2024年5月4日发(作者：台式机组装流程)

StudiesinEducationalEvaluation37(2011)3–14

ContentslistsavailableatScienceDirect

StudiesinEducationalEvaluation

journalhomepage:/stueduc

Whatisassessmentforlearning?

DylanWiliam

InstituteofEducation,UniversityofLondon,UnitedKingdom

ARTICLEINFOABSTRACT

Articlehistory:

Availableonline12April2011

Keywords:

Formativeassessment

Assessmentforlearning

Feedback

Theideathatassessmentisintrinsictoeffectiveinstructionistracedfromearlyexperimentsinthe

individualizationoflearningthroughtheworkofBenjaminBloomtoreviewsoftheimpactoffeedback

anyofthesereviewsdetailedtheadverseimpactofassessmenton

learning,theyalsoindicatedthatundercertainconditionsassessmenthadconsiderablepotentialto

ownthatunderstandingtheimpactthatassessmenthasonlearningrequiresa

broaderfocusthanthefeedbackinterventionitself,particularlythelearner’sresponsestothefeedback,

entdefinitionsoftheterms‘‘formative

assessment’’and‘‘assessmentforlearning’’arediscussed,andsubsumedwithinabroaddefinitionthat

focuseerconcludes

byexploringsomeoftheconsequencesofthisdefinitionforclassroompractice.

ßhtsreserved.

uction

Almosthalfacenturyago,DavidAusubelsuggestedthatthe

mostimportantfactorinfluencinglearningiswhatthelearner

alreadyknows,thatteachersshouldascertainthis,andteach

accordingly(Ausubel,1968).

Atthetime,andperhapsevennow,suchaprescriptionmight

seemsimple,

wheninstructionisplannedwithgreatcare,deliveredeffectively,

andinawaythatengagesstudents,thelearningoutcomesoften

astudent

learnsasaresultofaparticularsequenceofinstructionalactivities

isimpossibletopredict,evenintheunlikelyeventthatallthe

learnersinaninstructionalgroupareatthesameplacewhenthe

instructionstarts,withinminutes,studentswillhavereached

whyassessmentisa,perhapsthe,

lythrough

assessmentthatwecanfindoutwhetheraparticularsequence

ofinstructionalactivitieshasresultedintheintendedlearning

outcomes.

Formanyyears,theword‘‘assessment’’wasusedprimarilyto

describeprocessesofevaluatingtheeffectivenessofsequencesof

actionsthatguidedlearningprocessesbeforetheendofthe

sequenceweregenerallynotregardedaskindsofassessments.

WithintheFrenchlanguageliterature,theyweretypically

discussedasaspectsoftheregulationoflearningprocesses,and

withintheEnglishlanguageliterature,totheextentthatitwas

discussedatall,

recently,particularlyintheEnglish-speakingresearchcommunity,

however,therehasbeenanincreasingtendencytoseekto

understandactivitiesthatareintendedtoguidethelearning

towardstheintendedgoal,andthattakeplaceduringthelearning

process,paperIreviewthis

development,andattempttoclarifythemeaningsoftheterms

assessmentforlearningandformativeassessment.

mentforlearning:originsandantecedents

Formanyyears,itseemsthattheprevailingviewofeducation

wasthat,providedinstructionwasofreasonablequality,itneed

ssumedeitherthat

well-designedinstructionwouldbeeffectiveforthemajorityof

studentsforwhomitwasintended(withothersbeingassignedto

remedialactivities)orthatthecausesofanyfailurestolearnlay

withintheindividuallearner(thematerialwasjusttoohardfor

them,andtheyshouldinsteadpursueother,andgenerallyless

academic,avenues).However,inthe1960s,BenjaminBloomand

hisgraduatestudentsattheUniversityofChicagobegantoexplore

theideathatthenormaldistributionofstudentoutcomeswasnot

a‘‘natural’’outcome,butcausedbythefailureoftheinstructionto

recognisedifferencesinlearners.

The‘‘IndividualSystem’’oftenregardedasthefirsttruly

individualizedsystemofinstruction(see,forexample,Reiser,

1986),wasdevelopedbyFredericBurk,from1912to1913,foruse

intheelementaryschoolassociatedwiththeSanFranciscoNormal

StateSchool,aninstitutionprovidingpre-serviceeducationfor

urk’scolleagues,MaryWard,hadbeengetting

hertraineeteacherstodevelopself-instructionalmaterialsand

E-mailaddress:dylanwiliam@.

0191-491X/$–seefrontmatterßhtsreserved.

doi:10.1016/c.2011.03.001

/StudiesinEducationalEvaluation37(2011)3–14

Burkandothersdevelopedsimilarmaterialsthatcoveredmostof

er

individualswhohadworkedwithWardandBurkattheSan

FranciscoNormalStateSchool,CarletonWashburneandHelen

Parkhurst,developedtheseideasfurtheraftertheylefttheSchool.

In1919,WashburneimplementedtheWinnetkaPlan,whenhe

becamesuperintendentoftheWinnetkaPublicSchoolsinIllinois

andinthesameyear,ParkhurstimplementedtheDaltonPlanina

schoolfordisabledstudentsinDalton,Massachusetts(Parkhurst,

1922).

Bloomwasconvincedthatsuchindividualizationwasbenefi-

cial—indeedheregardedone-to-onetutorialinstructionasthe

‘‘goldstandard’’foreducationagainstwhichothersshouldbe

compared(Bloom,1984a)—butwasconcernedthatthisobviously

wouldnotbeaffordableformasspubliceducation,hence‘‘the

searchformethodsofgroupinstructionaseffectiveasone-to-one

tutoring’’(Bloom,1984b).

Oneofthemainreasonsthatone-to-onetutoringissoeffective,

accordingtoBloom,isthatthetutorisabletoidentifyerrorsinthe

student’sworkimmediately,andthentoprovideclarification,and

furtherfollow-upifnecessary(Guskey,2010).Bloomdescribed

thesetwoprocessesas‘‘feedback’’and‘‘correctives’’andthis

languagehasbecomepartofthestandardwayoftalkingabout

r,inaveryimportantsense,

Bloom’sdistinctionbetween‘‘feedback’’and‘‘correctives’’has

beencounterproductive,andhasservedtodistorttheoriginal

meaningoftheterm‘‘feedback’’inaparticularlyunfortunate

manner.

In1940,NorbertWienerandhiscolleagueshadbeen

developingautomaticrange-fi

realizedthateffectiveactionrequiredaclosedsystemthatallowed

theeffectsofactionstakenwithinthesystemtobeevaluated,and

inthelightofthatevaluation,tomodifyfutureactions(Wiener,

1948).Insuchsystems,thereweretwokindsofloops:thosethat

tendedtopushthesystemfurtherinthedirectioninwhichitwas

alreadygoing(whichhetermedpositivefeedbackloops)andthose

thatopposedthetendencyinthesystem(whichhetermed

negativefeedbackloops).Positivefeedbackloopsproduce

instability,drivingthesystemtowardseitherexplosionorcollapse.

Examplesoftheformeraresimplepopulationgrowthwith

plentifulfoodandnopredators,andinflationaryprice/wagespirals

ineconomics;examplesofthelatterincludeeconomicdepression,

foodhoardingintimesofshortage,andthelossoftaxrevenuein

urbanareasasaresultof‘‘middle-classflight’’.Theuseofthe

qualifier‘‘positive’’isnotintendedtoprovideanyassessmentof

thevalueofsuchfeedback—indeed,positivefeedbackalmost

theterm‘‘positive’’denotes

simplythealignmentbetweentheexistingtendencyofthesystem,

andtheeffectoftheimpetusprovidedbythefeedback.

Incontrast,negativefeedbackloopsproducestability,because

mple

ofsuchasystemispopulationgrowthwithlimitedfoodsupply,in

ingonthe

conditions,thesystemtheneitherapproaches,oroscillateswith

decreasingamplitudearound,asteadystate(thecarryingcapacity

oftheenvironment).Perhapsthemostfamiliarexampleisthe

etemperatureoftheroomdrops

belowthesettingonthethermostat,asignalissenttoturnonthe

eroomheatsupabovethesettingonthe

thermostat,asignalissenttoturnofftheheatingsystem.

TheimportantpointaboutWiener’sformulationisthat

informationdoesnotbecome‘‘feedback’’unlessitisprovided

withinasystemthatcanusethatinformationtoaffectfuture

ortanceofthinkingaboutfeedbacksystems,

ratherthanjustthenatureoftheinformationitself,particularly

withinthebehaviouralsciences,wasemphasizedbyRamaprasad

(1983)whonoted:‘‘Feedbackisinformationaboutthegap

betweentheactuallevelandthereferencelevelofasystem

parameterwhichisusedtoalterthegapinsomeway’’

(Ramaprasad,1983,p.4).Theuseoftheinformationwasreinforced

bySadler(1989):

AnimportantfeatureofRamaprasad’sdefinitionisthat

informationaboutthegapbetweenactualandreferencelevels

theinformationissimplyrecorded,passedtoathirdpartywho

lackseithertheknowledgeorthepowertochangetheoutcome,

oristoodeeplycoded(forexample,asasummarygradegiven

bytheteacher)toleadtoappropriateaction,thecontrolloop

cannotbeclosed,and‘‘danglingdata’’substitutedforeffective

feedback.(p.121)

ThisiswhyBloom’ribingthe

informationgeneratedaboutthegapbetweencurrentanddesired

performanceas‘‘feedback’’Bloomseparatedtheinformationfrom

ner,Ramaprasad,and

Sadler,formation

generatedwithinaparticularsystem,

iswhyfeedbackhastobedomain-specifiingplayers

improvetheirfree-throwpercentage,basketballcoachesdonot

justtelltheathletestomakesurethattheygettheballthroughthe

hoop;theyfocusonmechanicssuchasremindingtheathleteto

bendtheirknees,

providingfeedbacktostudentslearningmathematics,itisnot

helpfultotellthemthattheyneedtoimprovetheirwork,evenif

rehelpfultopointoutwhatkindsoferrorsthey

aremaking,andwhattheyneedtodotoimprove.

