Journal of Educational Psychology1998, Vol. 90, No. 4,715-729

Copyright 1998 by the American Psychological Association, Inc.0022-0663/98/53.00

Young Children's Self-Regulated Learning and Contexts That Support It

Nancy E. PerryUniversity of Michigan

Writing and portfolio activities provided a context for examining relations between classroomcontexts and young children's self-regulated learning (SRL). Data collection spanned 6months and included weekly visits to Grade 2 and 3 classrooms during regularly scheduledwriting and portfolio activities. Data included teacher questionnaires and observations andstudent questionnaires, observations, and interviews. Young children deliberated about how toregulate writing and demonstrated either mastery or performance orientations as a function ofclassroom-specific tasks, authority structures, and evaluation practices. Findings supportsociocognitive models of learning regarding how classroom contexts affect students' beliefs,values, expectations, and actions. Also they challenge assumptions that young children lackthe cognitive sophistication required for SRL and do not adopt motivational orientations thatundermine it.

Contemporary writings on school reform emphasize theneed for students to develop skills and attitudes for self-regulated lifelong learning (Palincsar & Klenk, 1992; Paris,Lawton, & Turner, 1992). The term self-regulated is typi-cally associated with learners who are metacognitive, intrin-sically motivated, and strategic (Zimmerman, 1990). Thesestudents are aware of task demands and their abilities tomeet them; they have high efficacy for learning and attributeoutcomes to factors under their control (e.g., effort andstrategy use); and they have a repertoire of effective learningand problem-solving strategies that they apply appropriately.The term self-regulated also is used to contrast an approachto learning that has been referred to as defensive (Paris &Newman, 1990) and self-handicapping (Covington, 1992).Students who adopt this latter approach typically have lowefficacy for learning and avoid failure and damage toself-esteem by seeking easy tasks, procrastinating, or avoid-ing work altogether. Not surprisingly, this approach tolearning impedes academic success (Winne, 1995; Zimmer-man &Bandura, 1994).

The preponderance of research on academic self-regulation has involved students in upper elementary gradesthrough college. Although studies have documented youngchildren's use of strategies to complete simple tasks (i.e.,memory and vocabulary learning), most self-regulated learn-ing (SRL) theorists assume that children under age 10 havedifficulty coordinating the cognitive and metacognitiveprocesses required to complete complex, multifaceted tasks,such as writing a research report (Pressley, Forrest-Pressley,

Nancy E. Perry, School of Education, University of Michigan.This article reports findings from Nancy E. Perry's doctoral

dissertation research. An earlier version of this article was pre-sented at the annual meeting of the American Educational ResearchAssociation in March 1997.

Correspondence concerning this article should be addressed toNancy E. Perry, who is now at the Department of EducationalPsychology and Special Education, University of British Colum-bia, 2125 Main Mall, Vancouver, British Columbia, Canada V6T1Z4. Electronic mail may be sent to nancy.perry@ubc.ca.

Elliott-Faust, & Miller, 1985; Winne, 1997; Zimmerman,1990). Also, theory and research concerning motivationsuggest that young children are protected from defensivemotivational patterns that undermine SRL by their tendencyto view ability in incremental terms, rate their ability highly,and expect to do well so long as they exert effort (Cain &Dweck, 1995; Paris & Newman, 1990). However, recentstudies have documented negative affect and nonpersistencein the face of failure, attributions linking outcomes to ability,and low expectations for future success in children age 4—7years (e.g., Cain & Dweck, 1995; Stipek, Feiler, Daniels, &Milbum, 1995; Turner, 1995). These findings support at-tempts to discover ways of encouraging students to adoptskills and attitudes associated with SRL early in their schoolcareers rather than waiting until the intermediate grades,when attitudes and actions associated with academicallyineffective behaviors are well formed.

Studies of instructional contexts offer insights concerningthis problem. For example, Stipek et al. (1995) comparedchildren age 4-6 years in child-centered classrooms withchildren in teacher-directed classrooms in terms of theirself-perceptions, expectations for success, preference forchallenge, and independence. In child-centered classrooms,children had choices within a set of diverse activities andmaterials in an atmosphere that encouraged peer interaction.Teachers embedded instruction in activities that were mean-ingful to students and were characterized as nurturing,respectful, and responsive in their interactions with children.In teacher-directed classrooms, learning was predominantlyteacher controlled and focused on basic skills presentedapart from authentic activities. There was an emphasis onperformance, and teachers used external evaluations andrewards and made social comparisons. Children in theteacher-directed classrooms rated their abilities significantlylower tban students in child-centered classrooms, had lowerexpectations for success on academic tasks, chose simplertasks, and were more dependent on adults to define tasks andevaluate outcomes. Also, these students evidenced less pride

715

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

716 PERRY

in their accomplishments and worried more about schoolthan students in child-centered classrooms.

These findings support sociocognitive models of learningthat prompt researchers and teachers to consider howclassroom task structures, authority structures, and evalua-tion practices influence students' beliefs about themselves aslearners, their goals and expectations, and their decisionsabout how to regulate their behavior in school (Pintrich,Marx, & Boyle, 1993). Proponents of these models suggestthat to develop effective forms of SRL, students need to beinvolved in complex meaningful tasks, choosing the prod-ucts and processes that will be evaluated, modifying tasksand assessment criteria to attain optimal challenge, obtain-ing support from peers, and evaluating their own work (Paris& Ayres, 1994; Pintrich et al., 1993; Turner, 1995). Also theysuggest that SRL is more likely to develop when teachersguide rather than direct students' learning, acting as facilita-tors or coaches rather than managers (Brown & Campione,1994). In this role, teachers provide instrumental support tostudents. They carefully orchestrate instruction to ensurethat students acquire the domain and strategy knowledgethey need to operate independently, help them make appro-priate choices, encourage them to expand their developingabilities by attempting challenging tasks, and use nonthreat-ening evaluation practices that encourage students to focuson personal progress and interpret errors as opportunities tolearn (Moll & Whitmore, 1993; Paris & Ayres, 1994).

To date, very few studies have investigated SRL or itscorrelates (e.g., intrinsic motivation) in naturalistic settings(i.e., classrooms). More typically, studies of SRL rely onstudents' self-reports without regard for contextual factors,or they design interventions that provide optimal but decon-textualized environments for SRL to occur. Stipek et al.'s(1995) findings and sociocognitive models of learningsuggest that research now needs to examine SRL occurringnaturally in classrooms and identify features of classroomcontexts that promote or impede it.

Along these lines, Turner (1995) has examined howreading tasks affect young students' intrinsic motivation, acomponent of SRL. She observed as students in 12 Grade 1classrooms engaged in literacy activities and distinguishedbetween open and closed task structures. Open task struc-tures provided students with opportunities to choose whatthey read (mainly trade books), as well as where and whenthey read. This enabled students to choose materials thatwere personally meaningful and at an appropriate level ofdifficulty. Open tasks often included partner reading andauthentic writing tasks (e.g., writing party invitations andmaking birthday cards after reading a story called "Clif-ford's Birthday Party," Bridwell, 1998). Students oftenworked in groups, where they could observe others workingon similar tasks and ask for help. Teachers who assignee-open tasks tended to teach skills associated with reading inthe context of authentic reading activities (e.g., teachingprint concepts as they appear in a book the class is reading),modeled strategic thinking and problem solving (e.g., mak-ing predictions), reminded students to use strategies, andexplained why they were useful. Also, they discussedmetacognitive strategies appropriate to complex reading

tasks (e.g., making good use of resources, monitoringprogress, and persisting when faced with difficulties).

In contrast, in closed task structures, the teacher or thetask indicated what materials and information would be usedby students and what would be the finished product.Therefore, students had limited opportunities to make deci-sions or control their learning. Teachers tended to assignidentical seat work to all students, encouraged them to workindependently, and provided instruction in the context ofteacher-directed reading groups. Typically, students prac-ticed basic reading skills apart from authentic readingactivities (e.g., completing matching or fill-in-the-blankexercises to automatize decoding skills).

Turner (1995) has observed more evidence of intrinsicmotivation in classrooms where open tasks were the norm.In these environments, students used more strategies, exer-cised more volitional control, and persisted longer whenfaced with difficulties. Her interviews indicated that studentswho experienced predominantly open tasks were moremetacognitive and more focused on the meaningful aspectsof reading tasks. These findings provide important evidenceconcerning relations between instructional contexts andyoung children's understanding about and engagements inliteracy tasks. Open task structures offered students moreopportunities to develop skills and attitudes associated withSRL (i.e., intrinsic motivation, metacognition, and strategicbehavior), and consequently, they did. According to Turner,the complex and varied tasks assigned in open task environ-ments required students to adapt strategies and use a range ofstrategies. Also, making students more responsible for theirlearning and giving them opportunities to adapt tasks to suittheir interests and control challenge resulted in greaterpersonal investment.

