DOCUMENT RESUME ED 303 790 AUTHOR Pinnell, Gay Su; And … · 2013. 8. 2. · DOCUMENT RESUME. CS 009 535. Pinnell, Gay Su; And Others Reading Recovery: Early Intervention for At-Risk

ED 303 790

AUTHORTITLE

INSTITUTIONPUB DATENOTE

AVAILABLE FROM

PUB TYPE

EDRS PRICEDESCRIPTORS

DOCUMENT RESUME

CS 009 535

Pinnell, Gay Su; And OthersReading Recovery: Early Intervention for At-RiskFirst Graders. ERS Monograph.Educational Research Service, Arlington, Va.88

87p.

Educational Research Service, 2000 Clarendon Blvd.,Arlington, VA 22201 ($18.00 subscriber; $36.00no'asubscribers).

Reports - Descriptive (141) -- Guides - Non-ClassroomUse (055)

MF01 Plus Postage. PC Not Available from EDRS.Case Studies; Grade 1; *High Risk Students;Instructional Effectiveness; Longitudinal Studies;Primary Education; *Reading Difficulties; *ReadingInstruction; Reading Strategies; Teaching Methods

IDENTIFIERS Ohio (Columbus); *Reading Recovery

ABSTRACT

This monograph presents information about ReadingRecovery, describes the latest research concerning the program, andsummarizes practical experience concerning the implementation of thisinnovation in reading instruction. Chapter 1 presents a generaldescription of Reading Recovery instructional procedures. Chapter 2contains three case studies that provide a more concrete look at howthe program works with individual children and teachers. Chapter 3discusses a longitudinal study conducted in the Columbus PublicSchools to determine both the short-range and the long-range effectsof Reading Recovery on a group of at-risk students. Chapter 4describes the studies of Reading Recovery at sites throughout thestate of Ohio during the years of 1985-86, 1986-87, and 198-88.Chapter 5 describes the Reading Recovery staff development component,along with studies of teacher training and development in programtechniques. Chapter 6 presents suggestions for school districts orstate agencies that wish to implement Reading Recovery. Thirty-threereferences and three appendixes containing a list of books used inReading Recovery, a description of the alternative .. nterventionprogram employed during the first year of the longitudinal study, andmeasures used to assess children in the Reading Recovery Program areattached. (MS)

*

*

* *

Reproductions supplied by EDRS are MIa best that can be madefrom the original dccument.

...

Monograph

"PERMISSION TO REPRODUCE THIS

MATERIAL IN MICROFICHE ONLY

HAS BEEN GRANTED BY

1A-01\ tcN4

TO THE EDUCATIONAL RESOURCES

INFORMATION CENTER (ERIC)"

r'.4411...

..e.gle.,-.61.,-

la

U S DEPARTMENT OF EDUCATIONOffice of Educafional Research and Improvement

EDUCATIONAL. RESOURCES INFORMATIONCENTER (ERIC)

C INS document haS been reproduced asreceived from the cerSon or organizationOnginahng .1

C Mono, changes have been made to improvereproduction Quality

Points of view or ommuns staled In this docu-ment do not necessarily represent cefiCoalOEM positron or policy

Reading Recovery:Early Intervention forAt-Risk First Graders

Gay Su PinnellDiane E. De Ford

Carol A. Lyons

Copyright 01988 Educational Research ServiceAll rights reserved.

Ns

lb

0 Educational Research Service2000 Clarendon BoulevardArlington, Virginia 22201Phone: (703) 243-2100

The Information Source for School Decisions

Sponsored by:

American Association of School AdministratorsAmerican Association of School Personnel AdministratorsAssociation of School Business OfficialsCouncil of Chief State School OfficersNational Association of Elementary School PrincipalsNational Association of Secondary School PrincipalsNational School Public Relations Association

Educational Research Service is an independent, nonprofit corporationserving the research and information needs of the nation's school systt ns,

their administrative staffs, school boards, and the public.

ERS serves as both a national source and a clear-inghouse for school research and information. Bycompiling, analyzing, and sharing information that isessential to effective decision making, ;;RS assistslocal school districts in both day-to-day operationsand long-range planning. It provides an invaluableapproach to gathering and reporting data and informa-tion, through services that are geared to prompt yetthorough response to requests from subscribers. Localschool districts that maintain a yearly ERS order re-ceive the full package of ERS information resources,including: full access to the ERS On-Call InformationService for the district's administrative staff and board

a copy of each new ERS Report as soon as it is pro-duced (12-15 each year) the monthly ERS Bulletin,summarizing research published by other agenciesthe quarterly ERS Spectrum, the journal of school re-search and information access to the successful in-structional approaches and school operations of other

districts through the ERS School Operations Informa-tion Bank ERS Status and Opinion Surveys a com-puter-generated Local School Budget Analysis reportindividually prepared for the district upon request. Anannual order that includes most of these resources mayalso be placed by regional educational service agen-cies, associations of school administrators and schoolboards, university departments of schc31 administra-tion, and state departments of education. Services maybe provided to other groups by special action of theERS Board of Directors.

To ensure that small as well as large school dis-tricts can benefit from ERS, graduated annual fees forERS services have been established. This enablessubscribing districts and agencies to share equitably inthe cost of compiling and reporting the research need-ed by all school administrators and governing boards.Subscription rates are available on request.

ERS Executive Staff

Glen E. Robinson John M. ForsythPresident Vice President

Director of Research

ERS is solely responsible for this publication; no approval or endorsement by specific ERS sponsoringorganizations is implied.

Ordering information: ERS subscriber price: $18.00. (Single copy routinely mailed as part of subscription )Nonsubscriber price: $36.00. Quantity discounts available for specific uses. Actual cost of postage and handlingwill be added to billed orders. Please add $2.50 for postage and handling to prepaid orders under $20.00.

Stock #222-00046

Iti

Contents

Foreword ii

Acknowledgments iii

About the Authors iv

Introduction: What Is Reading Recovery? 1

Chapter 1: Description of Reading Recovery Lessons 7

Chapter 2: Case Studies of Children 15

Chapter 3: Longitudinal Study, Columbus Public Schools 27

Chapter 4: Ohio Reading Recovery Project 45

Chapter 5: Reading Recovery Staffing and Staff Development Program 55

Chapter 6: Implementing a Reading Recovery Program 65

Bibliography 73

Appendices:

Appendix A: Some Small Books Used in Reading Recovery 75

Appendix B: Description of Alternative Intervention Program,Columbus Longitudinal Study 1985-86 81

Appendix C: Dependent ? leasures Used to Assess Children inReading Recovery Program 83

1 rti

.

Foreword

Early . e.ading difficulties can prevent children from achieving initial success in school, lockingmany of them into an early pattern of school failure. Even with extensive and costly remedialassistance throughout their school careers, such children often do not break out of this pattern.The dilemma of how to take children with early reading difficulties and put them on the road tosuccess is a major concern for school officials and teachers.

This monograph, Reading Recovery: Earl: intervention for At-Risk First Graders, describes aninnovative program that has achieved impressive results with a large percentage of faltering earlyreaders. Reading Recovery originated in New Zealand, and has been a nationwide program in thatcountry since 1979. It has been successfully adapted and tested for four years in Ohio, and is nowbeing disseminated to many other locations throughout the United States, Canada, and Australia.

I have had the opportunity to observe the Reading Recovery program in action first-hand in Ohio,in New Zealand, and in Australia. In each of these varied locations and with a variety of children,the program has consistently produced positivetesults by taking a large proportion of childret. whowere performing in the bottom 15 or 20 perce.it of their class in reading skills and raising thesechildren to the average range for their class in a very short time. Moreover, these gains were con-sistently sustained over the long term without further intervention.

Although the Educational Research Service, in accordance with its standard policy, does not en-dorse any particular program or instructional method, the Reading Recovery results and evalua-tions presented in this monograph deserve the special consideration of educators and concernedcitizens nationwide. The monograph describes these study results, how Reading Recovery oper-ates, and how it may be implemented in local school districts. The monograph is an example of therole that ERS plays in providing dependable information that school officials, other educators, andresponsible citizens need to make sound educational decisions in their states and school districts.As with all ERS monographs, the data and views presented in this publication are solely those ofthe authors, and should not be construed as those of ERS or any of its sponsoring organizations.

ERS wishes to thank the authors of this monograph, Drs. Gay Su Pinnell, Diane DeFord, andCarol Lyons, for the excellent work they have done in explaining in an interesting and under-standable way both the Reading Recovery process and the research on its immediate and long-termeffects on children having difficulty learning to read. In addition, I personally want to acknowl-edge and thank Dr. Marie Clay, Professor of Child Psychology, the University of Auckland, whoinitially researched and developed Reading Recovery, and Dr. Barbara Watson, who is Director ofReading Recovery in New Zealand, for their kind assistance in acquainting me with their researchand their long-term experience with Reading Recover/ in the land of its origin.

ii

Glen RobinsonDirector of ResearchEducational Research Service

6---

Acknowledgments

The story of Reading Recovery implementation is in itself a case study of the kind of inter-institutional and interpersonal collaboration needed to create positive change in education. The firstyear of the project was jointly funded by The Martha Holden Jennings Foundation and theColumbus Foundation, with additional resources contributed by The Ohio State University andColumbus Public Schools. Since summer 1985, funds for training and materials have been pro-vided by The Ohio General Assembly.

From the beginning, a number of exceptional individuals in different organizations have col-laborated to conceptualize and implement the Reading Recovery project. The Ohio Department ofEducation, under the leadership of Franklin Walter, Superintendent of Public Instruction, G.Robert Bowers, Associate Superintendent, Nancy N. Eberhart, Director, and Hilda Edwards of theDivision of Inservice Education, worked to communicate with school districts and other agencies,to create the necessary policy climate, and to solve implementation problems. James Hyre, whowas Superintendent of the Columbus Public Schools at the start of the project, made the first testand longitudinal research for the program possible; the current Superintendent, Ronald Etheridge,has provided the continuing strong administrative support so necessary to the program's long-termsuccess. Evelyn Luckey, Assistant Superintendent, played a critical leadership role in encouragingteachers and administrators to undertake the project. Stacey Woodford, Director of FederalPrograms, and Shirley Mann provided support and encouragement. Teachers in the project areespecially grateful to John Hilliard, Director of Chapter 1, for his problem-solving ability and hiscommitment to making Reading Recovery a quality program for children.

At The Ohio State University, the Office of Research and Graduate Studies, Thomas L. SweeneyAssociate Vice President, has contributed tuition costs for NDN Teacher Leaders. The Office ofthe Provost, David Boyne Associate Provost, has contributed continuing support for the project.The Office of the Dean of Education, Donald P. Anderson, Dean, and the OSU Research Foun-dation have provided valuable leadership. Faculty members involved in tile project especially thankour Department Chair, Frank Zidonis, and our colleagues in Language, Literature, and Reading.

The first teacher leaders, Mary Fried, Arlene Stuck, and Edwina Bradley, now Director of Read-ing, as well as those teacher leaders, site coordinators, and area administrators who have assumedthe responsibility for the state of Ohio program, took on extra heavy work loads and many newresponsibilities. Those leaders and the Reading Recovery teachers have been willing to learn newways of teaching and performing their roles, and they have done so with such high quality thatthey are responsible for the results described here. At The Ohio State University, the ReadingRecovery project staff over the five-year period has included Katie Button, Eleanor Ilanderhan,Kathleen Holland, Nancy Nussbaum, Andrea Mc Carrier, Barbara Peterson, Jim Schnug, KathyG. Shrrt, Susan Tancock, Nora White, and Daniel Woolsey. Special thanks are due to PhilYoung, Jim Rinehart, Will Place, and David Bates for data analysis, and to Linda Hopper,Adrianne Johnson, and Karen Kerr for preparation of reports.

