand qualityof evidence on the role of peer support 338 in this processlimited.Amundsenand Barnard(1989) workedwith bank One set metin employeesstudyingaccountingandbusinessadministration. had a second and also distance peer supportgroups peer supportgroups, while a both of third had these on a nominated and also learning studyskills, A fourthgroupwas mentorwho was a previousgraduateof the programme. a controlcondition.Outcomemeasuresincludedassignmentgrades,final examscores,finaldegreegradesandsubjectiveself-assessments. However, the studygroupswere formedinevitablyon a geographicbasis, and were thusself selectedandof doubtfulcomparability. the degreeof Furthermore, to intended was in the doubt and some conformity process subjectswere no from the excluded analysis.Virtually significantdifferenceswerefound betweenthe groups.However,the authorsareto be commendedfor a brave effortin a difficultarea. A programme as a partof continuingeducation foraudio-teleconferencing for nurseswasdevelopedin Australiaby Hart(1990).Thetopicsvariedfrom week to week andwere suggestedby the participants. Eachtele-conference involvedbetween6 and 12 nurses.Themajorityof participants werewomen andthe authordiscusseswhetherfemalesneedor seek groupsupportmore was reported,butthe thanmales.Subjectivefeedbackfromthe participants This does rate was 34%. a good discussion include only paper response of practicalproblemsinvolved.In summary,althoughthereis some weak in distance evidencethatbuildingin peercontactis likedby someparticipants evidence little there seems to be it that increases student satisfactory learning, achievement.However,furtherresearchin thisareais certainlyneeded. Summaryand conclusion Peer tutoringis alreadywidely used in furtherand highereducation,in a varietyof differentforms.Surveyssuggestseveralhundredinstitutions deploy this interactivemethodof teachingandlearning.Of course,the existenceof one smallpilot projectat one timein an institutiondoes not constitutepeer tutoringon a largescale acrossthe curriculumwhichis qualitycontrolled andembeddedwithintheorganizational culture.Of thedifferentformatsand andSupplemental Instruction methods,thePersonalisedSystemof Instruction thelatterscenario. havemostnearlyapproached A considerableamountis alreadyknownaboutthe effectivenessof peer tutoringin furtherand highereducation.Cross-yearsmall-grouptutoring, theformatleastdisparatefromtraditional methods,canworkwell. Studiesof achievementgainsalmostall indicateoutcomesasgoodas orbetterthangroup tutoringby faculty,andstudentsubjectivefeedbackis generallyverypositive. hasbeen widelyusedandevaluated The PersonalisedSystemof Instruction 339 in the US. Two thirds of studies found PSI involvement associated with higherclass marksand 93% of studiesfoundPSI associatedwith higherfinal examination performance,comparedto control groups. PSI also improved longer term retentionof the materiallearnt.SupplementalInstructionadopts a very differentmodel of operationand has become morepopularoutside the USA than PSI. There is very substantialand persuasiveevidence from the USA of impact on course grades, graduationoutcomes and drop-out rates. Research in the UK is improvingin quality and also demonstratingpositive outcomes. Same-yeardyadic fixed-roletutoringhas been the subjectof severalstudies over the years, researchof mixed quality yielding mixed results. However, two good qualitystudiesfound improvedachievementfromthis format,while three othersfound achievementthe same as with faculty teaching. Five out of 6 studies of same-yeardyadic reciprocaltutoringhave demonstratedincreasedattainment.Therewas also evidenceof reducedstudentstress and improvedtransferableskills. The degree of structurein the programme was positively relatedto outcomes. Dyadic cross-yearfixed-roletutoringhas been the subjectof threestudiesof poorquality.Same-yeargrouptutoringhas yielded positive