To date, no studies parallel Turner's (1995) investigationof young children's SRL in the context of writing. However,studies involving students in intermediate grades throughcollege indicate that writing activities, particularly whenthey engage students in all aspects of the writing process(planning, drafting, editing, and revising), present richcontexts for observing SRL (Graham & Harris, 1994;Zimmerman & Bandura, 1994). Process approaches towriting view writing as a complex problem-solving activityand emphasize the need to engage students in writing forauthentic, communicative purposes about topics that holdmeaning for them (Flower & Hayes, 1984; Graham, Harris,MacArthur, & Schwartz, 1991). In carrying out theseactivities, students are involved in high-level cognitiveactivities such as planning, allocating resources (e.g., time,space, and information), monitoring goals and purposes, andpersisting when faced with difficulties (e.g., writer's block,multiple revisions; Zimmerman & Bandura, 1994). Teachersguide rather than direct students' writing by discussing anddemonstrating strategies for planning and revising text,involving students in evaluating their writing products, andencouraging them to collaborate with one another (Englert,Raphael, Anderson, Anthony, & Stevens, 1991; Graham etal., 1991). Process approaches to writing contrast withskills-oriented models, which reflect beliefs that proficient

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

YOUNG CHILDREN'S SELF-REGULATED LEARNING 717

writing results from teaching specific skills in isolation andin a hierarchical sequence (Needles & Knapp, 1994).

Many primary school teachers are combining processapproaches to writing instruction with portfolio assessment.Portfolios are associated with nonthreatening, noncompeti-tive evaluation practices that are believed to encourageattitudes and actions associated with SRL (Pintrich et al.,1993). If implemented as intended, portfolio activitiesinvolve students in choosing what goes into their portfolios,generating criteria that will be used to evaluate, andevaluating the contents of their portfolios (Paris & Ayres,1994; Paris, Turner, Muchmore, & Perry, 1995). Further-more, because portfolio assessment is embedded in students'daily routines and targets their individual goals and needs, itis nonthreatening. Students have multiple and varied oppor-tunities to demonstrate their learning, and teachers areencouraged to interact with students during assessment,scaffolding their choices, challenges, and self-evaluations,emphasizing mastery goals, and helping them to interpreterrors as opportunities to learn (Paris et al., 1992; Wolf,Bixby, Glen, & Gardner, 1991).

The assumption that process approaches to writing instruc-tion and portfolio assessment present numerous opportuni-ties to observe young children's SRL needs to be testedempirically and in naturalistic contexts. One study thatdocuments teachers' independent implementations of portfo-lios suggests that their practices vary widely (Calfee, 1994).Similarly, wide variations in teachers' implementations ofprocess writing programs have been documented, even instudies where teachers have been guided by carefullydesigned interventions of researchers (Anderson, Raphael,Englert, & Stevens, 1991). Apparently, designing and imple-menting activities that are aligned with sociocognitivemodels of learning (i.e., process writing and portfolioassessments) pose formidable challenges for teachersschooled in skills-oriented environments (Needles & Knapp,1994). These findings underscore a need to document therange of writing and portfolio activities across classroomsand to distinguish between environments more and lesslikely to promote SRL, before beginning to characterizestudents' engagements in them.

My study examines relations between particular featuresof writing and portfolio activities and SRL in five Grade 2and 3 classrooms. Specifically, I investigate the nature ofwriting tasks, authority structures, and portfolio evaluationpractices in these classrooms and their effects on students'(a) perceptions of control and support, (b) beliefs, values,and expectations concerning writing, and (c) regulation ofwriting behaviors.

Method

Design and Participants

Appendix A outlines the design of my study. I collected data inthree phases, spanning 6 months (January through June 1995),using a multiple and embedded case study design (Yin, 1989). InPhase 1, 19 Grade 2 and 3 teachers in a large suburban schooldistrict in British Columbia responded to a questionnaire thatgauged the extent to which writing and portfolio activities in their

classrooms were consistent with those theorized to promote SRL(i.e., activities were complex and offered students opportunities tomake choices, control challenge, collaborate with peers, andevaluate their work; teachers' interactions with students resembleddescriptions of scaffolded instruction or guided discovery). Thisquestionnaire and the other instruments used in this investigationare described in detail in the next section. On the basis of teachers'responses, I ranked classrooms to reflect the degree to whichwriting and portfolio activities within them were consistent withthose thought to promote SRL. Then I selected five classrooms formore detailed observation (four Grade 2/3, one Grade 3). Selec-tions were based on classrooms' relative ranking concerningactivities thought to promote SRL, three high and two low, andteachers' willingness to allow researchers to observe writing andportfolio activities in their classrooms once or twice each week for6 months.

hi Phase 2,94 students from these five classrooms (47 boys and47 girls, 35 Grade 2 and 59 Grade 3 students) were surveyedconcerning their perceptions of control and support in theirclassroom and their beliefs, values, and expectations concerningwriting. Also, teachers ranked their students in terms of writingachievement. On the basis of these rankings, 10 students, 2 fromeach classroom, 1 high and 1 low achiever, were selected, and theirSRL was studied in Phase 3. Teachers of split-grade classroomschose to rank their Grade 2 and 3 students together because allstudents in their classrooms participated in the same writingactivities. Students in both grades appeared in the top and bottomquartiles of each teachers' rankings. Subsequently, 3 high-achieving Grade 2 students and 2 low-achieving Grade 3 studentswere selected for Phase 3. Six of the 10 selected students were inGrade 2.

Table 1 contains demographic information for the five selectedclassrooms and the larger pool of classrooms from which they weredrawn. Teachers in selected classrooms, like those in the largergroup, varied in terms of their teaching experience and the numberof years they had been using portfolios to evaluate students'writing. Kathy (all names are pseudonyms) had been teaching for 4years and was using portfolios for the first time. Laura had beenteaching for 26 years and had been using portfolios for 4 years. Theremaining three teachers, Nan, Linda, and Peter, had 9, 10, and 16years of teaching experience, respectively. Peter, like Kathy, wasimplementing portfolios for the first time, whereas Nan and Lindahad been using portfolios to evaluate students* writing for 5 and 8years, respectively. The average number of students in selectedclassrooms was 22, and no class had more than 25 students. Thesefigures are consistent with those for the larger pool of classrooms.Each of the selected classrooms was located in a middle- tohigh-middle-socioeconomic status neighborhood and included atleast 1 student with special needs and 1 student who spoke Englishas a second language. These characteristics also reflect those of thelarger pool of classrooms. Students with severe learning disabilitiesand those whose command of English was judged insufficient tocomprehend questionnaire and interview questions were excludedfrom the study.

Data Collection and Analyses

The 6 months of data collection included weekly visits to the fiveselected classrooms during regularly scheduled writing and portfo-lio activities.

Phase 1. The questionnaire used to select classrooms forin-depth observation was divided into four sections: Section Aasked for demographic information about teachers and theirstudents (as shown in Table 1). Section B asked for informationabout teachers' approaches to writing instruction (e.g., the extent to

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

718 PERRY

•5k.

I§

I

s

tu-gCO «

T3 xi

z so

1

i-H ©

V V A V V

sssss

r- 00

* CN cs -^ ^t —' rs

r~ r-«i

—< o —• o *n

11° 2•a a

'3 a> —•

n-o

Is

-S.9

oXJ

• s s - s

on-- .5

which teachers tailored instruction to meet the needs of students vs.their curricular goals and the extent to which instruction focused onmeaningful vs. mechanical aspects of writing). Sections C and Dasked for information about students' engagements in writing andportfolio activities, including the extent to which students partici-pated in making decisions about their writing and portfolios,evaluated their work, and consulted and collaborated with peers.Most questions in Sections B-D asked teachers to rate howfrequently particular actions occurred in writing and portfolioactivities in their classrooms (e.g., "How often do studentsevaluate their writing products/progress?"). Ratings correspondedto a 5-point scale, ranging from never (1) to always (5). Someopen-ended questions elaborated on teachers' ratings (e.g., "Ifstudents evaluate their writing, what form does that evaluationtake?"). Teachers' ratings in Sections B-D were grouped to formfive scales that reflected characteristics of high-SRL environments:choice (7 items), opportunities to control challenge (6 items),self-evaluation (2 items), peer support (4 items), and teachersupport (13 items). All scales obtained moderate to high degrees ofinternal consistency (alphas ranged from .72-.90). Teachers re-ceived a score for each scale corresponding to their average ratingsof items in the scale. High scores across scales suggested thatwriting and portfolio activities in their classrooms would promoteSRL. The five teachers selected from these data had eitherconsistently high or consistently low scores across scales.