Asecondpointaboutthefeedbacksystemisthatithasbeen

designedsothattheinformationgeneratediscapableofaffecting

astheterm

‘‘feedback’’isusedtodenotesimplyanydataaboutthegap

betweencurrentanddesiredlevelofperformance,orworse,

simplyfordescriptionsofthecurrentlevelofperformance,itloses

allconnectionwithitsoriginal,andpowerful,meaning.

ally,thesuffix‘‘back’’

wasintendedtodescribethedirectionofinformationfl

muchcurrentusage,itappearstobeusedtosignifychronology,

anditseemsthatanyinformationaboutpreviousperformance

asadandSadlersuggestthatthe

term‘‘feedback’’shouldnotbeusedunlessthereisanimpacton

(e.g.,Black&Wiliam,1998b)accept

theterm‘‘feedback’’asitiscommonlyused,butrequirean

additionalcondition,thatitactuallyimprovesstudentlearning,for

way,whatisimportantisthe

acknowledgementthattheuseofassessmentinformationto

improvelearningcannotbeseparatedfromtheinstructional

n1986and1998,nine

substantialarticlesreviewedtheimpactofassessmentpracticeson

studentsandtheirlearninginthecontextoftheclassroom,andthe

consistencyoffindingsfromtheseninereviewscreatedsignificant

interestamongstresearchers,policy-makersandclassroom

thefollowingsection,theseninereviews,andmorerecentworkin

thisarea,arebrieflyreviewed.

sofresearchonassessmentandclassroomlearning

FuchsandFuchs(1986)conductedameta-analysisof21

researchreports,whichyielded96differenteffectsizes,relatingto

learnersfrompre-schooltotwelfthgrade,mostofwhomhadmild

hestudiesfocusedonthe

/StudiesinEducationalEvaluation37(2011)3–145

useofthefeedbacktoandbyteachers,withfrequenciesofbetween

twoandfineffectsizewas0.70

standarddeviationswithslightlysmallereffects(0.63)forthe22

estudies

whereteachersworkedtosetrulesaboutreviewsofthedataand

actionstofollow(abouthalfofthestudiesreviewed),themean

effectsizewas0.92,whereaswhenactionswerelefttoteachers’

teworthywasthat

whereteachersundertooktoproducegraphsoftheprogressof

individualchildrenasaguideandstimulustoaction,themean

effectwaslarger(0.70)thaninthosewherethiswasnotdone

(meaneffectsize0.26).

Anotherreview(Natriello,1987)proposedamodelofthe

evaluationprocessasconsistingofeightstages:

ishingthepurposeoftheevaluation;

ingtaskstostudents;

gcriteriaforstudentperformance;

gsstandardsforstudentperformance;

nginformationonstudentperformance;

singstudentperformance;

ingfeedbacktostudentperformers;and

ringoutcomesoftheevaluationofstudents.

Hismajorconclusionwasthatlittlecouldbeconcludedfrom

theresearch,firstwasthatthe

availablestudiesresearchedwhatwasactuallyinplace,rather

thanwhatmightbe,andasaresulttendedtoconfirmtheproblems

withexistingpracticesratherthanprovidingabasisforimproved

ondwasthefactthatmostofthestudiesreviewed

‘‘

result,theyfailtoconsidertheimpactofotherkeyelementsin

determiningtheeffectsofevaluations’’(p.170).

Third,andperhapsmostimportantly,fewofthestudies

Natrielloreviewedconsideredexplicitlythefactthatevaluations

wereusedinschoolsforamultiplicityofpurposesandthat

comparisonswerelikelytobemisleadingwheredifferentkindsof

evaluationswerecomparedintermsoffunctionsforwhichthey

mple,afindingthatdifferentiated

feedbackhadmoreimpactondirectingfuturestudentlearning

thangradesmaybeshowingnothingmorethanthefactthat

systemsgenerallydomoreeffectivelythosethingstheyare

designedtodothanthosethingstheyarenotdesignedtodo.

Thethirdreview(Crooks,1988)hadarathernarrowerfocus—

reviewcoveredformalclassroom-basedassessmentssuchastests,

informalevaluationprocessessuchasadjunctquestionsintexts,

nconclusionwas

that‘‘Toomuchemphasishasbeenplacedonthegradingfunction

ofevaluationandtoolittleonitsroleinassistingstudentstolearn’’

(p.468).Healsonotedthatarebalancingoftheattentionpaidto

thesetworoleswasneeded,sinceanover-emphasisonthegrading

functionnotonlyusedtimethatcouldmoreprofitablyspenton

otheractivities,butwasactuallycounter-productive,resultingin:

reductionofintrinsicmotivation,debilitatingevaluation

anxiety,abilityattributionsforsuccessandfailurethat

underminestudenteffort,loweredself-efficacyforlearning

intheweakerstudents,reduceduseandeffectivenessof

feedbacktoimprovelearning,andpoorersocialrelationships

amongthestudents.(p.468)

Afourthstudyoftheimpactofclassroomassessmentpractices

onstudentsandtheirlearninginthecontextoftheclassroomwas

undertakenbyBangert-Drowns,Kulik,andKulik(1991),who

und

thatstudentswhotookatleastonetestovera15weekperiod

scored0.5standarddeviationshigherthanthosewhodidnot,and

thatmorefrequenttestingwasassociatedwithhigherlevelsof

achievement,althoughtestingmorefrequentlythanonceevery

twoweeksappearedtoconfernoadditionalbenefiated

study,Bangert-Drowns,Kulik,Kulik,andMorgan(1991)reported

theresultsofameta-analysisof58effectsizesfrom40research

reportsontheeffectsoffeedbackinwhattheycalled‘‘test-like’’

eventssuchasquestionsembeddedinprogrammedlearning

materials,reviewtestsattheendofablockofteaching,andsoon.

Theyfoundthatthecrucialmediatingvariableindeterminingthe

impactoffeedbackonlearningwasthedegreetowhichthenature

ofthefeedback,andthewayitwasprovided,encouraged

‘‘mindfulness’’,wherestudentscouldlookahead

totheanswersbeforetheyhadattemptedthequestions

themselves,achievementwasreduced,butwherethestudies

controlledforthis‘‘pre-searchavailability,’’theeffectsizewas0.26

er,wheretheinterventionsalso

providedfeedbacktostudentsintermsofdetailsofthecorrect

answer,ratherthanjustwhetherthestudents’responsewas

correctornot,themeaneffectsizewas0.58standarddeviations.

ThesebroadfindingswereconfirmedbyDempster(1991)ina

reviewofstudiesontheeffectsoffeedbackintests,althoughhe

pointedoutthatmanyoftherelevantstudiesmeasured

achievementintermsofcontentknowledgeandlow-levelskills,

soitwasnotclearthatsuchfindingswouldnecessarilygeneralize

sequentpaper(Dempster,1992),

hearguedthat,whilethebenefitsofintegratingassessmentwith

instructionwereclearandtherewasanemergingconsensusinthe

researchfortheconditionsforeffectiveassessment,including

frequenttestingsoonafterinstruction,cumulatingdemand,and

feedbacksoonaftertesting,assessmentwasneglectedinteacher

educationandtherewasevidencethatcurrentpracticesinschools

werefarfromtheseideals.

AreviewbyElshout-Mohr(1994),publishedoriginallyin

Dutch,andreviewingmanystudiesnotavailableinEnglish,

suggestedthatknowledgeofcorrectanswerswasmoreusefulfor

ting

whatiswrongcanbeaneffectiveapproachforthelearningof

simplematerial,butformorecomplexmaterial,learningrequires

thedevelopmentofnewcapabilitiesthat,accordingtoElshout-

Mohr,requiresamoredialogickindoffeedback,ratherthanthe

givingofcorrectanswers,andthereforealsorequiresthelearnerto

becomeactiveinmanagingtheprocess.

Muchofthisworkhadfocusedontheeffectsoffeedbackin

r,in1996,KlugerandDeNisipublishedareviewof

theeffectsoffeedbackinschools,

beganbydefiningfeedbackinterventionsas‘‘actionstakenby(an)

externalagent(s)toprovideinformationregardingsomeaspect(s)

ofone’staskperformance’’(p.255).Theyidentifiedover3000

relevantstudiespublishedbetween1905and1995,butnotedthat

manyofthesewereverysmallstudies(inmanycasesinvolving

onlyasingleparticipant),andwerereportedininsufficientdetail

ordertobesurethatpoorqualitystudieswerenotbeingincluded,

KlugerandDeNisiestablishedthreecriteriaforinclusionintheir

review:

ticipantshadtobedividedintotwogroups,theonly

differencebetweenthegroups,asfarascouldbejudged,being

whethertheyreceivedfeedbackornot.

dyinvolvedatleasttenparticipants.

cludedameasurementofperformancewithsufficient

detailsprovidedforthesizeoftheimpactoffeedbackon

performancetobecalculated.

6

/StudiesinEducationalEvaluation37(2011)3–14

Table1

Possibleresponsestofeedbackinterventions(Kluger&DeNisi,1996).

ResponsetypeFeedbackindicatesperformanceexceedsgoalFeedbackindicatesperformancefallsshortofgoal

ChangebehaviourExertlesseffort

ChangegoalIncreaseaspiration

AbandongoalDecidegoalistooeasy

RejectfeedbackFeedbackisignored

Surprisingly,only131ofthe3000relevantstudiessatisfied

thesecriteria,andtheseselectedstudiesreported607effectsizes,

involving23,663observationsof12,allthe

studies,theaverageeffectsizeforfeedbackwas0.41standard

deviations,buttheeffectsvariedconsiderablyacrossthedifferent

tably,,38%),

ingto

understandthis,theylookedfor‘‘moderators’’offeedbackeffects

andfoundthatfeedbackinterventionswereleasteffectivewhen

theyfocusedattentionontheself,moreeffectivewhenthey

focusedonthefocaltask,andmosteffectivewhentheyfocusedon

thedetailsofthefocaltaskandwhentheyinvolvedgoal-setting.

However,theyconcludedthatwhetherfeedback‘‘works’’ornot,

andifso,byhowmuch,

pointedoutthattherearetwokindsoffeedbackinterventions:those

thatindicatethatcurrentperformancefallsshortofthedesiredgoal

ceived,

therearefourkindsofresponsestheindividualcanmake—change

behaviourtoreachthegoal,modifythegoal,abandonthegoal,or

adstoeightpossibleeffectsoffeedback

interventions,asshowninTable1.