Gay Su PinnellDiane E. De FordCarol A. Lyons

iii

About the Authors

Gay Su Pinnell is Associate Professor of Educational Theory and Practice at The Ohio StateUniversity, where she teaches courses on language arts, literacy, and children's literature. She isthe author of Teaching Reading Comprehension, Discovering Language with Children, and nu-merous articles on language and literacy development. She is Director of the Reading Recoveryproject at The Ohio State University, and Principle Investigator for the Early Literacy ResearchProject. Dr. Pinnell has been an elementary classroom teacher in the California and Ohio publicschools and has served in state administration in Ohio. She received the Ph.D. degree from TheOhio State University, and since has been invoNed in administration, evaluation, teaching, and re-search related to school organization, teacher education, and the development of language andliteracy. She has served as an officer of the National Council of Teachers of English, the Inter-national Reading Association, and the National Conference on Research in English. She currentlyserves on the Commission on the English Language of the National Cornell of Teachers of En-glish. She is a member of the American Educational Research Association.

Diane E. DeFord is Associate Professor of Educational Theory ard Practice at The Ohio StateUniversity. She teaches courses in reading and writing methodology and evaluation. She has au-thored articles in Language Arts, Journal of Reading, and Reading Research Quarterly on the im-pact of instruction on the development of children's reading and writing strategies. Dr. DeFord isa co-investigator in the National Diffusion Network Reading Recovery Program and Early LiteracyResearch Project. She also coordinates the faculty of Language, Literature and Reading. Shereceived the Ph.D. at Indiana University, and has taught at elementary and secondary schoollevels. She is a member of the National Conference on Research in English, the InternationalReading Association, the National Council of Teachers of English, the National Reading Confer-ence, and the American Educational Research Association.

Carol A. Lyons is Assistant Professor of Educational Theory and Practice and Director of theMartha L. King Center for Language and Literacy at The Ohio State University. She is the editorof Literacy Matters and the author of numerous articles on learning, reading disabilities, and teach-er education. Dr. Lyons is Director of the National Diffusion Network Reading Recovery Programand collaborator in the Early Literacy Research Project. She has been an elementary classroomteacher in Tacoma, Washington, New Britain, Connecticut, and Boardman, Ohio. She receivedthe Ph.D. degree from The Ohio State University. She is a member of the National Council ofTeachers of English, the International Reading Association, the National Reading Conference, theNational Conference on Research in English, and the American Educational Research Association.

iv

Introduction:What isReading Recovery?

The early years are crucial in the process ofbecoming literate. Al, °ugh most children makethe "breakthrough" t literacy during their firstyears of school, many find it difficult to learn toread and write.

It is risky to wait and see whether suchchildren will "grow into reading" or "catch on"later in school. When a child cannot read, theproblem soon goes beyond reading. Childrenwho experience reading difficulties quickly fallbehind in school, meet failure repeatedly, andrequire continuous and expensive extra help formany years. Often, they never learn to read well.

Current efforts to help such children requirean enormous, long-term investment of resources.Ui.fortunately, the evidence shows that this re-mediation often fails to help many childrt n withdifficulties. The problem is not simply one ofimmaturity, to be solved by holding children backto give them time to grow. Nor is it a matter ofraising standards so that children are not pro-moted until they are "motivated" or master certainskills. Children who have early difficulty withreading need extra time and special help, and theyneed it in the initial stages of learning. We mnstfind ways to teach the skills children need so thatthey can make timely progress and can functionproductively in school.

Fortunately, research has provided the basisfor promising approaches which can now be ap-

1

plied it-. a variety of school settings. This mono-graph reports the implementation and evaluationof Reading Recovery, which is an early interven-tion Ciort to reach those first-grade children whoare having the most difficulty learning to read andto help them catch up before they fall into a pat-tern of school failure.

The Reading Recovery Program

Liters cy at age six or seven serves childrenthroughout school and frees them to continue toacquire knowlee -:,e and understanding all theirlives. It is crucial, therefore, to ensure that allchildren ha, e access to literacy in the early yearsof their education. That was the goal of the OhioDepartment of Education, The Ohio State Univer-sity, and the Columbus Public Schools when theydecided to try a new program of early interventionfor children who were at risk of reading failure iritheir first year of school.

Originally developed by New Zealand childpsychologist and educator Marie M. Clay, Read-ing Recovery has been successfully adapted andtested for four years in Ohio. It has won supportfrom teachers, principals, school boards, the stateeducation agency, and the state legislature. Stud-ies of the research and development phase dem-onstrate the program's effectiveness across eco-nomic and ethnic groups. Now, Reading Recov-

Introduction

ery is a statewide program in Ohio, existing in228 school districts. Separate school district proj-ects have begun in Arizona; Illinois; South Caro-lina; Texas; Ontario, Canada; and Victoria, Aus-tralia. Reading Recovery has been a nationwideprogram in New Zealand since 1979.

Reading Recovery is based on the premisethat early, high-quality help has the greatest po-tential for lasting impact and for reducing the needfor continued compensatory help.

The program is an intensive one-to-one in-tervention program for the poorest readers (lowest20 percent) in first -grade classrooms, as identifiedby teacher judgment and a Diagnostic Survey.The primary goals of Reading Recovery are toreduce reading failure through early interventionand to help children become independent readers.The program accomplishes this by: 1) bringingchildren who are "at risk" of reading failure up tothe average of their class within a short period oftime, so that they can profit from ongoing class-room instruction, and 2) helping these childrendevelop a self-improving system for continuedgrowth in reading, so that additional help is notnecessary.

Reading Recovery supplements but does notsubstitute for conventional classroom teaching.During daily, 30-minute lessons, teachers whoare specially trained in Reading Recovery tech-niques individually tutor these faltering readers tohelp them develop the kinds of strategies thatgood readers use. The power of Reading Recov-ery is in the framework of the lesson itself and inthe development of teacher knowledge and prob-lem-solving ability. The approach combines theuse of related reading and writing experiences,close interaction between teacher and child wi.hinthe lesson, and careful selection of materials forreading. In this instructional program, the teacherfollows and guides the child individually in his orher use of reading and writing strategies. Theteacher closely assesses and monitors progress

and makes appropriate decisions to accelerate thechild's progress.

Research to date indicates that Reading Re-covery has potential for substantially reducing thenumber of children with reading difficulties. As aresult of accelerated progress, children typicallyleave the program within 12 to 16 weeks and areable to perform at satisfactory levels in readingwithout continued extra help. The sustained suc-cess that Reading Recovery achieves with thepoorest performers in first-grade classes runscounter to the experience in most remedial edu-cation programs.

Unique Featuresof Reading Recovery

A number of specific aspects characterizeReading Recovery and distinguish it from otherprograms designed to help children who havereading problems. Some of these unique featuresare listed below.

1. Early Intervention.Reading Recovery is an early intervention

program rather than a remediation program. Theidea is to provide intensive and focused intervention while the child is in the process of learningthe early strategies of reading. The interventiontakes place before the emotional impact and con-fusion of failure occurs. The program attempts toget children on the right track in reading, thuspreventing further difficulty.

2. Short-Term Extra Help.The program provides temporary help that

enables children :o develop the self-generatingsystem they need to continue learning indepen-dently. Like most remedial programs, ReadingRecovery means taking the child out of the class-room for 30 minutes each day; however, this"pull-out" period lasts a relatively short time andyields a tremendous payoff 5y boosting the read-

Introduction 3

ing skills of a high percentage of at-risk readersup to the classroom average.

3. Building on Strengths.Reading Recovery supports the develop-

ment of reading strategies by helping children usewhat they already know. Some remedial "defi-ciency" models focus on drilling children on thevery items that confuse them. In contrast, theReading Recovery teacher assesses each child'sstrengths in great detail and builds on thosestrengths in daily, individual lessons. Childrengain confidence because they realize that whatthey already know and can do has value in thereading-writing process. More importantly, theylearn specific strategies for applying their ownknowledge.

4. Independence.In Reading Recovery, children learn how to

be independent because they are taught how tosolve problems using specific strategies such asself-monitoring, cross-checking, predicting, andconfirming. They are encouraged to use multiplesources of information while reading and writing;they learn to "orchestrate" strategies while attend-ing to the meaning of the text. The program em-phasizes learning "how to" rather than memoriz-ing any specific list of words.

5. Flexibility and Responsiveness.Unlike other programs, Reading Recovery

does not depend on particular materials. It is notbased on the use of any one set of reading texts orone teaching method. Instead, it depends onteachers developing a systematic knowledge ofthe reading-writing process and helping childrento acquire the strategies they need to constructmeaning from texts.

Once teachers are trained to work with chil-dren in Reading Recovery, they can effectivelyselect and use a wide range of books and can helpchildren use their own writing to assist in reading.

They can perform and record their own assess-ments. No prescribed, step-by-step kits or sets ofconsumable materials are necessary.

6. Action-Oriented.The program is based on the premise that

children are active learners. As they interact withotners and with 'ts, they bring their own mean-ing to the books they read. The instructional set-ting provides the opportunity for children to thinkand solve problems while reading. The teacherprovides choices and support, but the child mustdo the work and solve the problems.

7. Enable,". Participation.Reading Recovery is not specifically tailored

to match the classroom program. However, theteacher is constantly aware of the level the childmust reach to be released from Reading Recov-ery. The program goal is not a set criterion or"gain." Th' aim is to help each child reach theaverage range for the particular instructional set-ting (class or school, whichever makes senseprogrammatically) in which he or she is oper-ating.

Childien who enter this program at sometime during their first -grade experience generallyhave already fallen far benind. They may havedifficulty making sense of much that goes on inclassroom instruction. When a child has movedahead in the Reading Recovery program to thepoint where he or she can read texts equivalent tothe average group in the classroom, then the childcan begin to profit substantially from the ongoinginstruction and can continue to improve in readingwithout extra help.

8. Accelerated Progress.Reading Recovery children are expected to

make accelerated progress so that they can catchup with their peers in the regular classroom set-ting. Intensive individual tutoring by speciallytrained teachers supports the children so that they

4 Introduction

grow better and better at using various strategies.The child does the accelerating, supported andguided by a knowledgeable teacher.

9. Reading-Writing Connections.Every Reading Recovery lesson has both

reading and writing components; learning in eachsituation enhances learning in the other. Writing isused in lessons as a support to developing readingstrategies. Writing allows children to attend to thedetails of print and to deNelop strategies for hear-ing sounds in words, for generating new wordsfrom known words, and for monitoring, search-ing, and cross-checking.

10. Individual Tailoring of Instruction.The lesson provides a framework of acti-

vity; within this framework, however, the pro-gram differs for each child. The difference takesplace in the nature of the moment-to-moment in-teractions between teacher and child, in the rar-ticular texts selected and read, and in the writingwork on a message the child has composed.

11. Teacher Expertise and Judgment.Children are identified for the program by

their teachers rather than by specialists. Thesechildren are the lowest achievers in the first-gradeage cohort, excluding none. Thus, Reading Re-covery provides a good "first net" for childrenwho are most likely to have reading problems. Itenables good teachers to work with children inspecial ways. These teachers, who because of thenature and high intensity of the program workonly half of the day in Reading Recovery, can andusually do teach other subjects during the rest ofthe day.