subjective feedback in four studies, but no harderevidence on achievementoutcomes. Nine studies of peer assisted writing have shown generally favourable outcomes in termsof subjectivefeedback.Gains in writingcompetence were shown in two or threeof the four studies examiningthis, despite the inherent difficultyof this kind of research.There is little evidence thatpeer assistance in distance learning improves achievementoutcomes, but this area is even more difficultto research. In summary,three methodsof peer tutoringin furtherand highereducation have alreadybeen widely used, have been demonstratedto be effective, and merit wider use in practice- these are Cross-yearSmall-groupTutoring,the PersonalisedSystem of Instructionand SupplementalInstruction.Same-year dyadic reciprocaltutoringhas been demonstratedto be effective, buthas been little used, and merits much wider deployment.Same-yeardyadic fixed-role tutoringandpeerassisted writinghave shownconsiderablebutnot necessarily consistentpromiseand shouldbe the focus of continuingexperimentationand more researchof betterquality.In threeareasthereare barelythe beginnings of a satisfactory body of evaluationresearch:dyadic cross-year fixed-role tutoring,same-yeargrouptutoringand peer assisted distance learning. It is essential that subsequentresearchstrives to achieve adequatequality in design and execution, preferablyincluding control groups or comparison groupswhich are trulycomparable,and addressesissues of achievementgain and parametersof successful course completion as well as subjective par- 340 thatgo beyond ticipantfeedback.If achievementgainscan be demonstrated the narrowconfinesof the institutional assessmentsystemandendurein the longerterm,so muchthebetter.Thisimpliesthatimpactuponwidercognitive skills shouldalso be measured. abilitiesandtransferable However,peertutoringis usuallya relativelysmallcomponentof a wide rangeof teachingand learningstrategiesdeployedin highereducation,so theextentto whichit is realisticto expectassociatedgainsto be measurable, andgeneralisedis debatable. widespread,maintained Acknowledgement The supportof the ScottishHigherEducationFundingCouncilis gratefully acknowledged. References American River College (1993). A.R.C. Beacon Project: Student Catalyst Program - Peer Assisted Learning; First Semester SummaryReport. SacramentoCA: American River College (ED355995). Amundsen, C.L. and Bernard,R.M. (1989). 'Institutionalsupportfor peer contact in distance education:an empiricalinvestigation',Distance Education 10(1), 7-23. Annis, L.F. (1983). 'The processes andeffects of peer tutoring',HumanLearning2(1), 39-47. Arneman K. and Prosser M. (1993). 'The developmentof two peer tutoringprogrammesin the Faculty of Dentistry,University of Sydney', in Proceedings of Conferenceon Peer Tutoring at University of Auckland 19-21 August 1993. Auckland: Higher Education ResearchOffice and Universityof Auckland. Ashman, A.F. and Elkins, J. (1990). 'Co-operativelearningamong special students',in Foot, H.C., Morgan,M.J. and Shute, R.H. (eds.), ChildrenHelping Children.Londonand New York:John Wiley. Bargh,J.A. andSchul, Y.(1980). 'On the cognitivebenefitsof teaching',Journalof Educational Psychology 72(5), 593-604. Barnett,R. (1992). ImprovingHigher Education.Buckingham:Open UniversityPress. Beach, L. R. (1960). 'Sociability and academic achievement in various types of learning situations', Journal of EducationalPsychology 51(4), 208-212. Bell, E. (1983). 'The peer tutor:the writingcenter'smost valuableresource', TeachingEnglish in the Two-YearCollege 9(2), 141-144. Benware, C.A. and Deci, E.L. (1984). 'Quality of learning with an active versus passive motivationalset', AmericanEducationalResearchJournal21(4), 755-65. Bidgood, P. (1994). 'The success of supplementalinstruction:the statistical evidence', in Rust, C. and Wallace,J. (eds.), Helping Studentsto LearnfromEach Other:Supplemental Instruction.Birmingham:Staff and EducationalDevelopmentAssociation. Black, J. (1993). 