Académique Documents
Professionnel Documents
Culture Documents
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
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.
Taylor & Francis, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to The Journal of Educational
Research.
http://www.jstor.org
This content downloaded from 142.244.11.244 on Fri, 18 Sep 2015 00:58:11 UTC
All use subject to JSTOR Terms and Conditions
Processes
ofSixth-Grade
TheRevising
WithandWithout
Writers
PeerFeedback
VICKI L. BRAKEL OLSON
AugsburgCollege
The primary
ABSTRACT
purposeof thisstudywas to
examinethe effectof peer feedbackon the qualityof student
and the amountand kindof revisionbehavior.Ninetywriting
threesixthgradersin sixintactclassroomswroteand revisedsix
stories,thelastone beingused as data forthisstudy.Instruction
variedacrossgroupsin the followingmanner:RI/PP students
and revisedstorieswitha peer; PP
receivedrevisioninstruction
studentsrevisedstorieswitha peer but did not receiverevision
but revised
RI studentsreceivedrevisioninstruction
instruction;
nor
storiesalone; C studentshad neitherrevisioninstruction
helpfrompeers.Chi-squareanalysisindicatedthatrevisionbehaviorwas influenced
analysis
Qualityof writing
by instruction.
acrossgroupson bothroughand
differences
revealedsignificant
finaldrafts.Peer feedbackseemedto help studentswriteinilinkedto
tiallysuperiorroughdraftsbut was not consistently
of contentbetweenroughand finaldrafts.Sucimprovement
cessfulsurfacestructure
editingoccurredwithor withoutpeer
feedback.
to VickiL. BrakelOlson,4732ColAddresscorrespondence
MN 55409.
faxAvenue,S., Minneapolis,
22
This content downloaded from 142.244.11.244 on Fri, 18 Sep 2015 00:58:11 UTC
All use subject to JSTOR Terms and Conditions
23
In thisstudyI assumedthatyoungwriters
willhave revisionproblemsthat are unique to theirdevelopmental
and experientiallevels. Specifically,those writerswill
have difficultyrecognizingaudience need and consciouslycontrollingtheiruse of revisionstrategies.The
researchjust reportedsuggestedtwo instructional
stratclassroomto increasereviegiesforuse in the elementary
sion behaviorsand improvequalityof writing:(a) peer
feedbackand (b) directinstructionin specificrevision
strategies.
Overviewof Study
Purpose
Classroomteachers,beingpragmatists,
recognizerevision as a potentialally in theirstruggleto help children
become successfulwriters.In the upper elementary
gradeswhereexpectationsforqualityare oftenhigh,revisionbecomesan especiallyattractive
skill.
This studyexploredthe effectsof two instructional
on the revisionbehaviorand qualityof writing
strategies
of sixth-grade
students.Researchquestionsguidingthis
wereas follows:
investigation
1. Willtypeand amountof revisionbehaviorvarysignifinstructional
situations?
icantlyacrossfourdifferent
2. Will quality of writingvary significantly
across instructional
situations?
3. Will quality of writingvary significantly
between
situaroughand finaldraftswithineach instructional
tion?
Subjects
The subjectsin this studywere 93 sixthgradersfrom
fourdifferent
schools withinthe same middle-classsuburban school district.Of the 93 students,49 were girls
and 44 were boys. The subjectswere membersof four
heterogeneousinstructionalgroups, one group per
school- each group being an intact classroom. Each
classroomwas informally
groupedby classroomteachers
at thebeginningof the schoolyearto includetheexisting
in each school; no attempt
rangeof abilitiesrepresented
was made to isolate specificabilitygroups withinthe
classrooms.All the studentswithineach class received
but not all the studentswere
processwriting
instruction,
partof the finalanalysis.I eliminatedsome studentsbecause theywereeitherfifth-grade
childrenin a combination classroom,or they were sixth-grade
studentswith
incompletesetsof writing
samples.