Inotherwords,thereareeightpossibleresponsestoafeedback

intervention,andsixofthemarelikelytobeineffectiveorworse.

Onlytworesponses,highlightedinboldinTable1,arelikelyto

nwhenfeedbackisseentohavean

effect,thismaynotbesustained,ashasbeenfoundinsome

backworks

toincreasemotivation,thenincreasinglylargeeffortsneedtobe

ckfocusedontasklearning

canemphasizeinstrumentalgoals,andthusinhibitdeeplearning.

Insuchsituationsitmightbebettertoprovidemoretask

informationoreventoencourageatrialandimprovement

strategy,thusgeneratingfeedbackwithoutafeedbackinterven-

owedthatfeedbackinterventionswerelesseffective

whentheycuedattentionbeyondthetask(forexampleonthe

self),andmoreeffectivewhenthefeedbackcuedattentionontask

motivationortasklearning(thisistakenupinmoredetailinthe

discussionofthereviewbyHattieandTimperleyandofthe‘‘dual-

processing’’theoryofMoniqueBoekaertsbelow).Thismodel

accountsforthewell-knownearlierfindingsthatpraise(likeother

cuesthatdrawattentionawayfromthetaskandtowardstheself)

oftenhasnegativeeffects(Brophy,1981).

BlackandWiliam(1998a)soughttoupdatethereviews

blemtheyreportedwasageneral

difficultyindefiningthefitedthatthereviewsby

NatrielloandCrooksmentionedabovehadcited91and241

referencesrespectively,andyetonlyninereferenceswere

commontobothpapersandneithercitedthereviewbyFuchs

rownwork,BlackandWiliamfoundthat

attemptingtorelyonelectronicmethodseithergeneratedfartoo

rto

besureofreviewingthefieldthoroughly,theyphysically

examinedeachissueof76ofthejournalsconsideredmostlikely

tocontainrelevantresearchpublishedbetween1987and1997.

BlackandWiliam’sreview,whichcited250studies,foundthat

effectiveuseofclassroomassessmentyieldedimprovementsin

studentachievementofbetween0.4and0.7standarddeviations

Increaseeffort

Reduceaspiration

Decidegoalistoohard

Feedbackisignored

(althoughtheynotedproblemswiththeinterpretationofeffect

sizesacrossdifferentstudies—seediscussioninnextsection).

Inframingtheirreview,BlackandWiliam(1998a)first

presentedanumberof‘‘examplesinevidence’’thatillustrateda

sthe

mostimportantfeaturetheyidentifiedwasthat,tobeeffective,

formativeassessmenthadtobeintegratedintoclassroompractice,

requiringafundamentalreorganizationofclassroomoperations:

Itishardtoseehowanyinnovationinformativeassessment

h

workinvolvessomedegreeoffeedbackbetweenthosetaught

andtheteacher,andthisisentailedinthequalityoftheir

interactionswhichisattheheartofpedagogy.(Black&Wiliam,

1998a,p.16)

BlackandWiliam(1998a)alsonotedthatforassessmentto

functionformatively,thefeedbackinformationhadtobeused,and

thusthedifferentialtreatmentsthatareincorporatedinresponseto

er,forthese

differentiatedtreatmentstobeselectedappropriately,teachersneed

adequatemodelsofhowstudentswillreactto,andmakeuseof,the

enoud(1998)notedinhiscommentaryontheBlack

andWiliampaper,‘‘...thefeedbackgiventopupilsinclassislikeso

anbesurethatthe

messagetheycontainwillonedayfindareceiver’’(p.87).The

consequenceofthisisthatthedesignofeffectiveformative

assessmentcannotbedetachedfromthelearningmilieuinwhich

ivationsandself-perceptionsofstudents,

andtheirassessmenthistories,willallbeimportantinfluenceson

howfeedbackisreceived(Deci&Ryan,1994).

Inordertoaddresstheinfluencesonhowfeedbackisreceived,

theBlackandWiliam(1998a)reviewexaminedthestudent

perspective,theroleofteachers,andsomeofthesystemsforthe

organizationofteachinginwhichformativeassessmentwasa

ingoutimplicationsforthepolicyand

practiceofformativeassessment,theyconcluded:

Theredoesnotemerge,fromthispresentreview,anyone

s

emergeisasetofguidingprinciples,withthegeneralcaveat

thatthechangesinclassroompracticethatareneededare

centralratherthanmarginal,andhavetobeincorporatedby

Thatistosay,reforminthisdimensionwillinevitablytakea

longtimeandneedcontinuingsupportfrombothpractitioners

andresearchers.(p.62)

Mostoftheworkreviewedabovefocusedonschool-age

t(2003)focusedonstudiesof

ewedapproximately

3000studiesoftheeffectsoffeedback,ofwhich86metthe

followinginclusioncriteria:

(a)experimentalmanipulationofacharacteristicrelevantto

feedback;

/StudiesinEducationalEvaluation37(2011)3–14

7

(b)usedasampleofcollege-agedlearners;

(c)measuredacademicperformance;and

(d)providedsufficientquantitativeinformationforaneffectsize

tobecalculated.

Fromthe86studiesitwaspossibletoderive185effectsizes.

Theanalysisyieldedameaneffectsizeof0.40standard

deviations—almostidenticaltothatfoundbyKlugerandDeNisi

(1996).Weightingtheeffectssothattheircontributiontothe

meaneffectwasproportionaltotheirreliabilityreducedthismean

effectslightlyto0.35(SE=0.17),althoughtheeffectsthemselves

werehighlyvariable,rangingfromÀ0.6to1.6standarddeviations.

Inordertoinvestigatemoderatorsofeffect,Nyquistdevelopedthe

followingtypologyofdifferentkindsofformativeassessments:

Weakerfeedbackonly:studentsaregivenonlytheknowledgeof

theirownscoreorgrade,oftendescribedas‘‘knowledgeof

results.’’

Feedbackonly:studentsaregiventheirownscoreorgrade,

togetherwitheithercleargoalstoworktowards,orfeedbackon

thecorrectanswerstothequestionstheyattempt,often

describedas‘‘knowledgeofcorrectresults.’’

Weakformativeassessment:studentsaregiveninformation

aboutthecorrectresults,togetherwithsomeexplanation.

Moderateformativeassessment:studentsaregiveninformation

aboutthecorrectresults,someexplanation,andsomespecific

suggestionsforimprovement.

Strongformativeassessment:studentsaregiveninformation

aboutthecorrectresults,someexplanation,andspecific

activitiestoundertakeinordertoimprove.

Table2providestheaveragestandardizedeffectsizeforeach

typeofintervention,althoughthesearecorrectedvaluesthatdiffer

fromthoseintheoriginalthesis(t,personalcommuni-

cation,May7,2007).Nyquist’sresultsechothefindingsofBangert-

Drownsetal.(1991a,1991b)vingstudents

feedbackaboutcurrentachievementproducesonlymodest

benefits,butwherefeedbackengagesstudentsinmindfulactivity,

ectsizesfoundby

Nyquistalsounderscorethedomain-specificityofeffective

feedbackmentionedearlier.

FromthereviewsofresearchconductedbyNatriello(1987),

Crooks(1988),Bangert-Drownsetal.(1991a),Bangert-Drowns

etal.(1991b),andBlackandWiliam(1998a),itisclearthatnotall

kindsoffeedbacktostudentsabouttheirworkareequally

example,Meisels,Atkins-Burnett,Xue,Bickel,andSon(2003)

exploredtheimpactoftheWorkSampleSystem(WSS)—asystem

ofcurriculum-embeddedperformanceassessments—onthe

achievementof96thirdgradeurbanstudentsinreadingand

mathematics,

comparedwithasampleof116thirdgradersinmatchedschools

andwithstudentsintheremainderoftheschooldistrict

(Pittsburgh,PA),theachievementofWSSstudentswassignifi-

ematics,

however,thedifferencesweremuchsmaller,andfailedtoreach

Table2

Effectsizesfordifferentkindsoffeedbackinterventions(Nyquist,2003).

NEffectsize

Weakerfeedbackonly310.14

Feedbackonly480.36

Weakerformativeassessment490.26

Moderateformativeassessment410.39

Strongformativeassessment160.56

Total185

statisticalsignifidthereforeappearthatdifferent

schoolsubjectsmayrequiredifferentapproaches,againreinfor-

cingthedomain-specificityofeffectiveinterventions.

ClassroomassessmentsystemssuchastheWorkSampling

Systemarealsooftendesignedprimarilyforsummative

purposes,tomonitorandreportonstudentprogress,with

theirusetogenerateinformationforformativepurposesoften

eviewoftheresearch

literatureonclassroomassessment,Brookhart(2004)beganby

undertakingonlinesearcheswith‘‘classroomassessment’’asa

inghitsrelatednotrelevanttoK-12

education(forexample,studiesconductedinhighereducation

settings)generatedatotalof41empiricalstudieswithafocus

cludedthat

classroomassessmentoccursattheintersectionofthree

teachingfunctions:instruction,classroommanagement,and

assessment,andnotedthatthetheoryrelevanttoclassroom

assessmentcamefromseveraldifferentfields,including

individualdifferencespsychology,thestudyofgroups,and

onotedthatmanyofthe

studiesshecitedapproachedthephenomenaunderstudyfroma

singledisciplinaryperspective(oftenpsychology)orwere

tudies

hadmixedtwoormorepracticalortheoreticalperspectives,she

concludedthat‘‘theresultingpictureofclassroomassessment

wasricherandmoremultidimensional’’(p.454).