'Peer tutor support in nursing and midwifery at Otago Polytechnic', in Proceedings of Conference on Peer Tutoringat Universityof Auckland 19-21 August 1993. Auckland:HigherEducationResearchOffice and Universityof Auckland. Bobko, E. (1984). 'The effective use of undergraduatesas tutorsfor college science students', Journal of College Science Teaching 14, 60-62. Bridgham, R.G. and Scarborough,S. (1992). 'Effects of supplementalinstructionin selected medical school science courses', AcademicMedicine 67(10), 569-571. 341 Brookfield, S. (1983). Adult Learners,Adult Education and the Community.Buckingham: Open UniversityPress. Button, B.L., Sims, R. and White, L. (1990). 'Experience of proctoringover three years at NottinghamPolytechnic',in Goodlad,S. andHirst,B. (eds.), Explorationsin Peer Tutoring. Oxford:Blackwell. Cohen. P.A., Kulik, J.A. and Kulik, C-L.C. (1982). 'Educationaloutcomes of tutoring: a meta-analysisof findings',AmericanEducationalResearchJournal 19(2), 237-48. Cone, A.L. (1988). 'Low tech/high touch criterion-basedlearning', Psychological Reports 63(1), 203-207. Cornwall,M.G. (1979). Studentsas Teachers:Peer Teachingin HigherEducation.Amsterdam: C.O.W.O.,Universityof Amsterdam. Davis, C.S. (1978). 'Peertutors:theirutility and trainingin the personalisedsystem of instruction', EducationalTechnology18, 23-26. Doise, W. and Mugny,G. (1984). TheSocial Developmentof the Intellect.Oxford:Pergamon Press. Durling, R. and Schick, C. (1976). 'Conceptattainmentby pairsand individualsas a function of vocalization', Journalof EducationalPsychology 68(1), 83-91. Ellis, R. (1993). Quality Assurancefor UniversityTeaching.Buckingham:Open University Press. Entwistle, N.J. and Ramsden,P. (1983). UnderstandingStudentLearning.Beckenham:Croom Helm. Entwistle, N. (1992). The Impactof Teachingand LearningOutcomesin Higher Education: A LiteratureReview. Sheffield:Universitiesand Colleges Staff DevelopmentUnit, CVCP. Falchikov, N. (1990). 'An experimentin same-age peer tutoringin higher education: some observationsconcerningthe repeatedexperienceof tutoringor being tutored',in Goodlad, S. and Hirst,B. (eds.), Explorationsin Peer Tutoring.Oxford:Blackwell. Fantuzzo, J.W., Dimeff, L.A. and Fox, S.L. (1989). 'Reciprocal peer tutoring:a multimodal assessmentof effectivenesswith college students',Teachingof Psychology 16(3), 133-135. Fantuzzo,J.W., Riggio, R.W.,Connelly, S. and Dimeff, L. (1989). 'Effects of reciprocalpeer tutoringon academic achievementand psychological adjustment:a componentanalysis', Journalof EducationalPsychology 81(2), 173-177. Fineman, S. (1981). 'Reflections on peer teaching and peer assessment: an undergraduate experience', Assessmentand Evaluationin Higher Education6(1), 82-93. Foot, H.C., Shute, R.H., Morgan, M.J. and Barron, A. (1990). 'Theoreticalissues in peer tutoring',in Foot, H.C., Morgan,M.J. and Shute, R.H. (eds.), ChildrenHelping Children. London and New York:JohnWiley. Forman,E. (1994). 'Peercollaborationas situatedactivity:examplesfromresearchon scientific problem solving', in Foot, H.C., Howe, C.J., Anderson, A., Tolmie, A.K. and Warden, D.A. (eds.), Group and InteractiveLearning. Southamptonand Boston: Computational Mechanics. Freemouw,W.J. and Feindler,E.L. (1978). 'Peer versus professionalmodels for study skills training',Journal of CounselingPsychology 25(6), 576-580. Gartner,S., Kohler,M. and Riessman(1971). ChildrenTeachChildren:Learningby Teaching. New York:Harperand Row. Gere, A.R. (1987). WritingGroups:History,Theoryand Implications.CarbondaleIL:Southern Illinois UniversityPress. Gibbs, G. (1981). TeachingStudentsToLearn. Buckingham:Open UniversityPress. Goldschmid, B. and Goldschmid,M.L. (1976). 'Peer teaching in highereducation:a review', Higher Education5, 9-33. Greenwood,C.R., Carta,J.J. and Kamps,D. (1990). 