Teachers
All of the participating
teachersvolunteeredto be a
part of the study.The teachershad a specialinterestin
and weremotiimprovingtheirown writinginstruction
vatedto trythe materialsused in thisstudy.All werevet-
This content downloaded from 142.244.11.244 on Fri, 18 Sep 2015 00:58:11 UTC
All use subject to JSTOR Terms and Conditions
Journalof EducationalResearch
24
Procedures
This studywas conductedoverAVimonths
Instruction.
underroutineclassroomconditions.Classroomteachers
Studentsin two of
carriedout the prescribedinstruction.
in theuse of thefive
thegroupsreceiveddirectinstruction
revisiontactics prior to startingthe autobiographical
occurredapproximately
writinglessons.That instruction
twicea week forabout 1 month.Duringthattime,the
other groups continuedwith theirprescribedlanguage
artsprogram,whichfocusedon grammarand combining
sentences.Then all studentsparticipatedin the six process-basedwritinglessons. Instructionduringthe writing
lessonsvariedonlyat the pointof revision.The fourinsituationscan be summarizedas follows:
structional
Revision instruction/peer
partners(RI/PP): Instrucwas givenpriorto the
tion in specificrevisionstrategies
lessons.Duringthe writuse of six process-basedwriting
to respondto
inglessons,studentsmetwithpeerpartners
and reviseroughdrafts.
was
Peer partneronly (PP): No revisioninstruction
with
the
students
met
the
lessons,
writing
given.During
peerpartnersto respondto and reviseroughdrafts.
Revisioninstruction
only (RI): Instructionin specific
was givenpriorto the writinglessons.
revisionstrategies
During the writinglessons, studentsworked alone to
evaluateand revisetheirroughdrafts.
Control: Those studentscompletedthe same writing
nor
lessons,but theydid not receiverevisioninstruction
did theyworkwitha peer partnerto revisetheirrough
weregiventimebut no help in
drafts.Those participants
lessons.
duringthe writing
revising
Because peer collaborationwas integralto two of the
treatmentgroups and because of the durationof the
on the basis of
study,I attemptedto assigntreatments
their
teachers
how the participating
normallystructured
used
who
teachers
Those
classrooms.
peer colregularly
were
instruction
of
their
laborationas a part
assignedto
that
situations
instructional
incorporatedpeer collaboranot
tion. Treatments
requiringpeer collaborationwere
who
tendedto use wholegroupdisto
teachers
assigned
than peer collaboration.So
rather
cussion techniques
was giventhebestpossiblechanceto
thateach treatment
in
of treatment
succeed,I overruledrandomassignment
favorof mymorepragmaticmethod.I hoped thatby my
action teacher cooperation and enthusiasmwould be
the courseof the study.
maintainedthroughout
Data collection.Data consistedof the roughdraftsof
lesson6, collectedand copied beforeand afterthe formal
revisingsessionand the finaldraftsof the same lesson.
Those draftswereused in the analysisof revisionbehavior and writingquality.I keptsession6 as similarto the
previoussessions as possible and returnedall original
workto the studentsaftercopying.
This content downloaded from 142.244.11.244 on Fri, 18 Sep 2015 00:58:11 UTC
All use subject to JSTOR Terms and Conditions
Analysis
Revisionbehavior.The revisioncategorysystemin this
study was adapted from a systemused by Bridwell
(1980). That systemconsideredthe syntacticlevels of
contentof text:
revisionbehavioraffecting
Single-word
level("Run" is changedto "race.")
Multiple-word
level("In frontof myhouse" is deletedfroma sentence.)
Sentencelevel("I loved myhorseverymuch" is added to a paragraph.)
Multiple-sentence
level ("Grandma and Grandpa
werewaitingfor us when we arrived.They had supper
readyand our beds all made. Boy were we glad to see
fromthe beginningof the storyto
them"are rearranged
a spotnearthe middle.)
That systemalso consideredtypes of revisionbehavior-mechanics, spelling,additions,deletions,substitutions, and rearrangements
(referredto as an "order
shift" by Bridwell, 1980). All revisiontypes (except
mechanicsand spelling)could occurat anyof the syntacticlevels.
Revisioncategorizationwas done twice for each student.First,the revisedroughdraftwas read, and all apparentrevisionswere classifiedas to typeand syntactic
level. Second, the roughand finaldraftsof each paper
were compared,word by word. All changes between
roughand finaldraftswereclassifiedand talliedby category.The revisionbehaviorsforeach setof roughand final draftswere counted and recorded.Followingthis
procedure,I calculatedrevisionsper 100 wordsby categoryforeach student.The numbersof studentsfalling
withinspecificrangesof revisionsper 100wordsweredetermined.