I observed writing and portfolio activities in each of the fiveclassrooms five times, for the total amount of time allocated to anactivity on a particular day (up to 2 hr). A 3-month pilot study,conducted the winter before this study, indicated that five observa-tions were an adequate sampling of writing and portfolio activitiesto characterize classroom writing environments reliably. Theobservation instrument had three sections (see Appendix B). Thefirst provided space to record (a) whose classroom was beingobserved and (b) the nature and duration of the activity in whichteachers and students engaged during that observation (e.g., writinga story or selecting samples of writing for portfolios). The secondsection provided space to keep a running record of what was goingon, including verbatim samples of teachers' and students' speech.The third section listed categories, derived from previous investiga-tions of high-SRL environments, as follows: (a) types of choice(i.e., choice about what, where, when, and who), (b) control overchallenge, (c) opportunities for self-evaluation, (d) support frompeers, and (e) support from the teacher.

During each observation, I positioned myself so that I couldclearly see and hear the teacher and students without beingintrusive. The running record of events and actions included a listof times related to events and actions and, as much as possible,verbatim speech in teacher-snident and student-student interactions.

After each observation, I read and annotated my running record,adding details regarding events and actions that I did not have timeto record during the observation and filling in gaps in teachers'speech with paraphrases of what they said. (Paraphrases weremarked with square brackets.) Once I was satisfied that the runningrecord was as detailed and accurate as my memory would allow, Ireread it, noting incidents and examples that reflected the originallist of categories in the third section of the observation instrument,as well as events and actions that suggested refinements oradditions to those categories. Next, I assigned a rating of 0 ,1 , or 2to each category to indicate its presence and quality in the activity.This scale was derived from the data to reflect the variance itexhibited and the discriminations observers could accurately andconsistently make. For example, if students had no choice concern-ing what to write, that category was rated 0. If students had somechoice but their choices were constrained (e.g., pick one of twotopics), the choice category was rated 1. A relatively unconstrained

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

YOUNG CHILDREN'S SELF-REGULATED LEARNING 719

choice of topics received a rating of 2. These ratings were enteredin a summary table for each class, to generate a profile of theconsistency with which writing and portfolio activities in a givenclassroom were high SRL across my five visits. Finally, classroomswere assigned a consistency rating for each category, indicatinghow consistently, across my five visits, each category was presentin writing and portfolio activities at a level thought to promoteSRL. These ratings corresponded to a 5-point scale; ranging fromnever (1) to always (5).

A second rater observed a writing and a portfolio activity in twoof the five selected classrooms, using my observation instrumentand coding procedures. These observations were conducted on twoseparate occasions while I also observed in these classrooms.Before observing in classrooms, the second rater participated in a3-hr training session, in which I explained the kinds of informationto record on the observation instrument and how to code it. Thenthe second rater used copies of my running records to practiceidentifying and rating the quality of instances of choice, controlover challenge, self-evaluation, and support from teachers andpeers. We compared examples and ratings and discussed discrepan-cies before observing in classrooms. After observing in classrooms,we coded our own running records and then coded one another's.We achieved an 88% level of agreement across all ratings, a level ofagreement 79% greater than would be expected by chance alone(K = .79). When we discussed the differences between our ratings,we achieved 100% agreement.

Phase 2. In Phase 2, students in the five selected classroomscompleted two questionnaires. Both were administered in smallgroups of 5 or 6 students, enabling me to observe and help studentsif any had difficulty with particular questions. Both questionnaireswere administered during one 30-min session. The sequence inwhich the questionnaires were administered alternated acrossgroups. At the beginning of the session, I explained that there wereno right or wrong answers to questions I would be asking and thatstudents might have different answers than the person sitting besidethem, "since we don't all think the same things about writing andportfolios." Then I directed students' attention to the first questionand instructed them to circle the option they thought was typicallytrue in their classroom or about mem. For both questionnaires, Iread each item aloud and discussed the options in the rating scales.

One questionnaire targeted students' perceptions of control andsupport in writing activities by asking them to rate, on a 5-pointscale ranging from 1 {never) to 5 (always), 20 items concerninghow often they had opportunities during writing and portfolioactivities to (a) make choices (7 items, e.g., "How often do youdecide what to write about?"), (b) control challenge (6 items, e.g.,"How often do you decide how to revise your writing?"), (c)evaluate their work (2 items, e.g., "How often do you evaluate yourown writing?"), and receive support from their (d) peers (4 items,e.g., "How often do students help you to edit your writing?") and(e) teacher (1 item, "How often does the teacher help you decidewhat to put in your portfolio?"). These items were initially groupedto form five scales matching those used to analyze data from theteacher questionnaire in Phase 1. However, these scales yieldederratic and generally low internal consistency alphas, ranging from.17 to .69. Therefore, the five scales were regrouped to form twoconceptually linked and more internally consistent scales. The firstcombined the choice, challenge, and self-evaluation scales, toreflect students' perceptions of control over learning (15 items,a = .73). The second combined the two scales that reflectedstudents1 perceptions of support from their teacher and peers duringwriting and portfolio activities (5 items, a = .46).

The perceptions of control and support questionnaire alsoincluded an open-ended question, "What would you want me tonotice about your writing, if I looked through your portfolio?" I,

along with a second rater, derived a set of categories to summarizestudents' responses to this question. Then we coded students'responses independently and compared our results. We achieved anabsolute agreement of 93%, a level of agreement 91% greater thanchance (K = .91).

The second questionnaire asked students to indicate their beliefs,values, and expectations concerning writing by rating, on a 5-pointscale ranging from 1 (really not true) to 5 (really true), how muchstatements such as "I like writing" were true about them. Fourstatements described each of six learner characteristics: (a) beliefsabout ability (e.g., "I can improve my writing if I work hard") andlearning (e.g., "Learning to write well takes a long time and a lot ofhard work"); (b) goal orientation (e.g., "I try to write as little aspossible so I can get finished quickly"); (c) interest (e.g., "I likehaving a portfolio for my writing"); (d) efficacy (e.g., "It's easy forme to think of ideas to write about"); (e) outcome expectations(e.g., "I'm usually pleased with the final copy of the things Iwrite"); and (f) desire for challenge (e.g., "I try to choose writingprojects that are easy for me"). Six additional statements targetedstudents' degree of independence (e.g., "I like it when the teachertells me what to write"). These a priori scales yielded variable andgenerally low levels of internal consistency (alphas ranged from.15 to .63), but moderate levels of stability across two administra-tions: Students gave identical responses to 64% of the items acrosstwo administrations of this measure spaced 10 days to 2 weeksapart. The alpha for all 30 items on this questionnaire was .64. Ireport students' responses to the first administration in the Resultssection.

Finally, teachers were asked to rank students in their classroomswho were participating in my study into quartiles of writingachievement. I checked the reliability of their rankings in threeways. First, I compared teachers' definitions of good writing toensure consistency of standards across participating classrooms.Second, I compared teachers' definitions of good writing with theirdescriptions of specific students they ranked as high- and low-achieving writers. Third, I asked three of the five teachers to reranktheir students after an interval of between 10 days and 2 weeks.Their second rankings achieved 97% agreement across all fourachievement levels (K = .96) and 93% agreement for the high andlow achievement levels targeted in this study (K = .87). On thebasis of these data, 10 students, 1 student from each of the top andbottom quartiles in each classroom, were selected as case studiesfor Phase 3.

Phase 3. I observed selected students in each of the fiveclassrooms five times during Phase 3. Observations lasted between1 and 1.5 hr, except in one classroom, where observations lasted 2hr because this was the length of time devoted to writing andportfolio activities. The observation instrument used to documentstudents' SRL had three sections (see Appendix C). The firstrecorded information about the child and activity being observed.The second was a running record of what went on. The thirdrecorded (a) specific strategies students used while engaging in thewriting process (e.g., planning or drafting), (b) specific portfolioactivities (e.g., reviewing or selecting), (c) executive strategies(e.g., making choices or modifying tasks or the environment tocontrol challenge), and (d) evidence of persistence.