'Teacher-mediatedversus peer-mediated instruction:a reviewof educationaladvantagesanddisadvantages',in Foot, H.C., Morgan, M.J. and Shute, R.H. (eds.), ChildrenHelping Children.London and New York: John Wiley. 342 in continuing education Hart,G. (1990).'Peerlearningandsupportviaaudio-teleconferencing for nurses', Distance Education 11(2), 308-319. H.J.(1990). 'Factorsaffectingthetutoringprocess',Journalof Educational DevelHartman, opment14(2),2-6. Hay, I. (1993). 'Writinggroupsin geography',in Proceedingsof Conferenceon Peer Tutoring at University of Auckland 19-21 August 1993. Auckland: Higher Education Research OfficeandUniversityof Auckland. instruction: a modelfor supportingstudentlearning',in Healy,C.E. (1994). 'Supplemental Foot,H.C.,Howe,C.J.,Anderson,A., Tolmie,A.K.andWarden,D.A. (eds.),Groupand andBoston:Computational Interactive Mechanics. Learning.Southampton W.J.andBoss, M. (1986). 'Reciprocal Hendelman, peerteachingby medicalstudentsin the gross anatomy laboratory',Journal of Medical Education61(8), 674-80. Holladay, J. (1989). Monroe County CommunityCollege WritingAcross the Curriculum: AnnualReport1988-9. Michigan:MonroeCountyCommunity College(ED310820). Holladay, J. M. (1990). WritingAcross the Curriculum:Annual Report 1989-90. Michigan: MonroeCountyCommunity College(ED326260). on the perHouse,J.D. andWohlt,V. (1990). 'Theeffectof tutoringprogramparticipation formanceof academicallyunderprepared collegefreshmen',Journalof CollegeStudent Development31, 365-370. D.F.andBircher,J. (1992).'Studentsas tutorsin problem-based Johansen,M.L.,Martenson, learning:does it work?',MedicalEducation26(2), 163-165. Johnston,C. (1993). 'The integrationof traineeteachersin an undergraduate peertutoring projectat the Universityof Melbourne',in Proceedingsof Conferenceon Peer Tutoringat Universityof Auckland19-21 August 1993. Auckland:HigherEducationResearchOffice andUniversityof Auckland. Keller, F.S. (1968). 'Goodbye, teacher.. .', Journalof AppliedBehaviorAnalysis 1(1), 79-89. studiesontheeffectivenessof supplemental Kenney,P.A.andKallison,J.M.(1994).'Research instruction in mathematics', New Directions for Teaching and Learning 60(4), 75-82 Instruction). (specialissueon Supplemental of outcomestudiesof Kulik,J.A., Kulik,C.C. and Cohen,P.A. (1979). 'A meta-analysis AmericanPsychologist34(4),307-318. Keller'spersonalised systemof instruction', D. (1993).Rethinking Laurillard, University Teaching. Buckingham: OpenUniversityPress. Lawson,D. (1989). 'Peerhelpingprogramsin the collegesanduniversitiesof Quebecand Ontario', Canadian Journalof Counselling 23(1), 41-56. Lee, R.E. (1988). 'Assessingretentionprogramholdingpowereffectivenessacrosssmaller community colleges', Journalof College StudentDevelopment29(3), 255-262. Levine,J.R.(1990).'Usinga peertutorto improvewritingin a psychologyclass:one instructor's experience', Teachingof Psychology 17(1), 57-58. Levine,H.M.,Glass,G.V.andMeister,G.R.(1987).'Acost-effectiveness analysisof computerEvaluationReview11(1),50-72. assistedinstruction', Lidren,D.M.,Meier,S.E. andBrigham,T.A. (1991). 'Theeffectsof minimalandmaximal of collegestudents',Psychological peertutoringsystemson the academicperformance Record41(1), 69-77. Loh,H. (1993).'Peerassistedstudysessionsin anatomyfornursingstudents',in Proceedings of Conferenceon Peer Tutoringat Universityof Auckland19-21 August 1883. Auckland: HigherFlication ResearchOfficeandUniversityof Auckland. J.(1991).'Usingmultipleregression C. andShoemaker, toevaluatea peertutoring Longuevan, programfor undergraduates',Paper Presented at the Annual Meeting of the California EducationalResearch Association, San Diego CA, November 14-15 (ED341717). Louth,R., McAllister,C. andMcAllister,H.A.(1993). 'Theeffectsof collaborative writing Fducation61(3), techniqueson freshmanwritingandattitudes',Journalof Experimental 215-224. Lovell,R.B. (1980).AdultLearning.London:Routledge. 