Chi-squaretestswerethenused to analyzeeach
typeof revisionbehavior.
on raterreliability,
As a cross-check
a randomsample
of 10% of the paperswas categorizedby a second rater
trainedto use thecategorysystem.Of 344 totalrevisions,
both ratersnoted 266 revisions,a resultof 77% agreement.Of the266 revisionsnotedbybothraters,242 were
bythemin thesamewayfor91% agreement.
categorized
to
note
revisionbehaviorswas a problemforboth
Failing
raters.Those resultsindicatedthatnumbersof revisions
in the analysisof reviwerelikelyto be underestimated
sion behaviorbut that the categorizationsystemitself
could be consistently
applied.
The scale used in thisstudyto anaof
writing.
Quality
lyzewriting
qualitywas an adaptationof Cooper's (1977)
Personal NarrativeWritingScale. Because the current
of studentsyoungerthanthose
studyinvolvedthewriting
for whom this scale was originallydeveloped, minor
modifications
werenecessary.
The PersonalNarrativeWritingScale considersboth
rhetoricaland surface structurequalities. Rhetorical
audienceconsidqualityis dividedintosix subcategories:
overerations,voice,centralfigure,setting/background,
25
Treatment
groups
RI/PP
PP
RI
C
All
items
Rhetorical
items
.88
.91
.90
.89
Surfacestructure
items
.95
.90
.96
.93
.86
.77
.77
.67
Among
Within
Total
SS
4,906.09
39,554.90
44,460.99
df
3
81
84
MS
1,635.36
488.33
3.35
.02*
This content downloaded from 142.244.11.244 on Fri, 18 Sep 2015 00:58:11 UTC
All use subject to JSTOR Terms and Conditions
26
Journalof EducationalResearch
Groups
1-9
RI/PP
PP
RI
C
16
16
4
7
Content
10 or more
7
8
20
11
Surfacestructure
1-9
10 or more
23
24
24
18
17
15
19
7
23
26
24
18
6
11
5
11
Results
RevisionBehavior
Amount. Chi-squareanalysisof all contentrevisions
(additions, deletions, substitutions,rearrangements)
difmade on roughand finaldraftsindicatedsignificant
ferencesacross groups. The majorityof studentswho
workedwithpeerpartnersrevisedthe content9 or fewer
timesper 100 words,whereasthe majorityof students
who workedwithoutpeer partnersrevisedcontent10 or
moretimesper 100 words(see Table 3). Studentsin the
RI group seemed most discrepantbecause considerably
morestudentsthanexpectedfellintothe upperrangeof
revisions
per 100words(see Table 3).
The resultsfor surfacestructurerevision(mechanics,
spelling)showedonlytheC groupto havethemajorityof
itsstudentsin theupperrangeof revisionsper 100words
(see Table 3). Thus, the majorityof the studentswho
claimedtheupperfrequency
workedwithoutpeerpartners
rangeof revisionsper 100 words, whereasRI students
dominatedin contentrevisionsand C studentsled in surface structurerevisions.Studentsin both peer-partner
fortheirconsistenttendencyto fall
groupswerestriking
into the lower frequencyranges of revisionsper 100
wordsacrossall typesof revisionbehavior.
Type.All groupsshoweda majorityof studentsmaking more contentrevisionsthan surfacestructurerevisions,buttheRI groupshowedthispatternmoststrongly
syntac(see Table 4). Acrossall groups,thepredominant
tic levelforcontentrevisionswas wordlevel(67% of all
contentrevisions).
used revisiontypesfor RI/PP,
The most frequently
PP, and C studentsweremechanicalrevisionsand substiand additionspretutions.For RI students,substitutions
dominated.Between92 to 99% of all students,regardless
of group,made substitutions,
additions,and deletionsin
theirstories.Frequencyof use seemedto be affectedby
instructional
situation,but the abilityto revise using
- withor withoutinstruction.
thosetacticsexisted
Time of revising.The instructional
groupsalso were
timeof revising,that is, becontrastedby predominant
tweendraftsor duringfinaldrafting
(see Table 5). Once
again,the RI groupstood apart fromthe restwithonly
Groups
Content
Surfacestructure
RI/PP
PP
RI
C
13
15
22
11
10
9
1
6
23
24
23
17
Rough draft
Final draft
7
6
20
8
14
19
4
8
21
25
24
16
at p < .05.