12. Focus on Meaning.In Reading Recovery, children read f

meaning from the very beginning stages of 'h: itinstructional program. From a list of over 500very short and interesting story books, the teacher

selects those that suit the child's interests, thathave appealing language and stories, and that areat a relatively easy level for the child to read.Thus, at every leve. of text difficulty, childrenread fluently and for meaning and enjoyment.

13. Sound-Letter Relationships.Although the basic approach is to teach the

child to read fluently for meaning and enjoyment,each lesson includes writing, through which chil-dren learn the relationship between the soundscontained in problem words and their relationshipto specific letters and combinations of letters.Thus, the child is encouraged to use tI.J sound-letter relationships as one of the basic strategies insolving problems that he or she encounters whenreading. Unlike some other approaches, in Read-ing Recovery the child works from the sounds inwords to the letters representing those sounds,rather than from letters to sounds.

14. Staff Development.Initial training for teachers takes one aca-

demic year, but Reading Recovery teachers andTeacher Leaders begin to work with childrenimmediately. In the year-len staff developmentprogram, teachers learn to obscfve children's be-havior carefully and systematically, to draw in-ferences from their observations, and to make de-cisions based on that information. From theirwide repertoire of actions, they try to select themost powerful and the most supportive at theparticular time. A key feature of the staff devel-opment program is the extensive use of a one-wayglass through which teachers watch each otherand analyze the child and teacher interactiing invarious situations.

15. System Intervention.Reading Recovery is more than a program

for children and a staff development course. It isa carefully designed set of interlocking principlesand actions that require the long-term commitment

Introduction5

of an entire school system t.. order to ensure aqtntlity program and sustained results.

Contents of the Monograph

The purpose of this monograph is to presentinformation about Reading Recovery, to describethe latest research concerning Reading Recovery,and to summarize practices experience concerningthe implementation of this innovation in readinginstruction. The monograph has been preparedfor the use of school officials, teachers, parents,political leaders, and concerned citizens who areinterested in examining and perhaps implementinga Reading Recovery project.

Chapter 1 presents a general description ofReading Recovery instructional procedures.Iowever, it is not the purpose of this document

to provide specific instructions on how to applythe teaching procedures used in Reading Recov-ery. The procedures for diagnosis and instructionare discussed in detail in the text used in the year-long training program for teachers, The Early De-tection of Reading Difficulties by Marie M. Clay(Heinemann, 1988).

Chapter 2 contains three case studies thatprovide a more concrete look at how the programworks with individual chiluren and teachers.

In the Chapters 3 and 4, we report the re-sults of evaluation studies conducted in Ohio toassess the effectiveness of Reading Recovery.

Chapter 3 discusses a longitudinal studyconducted in the Columbus Public Schools todetermine both the short-range and the leitg-rangeeffects of Reading Recovery on a group of at-riskstudents. These children, who were first gradersin fall 1985, were in the bottom 20 percent oftheir class in reading skills, according to diag-nostic measures and their teachers' assessments.Evaluations through the enu of their third-gradeyear showed that the Reading Recovery interven-tion these children received in the first graderaised a large proportion of them (73 percent) u,

to the average reading level of other first-gradechildren. Most importantly, these gains weremaintained for a substantial number of these chil-dren through the end of grade three without fu.-titer intervention.

The chart below, from Charter 3 page 36,shows that the group of successfully discontinuedReading Recovery children (those who were suc-cessfully released from the program) made sus-tained gains compared with the band of averagescores of a random sample of all first-grade stu-dents, and also compared with a group of similarchildren who received an alternative interventionprogram.

Imen,noon Pared Mein 13 Weeks

El Random 5' ^pie Mean Avenge Band( Sum moon)

Sta,essfulty Dr.onenued Readin8Recovery Chddrea

cr Cala-rum Cluldral

September 1985 May 19136 May 1987 May 1988

Chapter 4 describes the studies of ReadingRecovery at sites throughout the state of Ohioduring the years 1985-86, 1986-87, and 1987-88.Children who received Reading Recovery in-struction at these sites were assessed on variousmeasures and compared with a random sample offirst graders at the end of their respective first-grade years. The results statewide supported thepositive findings of the longitudinal Columbusstudy.

Chapter 5 describes the Reading Recoverystaff development component, along with studiesof teacher training and development in programtechniques. Finally, Chapter 6 presents sugges-tions for school districts or state agencies thatwish to implement Reading Recovery

3

Chapter 1:Description ofReading RecoveryLessons

Up went the giraffe.

This chapter describes how the ReadingRecovery program works for children: how thelesson format was developed, how the lessons arestructured, and the theoretical assumptions onwhich the instruction is based. This general de-szription is followed in Chapter 2 by three casestudies showing the difference that Reading Re-covery has made in the lives of individual chil-dren.

Reading Recovery provides individuallydesigned lessons to help children who are havingreading difficulties, to develop the kinds of stra-tegies used by good beginning readers. This goalis accomplished through teachers interacting withchildren who are engaged in holistic reading andwriting activities.

Development of the Process

Marie M. Ciay, a professor in child psy-chology at the University of Auckland, who de-veloped Reading Recovery in New Zealand,began her research with detailed observations ofgood readers in the early stages of learning toread. After constructing knowledge of just whatthese good readers do, she looked at children whowere having difficulty, asking the question: "Canwe see the reading process going wrong in thefirst year of schooling?" As teachers of young

7

14

Up wtnt the zebra.

children can verify, the answer to that questionwas "yes" (1988).

Clay went on to design and experiment withintervention procedures based on her detailed ob-servations. Acting on their observations whir-working with children, teachers sensitively inter-vened to support children's development of stra-tegies. The goal was to help children expand therange of strategies available for their use.

Next, Clay's research team constructed alesson framework. The activities were selectednot as a "lesson plan" with a script to follow, butas a set of generative activities that would provideplenty of opportunity to read extended text, to talkabout what was being read, and to use the fullrange of ix.formation sources available for con-structing meaning. After pilot Reading Recoveryprocedures were further refined, the staff devel-opment program was created. Reading Recoveryhas been a nationwide program in New Zealandsince 1979.

Diagnostic Procedures

In Ohio, children are selected for Readir gRecovery in the middle or late September of theirfirst-grade year. All children selected for ReadingRecovery must be in the lowest 20 percentachievement group of their first-grade class in

8 Description of Reading Recovery Lessons

reading. The Reading Recovery teacher selectsstudents by using a combination of measures, in-cluding the classroom teacher's ranking, the kin-dergarten teacher's opinion if applicable, scoreson the six measures of the Diagnostic Survey, andany additional information, such as standardizedtest scores, that may be available. (In districtswhere Reading Recovery is supported by Chapter1 funds, all children served must also qualify un-der Chapter 1 criteria.)

Prior to beginning a Reading Recovery pro-gram, children are assessed using the compre-hensive set of individually administered instru-ments that make up the Diagnostic Survey. Toadminister the Diagnostic Survey, teachers in-volve the children in six assessments, each pre-senting a different aspect of reading and writing.The goal is to gain an understanding of what thechild already knows about reading and writing.

There are several. important points to noteconcerning the Diagnostic Survey. First, most ofthe measures involve children in reading andwriting tasks. Throughout the testing, whichtakes about one hour, the teacher and child inter-act in an informal way with books and throughwriting.

Second, no one of the measures is intendedto be used as the sole determinant of a child'sprogram. Reading Recovery teachers look at thechild's behavior across all measures to summarizerelevant information about the child. This sum-mary is only the beginning of the teacher's de-tailed observation of the child's behavior. He/shewill spend the first 10 days of the program inter-acting with the child and observing closely thereading and writing behavior that provides cluesto the chilfs additional knowledge.

Third, scores on the assessment instrumentsare less important than the. observations and notesmade during the assessment and teaching ses-sions. Teachers are prepared to notice significantbehavior and to draw inferences to build theirknowledge of the child's competence.

Fourth, these assessments should not beconfused with the instructional program. They areintended to provide a broad first look at the child.Several of the tasks for example, writing allthe words the child knows are not used ininstruction. '.successful release from the programdepends on a qualitative look at the documentationof the child's progress over time.

The Diagnostic Survey includes the follow-ing assessments:

1. Letter Ident!fication. Children are asked toidentify as many as they can of 54 characters(the entire upper-case and lower-case alpha-bets, plus the alternative printed forms of "a"and 'g"). They may identify the name of theletter or the sound the letter makes, or theymay suggest a word that starts with that let-ter. Any of these responses would be consi-dered correct. Teachers notice the kinds ofsubstitutions children make as well as theiraccurate responses; for example, calling a "j"by the name of "t" may indicate awareness ofdistinctive features. This assessment is usednot because children must be able to name allthe letters in order to read; rather, teachersmust find out what the child knows aboutletters to help integrate this information intothe instructional program.

2. Word Test. The word list used in Ohio wascompiled from the high frequency words ona Dolch word list. Clay (1988) advises thatthe list be made up of the most frequentlyoccurring words in whatever oasic readingtexts are being used in the system. This testhelps the teacher get an idea as to whether thechild is acquiring knowledge of frequentlyoccurring words; it does not provide infor-mation as to the child'i ability to read ex-tended text. Again, the assessment shouldnot be confused with instruction. At Ito timein the Reading Recovery program is the childasked to read isolated words.

16

Description of Reading Recovery Lessons 9

3. Concepts About Print. The teacher and childinteract as the teacher reads a little book withpictures. The teacher questions the child inorder to assess the child's development ofsignificant concepts about printed language.For example, the child is asked to show aletter or a word, the front of the book, wherewe start to read, avid which way we go whenwe read.

4. Writing. Children are asked to write all thewords they can write (on a blank piece ofpaper) during a may'-num of 10 minutes.After the child exhausts his/her supply ofknown words, perhaps beginning with thechild's name, the teacher prompts from a listof high frequency words. The teacher noteswords at which children make good attempts,because those show competence and knowl-edge.

Reading Recovery teachers and classroomteachers may also examine writing samplesproduced by the child in the classroom set-ting, to gain as much information as possibleabout the child's knowledge of writing. (Aninformal period of two weeks will follow theassessment, auring which the child will en-gage in writing, and the teacher will have achance to observe thel,rcerss.)

5. Dictation Test. The reads a simplesentence, containing : 7 tut* -1, , and asksthe child to try to ,.; I,. e, ,e are in-terested in the ,:-v u analyze aword and to repr- T sounds heard.Accurate spelling the Foal.

6. Text Reading. The teacher takes a "runningrecord" of the child's reading of an extendedpiece of text. For a child who cannot yetindependently read even very short books,the teacher does most of the reading aloudand asks the child to read predictable bookswith repeated language patterns. A childwho can read a little is a' 1:ed to read textswhile the teacher uses checks and other sym-

bols to record reading behavior. Th-n, theteacher analyzes the record, looking for evi-dence about how the child uses the cueingsystems in reading (meaning, languagestructure, or visual information) and gettinginformation about the complex processes go-ing on during reading.

The Text Reading level is a numericalscore and refers to the level of difficulty achild can read with 90 percent accuracy orabove. In addition to level, the teacher makesa qualitative assessment of the child's readingbased on the behavior observed in readingvarious texts, from hard to easy.

This list provides only a brief description ofthe Diagnostic Survey. For a full description anddirections for administration and use, see Clay,The Early Detection of Reading Difficulties,1988, third edition.