Whilemanyofthestudiesincludedinthereviewsdiscussed

abovefocusonolderstudents,itisapparentthatstudents’

attitudestolearningareshapedbythefeedbacktheyreceive

r-longstudyofeightkindergarten

andfirstgradeclassroomsinsixschoolsinEngland,Tunstalland

Gipps(1996a,1996b)identifiedarangeofrolesplayedby

rranceandPryor(1998),theyfoundthatmuch

ofthefeedbackgivenbyteacherstostudentsfocusedon

socialization:‘‘I’monlyhelpingpeoplewhoaresittingdown

withtheirhandsup’’(Tunstall&Gipps,1996b,p.395).Beyond

thissocializationrole,theyidentifiedfourtypesoffeedbackon

academicwork(seeTable3).TypeAincludedfeedbackthat

rewardedorpunishedthestudentsfortheirwork,suchas

studentsbeingallowedtoleaveforlunchearlywhentheyhad

donegoodwork,orthreatenedwithnotbeingallowedtoleave

feedbackwasalsoevaluativebut,whiletypeAfeedbackfocused

onrewardsandsanction,typeBfeedbackindicatedtheteacher’s

,‘‘I’mverypleasedwithyou’’or‘‘I’mvery

disappointedinyoutoday’’.

IncontrasttotheevaluativefeedbackclassifiedastypesAand

B,feedbackclassifi

focusedontheadequacyoftheworkintermsoftheteacher’s

criteriaforsuccess,rangingfromtheextenttowhichthework

alreadysatisfi,‘‘Thisisextremely

wellexplained’’)tothestepsthestudentneededtotaketoimprove

,‘‘Iwantyoutogooverallofthemandwriteyour

equalssignineachone’’).AdefiningcharacteristicoftypeC

feedbackisthatitfocusesontheideaofworkasproduct,while

typeDfeedbackemphasizesprocessaspectsofwork,withthe

teacherplayingtheroleoffacilitator,

TunstallandGipps(1996b)explain,teachersengagedinthiskind

offeedback‘‘conveyedasenseofworkinprogress,heightening

awarenessofwhatwasbeingundertakenandreflectingonit’’(p.

399).

From2002to2004,aspartofitsresearchprogramon‘‘What

worksininnovationineducation’’theOrganisationforEconomic

CooperationandDevelopment(OECD)undertookareviewofthe

practiceofformativeassessmentinlower-secondaryschool

classroomsineightcountries:Australia,Canada,Denmark,

8

/StudiesinEducationalEvaluation37(2011)3–14

Table3

Typologyofteacherfeedback.

EvaluativefeedbackTypeATypeB

PositiveRewardingApproving

NegativePunishingDisapproving

DescriptiveTypeCTypeD

feedback

AchievementSpecifyingConstructing

feedbackattainmentachievement

ImprovementSpecifyingConstructingthe

feedbackimprovementwayforward

Source:AdaptedfromTunstallandGipps(1996a).

England,Finland,Italy,NewZealandandScotland(Looney,2005).

Aswellasdetailedcasestudiesoftheeightsystemsincludedinthe

review,thereportoftheprojectalsocontainedreviewsofthe

researchonformativeassessmentpublishedinFrench(Allal&

Lopez,2005)andGerman(Ko

¨

ller,2005).AllalandLopezreported

thatworkbyresearchersinFranceandFrench-speakingpartsof

Belgium,CanadaandSwitzerlandhasfocusedmuchmoreon

theoreticalthanempiricalwork,withveryfewcontrolled

ggestthatthemostimportantfinding

ofthereviewofover100studiespublishedinFrenchoverthelast

thirtyyearsisthatthestudiesofassessmentpracticesinFrench

speakingclassroomshaveutilizedan‘‘enlargedconceptionof

formativeassessment’’(p.245),alongthelinesadoptedbyBlack

andWiliam(1998a).

Inparticular,AllalandLopezarguethatthecentralconcept

intheapproachtofeedbackespousedwithintheAnglophone

tradition,forexamplebyBloom,isthatof‘‘remediation,’’

whichtheysummarizeas‘‘feedback+correction.’’Incontrast,

withinmuchoftheresearchundertakeninFrancophone

countries,thecentralconceptis‘‘regulation’’,summarizedas

‘‘feedback+adaptation’’(p.245).Itisimportanttonotethat

theFrenchwordre

´gulation

hasamuchmorespecificmeaning

thantheEnglishword‘‘regulation’’.Therearetwowaysto

translatetheword‘‘regulation’’intoFrench—re

`glement

and

re

´gulation

.Theformeroftheseisusedinthesenseof‘‘rules

andregulations,’’whilethelatterisusedinthesenseof

adjustmentinthewaythatathermostatregulatesthe

temperatureofaroom.

Intheirreview,AllalandLopez(2005)identifyfourmajor

developmentsinthedevelopmentoftheconceptionofformative

assessmentintheFrench-languageliteratureoverthelastthirty

first,whichtheyterm‘‘Focusoninstrumentation’’the

emphasiswasonthedevelopmentofassessmenttoolssuchas

second,entitled‘‘Searchfortheoreticalframeworks’’,theemphasis

shiftedtoa‘‘searchfortheoriesthatcanofferconceptual

orientationforconductingassessment’’(p.249).Thethird

development—‘‘Studiesofexistingassessmentpracticesintheir

contexts’’—providesagroundingforthesearchfortheoretical

frameworksbyarticulatingitwiththestudyofhowformative

rthandmost

recentdevelopmenthasbeen‘‘Developmentofactivestudent

involvementinassessment’’whichhasexaminedstudentself-

assessment,peerassessment,andthejointconstructionof

assessmentbystudentsandteacherstogether.

Thenotionofformativeassessmentasbeingcentraltothe

regulationoflearningprocesseshasbeenadoptedbysomewriters

intheAnglophonecommunity(see,forexample,Wiliam,2007),

andthebroadeningoftheconceptionofformativeassessmentin

theEnglish-languageliteraturewasnotedbyBrookhart(2007).

Herreviewoftheliteratureon‘‘formativeclassroomassessment’’

chartedthedevelopmentoftheconceptionofformativeassess-

mentasaseriesofnestedformulations:

Formativeassessmentprovidesinformationaboutthelearning

process;

Formativeassessmentprovidesinformationaboutthelearning

processthatteacherscanuseforinstructionaldecisions;

Formativeassessmentprovidesinformationaboutthelearning

processthatteacherscanuseforinstructionaldecisionsand

studentscanuseinimprovingtheirperformance;

Formativeassessmentprovidesinformationaboutthelearning

processthatteacherscanuseforinstructionaldecisionsand

studentscanuseinimprovingtheirperformance,which

motivatesstudents.

Ingeneral,however,therewouldappeartobefewlinks

betweenthestrongtheoreticalworkintheFrancophonetradition

andthestrongempiricalworkundertaken,particularlyinthe

UnitedStates.

AllalandLopez(2005)concludedthat‘‘studiesof

practiceareepisodicanddispersedindifferentsettings,which

makesitdiffiary,the

theoreticalpromiseofFrench-languageworkonformative

assessmentisinneedofconsiderablymoreempiricalgrounding’’

(p.256).

ThereviewoftheGerman-languageliteraturebyKo

¨

ller

(2005)beganwithanapproachsimilartothatadoptedbyBlack

andWiliam,withsearchesofon-linedatabasessupplemented

byscrutinyofallissuesfrom1980to2003ofthesixmost

¨

llernotedthat,while

thereweremanydevelopmentsrelatedtoformativeassessment

reportedinacademicjournals,therewaslittleevaluationofthe

outcomesofformativeassessmentpracticesforstudents,

althoughtherewereimportantconfirmationsofsomefindings

smostnotably,Ko

¨

ller

reportstheworkofMeyerwho,likeKlugerandDeNisi,found

thatpraisecansometimeshaveanegativeimpactonlearning,

whilecriticism,orevenblame,cansometimesbehelpful.

AnotherimportantstrandofworkmentionedbyKo

¨

llerconcerns

differencesbetweenteachersintheiruseofreferencenorms.A

numberofstudies,notablythosebyRheinberg(1980),have

shownthatstudentslearnmorewhentaughtbyteacherswho

judgeastudent’sperformanceagainstthesamestudent’s

previousperformance(anindividualreferencenorm)rather

thanteacherswhocomparestudentswithothersintheclass(a

socialreferencenorm).

Mostrecently,threesubstantialreviewsonformativeassess-

first(Wiliam,2007),focusedspecifically

asreviewingtheresearch

evidenceonformativeassessment,Wiliamdrewoutsomeofthe

implicationsofthisresearchformathematicsteachingand

outlinedhowthecentralideasofformativeassessmentcouldbe

integratedwithinthebroaderideaoftheregulationoflearning

processesdevelopedfromtheFrench-languageliteraturesumma-

rizedabove.

Theothertworecentreviewsappearedinconsecutiveyearsin

ofabroader

researchprogramonthedevelopmentofintelligenttutoring

environments,Shute(2008)examinedtheresearchonfeedbackto

of141publicationsmettheinclusioncriteria(103

journalarticles,24booksandbookchapters,10conference

proceedingsandfourresearchreports).While,asmightbe

expected,Shute’sreviewidentifiedmajorgapsintheliterature

andconcludedthattherewasnosimpleanswertothequestion,

‘‘Whatfeedbackworks?’’,thereviewdidendorsethefindingsof

earlierreviewsonthesizeoftheeffectsthatcouldbeexpected

fromfeedback:standardizedeffectsizesrangedfrom0.4to0.8

lsoofferedanumberofpreliminary

guidelinesforthedesignofeffectivefeedback:

/StudiesinEducationalEvaluation37(2011)3–149

ckshouldfocusonthe

specificfeaturesofthetask,andprovidesuggestionsonhowto

improve,ratherthanfocusonthelearner;itshouldfocusonthe

‘‘what,howandwhy’’ofaproblemratherthansimplyindicating

tostudentswhethertheywerecorrectornot;elaborated

feedbackshouldbepresentedinmanageableunitsand,echoing

Einstein’sfamousdictum,shouldbe‘‘assimpleaspossiblebut

nosimpler.’’However,feedbackshouldnotbesodetailedand

specificthatitscaffoldsthelearningsocompletelythatthe

ckisalso

moreeffectivewhenfromatrustedsource(whetherhumanor

computer).

imum

timingoffeedbackappearstodependstronglyonthekindof

atefeedbackappearstobe

mosthelpfulforprocedurallearning,orwherethetaskiswell

beyondthelearner’scapabilityatthebeginningofthelearning,

whiledelayedfeedbackappearstobemoreappropriatefortasks

wellwithinthelearner’scapability,orwheretransfertoother

contextsissought.