Note, x2 = 19.62, df = 3, p < .01. /?judged significant
This content downloaded from 142.244.11.244 on Fri, 18 Sep 2015 00:58:11 UTC
All use subject to JSTOR Terms and Conditions
27
1990[Vol.84(No.1)]
September/October
instructional
situation,leadingone to concludethatreviused in thisstudydid not buildskillsthat
sioninstruction
Howinstruction.
wereuniqueto thosestudentsreceiving
those
tactics
received
that
the
may have
ever,
emphasis
caused the RI studentsto revisecontentmorefrequently
thantheotherstudentsdid.
Source
MS
3,462.54
359.95
9.62
.01
8.96
.01
df
Rough draft
3
83
86
10,387.62
29,875.58
40,263.20
Among
Within
Total
Writing
Quality
Final draft
effects.Analysisof covarianceindicated
Across-group
acrossgroupsin qualityof writing
differences
significant
scoreson both roughand finaldrafts.For total quality
scores,studentsin theRI/PP groupswroteroughand final draftsof significantly
higherqualitythan all other
groups(see Tables 6 and 7). Studentsin the PP group
wroteroughand finaldraftsthatrankedsecondin quality.
When rhetoricalqualityscores were consideredseparately,roughand finaldraftsof the RI/PP groupswere
superiorto those of all othergroups
again significantly
groupsshowedno
(see Table 8). Althoughthe remaining
rhetoricalquality,
in rough-draft
differences
significant
superiorto
by finaldraftsthe PP groupwas significantly
theRI group.
On the roughdraft,the C group scored significantly
lowerthanall othergroupson surfacestructure
quality,
lower than the
and the RI group scored significantly
RI/PP group(see Table 8). But by thefinaldraft,no sigwerefoundin surfacestructure
differences
nificant
quality.
effects.Rough and final draftquality
Within-group
scoresforstudentsin each instructional
groupwerecomt
The
results
of thoset tests
tests.
paredusingdependent
in Table 9.
are summarized
Those resultsconfirmedthat studentsin all groups
wereable to significantly
improvethe surfacestructure
drafts.Accordingto that
stories
across
their
of
quality
the
process,whereovertconcern
finding, multiple-draft
issuesis postponeduntillaterin the
forsurfacestructure
drafting
process,was comfortableforthose studentsreof
gardless group.
In contrastto surfacestructure
revising,only the PP
able
to
improverhetoricalqualsignificantly
groupwas
showedthatthose sturevision
behavior
ity.Analysisof
in
the
lower
fell
dentsconsistently
rangeof revisionsper
of
revision.
all
100 words for
Yet, theirrevising
types
withsome success
been
musthave
purposeful,targeting
the most.
theportionsof textthatneededrevising
RI/PP studentsdid not significantly
improverhetorical quality.At leasttwo factorsmayhave been operating
in this study.First,the RI/PP teacherevaluatedeach
storyas heavilyformechanicalaccuracyas forcommunicativequality.Students,accustomedto theteacher'shigh
foraccuracy,mayhave focusedmoreattenexpectations
and lesson improving
surfacestructure
tionon perfecting
contentas theyrevised.Second, RI/PP studentshad alstoriesof superiorquality.The contentof
readywritten
3
83
86
8,784.80
27,114.10
35,898.90
Among
Within
Total
2,928.27
326.68
Groups
Covariate
M
SD
Dependent
M
SD
Adjusted
M
Rough draft
RI/PP
PP
RI
C
34.65
35.87
38.17
34.28
5.40
5.81
5.31
6.18
RI/PP
PP
RI
C
34.65
35.87
38.17
34.28
5.40
5.81
5.31
6.18
112.87
98.78
94.63
82.78
18.16
26.81
17.08
18.41
114.61
98.76
91.26
85.07
21.74
19.09
21.56
23.25
119.71
106.79
92.40
101.16
Final draft
117.26
106.83
97.13
97.94
Draft
df
Rough
3, 83
8.39
.02
Final
3, 83
12.24
.01
Rough
3, 83
7.61
.03
Final
3, 83
1.11
.35
Rhetoricalquality
Set 1: RI/PP (75.75)
Set 2: PP (62.68); RI (56.30);
C (55.05)
Set 1: RI/PP (78.85)
Set 2: PP (68.32); C (62.92)
Set 3: C; RI (54.76)
Surfacestructurequality
Set 1: RI/PP (38.87); PP
(36.08)
Set 2: PP; RI (34.96)
Set 3: C (30.02)
NSD*
at p < .05.