In all of these assessments, teacher judg-ment and ability to analyze are the critical factors.The process produces a set of numerical scoresthat can be quantified and used as justification forproviding special help (for funding agencies, forexample) or as documentation of progress. Byadding the q'ialitative analysis, the teacher buildsthe foundation for the instructional program. Thisanalysis provides the basis for the DiagnosticSummary, a written document in which theteacher brings all the test results together. Theteacher looks across assessments to make a set ofsummary statements that will provide a startingpoint for Reading Recovery instruction.

"Roaming Around the Known"

For the first 10 days of the child's 30-minute daily program, the teacher does not teach,but rather, explores reading and writing with thechild. During this time they can talk together,enjoy boas and collaborative writing, and get toknow each other. The teacher has some basic in-


formation about what the child knows, and usesthis informaticn to involve the child in very easytasks that make the most of what the child can do.

In this very supportive situation, the childmay begin to take risks and to produce responsesthat have not been evident in the classroom ortesting situation. By the end of the "roamingaround the known" period, the teacher has a muchbroader knowledge of the child and a betterknowledge base on which to proceed. Addition-ally, a foundation of trust has been establishedand the teacher and child go into more intensivelessons with greater confidence.

The Reading Recovery Lesson

Each Reading Recovery lesson includesreading many small books and composing andwriting a story. The lesson framework includesthe following.

The child rereads familiar books.The child reads again several favorite books

that he/she has previously read. The materials arestory books with natural language rather thancontrolled vocabulary. Books within a lessonmay range from quite easy to more challenging,but the child is generally reading above 90 percentaccuracy. The accuracy at here guides theteacher in making sure that the texts selected areappropriate for the child; that is, they are easyenough for the child to use effective strategies,and difficult enough to provide opportunities forindependent problem solving.

In addition to the accuracy index, the teacheralso assesses the balance of strategies and cues.During this time, the child has a chance to gainexperience in fluent reading and in using stra-tegies "on the run" while focusing on the meaningof the text. The teacher interacts with the childduring and after the reading, not "correcting," buttalking with the child about the story and sup-porting the effective actions the child has taken.

The teachers analyzes reading using therunning record.

Each day the teacher takes a "running rec-ord" of a book that was new fin the child the pre-vious day. The running record is a proceduresimilar to miscue analysis (Goodman, Watson,and Burke, 1987). Using a kind of shorthand ofchecks and other symbols, the teacher records thechild's reading behavior during oral reading of theday's selected book. The teacher examines run-ning records closely, analyzing errors and payingparticular attention to behavior such as self-correction. In this way, he/she determines thestrategies the child is using to gain meaning fromtext. This assessment provides an ongoing pic-ture of the progress the child makes.

The Reading Recovery teacher does notconsider one record an adequate source of evi-dence about a child's reading. He or she looksacross records, taken daily over a period of time,to discern patterns ano change. During this time,the teacher acts as a neutral olrerver; the childworks independently. The accuracy check tellsthe teacher whether the text has been well selectedand introduced the day before.

The child writes messages and storiesand then reads them.

Every day the child is invited to compose amessage and to write it with the support of theteacher. Writing is considered an integral part ofgaining control over messages in print. The pro-cess gives the child a chance to closely examinethe details of written language in a message thathe/she has composed, supported by his/her ownlanguage and sense of meaning. Through writ-ing, the child also develops strategies for hearingsounds in words and using visual information tomonitor and check his/her own reading.

After the construction of the message, theteacher writes it on a sentence strip and cuts it upfor the child to reassemble and read. This activityprovides a chance to search, check, and notice

.7


visual information. Using plastic letters on amagnetic board, the teacher may take the oppor-tunity to work briefly with the letters to increasethe child's familiarity with the names of lettersand their use in known words, such as the child'sname. This work will vary according to theknowledge the child already has.

The child reads new books.Every day the child is introduced to a new

book that he/she will be expected to read withouthelp the next day. Before reading, the teachertalks with the child about the book as they look atthe pictures. The teacher helps the child build aframe of meaning prior to reading the text. Thepurpose of the introduction is not necessarily tointroduce new words, but to create understandingin advance of reading so that it will be easier tokeep a focus on meaning.

This basic framework for the Reading Re-covery lesson provides a guide, but the teacher'sown knowledge of the child and of the repertoireof possible variations make it possible to indi-vidualize the lesson. Within this framework, ev-ery child's program differs. Children do a greatdeal of reading, but not from a graded sequence.No child reads the same series of books. Thesmall books are carefully selected by the teacherfor uat child at that time. In writing, childrenwork on their own messages, so they are writingand reading works that are important to themindividually. The special techniques used in thewriting part of the lesson are most powerful whenused on the children's own produced text. Themajor difference within and across lessons lies inthe teacher's ability to follow each child and torespond in ways that support acceleration and thedevelopment of strategies.

Meeting the Child'sIndividual NE zds

Reading Recovery teachers recognize thatthe difficulties children experience in learning to

read differ greatly from child to child. Therefore,although all Reading Recovery lessons folly -I a

standard structure, within this structure the teach-er carefully selects the activities needed by eachchild at a particular time. Throughout the lessons,the teacher looks for effective reading strategiesthat the child needs to acquire or strengthen.Such strategies may include directional move-ment, one-to-one matching, self-monitoring,cross-checking, using multiple cue sources, andself-correction. The Reading Recovery teacheruses instructional techniques designed to help thechild develop and use such strategies.

As one example of the different instructionaltechniques that the teacher may weave through thebasic lesson to encourage a specific reading skill,a section of The Early Detection of ReadingDifficulties (1988) is reproduced below. Clayoutlines the following suggestions for teachingthe skill of self-monitoring.

The successful reader who is making no errors ismonitoring his reading at all times. Effective moni-toring is a highly skilled process constructed over ma-ny years of reading. It begins early but must be con-tinually adapted to encompass new challenges.

To encourage self-monitoring in the very early stagesask the child to go back to one to one pointing:Say 'point to each one.'Or 'Use a pointer and make them match.'Direct the child's attention to meaning:Say 'Look at the picture.'Or 'What happened in the story when. . .'

For particular attention to an error allow the child tocontinue to the end of the sentence:Say 'I like the way you did that.

But can you find the hard bit?'Or 'I like the way you did that.

You found the hard bit.Where was it?'

If the child gives signs of uncertainty hesitation,frowning, a little shake of the head even thoughhe takes no action:Say 'Was that OK?'Or 'Why did you stop?'Or 'What did you notice?'

These questions tell the child that you want him tomonitor his own reading. The operation to be learnedis checking on oneself. It is more important that thechild comes to check on his own behaviour than that

12 Descri Lion of Readin Recover Lessons

he be required to use all the sources of cues at thisstage.

Don't forget to reinforce the child for his self-mon-itoring attempts whether they are successful or not.Say '1 liked the way you tried to work that out.'

Cues from letter sequences. Let the child predict theword he expects it to be. Cover the problem wordand ask for something you know he knows aboutthat word. One of these questions might be useful.'What do you expect to see at the beginning?'

at the end?after the 'M' ?'

Then ask him to check as you uncover the work.Ask the child 'Where you right?' after both correctand incorrect words. Ask 'How did you know?' aftercorrect words.As the child becomes more skilled do less teachingand prompting and modelling. Merely say 'Try thatagain' but make sure that your voice carries twomessages. You are requiring him to search, becauseyou know he can, and you are confident he can solvethe problem. (Clay 1988, pp. 72-73)

Completion of the ProgramThere is no set time or sequence of activities

to finish in order for the child to be released fromthe Reading Recovery program. Instructioncontinues until the individual child has reachedabout the average level of text reading for his/herclass or school. In addition, analysis of thechild's reading behavior must provide substantialevidence that he or she has developed effectivereading strategies and will be able to continuelearning without extra help. This may happen anytime during the school year. A typical programcould last for 12 to 16 weeks. Clay's guidelinesstate the following:

There is no fixed set of strategies nor any requiredlevels of text nor any test score that must beattained to warrant discontinuing. It is essentialthat the child has a system of strategies whichwork in such a way that the child learns from hisown attempts to read. (Clay 1988, p. 82)

The goal of the program has been achieved whenthe child has developed the kind of independentreading system that good readers use. Then, thechild can profit from the ongoing instruction inthe re vlar classroom and stands a good chance ofsurviving without compensatory help.

MaterialsThe Reading Recovery program uses few

consumable materials. There are no workbooksor worksheets. Instead, teachers use blank writ-ing books and pencils or markers. They also havea set of magnetic letters and a small magneticchalkboard.

The major materials for the program are thehundreds of little books that the children read.Books come from many different publishers andhave been selected for their potential in supportingliteracy development for young children. Theyinclude many different stories that offer supportfor readers by using familiar language patternswithin the framework of a predictable story. Theeasiest levels include repetitive or patterned lan-guage; more difficult levels present a wider vari-ety of writing styles.

Books are organized into 20 reading levelsfor teachers to use in tracking children's progressand as a guide in selecting the daily new book.Level 1 approximates a pre-reading stage inclassroom instruction and indicates that the childcan read very little beyond his or her name in anunsupported situation; Level 20 approximatesmaterial that good first-grade readers can read bythe end of the year.

A committee of Reading Recovery TeacherLeaders has emphasized that "the Booklist is not alist of required or recommended books but a re-source for Reading Recovery teachers. Selectingthe appropriate new book is a decision-makingprocess that draws on a knowledgeable teacher'sunderstanding of a child's current use of stra-tegies and need for engaging in some readingwork" (Report of the Committee on Books,1988). Appendix A contains an abbreviatedbooklist.

Theoretical PrinciplesTo summarize the program description, we

would like to emphasize six theoretical principles

10


that serve as the foundation for the Reading Re-covery program for children.

1. Reading is a strategic process thattakes place inside the reader's mind.

Readers are required to monitor and to useinformation or "cues" from a range of sources,including meaning, language structure, features ofprint, visual information, letter-sound relation-ships, and connections with individuals' ownunique backgrounds. Through complex, "in thehead" processes, called "strategies," readers ac-cess the information they need to construct mean-ing from written text. The meaning is never con-tained only in the print; readers bring their ownmeanings to the text. Therefore, even beginningreaders need to go beyond simple decoding and tomaintain a focus on meaning throughout all lit-eracy activities.

Good readers have several ways of func-tioning according to the difficulty of the material.They "orchestrate" strategies, simultaneouslymonitoring cues while maintaining a focus on themessages. Poor readers, on the other hand, mayoperate on a narrow range of strategies, perhapsfocusing on one kind of information and ne-glecting others. The goal of Reading Recovery isto help children become good readers who canuse effective strategies in a flexible and integratedway. For those readers, skill improves wheneverthey read because they have developed a "self-improving system."

2. Reading and writing are intercon-nected, reciprocal processes.

As children read and write, they make theconnections that form their basic understandingsabout both processes. Learning in one area en-hances learning in the other. Discovering andusing reading-writing connections may be an im-portant part of the process of becoming literate.

Children becoming literate especiallychildren at risk need many opportunities for

exploring and relating the two processes.Throughout the Reading Recovery program,reading and writing are used flexibly to help chil-dren develop concepts and skills. The supportivesituation allows children many chances to makeconnections between reading and writing. Teach-ers actively support that process.