During each observation, I positioned myself in the classroom sothat I could clearly see and hear students without being intrusive. Ikept a running record of events and actions, focusing on andrecording evidence of students' SRL. Also, I continued to recordfeatures of activities that distinguished high- and low-SRL class-rooms (e.g., did students have choices, control over challenge, andopportunities to evaluate their processes and products?). As muchas possible, I recorded the verbatim speech of students to students,individual students to teachers, and teachers to the class. Often I

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

720 PERRY

was able to sit with selected students as they worked and ask themabout their thinking and strategy use while they engaged in writingor portfolio processes.

After each observation, I read and annotated my running record,adding details regarding events and actions that I did not have timeto record during the observation and filling in gaps in students1

speech (paraphrases were noted with square brackets) until I wassatisfied that it was as detailed and accurate as my memory wouldallow. Then I reread the running record, this time looking forexamples of students1 strategy use and persistence and noting thembeside the correct category in Section C of the observationinstrument. Finally, each student's behaviors across the fiveobservations were entered into a summary table to create a profileof behaviors during writing and portfolio activities. A check wasused to indicate evidence of a particular behavior, and an X wasused to represent occasions when it would have been appropriatefor students to engage in particular behaviors but they did not.These summary tables were used as references for writing descrip-tive accounts of individuals' behaviors and patterns of behavioracross classroom contexts and achievement levels.

Running records for 4 of the selected students were coded by asecond rater, whom I trained with the same procedures that wereused in Phase 1. We achieved absolute agreement for 97% of ourcodes (K = .91).

Also in Phase 3, students were interviewed immediately aftereach observation, so I could check the validity of my interpreta-tions of their behaviors and gain insights into aspects of their SRLthat were unobservable (e.g., monitoring and decision making).Questions were derived directly from my observations but alwaysrequired students to reflect on (a) the purpose(s) of the activity inwhich they were engaged, (b) what they were thinking as theyengaged in it, (c) difficulties they experienced, (d) how theyhandled these, and (e) the assessment value of the activity forthemselves and their teacher. The second rater and I used tran-scripts from these interviews to generate categories reflectingstudents' descriptions of their writing and portfolio processes andproducts, modifications to tasks and environment, collaborationswith peers, difficulties, self- and teacher-evaluations, and evidenceof their motivational orientations (e.g., goals, expectations, andwillingness to take risks). Then we coded 4 students' transcriptsand compared our findings. Absolute agreement was 95% (K = .85).

Results and Discussion

The results of my study are presented in two sections. Thefirst reports findings from Phase 1, focusing on selectedteachers' responses to the questionnaire and observations ofrepresentative lessons from classrooms more and less likelyto promote SRL. The second section reports findings fromPhases 2 and 3. Analyses of Phase 2 data focus on the extentto which students' perceptions of control and support andtheir beliefs, values, and expectations concerning writingvaried across classroom contexts. Analyses of Phase 3 datafocus on evidence that selected students were engaging inSRL and how their behavior differed as a function ofclassroom environment and achievement level.

Phase 1 Findings

Teacher questionnaires. Responses indicated that writ-ing and portfolio activities in Laura's, Kathy's, and Nan'sclassrooms were likely to promote SRL. Their averageratings for items concerning choice, control over challenge,student self-evaluation, peer support, and teacher supportwere, for the most part, above 3 on the 5-point scale (seeFigures 1 and 2), indicating that these features of high-SRLenvironments were present in more than half the writing andportfolio activities in their classrooms. Peter's and Linda'sresponses indicated that these features seldom or nevercharacterized writing activities in their classrooms (theiraverage ratings for most scales were below 3). Peter alsorated items concerning portfolio practices below 3, whereasLinda was more erratic in her ratings of portfolio items. Sheindicated that her students often made choices and collabo-rated with peers during portfolio activities but seldom hadopportunities to control challenge. However, my observa-tions and in-person conversations with Linda led me toconclude that portfolio activities in her classroom did notoffer students the kinds of choices or support associated withhigh-SRL environments (see below).

5

4.5

4

3.5

3

2.5

2

1.5

1

0.5

0

x

— • —

O

- • -

— * —

Group

Laura

Kathy

Nan

Linda

Peter

Choice Self-Evaluation

PeerSupport

TeacherSupport

Challenge

Criteria for Teacher Selection

Figure 1. Mean ratings for writing scales on the teacher questionnaire.

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

YOUNG CHILDREN'S SELF-REGULATED LEARNING 721

&I

5

4.5

4

3.5

3

2.5

2

1.5

1

0.5

0

Choice Challenge Self- Peer TeacherEvaluation Support Support

Criteria for Teacher Selection

Figure 2. Mean ratings for portfolio scales on the teacher questionnaire.

Observations. Appendix D summarizes the qualitative differ-ences in writing and portfolio activities between high- and low-SRL classrooms. In general, my observations were consistent withteachers' descriptions of writing and portfolio activities in theirclassrooms. Specifically, in the three high-SRL classrooms, stu-dents engaged in complex, meaningful activities that extended overmultiple writing sessions and engaged them in all phases of thewriting process. They were given opportunities to decide what towrite, where, and with whom; they controlled challenge bydetermining how much to write, at what pace, and with what levelof support; and they were involved in setting evaluation criteria,editing and revising their work, and selecting samples for theirportfolios. Support from teachers and peers in these classroomsincluded sharing ideas, resources, and strategies and collaborativeediting.

For example, in Laura's Grade 2/3 class, I observed as studentsconducted research and wrote reports, evaluated their learning andwrote about it in their "reflections journal," and prepared a table ofcontents and selected writing samples for their portfolio. Table 2contains the consistency ratings assigned to activities in Laura'sclass. These ratings indicate that students were given control and

Table 2Consistency Ratings for Writing and Portfolio Activitiesin Laura's, Peters, and Linda's Classes

Variable

ChoiceWhatWhoWhereWhen

ChallengeSelf-evaluationTeacher supportPeer support

Laura's class

55215555

Peter's class

21112221

Linda's class

21112211

support in keeping with environments hypothesized to promoteSRL during each of my five visits. Typically, writing blocks inLaura's class lasted for 1.5 to 2 hr and began with a 10-15-minclass meeting. The purpose of these meetings was to plan writingactivities (e.g., "How's the research going?" "Has anyone got adraft ready for the computer?"), solve problems (e.g., "When youhave something new to do, it helps to look at something you'vealready done"), or learn new skills for the next phase of theirproject (e.g., how to provide constructive feedback to a peer or howto use the cut and paste function on the computer).

After class meetings, students were responsible for managingtheir time toward the completion of their projects. Typically, theydivided their time among several tasks, perhaps beginning at thecomputer to revise a section of a research report that had beenedited last period and then returning to the classroom to illustratethe printed version of that section or to help a peer plan or edit asection of their report. Within the writing block, students chosewhat to work on (usually within a broad topic that the class shared,e.g., research on penguins), when, where, and with whom. Theycontrolled challenge by determining how much they would writeand the pace at which they would work, how much and what kindsof support they would seek, and how much and what kinds offeedback they would incorporate in revisions. Self, peer, andteacher evaluations were embedded in iterative processes of editingand revising, and students were involved in generating criteria usedto evaluate good writing and good research. Finally, there was anabundance of instrumental support in keeping with approaches towriting and reading instruction believed to promote SRL (Graham& Harris, 1994; Turner, 1995). Laura's carefully orchestratedminilessons and class discussions provided students with knowl-edge, skills, and attitudes for accomplishing complex writing andportfolio tasks and working constructively with peers. She wassensitive to the level of support needed by individual students andthe class as a whole and adjusted her support and evaluativefeedback according to her judgments about individuals' capabilitiesand threshold for discouragement. Reflecting on students' first

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

722 PERRY

research project, which occurred in the fall before my observations,Laura described how she "practically spoon-fed" students theprocess of doing research. In January, when these observationswere conducted, she expected more input and independence fromstudents as she primed them for a final project they would completeindependently (see Phase 3 findings section).