AreviewbyHattieandTimperley(2007)summarizesan

extensiveprogramofworkconductedbyHattieandhiscolleagues

onsystematicreviewsofinfl

earlierpaper(Hattie,1999)describedtheconstructionofa

databaseof500meta-analyses,reporting450,000effectsizes

from180,

analysisofthe74meta-analysesusedinthe1999studythat

specificallymentionedfeedbackfoundanaverageeffectsizeof

0.56across13,370effectsizesinthe74meta-analyses(Hattie&

Timperley,2007),butHattieandTimperleyfound,ashadKluger

andDeNisi(1996),thattherewassignificantvariabilityamongst

averageofthe5755effectsizesstudiesthatHattieandTimperley

summarizedas‘‘Feedback’’was0.95standarddeviations,topped

onlyby89studiescodedas‘‘Cues’’,whichaveraged1.1standard

deviations.

HattieandTimperleydefinethepurposeoffeedbackasreducing

discrepanciesbetweencurrentunderstandingsorperformanceand

adesiredgoal(asproposedbyRamaprasad,1983).Buildingonthe

workofDeciandRyan(1994)andKlugerandDeNisi(1996),their

modelpositsthatstudentscanreducethediscrepancyeitherby

employingmoreeffectivestrategiesorbyincreasingeffortonthe

onehand,orbyabandoning,blurringorloweringthegoalstheyhave

rscanreducethe

discrepancybychangingthedifficultyorthespecificityofthegoals,

elspecifies

threekindsofquestionsthatfeedbackisdesignedtoanswer(Where

amIgoing?HowamIgoing?Wherenext?)andeachfeedback

questionoperatesatfourlevels:feedbackaboutthetask(FT),

feedbackabouttheprocessingofthetask(FP),feedbackaboutself-

regulation(FR)andfeedbackabouttheselfasaperson(FS).They

demonstratethatFSistheleasteffectiveformoffeedback,thatFR

andFP‘‘arepowerfulintermsofdeepprocessingandmasteryof

tasks’’(pp.90–91)whileFTispowerfulwhenthefeedbackisused

eithertoimprovestrategyprocessing,orforenhancingself-

regulation(althoughtheynotethattheseconditionsarerarely

metinpractice).Theroleofself-regulationinformativeassessment

istakenupinmoredetailbelow.

finitionsofformativeassessmentandassessment

forlearning

Whiletheresearchreviewedabovesuggeststhattheuseof

assessmenttoinforminstructionmighthavesignificantimpact

onlearning,differentreviewsfindverydifferenteffectsizesfor

thebenefiandDeNisi(1996)

foundanaverageeffectsizeof0.41forfeedbackinterventions,

whileBlackandWiliam(1998a,1998b)estimatedthatthe

effectsofformativeassessmentwerearound0.4–0.7standard

(2008)suggestedasimilarrange(0.4–0.8)but

HattieandTimperleyproposedanaverageeffectsizeof0.96

ther

hand,inaclassroomsetting,carriedoutoverayear,with

ordinaryteachers,andwhereperformancewasmeasuredusing

externally-mandatedstandardizedtests,Wiliam,Lee,Harrison,

andBlack(2004)foundthatarangeofformativeassessment

strategiesintroducedbyteachershadaneffectsizeof0.32

asubstantialeffect(theauthors

estimatedthiswasequivalenttoanincreaseoftherateof

studentlearningof70%,oranextraeightmonthsoflearningper

year),butonlyone-thirdofthesizeofeffectssuggestedby

HattieandTimperley.

Partofthevariabilityis,nodoubt,causedbydifferencesinthe

sensitivityofthemeasuresusedinthedifferentstudiestothe

effectsofinstruction—seeWiliam(2010pp.20–22)fora

discussionoftherelationshipbetweensensitivitytoinstruction

sizeswillalsobeaffectedbydifferencesinthe

udiesincludedinreviewsof

researchareconductedonsub-populationsthatarenotrepresen-

mple,ifaneffectsizeis

calculatedinastudyofdifferentinterventionsforstudentswith

specialeducationalneeds,thenthateffectsizewouldnotbe

generalizabletothewholepopulation—wherethepopulationis

morevariable,thestandarddeviationthatisusedasthe

denominatorinthecalculationoftheeffectsizeislarger,leading

toasmallerestimateoftheeffectsize.

However,itseemslikelythatasignificantpart—perhapseven

most—ofthevariabilityiscausedbydifferencesinhowtheideasof

formativeassessmentorassessmentforlearningwereoperatio-

ett(2009)pointsout,inanimportantcritical

reviewofthefield,onecannotbesureabouttheeffectsofsuch

changesinpracticeunlessonehasanadequatedefinitionofwhat

thetermsformativeassessmentandassessmentforlearning

actuallymean,andaclosereadingofthedefinitionsthatare

providedsuggeststhatthereisnoclearconsensusaboutthe

meaningsofthetermsformativeassessmentandassessmentfor

learning.

Asnotedabove,Bloomappearedtoconceptualizeformative

assessmentasacombinationoffeedbackandinstructional

ndWiliam(1998a)definedformativeassess-

mentasfollows:

Weusethegeneraltermassessmenttorefertoallthose

activitiesundertakenbyteachers—andbytheirstudentsin

assessingthemselves—thatprovideinformationtobeusedas

assessmentbecomesformativeassessmentwhentheevidenceis

actuallyusedtoadapttheteachingtomeetstudentneeds’’

(Black&Wiliam,1998b,p.140)

Anumberofauthorshaveproposedsomewhatnarrower

definitions,mostcommonlybyrequiringthechangestoinstruc-

tiontotakeplaceduringtheinstruction,asthefollowingfour

quotationsillustrate:

‘‘theprocessusedbyteachersandstudentstorecogniseand

respondtostudentlearninginordertoenhancethatlearning,

duringthelearning’’(Cowie&Bell,1999,p.32)

‘‘assessmentcarriedoutduringtheinstructionalprocessforthe

purposeofimprovingteachingorlearning’’(Shepardetal.,

2005,p.275)

/StudiesinEducationalEvaluation37(2011)3–14

‘‘Formativeassessmentreferstofrequent,interactiveassess-

mentsofstudents’progressandunderstandingtoidentify

learningneedsandadjustteachingappropriately’’(Looney,

2005,p.21)

‘‘Aformativeassessmentisatoolthatteachersusetomeasure

studentgraspofspecifi’sa

‘midstream’tooltoidentifyspecificstudentmisconceptionsand

mistakeswhilethematerialisbeingtaught’’(Kahl,2005,p.11)

TheAssessmentReformGroup—agroupdedicatedtoensuring

thatassessmentpolicyandpracticeareinformedbyresearch

evidence—acknowledgedthepowerthatassessmenthadto

influencelearning,bothforgoodandforill,andproposedseven

preceptsthatsummarizedthecharacteristicsofassessmentthat

promoteslearning:

itisembeddedinaviewofteachingandlearningofwhichitis

anessentialpart;

itinvolvessharinglearninggoalswithpupils;

itaimstohelppupilstoknowandtorecognisethestandards

theyareaimingfor;

itinvolvespupilsinself-assessment;

itprovidesfeedbackwhichleadstopupilsrecognisingtheir

nextstepsandhowtotakethem;

itisunderpinnedbyconfidencethateverystudentcan

improve;

itinvolvesbothteacherandpupilsreviewingandreflectingon

assessmentdata(Broadfootetal.,1999,p.7).

Inlookingforatermtodescribesuchassessments,they

suggestedthatthetermformativeassessmentwasusedinsuch

differentwaysthatitwasnolongerhelpful:

Theterm‘formative’itselfisopentoavarietyofinterpretations

andoftenmeansnomorethanthatassessmentiscarriedout

assessmentdoesnotnecessarilyhaveallthecharacteristicsjust

identifieformativeinhelping

theteachertoidentifyareaswheremoreexplanationor

thepupils,themarksorremarkson

theirworkmaytellthemabouttheirsuccessorfailurebutnot

abouthowtomakeprogresstowardsfurtherlearning.(Broad-

footetal.,1999,p.7)

Instead,theypreferredthetermassessmentforlearning,which

theydefinedas‘‘theprocessofseekingandinterpretingevidence

forusebylearnersandtheirteacherstodecidewherethelearners

areintheirlearning,wheretheyneedtogoandhowbesttoget

there’’(Broadfootetal.,2002,pp.2–3).

Theearliestuseofthetermassessmentforlearningappearstobe

achapterofthattitlebyHarryBlack(1986).Itwasalsothetitleofa

papergivenatAERAin1992(James,1992)—thesameyearthata

bookcalledTestingforlearningwaspublishedintheUS(Mitchell,

1992)—andthreeyearslater,asthetitleofabookbyRuthSutton

(1995).IntheUnitedStates,theoriginofthetermisoften

mistakenlyattributedtoRickStigginsasaresultofhispopulari-

zationoftheterm(see,forexample,Stiggins,2005),although

Stigginshimselfhasalwaysattributedthetermtootherauthors.

Mostrecently,aninternationalconferenceonassessmentfor

learninginDunedinin2009,buildingonworkdoneattwoearlier

conferencesintheUK(2001)andtheUSA(2005),adoptedthe

followingdefinition:

AssessmentforLearningispartofeverydaypracticeby

students,teachersandpeersthatseeks,reflectsuponand

respondstoinformationfromdialogue,demonstrationand

observationinwaysthatenhanceongoinglearning.(Klenowski,

2009,p.264)

Thephraseassessmentforlearninghasanundoubtedappeal,

especiallywhencontrastedwithassessmentoflearning,butas

Bennett(2009)pointsout,replacingonetermwithanotherserves

merelytomovethedefiportantly,asBlack

andWiliamandtheircolleagueshavepointedout,thedistinctions

betweenassessmentforlearningandassessmentoflearningonthe

onehand,andbetweenformativeandsummativeassessmentonthe

other,merdistinctionrelatestothe

purposeforwhichtheassessmentiscarriedout,whilethesecond

relatestothefunctionitactuallyserves:

Assessmentforlearningisanyassessmentforwhichthefirst

priorityinitsdesignandpracticeistoservethepurposeof

promotingstudents’differsfromassessment

designedprimarilytoservethepurposesofaccountability,orof

ranking,ssmentactivity

canhelplearningifitprovidesinformationthatteachersand

theirstudentscanuseasfeedbackinassessingthemselvesand

oneanotherandinmodifyingtheteachingandlearning

sessmentbecomes

‘‘formativeassessment’’whentheevidenceisactuallyusedto

adapttheteachingworktomeetlearningneeds.(Black,

Harrison,Lee,Marshall,andWiliam,2004,p.10)

Bennett(2009)endorsestheideathatitisunhelpful,and

simplistic,toequateassessmentforlearningwithformative

assessmentandassessmentoflearningwithsummativeassess-

‘‘morenuanced’’(p.5)view,hesuggeststhat

assessmentsdesignedprimarilytoserveasummativefunction

mayalsofunctionformatively,whilethosedesignedprimarilyto

serveaformativefunctionmayalsofunctionsummatively.