3. In order to make accelerated progressin reading, children must actually engagein reading.

Almost every minute during the lesson,children actively engage in reading or writingmessages and stories. Familiar material helpschildren build fluency and experience success;new material challenges children to do indepen-dent problem solving. The teacher selects textsthat are clear, interesting, and easy for the childand that include language close to the child's na-tural way of talking. These texts shoOd provideopportunities for the child to apply new re-spoir , skills, or procedures.

4. Literacy instruction in school has apowerful influence on children's devel-oping concepts of what reading is allabout.

This statement is especially true for childrenat risk because they are vulnerable to the schoolexperience. Programs linked to abstract drill onsmall segments of language may not provide theexperience in reading whole texts that childrenneed. Children may become convinced that read-ing is only looking at words or letters and makingsounds; those readers may fail to integrate theisolated activities into the larger process of con-structing meaning from text.

On the other hand, programs that assume a"macro" view and emphasize only language andmeaning may not give ai-risk children the specialsupport they need in order to focus on details ofprint within a meaningful context. Those children

ti ')


may not know how to use what they alreadyknow to make sense of reading.

As they learn to read, children develop theirown theories about the process of reading; theyneed experiences that help them develop an impli-cit understanding of the whole range of informa-tion that they must use in reading and writing.

5. It is most educationally productive toirtervene early.

In the I ist, many educators have believedthat children need to "mature" into reading andthat given time and a rich environment, all chil-dren will become literate. In our view, develop-ment has a role, but children's experiences havean enormous impact. For the small group ofchildren who, no matter how good the classroomteaching, have great difficulty in learning 'o read,it makes sense to intervene early, before the childis locked into unproductive responses and expel:-ences the frustration of failure and accompanyingdeficits in other areas of the curriculum. Th --_:.children need sensitive early intervention with .---high quality program that involves the child in realreading and writing.

We can see the reading process goingwrong in the first year of instruction; we canidentify causes of the problem; we can identify thechild's strengths; we can trace subtle shifts thatindicate progress; and we can assist those childrento develop the same abilities that good readershave. Since we can do it, we are obliged to do it.

6. Accelerated progress is possible.Children at risk can make the accelerated

progress needed in order to catch up with theirpeers. Acceleration is not achieved by applyingpressure or making the child struggle. It is notReading Recovery's goal to force fast pacing oraccuracy. In our experience, young children dowant to read some want it desperately andwith the right support they will learn.

Because the teacher and child are workingtogether, the at-risk child can achieve more thanwould be possible alone or in a group. Accel-eration is achieved as the child takes ,ver thelearning process and works indepen dly, dis-covering new things and pushing the uoundariesof his/her own knowledge.

Chapter 2:Case StudiesOf Children

The following three case studies, as told bythe children's Reading Recovery teachers, give anidea of how the program works for real children.The case studies also allow insights into the viewsof the teachers involved. These three childrenprovide "living images" that represent importantinformation for program implementers. Theyattend both urban and suburban schools, and theyrepresent different races. All three had great dif-ficulty in beginning reading.

As these case histories illustrate, traditionalprogram evaluation is only part of the data to beexamined in implementing a Reading Recoveryprogram. Raising test scores is important, but itis also important to learn more about how indivi-dual children develop their own successful read-ing strategies. Detailed analyses of individual les-sons provide insights into teacher-child interac-tions that produce success as well as ways teach-ers can tailor instruction to meet individual needs.

KEY TO NOTATIONS USED IN RUNNING RECORDS

The figures that appear in the following pages illustrate teacher running records made during Reading Recovery lessons inthese case studies. Notations in the running records include:

Notation

6144'5 refPonse-eerrea word

Sc

Meaning

Child read word correctly.

Child read work incorrectly. Child's response is wrinen aboveline, and correct stJrd is written below line.

Child self-corrected.

Child returned to beginning of arrow and reread.

Child skipped word.

The columns on the right of some running records (. ginning in Figure 3) are used to tabulate (first two columns) and thento analyze (second two columns) errors and self corrections. The letters are codes representing the probable source ofinforms-ion being used by the child:

M=using information from the meaning system.S=using information related to the structure of language.V=using visual information, including letter/sound corr, once.

15

16 Case Studies of Children

MELANIEby Andrea Mc Carrier

Melanie was the lowest reader in her first-grade class when I began working with her in theReading Recovery program. She was able toremember and use language patterns in reading,but she appeared uncertain about whether picturesor print carried the message. She did not havecontrol of left-to-right dhectionality or one-to-onematching in reading. Consequently, she tended toinvent the text rather than attend to the printedmessage. She had a strong sense of story struc-ture and could predict a meaningful message, butshe did not notice discrepancies between her ownreading and the written text.

A good example occurred during her thirdlesson. (See the running record in Figure 1, page17.) She was reading The Tree House, a bookthat she had read for the first time the day before.When she came to the last sentence, one with aninverted structure, Melanie's strong sense of orallanguage patterns overrode the visual information,so that she read the sentence as it would morecommonly occur.

Based on the diagnostic tests and on manyexamples such as the one above, I decided thatMelanie needed to learn to check her predictionswith the visual information in the text. I wouldcontinue supporting her sense of meaning and useof language to predict, but I would also encourageher to attend to the pri.it. It would also be impor-tant to encourage her to point to words as she readto build her knowledge of early strategies such asdirectionality and one-to-one matching.

As Melanie gained experience in reading,she became more aware of 1) a mismatch betweenthe number of words in the text and her reading;and 2) discrepancies between her oral renditionand the print on the page. She began to monitor

her own reading and to hesitate and self-correctwhen appropriate.

Melanie also began to attend more closely toprint as she wrote her own messages and thenreassembled them after they had been written on apaper strip and cut apart. This activity gave her achance to notice visual details within a languagecontext that was particularly her own. She beganto look more closely at initial letters and to use herknowledge of the alphabet. Writing helped her toslow down the process so that noticing visualdetails would not distract her from the meaning ofthe language.

She quickly gained control over early stra-tegies, and soon she did not need to point towords hik reading, although she continued touse this technique in a flexible way when she readdifficult pieces of text. As her lessons pro-gressed, she continued to learn more about howto integrate visual cues with other sources of . '-formation. For example, when she read You 'iSoon Grow Into Them, Titch, she sut,c!:::::.,'"socks" for "sweater." Both items ut clothingwere depicted in the illustration, and the twowords started with the same letter. By carefullychecking her prediction with the details of print,Melanie was able to self-correct, therefore gainingan understanding of the text.

Even though she was working in a class-room where invented spellings were acceptableand writing was a daily activity Melanie wasreluctant to attempt writing her own stories. Bywriting every day in Reading Recovery, Melaniediscovered how to make her own sound analysesof words and to represent sounds with letters inwriting. She began to take more risks and to

Case Studies of Children 17

produce more writing both in the individual ses-sion and in classroom work.

By the end of her program, Melanie wasreading fluently in materials comparable with theaverage reading group in class. She continued tocheck the illustrations for information but did nothave to depend on them totally. She displayed heability to use multiple sources of information asshe read, and she showed evidence of being anindependent reader. In the example illustrated ir.Figure 2, Melanie showed self-correction and the

ability to cross-check one source of informationwith another.

In the last cunning record taken in the pro-gram (Figure 3) Melanie showed her ability toread a difficult text She focused on meaning andused her ability to predict; as competent readersdo, she made meaningful miscues that did notneed to be corrected, but she was able to solve herown problems when she had difficulty in gettingmeaning from her reading.

FIGURE 1.In her third lesson, Melanie's strong sense of oral language patternsoverrode visual information.

The. -f-ree.Down c.aryle.

house came clownthe tree- house

V

FIGURE 2. Melanie showed self - correction and ability to cross-check.

Ithe next dayhe said to his friendsyesterday I could read.

7R lesier41 Ada" neX + .5

V V V.../..e sc.

yesterday lk

SC

0

18 Case Studies or Children

Melanie was discontinued from the programafter 55 lessons over a span of about 16 weeks(with some time gaps because of vacation periodsand absences). At that time, she was able to readbeginning second-grade material. Her motherreported that her reading grade had also improved

from a "D" at tlit beginning of first grade to an"A" at the beginning of spring and that Melanieenjoyed reading at home. Melanie thinks her twoyounger sisters should not have to wait until firstgrade to learn how to read. According to hermother, Melanie is giving them lessons now!

FIGURE 3.Last running record taken in program: Melanie read a difficult text.

There's et. N n my c/oret Accvnitofr. %7uSel,

,...- ,--- ..- ...-- ..... ., ,.,--There used is, be- a ni,Pd-rnare. in my closet.Sere, 90749 -kg sleep;

...-

I 'always dosed 4-he. c/o sel" door,...- v-

turnwas even c4rotrd 1 "furnarogeld and /crok.

,

bed,When I was $a4-e in ed)S did ./vL'a peek . . . sorrel-in-1es

r..0

The. Crest 8t9 Enormous 7Ur ci ipPage Level la.

3

Accwucy= 99%c ties inn

v_ ..- w- ,...- .-- to."" I sOnce .vpon ta. -lime an old man f o Aifel4;1 /;/titltle h.-bnie

.--, ..- v. , e- .1 ...-The old Alan -aia, o-rb4A.), 9 ro64J bifie -knit,.

V le' 1Crew s weer....- ,....-

..-3---sc,r,,..07 orcria.); it Pie #Urn i p .

Crtv tAJ 5hwho.u

o22kSC itArid Ae. turnip y-ew up sweet and

,..- 1,- ., ota = ....-

Sfirm9 elhel lo an enormou S .

493086V

'Of

iv&

MC)

49


TIMby Carol A. Lyons

The principal of Tim's school pulled a thickfolder from her top desk drawer and began todescribe this first-grade boy who had just spentone year in a kindergarten class for learning dis-abled children. The principal said she had neverreceived so many records on a young child in her10 years of experience as an administrator. As Iexamined the records, I wondered how a six yearold could have endured so many standardizedtests. Two teams of evaluators, school psychol-ogists at a private clinic, and an interdisciplinaryteam of professionals had diagnosed him aslearning disabled. The problem was that no LDclassroom would be available until second grade.Tim's parents felt sure that he could not survive ina regular first-grade classroom. We ended theconference with my agreeing to work with Tim inthe Reading Recovery program.

After administering the Diagnostic Survey, Ipredicted that Tim would have a good chance forsuccess in the program. Pe had control of someearly strategies such as word-by-word matchingand directionality, and he knew most of the al-phabet letters. He could represent 16 of the 37phonemes on the dictation test. On the otherhand, he seemed confused about how to use hisknowledge when he was actually reading text,and he could not recognize words 'Ai isolation. Ihad discovered, however, that Tim had quite a bitof knowledge about reading and writing. Heneeded to learn how to make the most of hisstrengths.

For the first two weeks of Tim's program,I watched for and recorded what he knew aboutreading. I read many books to him, and soon wewere reading favorite books together. He alsoread many very simple books to me. We collab-

oratively wrote messages and whole stories thatTim read in subsequent lessons.

During this two- week "roaming around theknown" period, Tim demonstrated many morestrengths that \ 'ere not evident even in the indivi-dualized testing in the Diagnostic Survey. Forexample, Tim was well aware of the special lan-guage used in books. When we read together, hecould easily move his finger under the words.However, when he tried to read more than oneline of print independently, he was not able toattend to the print. Instead, he invented a text thatwas meaningful and corresponded to the illustra-tion, but he did not monitor his reading usingvisual information.