In contrast, students in Peter's and Linda's classrooms engagedin writing activities that were brief (10 min) and focused on specificskills apart from composition (e.g., correcting spelling and punctua-tion errors in a sentence the teacher wrote on the board). In theseactivities, students' choices were limited and challenge and criteriafor evaluation were controlled by the teacher and were typically thesame for all students. Support typically targeted procedural aspectsof task completion (e.g., giving directions, distributing materials,or correcting an answer) rather than constructing meaning andstrategic learning. The consistency ratings for these classroomsindicate the extent to which activities observed in these classroomswere consistent with those believed to promote SRL (see Table 2).

The "sentence fix-up" activity that occurred each morning inPeter's classroom is a typical example. Students copied a sentenceof Peter's that contained some number of errors (the number oferrors was posted beside the sentence) from the blackboard intotheir language notebooks. Their task was to find Peter's mistakesand correct them. Later, the class corrected the sentence together.Peter called on students to offer corrections and explain why theywere needed. Once all the errors had been identified and thesentence on the board was correct, Peter asked students todouble-check their version of the sentence with the one on theboard. At the end of this activity, students submitted their note-books to Peter so he could check their work. The next day, students'notebooks were returned, and if they had successfully corrected theerrors in the sentence and done "neat printing," there would be asticker beside the entry.

Activities like these offered students few choices and littlecontrol over challenge, and although students were involved inevaluating their work, their assessment was limited to mechanicalfeatures of writing (e.g., spelling, punctuation, and neat printing)and being "correct." On rare occasions, Peter engaged students instory writing. During these activities, students had more controlover content and opportunities to seek support from peers. How-ever, they were encouraged to limit help seeking to students in then-group of 4, and to procedural aspects of tasks (e.g., clarifyingdirections, spelling, or cutting and pasting). When it came tosharing ideas, the message was, "If you have an idea, I know youare excited, but keep it to yourself." Also, story writing in Peter'sclass did not engage students in all phases of the writing process orin discussing the role of each phase in the process as a whole.Although he did involve students in prewriting activities, he did notask them to consider possible purposes for those activities or makeexplicit links between the planning phase and the drafting phasethat would follow. Furthermore, students in Peter's class did notedit their own or each other's writing. This resulted in lostopportunities to apply editing skills they were acquiring throughthe sentence fix-up activity and to reflect on and revise othermeaning-related qualities of their writing (e.g., organization andclarity of ideas). In this context, students had fewer opportunities todevelop skills and awareness that would enable them to direct theirown writing projects in the future.

Similarly, portfolio activities in Linda's and Peter's classroomscurtailed students' SRL. On two occasions, March 7, 1995, March14, 1995 I observed as Linda met with students about theirportfolios. Linda's plan was to have students sort the contents oftheir portfolios into piles and then "pick some things that you'dlike your parents to see." As students made selections, Linda wouldattach a "star quality slip" to the sample and ask the student to

explain, "Why did you like this?" Their comments were recordedon the slip. Unfortunately, these portfolio meetings were hurriedbecause Linda had only one 30-min block each week to meet withstudents while the counselor supervised her class. She felt rushed tosee as many students as possible (typically 3) and, consequently,often began sorting with or for the student, "That goes in thelanguage arts pile, that goes there . . . . " Sometimes sorting wasincomplete when Linda began prompting students to begin select-ing samples to show their parents, and she often followed anopen-ended prompt, such as "Is there anything here you are proudof?" with a more specific probe like "What about this one?" or "Ilike this one" before students could respond. Finally, to expeditethe commenting task, Linda often provided students with their textby summarizing their oral comments, "OK, you can wri te . . . ." Forsome students, Linda wrote the comment "cause [they were]running out of time."

All five of the selected teachers commented that one of the mostdifficult aspects of using portfolios was finding time to interact withindividual students about their work. However, teachers in high-SRL classrooms found ways around individual conferences thatmet the objectives of involving students in sorting, selecting, andevaluating their work. For example, they provided the kinds ofsupport teachers might provide to students, especially youngstudents, during conferences during classroom discussions andminilessons. Then as students worked independently, which theydid more often in high-SRL classrooms, teachers interacted withindividuals. Also, these teachers used peer collaboration to distrib-ute responsibility for supporting students.

Phase 2 and 3 Findings

Students' perceptions of control and support. Statisticsfor the control and support scales of the student perceptionsquestionnaire are displayed in Table 3. Given my field notes,which indicate that some students had difficulty makingrelative judgments about how often they engaged in behav-iors relating to SRL (e.g., choosing writing topics orself-evaluating) and the generally low levels of internalconsistency among items in the original five scales, thesedata should not be overinterpreted. However, some patternswere consistent with the teacher questionnaire data fromPhase 1. On average, students in the three high-SRLclassrooms rated items having to do with student control

Table 3Students' Perceptions of Control and Support and TheirBeliefs, Values, and Expectations Regarding Writingand Portfolio Activities

Classroom content

High-SRL classroomsMSD

Low-SRL classroomsMSD

High- vs. low-SRLclassrooms

Effect size

Controlover

learning

3.000.54

2.530.50

.94

Supportfor

learning

2.490.65

2.230.54

.48

Beliefs, values,and expectations

concerning writing

3.360.353

3.320.402

.11

Note. SRL = self-regulated learning.

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

YOUNG CHILDREN'S SELF-REGULATED LEARNING 723

higher than did students in the other two classrooms, t(92) =4.12, MSE = .116, p < .001, effect size = 0.94. On specificitems, they rated themselves as having more choices aboutwhat to write (effect size = 0.84), where to work (effectsize = 0.51), what to put in their portfolio (effectsize = 1.06), and who would see the contents of theirportfolio (effect size = 0.75). Also, they rated themselves ashaving more control over the number of drafts they wouldproduce (effect size = 0.71) and more opportunities toevaluate their writing (effect size = 0.42). Students inhigh-SRL classrooms perceived higher levels of supportfrom their teacher and peers than students in the low-SRLclassrooms, f(93) = 1.96, MSE = .136, p = .05, effectsize = .48. On specific items, students in high-SRL class-rooms perceived less support from their teachers for portfo-lio activities than students in low-SRL classrooms (effectsize = —0.39) but more opportunities to critique (effectsize — 0.44) and edit (effect size = 0.83) one another'swork. Students' ratings regarding teacher support for portfo-lio activities differed from their teachers' ratings and myobservations. Perhaps students in high-SRL classroomswere unaware of the subtle support their teachers providedduring portfolio activities, or the support they received fromtheir teacher was proportionally less than the responsibilitythey assumed for reviewing, sorting, and selecting thecontents of their portfolios. Another possibility is that thesupport from teachers they were currently receiving seemedless than the support they had received in past years fromother teachers.

Students' beliefs, values, and expectations. Students'average ratings for items on the beliefs, values, and expecta-tions questionnaire are shown in Table 3. Tests for differ-ences between means suggest that students in high- andlow-SRL contexts have similar beliefs, values, and expecta-tions concerning writing, f(93) = - .44, MSE = .081, p —.660, effect size = 0.11. On average, students in bothcontexts indicated beliefs associated with incremental viewsof ability, mastery goals, and high efficacy. These findingswere consistent with previous research about young children(see Paris & Newman, 1990, for a review). However,students' responses to the open-ended question on theperceptions questionnaire suggested some motivational dif-ferences across contexts. This question asked students toindicate what they would want me to notice about theirwriting if I looked through their portfolio. Although a largeproportion of students in both contexts indicated that themechanical aspects of writing were a focus for them (37% inhigh-SRL classrooms and 58% in low-SRL classrooms),students in high-SRL classrooms also alluded to the meaning-ful aspects (27% vs. 6% in low-SRL classrooms) andintrinsic value (21% vs. 3% in low-SRL classrooms) of then-writing. One student indicated, "[I want] to see youinterested in my writing." No student in a high-SRLclassroom referred to the extrinsic value of their writing,compared with 6% of students in the other two classroomswho indicated that they would want me to notice that "mostof it is right" or "my award for doing everything right.'*Students in the low-SRL classrooms also were more likely torespond "I don't know" to this question or with an answer

that suggested a lack of caring on their part about what Imight notice (e.g., "nothing" or "I don't care"). Finally,22% of the students in the high-SRL classrooms gaveresponses that included what I initially took to be a vaguedescriptor, "good writing." In thinking more about thislabel, however, I realized that students in these classes spenttime developing and displaying elaborate lists of what goodwriting means. It is possible that these students associatedthis descriptor with an elaborate array of attributes that theywanted me to notice while examining their work.