343 McDonnell, J.T. (1994). 'Peer tutoring:a pilot scheme among computerscience undergraduates', Mentoringand Tutoring2(2), 3-10. McKellar,N.A. (1986). 'Behaviorsused in peer tutoring',Journal of ExperimentalEducation 54(3), 163-167. Magin, D. and Churches,A. (1993). 'Studentproctoring:who learns what?', Proceedings of Conference on Peer Tutoringat Universityof Auckland19-21 August 1993. Auckland: Higher EducationResearchOffice and Universityof Auckland. Mallatrat,J. (1994). 'Learningaboutthe learners- the impact of a peer tutoringscheme', in Foot, H.C., Howe, C.J., Anderson,A., Tolmie, A.K. and Warden,D.A. (eds.), Groupand InteractiveLearning.Southamptonand Boston: ComputationalMechanics. Martin,D.C. and Arendale,D.R. (1990). SupplementalInstruction:ImprovingStudentPerformance, IncreasingStudentPersistence.KansasCity MO: Universityof Missouri. Martin,D.C. and Arendale,D.R. (1992). 'Supplementalinstruction:improvingfirst-yearstudent success in high-riskcourses', The FreshmanYearExperience:MonographSeries No. 7. ColumbiaSC: South CarolinaUniversity(ED354839). Martin,D.C., Blanc, R.A. and DeBuhr,L. (1983). 'Breakingthe attritioncycle: the effects of supplementalinstructionon undergraduateperformanceand attrition',Journal of Higher Education54(1), 80-89. Marton,E, Hounsell, D. and Entwistle, N. (1984). The Experienceof Learning. Edinburgh: Scottish Academic Press. Maxwell, M. (1990). 'Does tutoringhelp? a look at the literature',Review of Research in DevelopmentalEducation7(4), 3-7. Meredith, G.M. and Schmitz, E.D. (1986). 'Student-taughtand faculty-taughtseminars in undergraduateeducation:anotherlook', Perceptualand Motor Skills 62(2), 593-594. Merriam,R. and Caffarella,R.S. (1991). Learningin Adulthood.San Francisco:Jossey-Bass. Moody, S. and McCrae, J. (1994). 'Cross year peer tutoring with law undergraduates',in Foot, H.C., Howe, C.J., Anderson,A., Tolmie, A.K. and Warden,D.A. (eds.), Groupand InteractiveLearning. Southamptonand Boston: ComputationalMechanics. Moore-West,M., Hennessy, A., Meilman,P.W.and O'Donnell, J.F. (1990). 'The presence of student-basedpeer advising, peer tutoringand performanceevaluationprogramsamong U.S. medical schools', AcademicMedicine65(10), 660-661. Moust, J.H.C., De Volder,M.L., andNuy, H.J.P.(1989).'Peerteachingand higherlevel cognitive learningoutcomes in problem-basedlearning',Higher Education 18(6), 737-742. studentsas tutors:aretheyas effective as Moust,J.C. andSchmidt,H.G. (1992). 'Undergraduate facultyin conductingsmall-grouptutorials?',Paperpresentedat theAmericanEducational Research Association Symposiumon RewardingTeachingat Research Universities, San FranciscoCA, April 23 (ED 346774). Moust, J.H.C.and Schmidt,H.G. (1994a). 'Facilitatingsmall-grouplearning:a comparisonof studentand staff tutors' behavior',InstructionalScience 22, 287-301. Moust, J.H.C. and Schmidt, H.G. (1994b). 'Effects of staff and student tutors on student achievement',Higher Education28, 471-482. National Center for SupplementalInstruction(1994). Review of Research Concerning the Effectivenessof SI. KansasCity MO: NCSI, Universityof Missouriat KansasCity. O'Donnell, A.M., Dansereau,D.F., Rocklin,T., Lambiotte,J.G., Hythecker,V.I. and Larson, C.O. (1985). 'Co-operativewriting:direct effects and transfer', WrittenCommunication 2(3), 307-315. Oley, N. (1992). 'Extracreditandpeertutoring:impacton the qualityof writingin introductory psychology in an OA college', Teachingof Psychology 19(2), 78-81. Olson, G.A. (ed.) (1984). Writing Centers: Theory and Administration.Illinois: National Council of Teachersof English. Quintrell, N. and Westwood, M. (1994). 'The influence of a peer-pairingprogramon international students' first year experience and use of student services', Higher Education Researchand Development13(1), 49-57. Ramsden,P. (ed.) (1986). ImprovingLearning.London:KoganPage. 344 J.T.,Eysenck,M.W.andPiper,D.W.