Initially, Tim did not want to write any-thing. He had no trouble creating sentences; infact, he usually composed several sentences. Buthe hesitated to write. As we worked together, hewas persuaded to provide pats of the messageshe composed and his strong awareness of letter/sound relationships was evident in his writing.He was particularly good at analyzing words forinitial sounds, and his reading showed evidencethat he could use this knowledge.

In the first two weeks, when we were read-ing together and the texts were very simple, Timwas growing in confidence, and his mother re-ported that Tim was developing "a new attitudetoward reading." Although he did not give a greatdeal of attention to visual information, he freelyinvented meaningful text based on his own lan-guage sense.

When we began more structured lessons,however, I had a surprise. Tim seemed to aban-don his own language sense and meaning as asource of information in reading. He appearednot to be attending to meaning. Instead, he tried

-U


to look for the individual words he knew. Onunfamiliar words, he guessed, using the first let-ter as the main source of information. He almosttotally disregarded the pictures which would havebeen good sources for prediction. Early in hisprogram, he provided an example of reading (seeFigure 4 on page 21).

When Tim finished reading The Bicycle, Isaid, "Tim, you said 'the lake got on.' Check thepicture. Did a 'lake' get on the bicycle?"

"Oh," Tim said, carefully looking at thepicture. "That's silly. How could a lake get on abicycle? It's a girl."

Here, I asked him to predict what letters hewould expect to see in "girl" and check (tie wordagain. He was able finally to select "lady" as theword in the text and to reread the section to makesure that it made sense.

Tim continued in this kind of problem-solving work while reading. He began to regainhis expectation that reading should make sense,and was not content to produce nonsense even ifit did match the visual information. He becamemore consistent in reading for meaning and inusing his own strong sense of language to predictwhat he thought the text might say. He couldcheck those predictions with his knowledge ofletters and sounds. During the next period I pro-vided increasingly more difficult texts to try togive him more chances r problem solving. Therunning record shown in Figure 5 reveals Tim'sprogress as he independently read Mouse. At thispoint, he had shifted from an over-reliance onvisual information to a more balanced set of stra-tegies. In this example, Tim often reread, gettinga "running start" to establish the language patternsand use them to predict. He was self-correctingand reading for comprehension.

Tim also made good progress in writing.He had no trouble composing a message; his sen-tences often contained high frequency words thathe knew how to write. With unfamiliar words,he would produce all the letters he could, and I

would fill in the rest. He was good at analyzingsounds in words, although he could not neces-sarily represent them in sequence. The underlinedwords and letters in Figure 6 show what Timwrote independently The numbered boxes at thetop indicate letter sounds that he produced whenas_ed to sound out the word. Note that he pro-duced these sounds out of sequence; the endingsound of each word was the first one produced.

Sevaal weeks later, Tim could analyzesc ,r. ds in sequence and write much more inde-pendently, as illustrated in Figure 7.

Tim continued to make good progress andwas discontinued from the program in less than10 weeks, when he was placed in the middlereading group in his first-grade class. In secondgrade, he was placed in the highest reading groupand has remained a good reader through the thirdgrade. In September, he will enter fourth gradeas one of the best readers in his class.

I suspect that Tim believed that reading wasonly sounding out letters and visually analyzingwords. He seemed to rely on visual informationand to ignore his own sense of meaning andknowledge of language.

Tim's early responses to books, in protectedhome situations, may have been like those he dis-played during the "roaming around the known"period. According to his parents, the LD kinder-garten curriculum focused on isolated letter nam-ing and letter/sound relationships. Tim rarely readany books at school, although his parents read tohim at home. Could his 1; ter reading behaviorhave been influenced by the instructional programin kindergarten?

If the answer is yes. then Tim may havelearned to be "learning disabled." His develop-ment of a broader range of strategies illustratesthat he could learn. My hunch is that Tim wasnever really learning disabled. A more appropriateterm might have been "ir uctionally disabled."

Case Studies of Children21

FIGURE 4.-In Tim's early reading, he disregarded his own language sense and meaning.

week. 01e. Level 3 Bic Cie. /4CCUrac sg 7. cads used

3

7 ,7

b/-The clown 9trt-

tiakeand The 74.- d: .9;;;.- on

bLand the boyo901L on

.1.1122

and nie 9of ors

ca'"nel eelt 9;4' onoriel the bicYcle jot s;a7Leet.

ScE

I

sis0

St,

FIGURE 5.-Later, Tim used a balanced set of strategies to read for comprehension.

mouse Leve/

3

1/

6

&conic co7. tuaki9 Cues uud0"17- v VOut of the. e.ff.j...Aise

tree, treev Nv

E StI

1

1

i

I

1

e80v

one

SOY

eitv

MYpv

s e.

ThsQ

&Dv

in 90

his@

MsNIS@

fr:mu)04.36, t ke - "-ass

creep creep JR/sc. 1'f/10-doesn'tmake

nses+ the !RI S e

creep creepV .... ,../

tinder +Ae "lourv- ....-

ereep creep

muld ., 1

Across the. floor 'RISC.ere!! creep_rnliSc C cli 7 j ± .N 1 4 Z ) S cInto +he. Cue board 1

1.--Creep Creep

v.- v-- ..- v-tAp +o the cheesetree, Creep

22Case Studies of Children

FIGURE 6.Tim's writing, Week 1. He produced letter sounds out of sequence.

F1-1_ erd-L

funon my

le 4

TD

FIGURE 7. Tim's writing, Week 4. He wrote more Independently and produced letter

sounds In sequence.

Z WOs In tie fourth

oc J',,/i/

Gls.

f Ora de

yeo r,


SARAHby Diane DeFord

Sarah was seven when she was evaluatedfor Reading Recovery. She was among the low-est in her first -grace class, and her parents, bothteachers, were concerned about the possibility thatshe might be retained.

Sarah had been read to by her parents sinceshe was a toddler, and by her teachers in kinder-garten and first grade. She was in a rich literacyenvironment, and had caring adults who sup-ported her learning. However, by the beginningof November, she was still being considered forlearning disabilities tutoring and was at risk with-in the classroom setting. After initial testing,Sarah was identified as one of the children to beincluded in the Reading Recovery program.

The early diagnostic testing indicated thatSarah had many strengths. She identified 53 ofthe 54 letters, was able to read 8 words out of alist of 20 basic vocabulary items (and, the, down,am, there, little, them, yes), and was familiar withmany book handling concepts. She exhibited allof the early strategies, but did not attend toletter/word/lire order information in the ConceptsAbout Print Test. She could easily identify lettersand specific words (was, no), and she understoodthe difference between the concepts of "word,""letter," "first," and "last."

In the writing portions of the DiagnosticSurvey, she was able to write 13 words and torepresent 19 of the 37 phonemes in the sentence"The bus is coming. It will stop here to let me geton."

As I began working with Sarah, it was evi-dent thai she was a risk taker. She made manyattempts during the testing and early lessons thatshowed her willingness to try.

During the first 10 sessions, "roamingaround the known," Sarah produced three books,exhibited another 26 written vocabulary items,and read 26 different books. She was activelyparticipating in writing and reading experiences,and rapidly developing the necessary strategiesthat allowed her to quickly begin to acceleratewithin the program. I decided to 1. egin her in-structional program at a Level 2.

On an early reading at Level 6 (five weeksinto lessons), I was pleased to see that she wasreading for meaning, using language cues andchecking these sources of information againstvisual information. (See Figure 8, page 24.)

Sarah clearly enjoyed our writing sesions.A writing sample collected in the last week beforeChristmas indicated that Sarah was independentlygenerating stories and writing all of the text. (SeeFigure 9.) Her language was rich, and her storiescreative. Her attitude in class had changed, andshe was rapidly becoming one of the better read-ers within the lowest reading group. She was stillhaving difficulty with writing during independentwriting times, but was more often able to com-plete most of her seatwork in the time provided.

Sarah was discontinued within 47 lessonsacross 12 weeks. She grew to be an independentreader and writer, and was functioning within theaverage of her class in the middle reading groupby the end of February. The example of herreading of The Little Red Hen (Figure 10) illus-trates the type of strategies she exhibited at thepoint of being discontinued. She was reading atLevel 12, which was typical of first-grade readingmaterial with the average of her class. She wasable to solve} .oblems independently as a readerand a writer.

33


The dictation and writing subtests of thediscontinuing testing indicate her progress inwriting. She had a core of at !east 35 high-fre-quency words she could write with ease, and shewas able to represent 36 out of 37 phonemes inthe sentences "I have a big dog at home. Today Iam going to take him to school." In short, Sarahhad developed into an independent reader andwriter, and indicated she had a self-improvingsystem.

In a follow-up of children in Sarah's school,I found that she was still progressing as a reader,and going beyond the average of her classmates.At the end of second grade, Sarah was reading ata Level 18, or at the end of the second-gradereading material. At the end of third grade, shehad progressed to a Level 28, or fifth-grade ma-terial. When we talked with her classroom teach-er, she indicated her surprise that Sarah had everhad problems in reading as a first grader.

FIGURE 8.In an early reading, Sa h Is reading for meaning, used language cues,and checked these sources of in;armation against visual Information.

rues UsedPAGE Cats and Kittens E SC E SC

I2 Cats play.

3 Kittens play too.,/

4 Cats like milk.

5 Kittens like milk too.,./

6 Cats keep clean.1221,_

They lick their fur. I /11051 KiNers 'N sC .e' ..... u..'

7 Trgrg keep their kittens clean too.... ,- i 05v 0"

8 Cats hunt.v

9 Kittens hunt too.

10 Cats get mad. /Their fur stands on end.

,,,.. 1.-.;/ 2,,PJA

1/

Their tails puff up.v v ,v

11 Kittens get mad too.v ,./

12 Cats sleep.,./

13 Kittens sleep too.


FIGURE S.Writing sample: Sarah Independently generated story and wrote text.

=Lk -r_ b W0-408\Vegete Seti/02)

GhArli5an-'etc) -AS frtcatilt pre5e1 I.

50 kl-" L.

Thvue °I' BP er\)VesevitS ktkinie0-C-1 14-kesMoNi i e. pici-viz's .

Pr6emstfl.Me T°°Fey\

Ptoc<)eruta_561\41x,

leNegYl

At5 'N iV,c66

stagiAl

1-1"beeirey14cy... Tii-eu

crtve Pre seruts Jbin

t co. VI,

otjAhe Pre561'71"

o

E- \,,ey&- ThecAttlse-Dl snit6-1C\C61\ b 1 5c4vil,

ck,ris'i as- _isotIN. it

32


FIGURE 10.Running record showing strategies Sarah was using In reading typicalfirst-grade material at the time she was successfully discontinued.

PAGE

I

3

45-

b

7

The Little Red Hen I E SC

Cues UsedE SC

VOnce upon a timei v V ,/ V / ,/a cat and a dog and a mouse

an0 a little red hen.1----- v ,... ,, .

1.,' -kall lived togefra7in a cozy little huuse.5C

The cat liked to sleep all day

on the soft couch.

The dog liked to nap all day

on the sunny back porch. SE.

v vThe mouse liked to snooze all day

4i re.

in the warm chair by the TIT=

../ ...-- u., v..

So the little red hen had to do allV

the housework...-

She cooked the meals and washed the lishes,v---16.37)sc ../

and made the beds. She swept the floor

and washed the windowsvn ad t.

and mended the clothes.