Similarly, in Phase 3 interviews, the 6 students I observedin the high-SRL classrooms indicated an approach tolearning that reflected intrinsic theories of motivation (Pin-trich & Schrauben, 1992). They indicated a task focus whenchoosing topics or collaborators for their writing andfocused on what they had learned about a topic and howtheir writing had improved when they evaluated theirwriting products. For example, Carol, the low-achievingstudent selected in Laura's class, chose to do research ontigers because "they are interesting . . . and there's lots tolearn about them." She had teamed up with a classmate whowas studying cougars "because cougars and tigers are in thesame family." Comparing her work on this project withwork completed earlier in the year, Carol commented, "Ilearned a lot from when I was in Grade 1 because I had a lot

of trouble then [Writing is] getting a lot easier and . . .quicker." Also, the students in high-SRL classrooms did notshy away from challenging tasks and topics, and the reasonsthey gave for seeking challenge reflected a desire forpersonal mastery. Susan, the low-achieving student selectedin Kathy's class, was not daunted by how hard it is to "getsome of the sentences down [to note facts from a book inpoint form] and think of sentences for that [transform pointsinto complete sentences in a research report]." She waspleased with what she was learning "about different [places]and what kind of animals they've got there . . . what kind ofpeople." Also, she was convinced that she was learning"harder words" and that "the more she used them, the easier. . . the less [she had] to check."

In contrast, the 4 students in the low-SRL classroomswere more focused on their teachers' evaluations of theirwriting and how many they got right on a particularassignment. Both the high- and low-achieving students inthese classrooms were concerned with getting a "goodmark." Furthermore, the 2 low-achieving students in theseclassrooms made comments that reflected a desire to avoidfailure and perceptions of low ability and low efficacyregarding writing. For example, Rebecca, in Peter's class,described choosing work that is "easier" and often hid herwork, "so no one else would see." Natalie, in Linda's class,also avoided difficult writing tasks and commented on morethan one occasion, "I'm not very good at writing."

Self-regulated learning. Qualitative differences in stu-dents' behaviors across high- and low-SRL classrooms aresummarized in Appendix E. Tasks, authority structures, andevaluation practices in high-SRL classrooms offered stu-dents more opportunities to think and act as self-regulatedwriters. In these classrooms, the 6 students I observedengaged in all phases of the writing process and moved

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

724 PERRY

among the phases independently, flexibly, and recursively.In Laura's class, I observed Carol and Nand as theyconducted research. They planned for writing by creating asemantic web to organize facts gathered from multiplesources. They drafted each section of their report multipletimes, editing each version with a peer or the teacher. Theypublished their final version in a spiral-bound booklet,complete with a table of contents and a bibliography. Eachselected a page from their reports to go into their portfolios.Like skilled writers, their path through the writing processwas not linear (Graham & Harris, 1994). They juggled thevarious phases in the writing process depending on whatinformation and support were available to them, theirinterests, and deadlines. On one occasion, (May 10, 1995) Iobserved Nand as he drafted a section of his report thatdescribed what owls eat. He mentioned that he hadn'tfinished his web but would fill in missing facts "when Icome to them . . . I'm just doing what I can." He explainedthat he was writing the section on food "because I have lotson food." In this instance, Nand's decisions about how toallocate time and resources reflected efforts to cope with adifficult aspect of writing, gathering information. On otheroccasions, (May 17, 1995) Nand's decisions about how tomanage time were based on his interest in spending somepart of every writing session at the computer or on pragmaticconcerns (e.g., seeing that Laura was busy with anotherstudent, he began working on another section of his reportuntil she was available).

My observations and interviews also offer evidence thatthe 6 students in the high-SRL classrooms monitored andevaluated their writing progress in productive ways. Forexample, the rough drafts in Carol's portfolio had linesdrawn under some words and arrows linking one place in thetext to another. Carol explained that she underlined wordsshe was not sure how to spell while writing and then askedLaura to help her correct spelling errors when they met to goover her work. The arrows were the result of editing withLaura and peers. "We saw some corrections and movedthings around." (May 26, 1995) While Carol reviewededited documents in her portfolio, (May 17, 1995) shecommented on improvements in her writing from Septemberto May ("I learned a lot from when I was in Grade 1 . . . . I'mlearning more words . . . finding commas and that stuff.").Improvements were not so obvious to Nand when hereviewed drafts of his writing, (May 10, 1995) perhapsbecause writing had presented fewer difficulties for himthroughout the year. However, he did comment on theprogress he and the class had made toward the goal ofconducting research independently. "Normally, [Laura]helped us with the whole research but this is our own . . . thefirst two times we had to see how it works, how to do it."

Finally, my observations and interviews suggest that the 6students in the high-SRL classrooms sought instrumentalsupport from one another and their teacher. In selecting apartner for collaborative activities, these students demon-strated a task focus by choosing someone who sharedinterests with them or had skills they needed to develop. InLaura's class, Carol chose a student who was working on arelated research topic and helped her "understand things."

(May 10, 1995) She explained that because "[tigers] arehard to find . . . there's not that many books that have tigersin them," she needed to learn to look through books about"all kinds of animals . . . and, if the book has a table ofcontents, I look through the contents." Her buddy helped herlearn that because he was "in Grade 3." The 3 high-achieving students in the high-SRL classrooms did not relyon others to complete tasks successfully. Nonetheless, theyinteracted with one another about the substance of theirwriting, "Do you know why owls eat skunks? . . . Do youthink this sentence is too short?" (May 17, 1995) Or theyshared discoveries about the writing process, "I don't need adraft but it's better if you have a draft . . . because it's fasterto write on paper than I am typing on the computer." (May 3,1995). These behaviors contrast sharply with collaborationsamong the students in the low-SRL classrooms. In thoseclassrooms, the 4 students I observed chose partners mainlyfor social reasons (e.g., "I like playing with her and I hadn'tseen her for a long time"; May 3,1995) and then spent muchof their writing time off task.

The 6 students in the high-SRL classrooms viewedteachers as collaborators rather than managers of theirwriting process. They asked teachers to help with challeng-ing problems and to comment on their work in progress butdid not appear to perceive or rely on them as the ultimatesource of knowledge. When asked what their teachers mightlook for when they reviewed students' work, the students inthe high-SRL classrooms typically indicated improvementsor information. Carol commented that Laura "might learnabout tigers too, because I wrote it and she reads it and shegets information too" (May 10, 1995). In contrast, the 4students in the low-SRL classrooms perceived that teacherscontrolled their writing (e.g., "She didn't ask me, I knew Ihad to do it"; May 16,1995) and looked for mistakes whenreviewing students' work. In these classrooms, the studentsappeared to focus on superficial indicators of performance(e.g., stickers and scores). They missed opportunities toimprove their knowledge and skill in writing by not evaluat-ing their writing before submitting it to the teacher and notreviewing their work when it was returned. During thesentence fix-up task in Peter's class, I observed manynotebooks turned face down on desks while the errors in thesentence on the board were identified and discussed. Simi-larly, the low-achieving student selected in Peter's classindicated that she did not check her work before handing itin because she "didn't have to" and because "mistakes arechecked by the teacher" (May 3,1995).

As was documented in Phase 1, students in the low-SRLclassrooms had fewer opportunities and incentives to behavelike self-regulated writers. Many activities in which thesestudents engaged did not require them to produce extendeddiscourse. For activities that did, their teachers assumedmuch of the responsibility for navigating phases of thewriting process, controlling challenge, editing, and revising.In these contexts, I observed students not only passing over,even rejecting, opportunities to take responsibility for someaspect of their writing (e.g., reviewing and self-correcting)but also choosing ineffective, even self-handicapping, strate-gies (Covington, 1992). Natalie, the low-achieving student

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

YOUNG CHILDREN'S SELF-REGULATED LEARNING 725

selected in Linda's class, avoided journal writing by engag-ing in conversations with students around her and thenhiding her journal rather than submitting it to her teacher.She was able to verbalize several strategies for coping withwriting difficulties, including asking "other people" forhelp, using one of Linda's story starters, and looking throughbooks. However, I observed that Natalie, like many low-achieving writers, did not use these strategies optimally(Graham & Harris, 1994). I never observed Natalie approach-ing her peers for help. Instead, when she experienceddifficulties with writing, she waited for Linda to becomeavailable. Also, when making use of story starters or books,Natalie tried to choose "easy" topics over interesting onesand then found it difficult to stick with them.