Considerthefollowingsevenassessmentscenarios.

fmathematicsteachersfromthesameschoolmeetto

alyzethe

scoresobtainedbytheirstudentsonnationaltestsandseethat

whiletheirscoresare,overall,comparabletonationalbench-

marks,theirstudentstendtoscorelesswellonitemsinvolving

cidetomakeratioandproportion

thefocusoftheirprofessionaldevelopmentactivitiesforthe

comingyear,meetingregularlytodiscussthechangestheyhave

rslater,theyfind

thattheirstudentsarescoringwellonitemsonratioand

proportioninthenationaltests,whichtakestheirstudents’

scoreswellabovethenationalbenchmarks.

ar,agroupoffourth-gradeteachersmeettogetherto

reviewstudents’performanceonastandardizedreadingtest,

andinparticular,lookatthefacility(proportioncorrect)for

temfacilitiesare

lowerthanexpected,theylookathowtheinstructiononthat

aspectofreadingwasplannedanddelivered,andtheylookat

waysinwhichtheinstructioncanbestrengthenedinthe

followingyear.

evenweeks,teachersinaschooluseaseriesof‘‘interim’’

dentwhoscores

belowathresholdjudgedtobenecessarytomakeadequate

dent

whoscoresbelowthethresholdontwosuccessiveoccasionsis

requiredtoattendadditionalinstruction.

erdesignsaninstructionalunitonPulleysandlevers.

FollowingthepatternthatiscommoninmiddleschoolsinJapan

/StudiesinEducationalEvaluation37(2011)3–1411

(Lewis,2002p.76),although14periodsareallocatedtotheunit,

theteachermakessurethatallthecontentiscoveredinthefirst

od12,thestudentscompleteatestonwhat

theyhavecoveredintheprevious11periods,andtheteacher

collectsinthestudentresponses,readsthem,and,onthebasisof

whatshelearnsabouttheclass’sunderstandingofthetopic,

planswhatsheisgoingtodoinlessons13and14.

erhasjustbeendiscussingwithaclasswhyhistorical

essonisdrawing

toaclose,eachstudentisgivenanindexcard(8cmÂ13cm)and

isaskedtowriteananswertothequestion‘‘Whyarehistorians

concernedaboutbiasinhistoricalsources?’’Astheyleavethe

classroom,thestudentshandtheteacherthese‘‘exitpasses’’and

afterallthestudentshaveleft,theteacherreadsthroughthe

cards,andthendecideshowtobeginthenextlesson.

-gradeclasshasbeenlearningaboutdifferentkindsof

firtocheckontheclass’sunderstand-

ing,theteachergiveseachstudentasetofsixcardsbearingthe

lettersA,B,C,D,nteractivewhiteboard,she

displaysthefollowinglist:

ration

opoeia

ole

ification

Shethenreadsoutaseriesofstatements:

ikeabullinachinashop.

ckpackweighsaton.

stallasahouse.

edhishornatthecyclist.

Aseachstatementisreadouttothem,eachmemberofthe

classhastoholduplettercardstoindicatewhatkindoffigurate

cherrealizesthatalmostall

thestudentshaveassumedthateachsentencecanhaveonlyone

kindoffintsoutthatthethird

sentenceisasimile,butisalsohyperbole,andshethenre-polls

theclassonthelasttwostatements,andfindsthatmost

studentscannowcorrectlyidentifythetwokindsoffigurative

tion,shemakesa

mentalnoteofthreestudentswhoanswermostofthequestions

incorrectly,sothatshecanfollowupwiththemindividuallyat

somelaterpoint.

-schoolchemistryteacherhasbeenteachingaclasshow

rtotesttheclass,she

writesuptheunbalancedequationforthereactionofmercury

ninvitesstudentsto

changethequantitiesofthevariouselementsintheequation,

andwhentherearenomoresuggestionsfromtheclass,sheasks

e

intheafficherconcludesthattheclasshas

understood,andmoveson.

Ineachofthesesevenscenarios,assessmentinformationwas

usedtomakeabetterdecisionaboutinstructionthanwouldhave

firsttwo

scenarios,theassessmentinstrumentusedhadbeendesigned

entirelytoserveasummativefunction,buttheteachersinvolved

foundawayofusingtheevidenceaboutstudentachievement

elicitedbytheassessmenttoimprovetheirinstruction.

Inthefirstsixscenarios,theuseoftheevidencechangedthe

instructionforthebetterwhileinthelast,theassessment

informationconfirmedthatwhattheteacherhadplannedtodo

sense,itwasa

betterdecisionthanitwouldhavebeenintheabsenceofany

rwords,evidence

fromassessmentswasusedfortheimprovementoflearning,even

thoughmanyoftheauthorscitedabovewould,inallprobability

notregardthefirstthreeasassessmentforlearningorformative

assessment.

Oneresponsetothiswouldbetotrytorestrictthemeaningof

formativeassessmentorassessmentforlearningtothekindsof

assessmentsthatareclosetoinstruction,whichwouldruleout

thefirstthreescenarios,andforsomeauthors,thefourthalso.

However,restrictingthemeaningoruseofthetermsassessment

forlearningandformativeassessmentsimplytotrytoensure

thatthetermsapplyonlytopracticesthatareregarded

therliketheapproachused

byacharacterinLewisCarroll’sThroughtheLookingGlass:‘‘When

itmeansjustwhatIchooseittomean—neither

morenorless’’(Carroll,1871).Thevalueofusingtheterm

assessmentinphraseslikeassessmentforlearningandformative

assessmentisthatitisilluminatingtodrawattentiontothefact

thattheprocessesunderconsiderationcanbethoughtof

assessmentprocesses.

Forthisreason,itseemsmorehelpfultoacknowledgethatin

eachofthesecases,assessmentwasconductedwiththeintention

ofimprovinglearning(althoughthatmaynothavebeentheonly

reasonfortheassessment),andthattheevidencefromthe

sreason,

BlackandWiliamrestatedtheiroriginaldefinitioninaslightly

differentway,whichtheysuggestedwasconsistentwiththeir

originaldefinition,andthoseothersgivenabove,includingthatof

theAssessmentReformGroup:

Practiceinaclassroomisformativetotheextentthatevidence

aboutstudentachievementiselicited,interpreted,andusedby

teachers,learners,ortheirpeers,tomakedecisionsaboutthe

nextstepsininstructionthatarelikelytobebetter,orbetter

founded,thanthedecisionstheywouldhavetakeninthe

absenceoftheevidencethatwaselicited.(Black&Wiliam,

2009,p.9)

Athoroughexplorationoftheconsequenceofthisdefinitionis

beyondthescopeofthispaper,butonepointaboutthisdefinition

requiresclarifiainingthisdefinition,Blackand

Wiliammakeclearthattheterm‘‘instruction’’isusedinthesense

inwhichitisusedintheUnitedStates—thedesignoflearning

environments—andthe‘‘nextstepsininstruction’’canbetakenby

teachers,learners,ortheirpeers,oranycombinationofthesethree.

Thejustificationforeachofthecomponentsofthedefinitioncanbe

foundinBlackandWiliam(2009)andWiliam(2010),and

explorationsofhowformativeassessmentrelatestoother

theoreticalperspectivesonteachingandlearningcanbefound

inBlackandWiliam(2004,2011)andWiliam(2007).Inthefinal

sectionofthispaper,Iexploresomeoftheconditionsthatneedto

beinplaceforassessmenttosupportlearning.

esassessmentsupportlearning?

WhilethedefinitionproposedbyBlackandWiliamaboveis

relativelyprecise,itismuchmoreameansfordetermining

whetheranassessmenthas,infact,functionedformativelythanit

isaprescriptionforgeneratingassessmentsthatwill,orarelikely

to,eresearchstudiedabove,two

featuresappeartobeparticularlyimportantindesigningassess-

hattheevidencegenerated

is‘‘instructionallytractable’’(Wiliam,2007

).Inotherwords,the

evidenceismorethaninformationaboutthepresenceofagap

12

/StudiesinEducationalEvaluation37(2011)3–14

Teacher

Where the learner is going

Clarifying learning

intentions and sharing and

criteria for success

Understanding and sharing

learning intentions and

criteria for success

Understanding learning

intentions and criteria for

success

Where the learner is right now

Engineering effective

classroom discussions,

activities and tasks that elicit

evidence of learning

How to get there

Providing feedback that

moves learners forward

Peer

Learner

Activating learners as instructional

resources for one another

Activating learners as the owners of their own learning

sofformativeassessment(Wiliam&Thompson,2008).

dencemustalso

provideinformationaboutwhatkindsofinstructionalactivities

mple,alow

scoreonamathematicstestislikelytoindicatenothingmorethan

y

guidancethisprovidestotheteacheristhatmoreinstructionis

ssessmenthasbeendesignedtosupportvalid

inferencesaboutspecificaspectsofperformance,thentheteacher

mightalsorealizethatthestudentishavingparticulardifficulties

lowstheteachertofocusthe

remedialinstructionmorenarrowly,butprovideslittleinsightinto

whythestudentishavingdiffi,however,theassessment

revealsaspecificissue—forexamplethatthestudentbelievesthat

onlythesizeofthedenominatormatterswhencomparing

fractions(Vinner,1997)thenthisprovidesclearguidancefor

theteacheraboutwhatkindsofinstructionalactivitiestoprovide

forthelearner.