YV

MSMSS

m010

mS.0

Chapter 3:Longitudinal Study,ColumbusPublic Schools

Columbus

Studies of Reading Recovery in Ohio

1984-85

Columbus Pilot Study

Longitudinal Study

Ohio State Study, 1st Year

Ohio State Study, 2nd Year

Ohio State Study, 3rd Year

1985-86 1986-87 1987-88

In Chapter 3 and Chapter 4, we describe theresults of detailed studies of Reading Recoverythat have been conducted in the Columbus PublicSchools and in many school districts throughoutthe state of Ohio in the last few years. The resultsindicate that Reading Recovery can produce ma-jor, sustained improvement in the reading abilityof a high proportion of at-risk first-grade stu-dents, rescuing these children from a future ofschool failure.

We present these research results asevidence of the unusual effectiveness that distin-guishes Reading Recovery from the many othercompensatory reading programs in use

Chapter 3 discusses the longitudinal studyof Reading Recovery conducted in the ColumbusPublic Schools. This study followed the progressof a cohort of students who were in the bottom 20percent of their classes in reading skills and whoreceived Reading Recovery instruction in firstgrade during the 1985-86 school year.

The results showed that the short-termReading Recovery intervention had greatly im-proved the reading skills of these children, andbefore the end of their first-grade year had en-abled 73 percent of them to be successfully re-leased to regular insruction at the average level oftheir first-grade classmates. The two follow-up

27

studies show that the initial gains of a high per-centage of these children were sustained throughthe second grade and on through the third gradewithout any further intervention.

Chapter 4 reports on the implementation ofReading Recovery at sites throughout the state ofOhio. In 1985-86, the same year that the longi-tudinal study began in Columbus, Reading Re-covery was implemented at 18 regional trainingsites throughout Ohio. The statewide programwas expanded to 23 sites in 1986-87 and .1_987-88. In Chapter 4 we describe studies of OhioReading Recovery students at the end of firstgrade in each of the three "ears. The results of thethree separate first-grade studies confirm the pos-itive findings documented in the first year of theColumbus Public Schools longitudinal study.

Background:From New Zealand to Ohio

The Reading Recovery program was pilotedin New Zealand in 1979 with remarkable results.In Clay's study (1982, 1988), New Zealand chil-dren in the Reading Recovery program madeaccelerated progress and were able to reach theaverage levels for their classmates in an averageof 14 to 16 weeks. Over 90 percent of the chil-

28 Longitudinal Study, Columbus Public Schools

dren served were successfully discontinued; thatis, they reached average levels and displayed evi-dence of having developed an independent systemof reading.

After being successfully released from theReading Recovery program, children received nofurther special help. Three years later, a high per-centage continued to make progress and to perform at average reading levels. Based on theseresearch results, Reading Recovery was made anational program in New Zealand.

In 1984-85, Marie Clay and BarbaraWatson, National Director of Reading Recoveryin New Zealand, came to The Ohio State Univer-Oty to train Reading Recovery teachers andTeacher Leaders. As part of the training, theseteachers piloted Reading Recovery in six Colum-bus Public Schools from January through May1985. Positive results of the pilot project encour-aged the Columbus Public Schools to proceedwith implementation of the Reading Recoveryprogram in the 1985-86 school year and to initiatea longitudinal study of the effects of ReadingRecovery.

Columbus Longitudinal Study,Initial Year

In 1985-86, the ii,itial year of the longitu-dinal study, Reading Recovery was implementedin 12 schools in Columbus.

Thirty-two teachers were involved in the1985-1986 Reading Recovery project. Of these,12 had received their Reading Recovery trainingduring the previous (pilot) year; another 20 werenew Reading Recovery teachers who receivedtheir training during 1985-86. (For a descriptionof the teacher training program, see Chapter 5.)

These 32 teachers began to teach ReadingRecovery children in October, 1985. In somecases, a sharing arrangement was used, in whichtwo teachers trained in Reading Recovery werepaired and shared one first-grade class; each

teacher spent half the day teaching the wholeclass, and the other half tutoring Reading Recov-ery children. In other cases, teachers who taughtthe whole class were not trained in Reading Re-covery.'

R.-search GroupsIn September 1985, the lowest 20 percent of

children in the classrooms taught by Reading Re-covery teachers (as determined by the DiagnosticSurvey and the clL ssroom teachers' judgment)were selected for Reading Recovery. The lowest20 percent of children were also identified in otherclassrooms in the same schools: half of thesechildren were randomly assigned to receiveReading Recovery intervention, and the other halfwere randomly assigned to an alternative com-pensatory program. A total of 136 were assignedto receive Reading Recovery tutoring, and a totalof 51 were assigned to the alternative compen-satory program. (Se, Appendix B for a descrip-tion of the alternative program.)

For research purposes, Reading Recoverychildren were defined as those children who atsome time during their first-grade year had 60 ormore lessons or werc successfully discontinued(released) from the program. Comparison

*The question arose: Would children receivingseparate Reading Pecovery tutoring and alsotaught reading in first-grade classes taught byReading Recovery teachers achieve more than didchildren in first-grade classes taught by regularteachers who were not Reading Recovery-trained?A rather elaborate design was established to de-tect any such possible impact. The results ofthis research have been reported in Pinnell andothers (1984-86) and Deford and others (1986-88). A slight difference was found in favor ofstudents taught in the whole class by ReadingRecovery teachers, but the difference was not sta-tistically significant. For purposes of brevity inthis monograph, all children receiving ReadingRecovery instruction will be treated as a singlegroup.

Longitudinal Study, Col'" bus Public Schools 29

children were those initially identified as being inthe lowest 20 percent of regular classrooms whowere served all year by the alternative compen-satory program. No children were served by bothprograms.

Figure 11 on page 30 summarizes the com-position of the different research groups used inthe longitudinal study during the initial year(1985-86) and the two follow-up years, as thiscohort of first graders moved on tc the secondgrade and then to the third grade.

Research Questions, Initial YearThe first evaluation of the effects of Reading

Recovery on 1985-86 first graders was conductedin May 1986, at the end of their first-grade year.Reading Recovery children had received tutoringfor various lengths of time during the year. Mostof them had been successfully discontinued fromthe nrogram; some had not been successfully dis-cor .inued by the end of the year. The followingresearch questions were addressed:

1) How did Reading Recovery children andComparison children compare at the end ofgrade one on a variety of measures of read-ing ability?

2) How did Reading Recovery and Compari-son children perform at the end of grade oneon nationally normed standardized tests?

3) What pro2ortion of successfully discon-tinued Reading Recovery children (thosewho were successfully released from theprogram) achieved end-of-year scores equi-valent to the average band of achievement ofa Random Sample of first-grade students?

Procedures, Initial YearIn October and May, the Reading Recovery

and Comparison children were assessed on eightdependent measures. (For a description of eachof the dependent measures, see Appendix C.)The measures were:

1) Text Reading;

2) Letter Identification;3) Word Test;4) Concepts About Print;5) Writing Vocabulary;6) Dictation;7) Two subtests of the Comprehensive Tests of

Basic Skills (Reading Vocabulary andReading Comprehension); and

8) A writing sample.In addition to the Reading Recovery and Com-parison groups, a random sample of 102 first-grade students in Columbus project schools wasalso tested on the first seven dependent measureslisted above. This Random Sample group pro-vided a per ,ctive for comparing the achieve-ment of the two groups of Research children withthe average achievement of other children at thesame grade level.

Means and standard deviations were calcu-lated on all measures.

Results and Analysis, Initial YearOf the 136 children assigned to Reading

Recovery in September 1985, 73.5 percent weresuccessfully discontinued from the program atvarious times during the school year and receivedno further treatment. (Three of these successfullydiscontinued children moved from te districtbefore the end of the year, and therefore do notappear in the May testing results). These suc-cessfully discontinued children received an aver-age of 67 thirty-minute Reading Recovery les-sons. The other 26.5 percent of children had notbeen discontinued by the end of the school year.

Question #1 How did Reading Recoverychildren and Comparison children compare at theend of grade one on a variety of measures ofreading ability?

In Table 1 on page 38, September and Mayscores on the first seven measures are presented

30

30 Lon itudinal Study, Columbus Public Schools

FIGURE 11.

Summary of Groups: Longitudinal Study, Columbus Public Schools

Initial Year (1985-86)

FALL 1985 (BEGINNING OF FIRST GRADE)

Reading Recovery childrenfrom, bottom 20% in readingassignee lo receive Reading Recovery program 136

Comparison Childrenfrom bottom 20% in reading, assigned to receive regular remedial instruction for full school year 51

MAY 1986 TESTING (END OF FIRST GRADE)

All Reading Recovery ChildrenReceived at least 60 Reading Recovery lessons or were successfully discontinuedfrom the program during ute school year. (Three children from fall cohort who had beensuccessfully discontinued moved from district befoye spring testing) 133

Number successfully discontinued from program (100 minus 3 who moved) 97

Number not discontinued from program 36

Percent of all Reading Recovery clf.idren (including three who left school system)who were succr ssfully discontinued 73.5%

Average number of daily, 30-minute sessions for children successfully discontinued 67 sessions

Comparison Children (same group as in fall 1985) 51

Random sample of first-grade students in project schools, excluding Reading Recoveryand Comparison children (base for first-grade average) 102

Follow-up (1986-87 and 1987-88)

MAY 1987 TESTING (END OF SECOND GRADE)Reading Recovery Children

successfully discontinued and not-discontinued children from fall 1985 cohort who werestill in district in spring 1987 116

Comparison Childrenfrom fall 1985 cohort who were still in district in spring 1987 43

Random sample of second-grade students in project schools, excluding Reading Recoveryand Comparison children (base for 2nd-grade average) 68

MAY 1988 TESTING (END OF THIRD GRADE)Reading Recovery Children

successfully discontinued and not-discontinued children from fall 1985 cohortwho were still in district in spring 1988 105

Comparison Childrenfrom fall 1985 cohort who were still in district in spring 1987 42

Random sample of third-grade students in project schools, excluding Reading Recoveryand Comparison children (base for third-grade average) 67

37

Longitudinal Study, Columbus Public Schools 31

for three groups: 1) Reading Recovery children,2) Comparison children; and 3) Random Sampleof all first-grade children in project schools.

These data indicate that in May of 1986,Reading Recovery children as a total group (suc-cessfully discontinued and not-discontinued com-bined) scolcd higher than Comparison children onall measures. In fact, the scores of the totalReading Recovery children were very similar tothose of the Random Sample group of first-gradestudents. Specifically, the Reading Recoverygroup scored slightly higher on Letter Identifica-tion, Concepts About Print, Writing Sample, andDictation, and slightly lower on Writing Vocabu-lary, Text Reading, and Word Test. As a group,the Reading Recovery children, who were in thelowest 20 percent of their class in reading skills atthe beginning of the year, now scored about thesame as a group of average first graders.

Table 2 (page 39) shows the May 1986 end-of-year scores broken into two groups: 1) suc-cessfully discontinued Reading Recovery chil-dren, who had made sufficient progress to bereleased from Reading Recovery; and 2) not-discontinued Reading Recovery children. Thescores of these two groups and the scores for theComparison group are compared with the scoresfor the Random Sample of all first graders. Thefigures show successfully discontinued ReadingRecovery students scoring higher than the Ran-dom Sample of all students on all seven mea-sures, while not-discontinued Reading Recoverystudents and Comparison students score lower onevery measure.