Note, p judged significant
^Adjustedgroupmeans: RI/PP = 40.83; PP = 38.47; RI = 37.68; C
= 37.76.
This content downloaded from 142.244.11.244 on Fri, 18 Sep 2015 00:58:11 UTC
All use subject to JSTOR Terms and Conditions
28
Journalof EducationalResearch
Group
RI/PP
Total
Rhetorical
Surfacestructure
PP
Total
Rhetorical
Surfacestructure
RI
Total
Rhetorical
Surfacestructure
C
Total
Rhetorical
Surfacestructure
df
22
22
22
1.58
1.11
1.77
.06
.14
.04*
25
25
25
2.20
1.81
2.22
.02*
.04*
.02*
23
23
23
.71
-.14
2.29
.25
.44
.02*
18
18
18
2.2
1.16
4.13
.02*
.13
.05*
Judgedsignificant
at/7 < .05.
This content downloaded from 142.244.11.244 on Fri, 18 Sep 2015 00:58:11 UTC
All use subject to JSTOR Terms and Conditions
29
REFERENCES
Applebee,A. N. (1986). Problemsin processapproaches:Towarda reof processinstruction.
In A. R. Petrosky& C. Barconceptualization
tholomae(Eds.), The teachingof writing:
85th Yearbookof theNational Societyfor the Study of Education. Part II, Chicago, IL:
NSSE.
Ban, M. (1983). The new orthodoxyabout writing:Confusingprocess
and pedagogy.LanguageArts,60(7). 829-840.
E. J. (1982). Learningto revise:Some componentprocesses.In
Bartlett,
MartinNystrand(Ed.). What writersknow. New York: Academic
Press.
Bereiter,C. (1980). Developmentin writing.In L. W. Gregg& E. R.
Steinberg(Eds.), Cognitiveprocessesin writing.
Hillsdale,NJ: LawrenceErlbaumAssociates.
Birnbaum,J. C. (1982). The readingand composingbehaviorof selectedfourthand seventhgradestudents.Researchin the Teachingof
English,76(3), 241-260.
in twelfth
Bridwell,L. (1980). Revisingstrategies
gradestudents'transactionalwriting.
Researchin the Teachingof English,74(3), 197-222.
Burtis,P. J., Bereiter,
C, Scardamalia,M., Stero,J. (1983). The develIn B. Kroll& G. Wells(Eds.), Exploopmentof planningin writing.
rationsin thedevelopment
New York: Wiley.
of writing.
Calkins, L. (1980). Children'srewriting
strategies.Research in the
Teachingof English,14(6).
Calkins,L. (1983). Lessonsfroma child.Exeter,NH: Heinemann.
Carter,R. (1982). By itself,peer group revisionhas no power. ERIC
DocumentReproductionServiceNo. 226 350
interCohen,E., & Scardamalia,M. (1983). The effectsof instructional
ventionin the revisionof essaysby grade six children.Paper presentedat the annual meetingof the AmericanEducationalResearch
Association,Montreal.
Cooper, C. R. (1977). Holisticevaluationof writing.In C Cooper &
L. Odell (Eds.), Evaluatingwriting:
Describingmeasuring,
judging.
Urbana,IL: NationalCouncilof Teachersof English.
This content downloaded from 142.244.11.244 on Fri, 18 Sep 2015 00:58:11 UTC
All use subject to JSTOR Terms and Conditions