My observations revealed some similarities in students'SRL across contexts that might be attributable to skill andinterest in writing. The 5 students rated as high-achievingwriters reported fewer difficulties than the 5 students rated aslow-achieving writers and different strategies for copingwith them. The high-achieving students indicated that gener-ating ideas for writing and editing their drafts was quite easy.When they experienced writer's block, they understoodwhat caused the problem (e.g., "I didn't do much"; May 26,1995) and used a range of strategies to solve it indepen-dently. One student described how "I just think of stuff I'vedone, read . . . if there's anything coming up like mybirthday...." (May 9, 1995). Another described how shepiggybacked on other students' ideas, "sometimes whenother people say their title, it will give me an idea," and usedother people's experiences, "I think about other people'sl ives . . . what I've heard about" (May 18,1995). In contrast,the low-achieving writers tended to turn to their teacher orclassmates for help when problems arose. None of thehigh-achieving writers indicated that they experienced diffi-culties with the mechanical aspects of writing, whereasmechanical problems were the most common frustration forthe low-achieving writers. Each of these writers mentioneddifficulties with spelling, typing, or reading his or herhandwriting. Two of the 3 low achievers in the high-SRLclassrooms also described the difficulties they had transform-ing the facts they gathered on their research topics into"good sentences." Finally, the high-achieving writers dem-onstrated more sophisticated ways of navigating the writingprocess (e.g., distributing rather than duplicating tasksduring collaborative activities and recognizing when onesentence, "Owls live all over the world except . . . , " couldconvey the same information as many: "I could have made it15 sentences, by writing where they do live, but I decided tojust put that"; June 7, 1995) and reported more instances ofwriting in their spare time in school and at home.

These findings were consistent with previous compari-sons of skilled and unskilled writers (Graham & Harris,1994). However, the 3 low-achieving writers in the high-SRL classrooms were not discouraged by their difficulties.Their comments indicated that although writing was difficultfor them, they were satisfied with the support they receivedand encouraged by improvements in their writing from thebeginning of the year. Hard work was paying off for thesestudents. In contrast, the defensive tactics used by the 2

low-achieving writers in the low-SRL classrooms, as well asstatements they made reflecting perceptions of low-ability inwriting, suggested that they were discouraged. They weremore likely than their low-achieving peers in high-SRLclassrooms to adopt an academically self-handicapping viewof the challenges writing presented (Covington, 1992;Rohrkemper & Corno, 1988).

Conclusion and Future Directions

My study provides evidence of young children's SRL innaturalistic contexts and empirical support for sociocogni-tive models of learning that emphasize the need for research-ers to attend to contextual factors such as tasks, authoritystructures, and evaluation practices to explain decisionsstudents make about how to behave in school. Specifically,Grade 2 and 3 students in high-SRL classrooms adoptedskills and attitudes that are characteristic of self-regulatedlearners, whereas students in low-SRL classrooms adoptedattitudes and actions associated with defensive, even self-handicapping, approaches to learning (Covington, 1992).Finally, my observations indicate that teachers' use of theparticular educational innovations (e.g., the writing processand portfolios) does not guarantee a high-SRL environment.This highlights the value of characterizing learning contextsbefore characterizing students' engagements in them.

The findings of my study have important implications forpractice. In particular, they highlight the importance ofattending to young children's motivations and designingprimary classroom environments that promote academicallyeffective forms of SRL. However, my observations of andinteractions with the teachers who participated in my studyindicate that aligning classroom practices with the goal offostering SRL requires more than exposing teachers tocontemporary approaches to assessment and instruction.More likely, teachers require guided and sustained activitiesaimed at helping them to develop deep understanding ofhow their current practices influence students and supportingtheir experimentation with alternative teaching and assess-ment strategies. Participatory models of teacher develop-ment, operationalized in the communities of practice litera-ture (Lave & Wenger, 1991; Palincsar, Magnusson, Marano,Ford, & Brown, 1998), offer useful frameworks for accom-plishing these goals. Also, in the interest of designingeffective development activities for teachers that will en-hance students' learning, I believe research on SRL needs tomake several advancements. In particular, more in-depthinvestigations of writing and portfolio activities in class-rooms are needed to increase the generalizability of studieslike mine (e.g., Pressley, Wharton-McDonald, Mistretta, &Echevarria, 1997; Turner, 1995). These investigations shouldextend throughout an entire school year, documentingimportant foundational activities to inform understandingabout how much and what kinds of support young childrenrequire to behave as students in the high-SRL classrooms inmy study did in January through June.

Researchers need to develop more valid measures ofyoung children's motivation and SRL, as well as measuresthat enable qualitative comparisons of students' writing

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

726 PERRY

achievement across high- and low-SRL classrooms. Inkeeping with the experiences of other researchers (Pintrich& Schunk, 1996), the questionnaires I developed to assessyoung students' perceptions of their classroom environ-ments and their beliefs, values, and expectations regardingwriting were not psychometrically strong and, in the case ofthe beliefs, values, and expectations questionnaire, did nottease out differences among young children's motivation. Inthe past, this methodological shortcoming has been inter-preted to mean that young children are unclear about thecorrelates of motivation (Cain & Dweck, 1995). However,my open-ended question on the perceptions questionnaireand my semistructured interviews provided more detailedand differentiated information about students' perceptions oftheir classroom contexts and motivational orientations. Thisfinding supports claims that methods used to assess olderstudents' motivation and SRL may not be appropriate foryoung children. According to Cain and Dweck, new mea-sures that target issues young children value, use languageand response formats they understand, and assess theirmotivational responses in the context of activities relevant totheir daily life need to be developed. Finally, I collectedsamples of students' writing in each of the classrooms Istudied. However, writing products varied considerablyacross classrooms, and because no standardized measures ofstudents* writing achievement were available, I was limitedin the kinds of comparisons 1 could make concerningrelations between classroom contexts and students' writingcompetence. In future research, this limitation might beaddressed by administering generic writing prompts (Needles& Knapp, 1994) to all participants and then evaluating theirproducts with criteria that address both meaningful andmechanical aspects of writing.

References

Anderson, L. M., Raphael, T. E., Englert, C. S., & Stevens, D. D.(1991, April). Teaching writing with a new instructional model-Variations in teachers' practices and students' performances.Paper presented at the annual meeting of the American Educa-tional Research Association, Washington, DC.

Bridwell, N. (1988). Clifford's Birthday Party. New York: Scholastic.Brown, A. L., & Campione, J. C. (1994). Guided discovery in a

community of learners. In K. McGilly (Ed.), Classroom lessons:Integrating cognitive theory with classroom practice (pp. 229-270). Cambridge, MA: MIT Press.

Cain, K. M., & Dweck, C. S. (1995). The relation betweenmotivational patterns and achievement cognitions through theelementary school years. Merrill-Palmer Quarterly, 41, 25-52.

Calfee, R. C. (1994, April). Teacher responses to survey questionsabout portfolio design. Paper presented at the annual meeting ofthe American Educational Research Association, New Orleans.

Covington, M. V (1992). Making the grade: A self-worth perspec-tive on motivation and school reform. New York: CambridgeUniversity Press.

Englert, C. S., Raphael, T. E., Anderson, L. M., Anthony, H. M., &Stevens, D. D. (1991). Making strategies and self-talk visible:Writing instruction in regular and special education classrooms.American Educational Research Journal, 28, 331-312.

Flower, L., & Hayes, J. R. (1984). The representation of meaning inwriting. Written Communication, 1, 120-160.

Graham, S., & Harris, K. R. (1994). The role and development ofself-regulation in the writing process. In D. H. Schunk & B. J.Zimmerman (Eds.), Self-regulation of learning and perfor-mance: Issues and educational applications (pp. 203-228).Hillsdale, NJ: Erlbaum.

Graham, S., Harris, K. R., MacArthur, C , & Schwartz, S. (1991).Writing instruction. In B. Y. L. Wong (Ed.), Learning aboutlearning disabilities (pp. 309-343). San Diego, CA: AcademicPress.

Lave, J. & Wenger, E. (1991). Situated learning: Legitimateperipheral participation. Cambridge, England: Cambridge Uni-versity Press.

Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry.Newbury Park, CA: Sage.

Moll, L. C , & Whitmore, K. F. (1993). Vygotsky in classroompractice: Moving from individual to social transaction. In E. A.Forman, N. Minick, & C. A. Stone (Eds.), Contexts for learning:Sociocultural dynamics in children's development (pp. 19-42),New York: Oxford University Press.

Needles, M. C , & Knapp, M. S. (1994). Teaching writing tochildren who are underserved. Journal of Educational Psychol-ogy, 86, 339-349.