Thesecondrequirementisthatthelearnerengagesinactionsto

improvelearning;thismaybeundertakingtheremedialactivities

providedbytheteacher,askingapeerforspecifichelp,orreflecting

ondifferentwaystomoveherownlearningforward—afterall,the

r

words,feedbackcannotbeevaluatedwithoutalsotakinginto

accounttheinstructionalcontextinwhichitisprovided,andused.

Inthesamewaythatengineersdesignfeedbacksystemsrather

thansimplywaysofgeneratingdata,tounderstandfeedbackwe

tful

feedbackgiventostudentswhohavecometobelievethattheyare

‘‘nogood’’ataparticularsubjectislikelytobeignoredorrejected,

orappropriatedinsomeotherwaytoallowthelearnertopreserve

asenseofwell-being.

Theinvolvementoflearners,andtheirpeers,wasexplicitly

incorporatedbyWiliamandThompson(2008)intheirproposal

thatformativeassessmentcouldbeconceivedofasinvolvingthree

mainprocesses(identifyingwherelearnersareintheirlearning,

wheretheyaregoing,howtogetthere)exercisedbythree

categoriesofactors(teacher,learner,peer).

Theresultingmatrixofninecells,theysuggested,couldbe

organizedasfive‘‘keystrategies’’offormativeassessment,as

showninFig.1.

Whileeachofthesefive‘‘keystrategies’’hasgenerateda

substantialresearchbasisindividually(seeWiliam,2007,fora

summary)theycanalsobeviewedcollectivelyasstrategiesforthe

,BlackandWiliam(2009)

pointoutthatformativeassessmentinvolves‘‘thecreationof,and

capitalizationupon,‘momentsofcontingency’ininstructionfor

thepurposeoftheregulationoflearningprocesses’’(p.12).

Anyattemptattheregulationoflearningprocessesrequiresat

leastsomeideaofagoal,whetherthisisconceptualizedasasingle

learningdestination,orabroad‘‘horizon’’oflearninggoalsanyof

cher’sroleisthentoelicit

evidenceofachievement,andundertakeappropriateactionto

direct,this

formulation,theroleofpeersisanalogoustothatofteachers—while

peersmaylackthetrainingandexperienceofteachers,theyhave

uniqueinsightsintolearning,andbecausethepowerrelationships

betweenpeersaredifferentfromthosebetweenteachersand

students,therewillbeinstructionalstrategiesopentothemthat

wouldnotbeopen,orwouldbelesseffective,whenusedby

teachers.

Thefinalstrategy,‘‘Activatingstudentsasownersoftheirown

learning’’clearlydrawstogetheranumberofrelatedfieldsof

research,suchasmetacognition(Hacker,Dunlosky,&Graesser,

1998),motivation(Deci&Ryan,1994),attributiontheory(Dweck,

2000),interest(Hidi&Harackiewicz,2000)and,mostimportantly,

self-regulatedlearning,definedbyBoekaerts(2006)as‘‘amultilevel,

multicomponentprocessthattargetsaffect,cognitions,andactions,

aswellasfeaturesoftheenvironmentformodulationintheservice

ofone’sgoals’’(p.347).Whilemuchoftheresearchonself-

regulationhastendedtoprioritizeeithercognitiveormotivational

approaches,inrecentyearstherehavebeenseveralsignificant

attemptstodrawthesetwostrandsmorecloselytogether,because,

asBoekaerts(2006)argues,self-regulatedlearningisboth

metacognitivelygovernedandaffectivelycharged(p.348).

Boekaertshasproposedadeceptivelysimple,butpowerful,

modelforunderstandingself-regulatedlearning,termedthedual

processingtheory(

Boekaerts,1993).Inthemodel:

Itisassumedthatstudentswhoareinvitedtoparticipateina

learningactivityusethreesourcesofinformationtoforma

mentalrepresentationofthetask-in-contextandtoappraiseit:

(1)currentperceptionsofthetaskandthephysical,social,and

instructionalcontextwithinwhichitisembedded;(2)

activateddomain-specificknowledgeand(meta)cognitive

strategiesrelatedtothetask;and(3)motivationalbeliefs,

includingdomain-specificcapacity,interestandeffortbeliefs.

(Boekaerts,2006,p.349)

Dependingontheoutcomeoftheappraisal,thestudent

activatesattentionalongoneoftwopathways:the‘‘growth

pathway’’wherethegoalistoincreasecompetenceorthe‘‘well-

beingpathway’’whereattentionisfocusedonpreventingthreat,

heformerisobviouslypreferable,thelatteris

notnecessarilycounter-productive—byattendingtothewell-

beingpathway,thestudentmayfindawaytorestorewell-being

(forexamplebyloweringthecostoffailure)thatallowsashiftof

energyandattentiontothegrowthpathway.

Studentswhoarepersonallyinterestedinataskareobviously

likelytoactivateenergyalongthegrowthpathway,butwhere

studentsarenotpersonallyinterestedinatask,anumberof

featuresofthetask-in-contextmayneverthelesssparksituational

erationsofthetrade-offbetweentaskvalueand

costwillalsoinfl

particular,studentsaremorelikelytofocusongrowthratherthan

wellbeingwhentheyseeabilityasincrementalratherthanfixed

(Dweck,2000),whentheyhaveamasteryratherthana

performanceorientation(Dweck,2000)andwhentheyidentify

withthegoal(Deci&Ryan,1994).

/StudiesinEducationalEvaluation37(2011)3–1413

Tosummarize,becauselearningisunpredictable,assessmentis

necessarytomakeadaptiveadjustmentstoinstruction,but

assessmentprocessesthemselvesimpactthelearner’swillingness,

desire,andcapacitytolearn(Harlen&Deakin-Crick,2002).For

assessmenttosupportlearning,itmustprovideguidanceaboutthe

nextstepsininstructionandmustbeprovidedinwaythat

encouragesthelearnertodirectenergytowardsgrowth,rather

thanwell-being.

sion

Theideathatassessmentcansupportlearningisnotanewidea.

Itisinconceivablethatthoseinvolvedintheearliestattemptsto

communicateideas,skills,orpracticestoothersdidnotrealizethat

suchattemptscouldnotbeguaranteedtobesuccessful,andthat

effectiveinstructionthereforerequiredevaluation,andadjust-

r,itisonlyfortyyearssinceBenjaminBloomfirst

suggestedthatitmightbeusefulorilluminativetoexaminethese

ime,Bloomindicatedthatsuch

processeswouldbemoreeffectiveiftheywereseparatedfromthe

useofassessmenttorecordtheachievementoflearners,butforthe

nexttwentyyears,thedominantroleofassessmentwasseenas

therecordingofstudentachievement,althoughtherewerea

numberofattemptstouseevidencecollectedforthepurposeof

summarizingachievementinotherways,notablyforthe

r,itwasnotuntilthelate

1980sthattheideathatclassroomassessmentpracticescouldboth

affordandconstrainstudentlearningbegantogainwidespread

acceptance;usedappropriatelyassessmentcouldsubstantially

improvelearning,butthatmostofthetime,theimpactof

assessmentpracticeswastolimit,andeventoreduce,student

learning.

Duringthe1990s,anumberofstudiesexploredtheideathat

attentiontoassessmentasanintegralpartofinstructioncould

improvelearningoutcomesforstudents,andatthesametime,

attemptsweremadetoconnectclassroompracticetorelated

bodiesofresearch,notablyfeedback,motivation,attribution,and

tofthistime,theterm‘‘formative

assessment’’wasnotpreciselydefined,and,asaresult,research

studiesononeaspectoftheuseofassessmenttoimprove

instructionwereusedasevidencesupportingtheefficacyofquite

inresponsetothis,manyauthorsstopped

usingtheterm‘‘formativeassessment’’preferringinsteadthe

phrase‘‘assessmentforlearning’’althoughagainitsprecise

meaningwasrarelydefined,beyondtheideathatassessment

shouldbeusedduringinstructiontoimprovelearningoutcomes.

Thispaperhasreviewedthesedevelopments,anddescribed

morerecentattemptsthathavebeenmadetotheorizeformative

assessmentandassessmentforlearninginanumberofways,

specificallyintermsofclassroomstrategiesandpractical

techniquesthatteacherscanusetoimprovethequalityof

evidenceonwhichtheinstructionaldecisionsthey,andtheir

students,hereremainsmuchmoreworktobedone

tointegrateresearchonassessmentforlearningwithmore

fundamentalresearchoninstructionaldesign,feedback,self-

regulatedlearning,andmotivation,thereisnowastrongbodyof

theoreticalandempiricalworkthatsuggeststhatintegrating

assessmentwithinstructionmaywellhaveunprecedentedpower

toincreasestudentengagementandtoimprovelearningout-

comes.

Acknowledgement

Fig.1,whichwasfirstpublishedinLeahy,S.,Lyon,C.,

Thompson,M.,&Wiliam,D.(2005).Classroomassessment:

ionalLeadership,63(3),

18–24,isusedwithpermissionoftheEducationalTestingService.

References

Allal,L.,&Lopez,L.M.(2005).Formativeassessmentoflearning:Areviewofpub-

(Ed.),Formativeassessment:Improvinglearningin

secondaryclassrooms(pp.241–264).Paris,France:OrganisationforEconomic

CooperationandDevelopment.

Ausubel,D.P.(1968).Educationalpsychology:k,NY:Holt,

Rinehart&Winston.

Bangert-Drowns,R.L.,Kulik,C.-L.C.,Kulik,J.A.,&Morgan,M.(1991).Theinstructional

ofEducationalResearch,61(2),213–238.

Bangert-Drowns,R.L.,Kulik,J.A.,&Kulik,C.-L.C.(1991).Effectsoffrequentclassroom

lofEducationalResearch,85(2),89–99.

Bennett,R.E.(2009).Acriticallookatthemeaningandbasisofformativeassessment(ETS

RM-09-06).Princeton,NJ:EducationalTestingService.