(eds.)(1987).StudentLearning:Research Richardson, in Educationand CognitivePsychology. Buckingham:Open UniversityPress. Riggio,R.E.,Fantuzzo,J.W.,Connelly,S. andDimeff,L.A.(1991).'Reciprocal peertutoring: a classroomstrategyfor promotingacademicand social integrationin undergraduate students', Journal of Social Behavior and Personality6(2), 387-396. Rizzolo,P.(1982).'Peertutorsmakegoodteachers:a successfulwritingprogram',Improving College and UniversityTeaching30(3), 115-119. Robin,A.L. andHeselton,P. (1977). 'Proctortraining:the effectsof a manualversusdirect training',Journal of Personalised Instruction2, 19-24. OpenUniversityPress. Rogers,J. (1977).AdultsLearning(secondedition).Buckingham: Rogoff, B. (1990). Apprenticeshipin Thinking:Cognitive Developmentin Social Context. OxfordandNew York:OxfordUniversityPress. outcomesasinfluenced Rosen,S., Powell,E.R.andSchubot,D.B.(1977).'Peer-tutoring bythe Journalof Educational Psychology69(3),244-252. equityandtypeof roleassignment', atOxfordBrookesUniversity', instruction Rust,C. (1993).'Supplemental Papergivenat Peer TutoringConsortiumConference,Universityof Glamorgan,23 June 1993. Rust, C. and Wallace, J. (eds.) (1994). Helping Studentsto Learnfrom Each Other: Supple- StaffandEducational mentalInstruction. Association. Birmingham: Development J. and P. (1993). 'Instruction in learningskills:an integrated P.D., Wallace, Bidgood, Rye, approach',Medical Education27, 470-473. Schaffer,J.L., Wile, M.Z. andGriggs,R.C. (1990). 'Studentsteachingstudents:a medical schoolpeertutorialprogramme', MedicalEducation24(4),336-343. Schmidt,H., Arend,A.V.D.,Kokx,I. and Boon, L. (1994). 'Peerversusstaff tutoringin Science22, 279-285. learning',Instructional problem-based Schunk,D.H. (1987). 'Self-efficacyandmotivatedlearning',in Hastings,N. andSchwieso, J. (eds.), New Directions in Educational Psychology: Behaviour and Motivationin the Classroom.LondonandNew York:FalmerPress. Schunk, D.H. and Zimmermann,B.J. (eds.) (1994). Self-Regulationof Learning and Perfor- mance.New York:LawrenceErlbaum. withinspecialeducation R.T.(1986). 'Tutoringinterventions Scruggs,T.E.andOsguthorpe, settings:a comparisonof cross-ageandpeertutoring',Psychologyin theSchools23(2), 187-193. C.F.(1981).'Peertutoring:a reviewof theliterature', Collected Sharpley,A.M.andSharpley, Original Resources in Education5(3), 7-C11 (fiche 7 and 8). Sherman,J.G.(1992). 'Reflectionson PSI:goodnewsandbad',Journalof AppliedBehavior Analysis 25(1), 59-64. andNew York:Cambridge Press. Stemberg,R.J.(1985).BeyondI.Q.Cambridge University P.(ed.)(1996).AdultLearning:A Reader.London:KoganPage(in press). Sutherland, Tight, M. (ed.) (1983). Adult Learningand Education.London:Routledge. anoverview',ThePsychologist Topping,K.J.(1992).'Co-operative learningandpeertutoring: 5, 151-161. Topping,K.J., Hill, S., McKaig,A., Rogers,C., Rushi,N. andYoung,D. (1996). 'Paired forpublication). economics',(submitted reciprocalpeertutoringin undergraduate of faculty-wide Topping,K.J.,Simpson,G., Thompson,L. andHill, S. (1996). 'Evaluation studentsupportedlearningat theUniversityof CentralLancashire', forpubli(submitted cation). Topping,K.J.,Watson,G.A.,Jarvis,R.J.andHill,S. (1996).'Same-year pairedpeertutoring in undergraduate forpublication). mathematics', (submitted of higherpsychological Vygotsky,L.S.(1978).'Mindin society:thedevelopment processes', MA: V.,Scribner,S. andSouberman, (editedby Cole, M., John-Steiner, E.). Cambridge MITPress. instructionat KingstonUniversity',Papergivenat Peer Wallace,J. (1993). 'Supplemental TutoringConsortiumConference,Universityof Glamorgan,23 June 1993. 345 Webb, N.M. (1982). 'Peer interactionand learningin co-operativesmall groups', Journal of Educational Psychology 5(74), 642-655. Whitman, N.A. (1988). Peer Teaching:To Teach Is To Learn Twice (ASHE-ERIC Higher EducationReport).WashingtonDC: ERICClearinghouseon Higher Education.