Palincsar, A. S., & Klenk, L. (1992). Examining and influencingcontexts for intentional literacy. In C. Collins & J. N. Mangieri(Eds.), Teaching thinking (pp. 297-315). Hillsdale, NJ: Erlbaum.

Palincsar, A. S., Magnusson, S. J., Marano, N., Ford, D., & Brown,N. (1998). Designing a community of practice: Principles andpractices of the GIsML Community. Teaching and TeacherEducation, 14, 5-19.

Paris, S. G., & Ayres, L. R. (1994). Becoming reflective studentsand teachers with portfolios and authentic assessment. Washing-ton, DC: American Psychological Association.

Paris, S. G., Lawton, T. A, & Turner, J. C. (1992). Reformingachievement testing to promote students' learning. In C Collins& J, N. Mangieri (Eds.), Teaching thinking (pp. 223-241).Hillsdale, NJ: Erlbaum.

Paris, S. G.( & Newman, R. S. (1990). Developmental aspects ofself-regulated learning. Educational Psychologist, 25, 87-102.

Paris, S. G., Turner, J. C , Muchmore, J. & Perry, N. (1995).Teachers' and students' perceptions of portfolios. Journal ofCognitive Education, 5, 7-40.

Pintrich, P. R., Marx, R. W., & Boyle, R. A. (1993). Beyond coldconceptual change: The role of motivational beliefs and class-room contextual factors in the process of conceptual change.Review of Educational Research, 63, 167-199.

Pintrich, P. R., & Schrauben, B. (1992). Students' motivationalbeliefs and their cognitive engagement in classroom academictasks. In D. Schunk & J. Meece (Eds.), Student perceptions in theclassroom (pp. 149-183). Hillsdale, NJ: Erlbaum.

Pintrich, P. R., & Schunk, D. H. (1996). Motivation in education:Theory, research and applications. Englewood Cliffs, NJ: CharlesE. Merrill.

Pressley, M., Forrest-Pressley, D. L., Elliott-Faust, D., & Miller, G.(1985). Children's use of cognitive strategies, how to teachstrategies, and what to do if they can't be taught. In M. Pressley& C. J. Brainerd (Eds.), Cognitive learning and memory inchildren (pp. 1-47). New York: Springer-Verlag.

Pressley, M., Wharton-McDonald, R., Mistretta, J., & Echevarria,M. (1997). The nature of literacy instruction in ten Grade 4/5classrooms in upstate New York. Unpublished manuscript.

Rohrkemper, M., & Corno, L. (1988). Success and failure onclassroom tasks: Adaptive learning and classroom teaching.Elementary School Journal, 88, 297-312.

Stipek, D., Feiler, R., Daniels, D., & Milburn, S. (1995). Effects of

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

YOUNG CHILDREN'S SELF-REGULATED LEARNING 727

different instructional approaches on young children's achieve-ment and motivation. Child Development, 66, 209-223.

Turner, J. C. (1995). The influence of classroom contexts on youngchildren's motivation for literacy. Reading Research Quarterly,50,410-441.

Winne, P. H. (1995). Inherent details in self-regulated learning.Educational Psychologist, 30, 173-187.

Winne, P. H. (1997). Experimenting to bootstrap self-regulatedlearning. Journal of Educational Psychology, 89, 397-410.

Wolf, D., Bixby, J., Glen, J., & Gardner, H. (1991). To use their

minds well: Investigating new forms of student assessment.Review of Research in Education, 17, 31-74.

Yin, R. K. (1989). Case study research: Designs and methods.Newbury Park, CA: Sage.

Zimmerman, B. J. (1990). Self-regulated learning and aca-demic achievement: An overview. Educational Psychologist, 25,3-17.

Zimmerman, B. J., & Bandura, A. (1994). Impact of self-regulatoryinfluences on writing course attainment. American EducationalResearch Journal, 31, 845-862.

Appendix A

Outline of Three Phases of Study

Phase 1 Phase 2 Phase 3

Surveyed 19 Grade 2 and 3classrooms about writingand portfolio activities intheir classrooms.

Ranked classes to reflect degreeto which activities pro-moted or curtailed SRL.

Selected five classrooms forin-depth observation:

3 high SRL2 low SRL

Surveyed students in the fiveselected classrooms abouttheir perceptions of controland support in their class-rooms and their beliefs,values, and expectationsregarding writing.

Teachers ranked students fromhighest to lowest achievingwriters.

Selected 10 students forin-depth observation:

5 high-achievingwriters

5 low-achieving writers

Observed and interviewedselected students, docu-menting differences in SRLacross classroom contextsand achievement levels.

Note. SRL = self-regulated learning.

Appendix B

Classroom Observation Instrument

Teacher _

Activity.

Section A

Date,

Start time. End time.

Running record

CategoryChoiceControl over challengeSelf-evaluationPeer supportTeacher support

Section B

Section C

Description/examples

(Appendixes continue)

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

728 PERRY

Appendix C

Observation Instrument for Documenting Students' Self-Regulated Learning

Child.

Section A

Classroom.

Activity. Date.

Start time. End time.

Section B

Running record

Section C

BehaviorsWriting processes

PlanningDraftingEditing (self/peer)Revising

Portfolio processesReviewingOrganizingSelecting

Executive processesMaking choicesControlling challengeSelf-evaluatingGetting support from peersGetting support from teacher

PersistingIn the face of difficultyThrough all stages of the writing processThrough many iterations of the process

(e.g., multiple revisions)

Description/examples

Appendix D

Qualitative Differences Between Writing and Portfolio Activitiesin High- and Low-SRL Classrooms

Variable High-SRL classrooms Low-SRL classrooms

Tasks Tasks were complex, focused on largechunks of meaning, extended overlong periods of time, and engagedstudents in all phases of the writingprocess.

Student control Students had a range of relativelyunconstrained choices and opportu-nities to control challenge and wereinvolved in evaluation.

Support Students received instrumental supportfrom peers and teacher (support thatled to self-regulation).

Evaluation Evaluation was embedded in ongoingactivities, interactive, emphasizedprocesses as well as products,focused on personal progress, andinterpreted errors as opportunities tolearn.

Tasks focused on specific skills apartfrom composition, were short induration, focused on prewriting anddrafting processes, and focused onspecific strategies versus beingstrategic.

Students choices were limited.Teachers controlled challenge and set

criteria for evaluation.

Students received procedural supportfrom peers and teachers.

Evaluation was a separate activity doneby the teacher, focused on themechanical aspects of writing,applied the same standards to all stu-dents, and highlighted number cor-rect and student differences.

Note. SRL = self-regulated learning.

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

YOUNG CHILDREN'S SELF-REGULATED LEARNING 729

Appendix E

Qualitative Differences in Students* Self-Regulated Learning (SRL)Across High- and Low-SRL Classrooms

High-SRL classrooms Low-SRL classrooms

Students engaged in all phases of the writingprocess (planning, drafting, editing,revising, and publishing)

independentlyflexiblyrecursively

Students managed time and resourcesto control challengeto accommodate intereststo maximize efficiency

Students monitored and evaluated theirprogress.

Students sought instrumental support fromteachers and peers.

Students engaged in prewriting and draftingphases of the writing process

under their teacher's directionin a linear fashion

Teachers managed time and resources.

Students overlooked or ignored opportunities toevaluate their work.

Some students used academically ineffective orself-handicapping strategies.

Received April 7, 1997Revision received March 20,1998

Accepted March 23,1998

Low Publication Prices for APA Members and Affiliates

Keeping you up-to-date. All APA Fellows, Members, Associates, and Student Affiliatesreceive—as part of their annual dues—subscriptions to the American Psychologist andAPA Monitor. High School Teacher and International Affiliates receive subscriptions tothe APA Monitor, and they may subscribe to the American Psychologist at a significantlyreduced rate. In addition, all Members and Student Affiliates are eligible for savings of upto 60% (plus a journal credit) on all other APA journals, as well as significant discounts onsubscriptions from cooperating societies and publishers (e.g., the American Association forCounseling and Development, Academic Press, and Human Sciences Press).

Essential resources. APA members and affiliates receive special rates for purchases ofAPA books, including the Publication Manual of the American Psychological Association,and on dozens of new topical books each year.

Other benefits of membership. Membership in APA also provides eligibility forcompetitive insurance plans, continuing education programs, reduced APA convention fees,and specialty divisions.

More information. Write to American Psychological Association, Membership Services,750 First Street, NE, Washington, DC 20002-4242.

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.