Black,H.(1986).l(Ed.),Assessingeducational

achievement(pp.7–18).London:FalmerPress.

Black,P.J.,&Wiliam,D.(1998a).mentin

Education:Principles,PolicyandPractice,5(1),7–74.

Black,P.J.,&Wiliam,D.(1998b).Insidetheblackbox:Raisingstandardsthrough

taKappan,80(2),139–148.

Black,P.J.,&Wiliam,D.(2004).Theformativepurpose:assessmentmustfirstpromote

on,M.(Ed.).Towardscoherencebetweenclassroomassessmentand

accountability:103rdYearbookoftheNationalSocietyfortheStudyofEducation

.

(part

2)(,pp.20–50).Chicago,IL:UniversityofChicagoPress

Black,P.J.,&Wiliam,D.(2009).Developingthetheoryofformativeassessment.

EducationalAssessment,EvaluationandAccountability,21(1),5–31.

Black,P.,&Wiliam,D.(2011)..

Gardner(Ed.),Assessmentandlearning(2nded.).London,UK:Sage.

Black,P.,Harrison,C.,Lee,C.,Marshall,B.,&Wiliam,D.(2004).Workinginsidetheblack

box:taKappan,86(1),8–21.

Bloom,B.S.(1984a).Thesearchformethodsofinstructionaseffectiveasone-to-one

ionalLeadership,41(8),4–17.

Bloom,B.S.(1984b).The2-sigmaproblem:Thesearchformethodsofgroupinstruc-

ionalResearcher,13(6),4–16.

Boekaerts,M.(1993).ional

Psychologist,28(2),149–167.

Boekaerts,M.(2006).ger&I.E.

Sigel(Eds.),Handbookofchildpsychologyvolume4:Childpsychologyinpractice(6th

ed.,pp.345–377).NewYork,NY:Wiley.

Broadfoot,P.M.,Daugherty,R.,Gardner,J.,Gipps,C.V.,Harlen,W.,James,M.,etal.

(1999).Assessmentforlearning:dge,UK:Universityof

CambridgeSchoolofEducation.

Broadfoot,P.M.,Daugherty,R.,Gardner,J.,Harlen,W.,James,M.,&Stobart,G.(2002).

Assessmentforlearning:dge,UK:UniversityofCambridge

SchoolofEducation.

Brookhart,S.M.(2004).Classroomassessment:Tensionsandintersectionsintheory

rsCollegeRecord,106(3),429–458.

Brookhart,S.M.(2007).Expandingviewsaboutformativeclassroomassessment:A

an(Ed.),Formativeclassroomassessment:

Theoryintopractice(pp.43–62).NewYork,NY:TeachersCollegePress.

Brophy,J.(1981).Teacherpraise:ofEducationalResearch,

51(1),5–32.

Carroll,L.(1871).,UK:

Macmillan.

Cowie,B.,&Bell,B.(1999).Amodelofformativeassessmentinscienceeducation.

AssessmentinEducation:Principles,PolicyandPractice,6(1),32–42.

Crooks,T.J.(1988).of

EducationalResearch,58(4),438–481.

Deci,E.L.,&Ryan,R.M.(1994).navian

JournalofEducationalResearch,38(1),3–14.

Dempster,F.N.(1991).ional

Leadership,48(7),71–76.

Dempster,F.N.(1992).Usingteststopromotelearning:Aneglectedclassroom

lofResearchandDevelopmentinEducation,25(4),213–217.

Dweck,C.S.(2000).Self-theories:Theirroleinmotivation,personalityanddevelopment.

Philadelphia,PA:PsychologyPress.

Elshout-Mohr,M.(1994).anEducation,26(2),58–73.

Fuchs,L.S.,&Fuchs,D.(1986).Effectsofsystematicformativeevaluation—Ameta-

ionalChildren,53(3),199–208.

Guskey,T.R.(2010).Formativeassessment:

e&(Eds.),Handbookofformativeassessment(pp.106–124).

NewYork,NY:Taylor&Francis.

Hacker,D.J.,Dunlosky,J.,&Graesser,A.C.(Eds.).(1998).Metacognitionineducational

,NJ:LawrenceErlbaumAssociates.

Harlen,W.,&Deakin-Crick,R.(2002).Asystematicreviewoftheimpactofsummative

assessmentandtestsonstudents’-Centre(Ed.),

Researchevidenceineducationlibrary(1.1ed.,pp.153–).London,UK:Universityof

LondonInstituteofEducationSocialScienceResearchUnit.

Hattie,J.(1999,August).Inflvedfrom:www.

/uoa/education/staff//papers/infl

Hattie,J.,&Timperley,H.(2007).Thepoweroffeedback.

ReviewofEducationalResearch,

77(1),81–112.

/StudiesinEducationalEvaluation37(2011)3–14

Hidi,S.,&Harackiewicz,J.M.(2000).Motivatingtheacademicallyunmotivated:A

ofEducationalResearch,70(2),151–179.

James,M.(1992,April).resentedattheAnnual

ConferenceoftheAssociationforSupervisionandCurriculumDevelopment

(Assemblysessionon‘CritiqueofReformsinAssessmentandTestinginBritain’)

heldatNewOrleans,LA.

Kahl,S.(2005).Whereintheworldareformativetests?Rightunderyournose!.

EducationWeek,25(September(4)),11.

Klenowski,V.(2009).Editorial:Assessmentforlearningrevisited:AnAsia-Pacific

mentinEducation:Principles,Policy,andPractice,16(3),263–268.

Kluger,A.N.,&DeNisi,A.(1996).Theeffectsoffeedbackinterventionsonperformance:

Ahistoricalreview,ameta-analysis,andapreliminaryfeedbackintervention

logicalBulletin,119(2),254–284.

Ko

¨

ller,O.(2005).Formativeassessmentinclassrooms:Areviewoftheempirical

(Ed.),Formativeassessment:Improvinglearningin

secondaryclassrooms(pp.265–279).Paris,France:OrganisationforEconomic

CooperationandDevelopment.

Lewis,C.C.(2002).Lessonstudy:Ahandbookofteacher-ledinstructionalchange.

Philadelphia,PA:ResearchforBetterSchools.

Looney,J.(Ed.).(2005).Formativeassessment:Improvinglearninginsecondaryclassrooms.

Paris,France:OrganisationforEconomicCooperationandDevelopment.

Meisels,S.J.,Atkins-Burnett,S.,Xue,Y.,Bickel,D.D.,&Son,S.-H.(2003).Creatinga

systemofaccountability:Theimpactofinstructionalassessmentonelementary

children’ionPolicyAnalysisArchives11(9).

Mitchell,R.(1992).k,USA:FreePress,Macmillan.

Natriello,G.(1987).ional

Psychologist,22(2),155–175.

Nyquist,J.B.(2003).Thebenefitsofreconstruingfeedbackasalargersystemofformative

assessment:ishedMasterofSciencethesis,Vanderbilt

University.

Parkhurst,H.(1922).,UK:GBellandSonsLtd.

Perrenoud,P.(1998).Fromformativeevaluationtoacontrolledregulationoflearning.

TowardsawiderconceptualfimentinEducation:Principles,Policyand

Practice,5(1),85–102.

Ramaprasad,A.(1983).OnthedefiouralScience,28(1),4–13.

Reiser,R.A.(1986).Instructionaltechnology:

´

(Ed.),

Instructionaltechnology:Foundations(pp.11–48).Hillsdale,NJ:Lawrence

ErlbaumAssociates.

Rheinberg,F.(1980).ementevaluation

andlearningmotivation].Go

¨

ttingen,Germany:Hogrefe.

Sadler,D.R.(1989).Formativeassessmentandthedesignofinstructionalsystems.

InstructionalScience,18,119–144.

Shepard,L.A.,Hammerness,K.,Darling-Hammond,L.,Rust,F.,Snowden,J.B.,Gordon,

E.,etal.(2005).g-Hammond&ord(Eds.),Preparing

teachersforachangingworld:Whatteachersshouldlearnandbeabletodo(pp.275–

326).SanFrancisco,CA:Jossey-Bass.

Shute,V.J.(2008).ofEducationalResearch,78(1),

153–189.

Stiggins,R.J.(2005).FromformativeassessmenttoassessmentFORlearning:Apathto

taKappan,87(4),324–328.

Sutton,R.(1995).d,UK:RSPublications.

Torrance,H.,&Pryor,J.(1998).gham,UK:

OpenUniversityPress.

Tunstall,P.,&Gipps,C.(1996a).‘Howdoesyourteacherhelpyoutomakeyourwork

better?’Children’sunderstandingofformativeassessment.

TheCurriculumJournal,

7(2),185–203.

Tunstall,P.,&Gipps,C.V.(1996b).Teacherfeedbacktoyoungchildreninformative

assessment:hEducationalResearchJournal,22(4),389–404.

Vinner,S.(1997).Fromintuitiontoinhibition—Mathematics,educationandother

en(Ed.),InProceedingsofthe21stconferenceofthe

internationalgroupforthepsychologyofmathematicseducation,Vol.1(pp.63–78).

Lahti,Finland:UniversityofHelsinkiLahtiResearchandTrainingCentre.

Wiener,N.(1948).Cybernetics,orcontrolandcommunicationintheanimalandthe

k,NY:JohnWiley&SonsInc.

Wiliam,D.(2007).Keepinglearningontrack:Classroomassessmentandtheregulation

,Jr.,(Ed.).Secondhandbookofmathematicsteachingand

learning(pp.1053–1098).Greenwich,CT:InformationAgePublishing.

Wiliam,D.(2010).Anintegrativesummaryoftheresearchliteratureandimplications

e&(Eds.),

Handbookofformativeassessment(pp.18–40).NewYork,NY:Taylor&Francis.

Wiliam,D.,&Thompson,M.(2008).Integratingassessmentwithinstruction:Whatwill

ittaketomakeitwork?(Ed.),Thefutureofassessment:Shaping

teachingandlearning(pp.53–82).Mahwah,NJ:LawrenceErlbaumAssociates.

Wiliam,D.,Lee,C.,Harrison,C.,&Black,P.J.(2004).Teachersdevelopingassessmentfor

learning:mentinEducation:Principles,Policy

andPractice,11(1),49–65.