Discontinued and not-discontinued childrenare considered separately because the expectationsare different for these two groups of children.Discontinued children not only have made accel-erated progr ;s in the program; an analysis oftheir reading liehaviors indicated that they havedeveloped ti strategies necessary to keep onlearning to reau better and eventually to learn fromtheir reading. They have made the "breakthrough"

to literacy. For children in some classrooms,meeting this criterion also may mean scoring atthe high end of "average" or even above average.In other classrooms, a child might score at thelow end of "average" yet show the necessaryevidence of effective reading strategies. Barringextraordinarily negative school environments, wewould expect those children to keep on makingprogress, and to the extent that they do, the dis-continuing judgments have been made success-fully.

Not - discontinued children are those whohave not qualified for release either by score or bythe analysis of reading and writing behaviors.Perhaps they needed a longer than average time ofindividual tutoring; or there may be physical oremotional difficulties that indicated the need forcontinued extra help. Those children may havemade progress, but the system is probably not inplace for them to continue that progress. Theywill tend to do what is typical of children in re-medial programs; they will make very slow prog-ress even with extra help.

Question #2 How did Reading Recoveryand Comparison Children perform at the end ofgrade one on nationally normed standardizedtests?

Results from the Reading Vocabulary sub-test of the Comprehensive Tests of Basic Skills, anationally normed standardized test, show thatReading Recovery children had a Normal CurveEquivalent (NCE) gain score of 7.4 compared to-2.6 for Comparison children. On the ReadingComprehension subtest, Reading Recovery chil-dren had an NCE gain of 7.0, compared to -4.5for Comparison children. On the Total Readingcombined score, the NCE gain was 8.6 for Read-ing Recovery children and -2.4 for Comparisonchildren. Thus, Comparison children lost ground

.13


while Reading Recovery children gained groundrelative to the expected achievement for theirgrade level. Table 3 (page 40) shows these data.

Question #3 What proportion of success-fully discork:',nued Reading Recovery children(those wno were successfully released from theprogram) achieved end-of-year scores equivalentto the average band of a Random Sample of first-grade students?

The May 1986 end-of-year scores of suc-cessfully discontinued Reading Recovery childrenwere compared with those of the average band ofscores of the Random Sample of first-grade stu-dents. The average band of scores was calculatedby computing both a mean and a standard devia-tion and by using plus or minus .5 standard devi-ation from the mean as the upper and lowerboundaries of average performance.

Table 4, page 41, shows that over 90 per-cent of the successfully discontinued ReadingRecovery students met or exceeded the averagerange on Text Reading, Letter Identification, Ba-sal Word Test, and Dictation. On the ConceptsAbout Print assessment, over 88 percent met orexceeded the criteria. Over 79 percent met orexceeded the criteria on Writing Vocabulary, andover 70 percent met or exceeded the criteria on theWriting Sample.

These very high percentages of successfullydiscontinued Reading Recovery children scoringwithin the average band at the end of first gradeshould be viewed in perspective. We must re-member that the criterion for successfully discon-tinuing a child from Reading Recovery was thathe or she attain a level of performance at or abovethe average of his or her peers. Therefore, bydefinition, almost all of the discontinued childrenwould logically be expec.ed to fall above or with-in the average band at the time of their releasefrom the Reading Recovery program. However,

we would not expect the: to sustain this closelyclustered distribution within the average band af-ter they returned to regular classroom instruction;rather, we would expect them, while maintainingat least minimum expectations for grade level, torearrange themselves into a distribution moreclosely resembling that of the Random Sample oftheir grade-level peers. In fact, as the discussionof follow-ap testing at the end of second and thirdgrades will show, this is what happened.

Figure 12 (page 33) indicat.:s that the meanscores for Reading Recovery children on fourmeasures of reading skills in May 1986 fell withinthe average band for first graders. It also showsthat the Comparison group mean was below thisverage band in each case, although both Com-

parison and Reading Recovery children had be-gun the year in the bottom 20 percent of theirclass, far below the average band.

Summary of Initial-Year ResultsThe Reading Recovery program, in its first

full year of operation in the Columbus PublicSchools (1985-86), produced positive outcomes.Of the 136 children who were provided with theReading Recovery intervention, 73.5 percentwere successfully discontinued; that is, they madesufficient progress to be released from the Read-ing Recovery program.

The group of all Reading Recovery children(successfully discontinued and not-discontinuedcombined) scored significantly higher than a con-trol group of Comparison children on a widevariety of measures of reading performance.Successfully discontinued Reading Recoverychildren achieved mean scores within the averageband for a random sample of all first-grade chil-dren on all measures.

These first-year results provided evidencethat Reading Recovery was an effective programfor Ohio children and that it could be used suc-cessfully in American public schools.

33


FIGURE 12.Text Reading Scores of Total Reading Recovery Group andComparison Groups, Compared with Average Band of First-Grade Children,

September 1985 and May 1986

Total Reading Recovery Children

Comparison Children9 Random Sample Mean +1- .5 standard deviations

Dictation

40 Max=37

35

30

25

20

15

10

5

0Septunix.1-

,

May

1

Writing Vocabulary

50 (10 minute limit)

45

40

35

30

25

20

15

10

5

0September May

Text Reading16 Max=26

14

12

10 Primer

8

6

4

2

0September

5.0...

.0.'

..5

May

34 Lon itudinal Study, Columbus Public Schools

Follow-up Studies:Second and Third Years

The initial evaluation of the 1985-86 first-grade cohort of Reading Recovery children in:olumbus demonstrated the success of Reading

Recovery in taking students with severe readingdifficuities and raising them up to the average le-vel of their class in a short time. The next step inthe longitudinal study wa3 to determine whetherthese children would maintain their dramatic gainsin subsequent school years, without further inter-vention.

The cohort of 1985-86 first-grade ReadingRecovery children who were still attending theColumbus public schools were -allowed duringthe first and second full years after the inter-vention. The purpose of the follow-up studieswas to gather information about the long-rangeeffectiveness of the Reading Recovery program.

Research Questions,Follow-up Studies

1. How did the performance of Reading Re-covery children (successfully discontinuedand not-discontinued combined) comparewith the performance of Compariscn childrenon text reading ability at the , d of secondand third grades?

2. What proportion of Reading Recovery chil-dren (successfully discontinued and not-discontinued) and Comparison childrenachieved end-of-year scores equivalent to tl: ..average band for all second-grade children in1987 and for third-grade children in 1988?

3. Did successfully discontinued Reading Re-covery children sustain the gains theyachieved in first grade through the end ofsecond and third grades, without any furtherintervention?

Procedures, Follow-up StudiesIn May 1987, Reading Recovery and Com-

parison children most of w!!Im were then

completing the econd grade were a.sessed ontheir Text Reading level.* To provide perspectivefor the scores of both groups, a random sample ofsecond-grade children was selected from regularclassroc s in prCect schools and administeredthe Text . eating t st. A similar testing processwas conducted in May 1988, at the end of thethird grade.

The Text Reading test for these grade levelshas material representing 1-30 levels of readingdifficulty. The student must read material with 90percent accuracy to be assessed as competent inreading at a specific level . (See page 9, Chapter 1for a description.)

All testing was carried out by "blind" testerswho were given lists of children to test individu-ally. Testers were sent to schools where they didnot know the children. Children on their listswere not designated as to group (Reading Recov-ery, Comparison, or Random Sample).

An average band for Text Reading level wascalculated for second graders in 1987 and forthird graders in 1988. As before, the averageband was defined as plus or minus .5 standarddeviation from the mean of the Random Sample.

Results and Analysis,Follow-up Studies

Question #1 How did the performance ofReading Recovery children (successfully discon-tinued and not-discontinued) compare with theperformance of Comparison children on textreading ability at the end of second and thirdgrades?

*The Text Reading test, was the most appropriate

and virtually the only measure available withinthe financial constraints and the regular testingprogram of the Columbus schools in these years.Standardized tests were administered to studentsonly in the first and fourth grades.

Longitudinal Stud Columbus Public Schools 35

One year after . e end of the treatment year,in May, 1987, both groups of students (most inseconc' ade) were assessed on Text Reading.Table J on page 42 shows that the Reading Re-covery children (total group) performed betterthan the Comparison children.

Again in May of 1988, two years after theend of the treatment year, w'.en most were inthird grade, the Reading Recovery children per-formed better than the Comparison children.

These differences were even greater for thegroup of successfully discontinued Reading Re-covery children, who far outperformed the Com-parison group at each testing. At the end of theirthird-grade year, the successfully discontinuedReading Recovery group averaged slightly abovethe Random Sample of all third-grade children inText Reading.

Scores comparing Reading Recovery chil-dren and Comparison children are graphicallyillustrated in Figure 13. As shown here, the group

of all Reading Recovery children still maintainedthe advantage that they had achieved by the end ofthe initial first-grade intervention.

Question #2 What proportion of ReadingRecovery children (successfully discontinued andnot-discontinued) and Comparison childrenachieved end-of-year Text Reading scores equi-valent to the average band for second gradechildith in 1987 and for third grade children in1988?

One year after the intervention, both groupswere compared with a Random Sample of secondgraders; two years after the intervention, bothgroups were compared with a Random Sample ofthird graders. The proportion of children whoperformed at average or above-average le /els wascalculated for the total group of Reading Recoverychildren (discontinued and not-discontinued) and

FIGURE 13. Mean Text Reading Level for the 1985-86 First GradeCohort: All Reading Recovery Children Versus Comparison Children

20

18

V

14

12

Primer 10

8

6

4

2

0

0 Intervention Period Mean = 13.4 Weeks

iAll Reading Recovery Children

1 Comparison Children

September 1985 May '986 May 1987 May 1988


for the total group of Comparison children. Thesedata are summarized in Table 6 on page 43.

Results of the average-band analysis indi-cate that substantially larger percentages of thetotal group of Reading Recovery children were ataverage or above-average levels on the measure ofText Reading compared to Comparison children.Successfully discontinued children had thn high-est proportions of children at average or above-average levels on Text Reading. These propor-tions compared favorably to those of the RandomSample. In looking at these results, we shouldremember that all of the Reading Recovery andComparison children began their first-grade yearin the bottom 20 percent of their classes in readingskills. It should also be pointed out that the greatloss of children due to migration and retention forreasons other than reading are limitations of thelongitudinal study.

Question #3 Di' successfully discontinuedReading Recovery children sustain the gains theyachieved in first grade through the end of secondand third grades, without any further interven-tion?

To address this question, the mean scoreson Text Reading levels of successfully discon-tinued Reading Recovery children were examinedat four points, from October, 1985, through May,1988. Results of the , ...)1low-up data on success-fully discontinued children are graphically illus-trated in Figure 14. Their progress and that of theComparison group are compared with the averageband of Text Reading level achieved by RandomSamples of all first-grade, second-grade, andthird-grade students.

FIGURE 14. Mean Text Reading Level for the 1985-86 First Grade Cohort:Successfully Discontinued Reading Recovery Children and

Comparison Children in Relationship to the Random Sample Average Band

30

25

20

15

Primer 10

0 Intervention Period Mean = 13.4 Weeks

HRandom Sample Mean Average Band(± 3 Standard Deviation)

1 Successfully Discontinued ReadingRecovery Children

1 Comparison Children

September 1985 May 1986 May 1987 May 1988


These results provide evidence that a highproportion of successfully discontinued childrencontinued to make progress for at least two fullyears after their individual Reading Recovery in-tervention had taken place. At the end of firstgrade, successfully discontinued children as agroup scored within the average band of the Ran-dom Sample of all first graders in their projectschools. At the end of second grade and again atthe end of the third grade, the mean Text Readinglevel of discontinued Reading Recovery childrenwas still within the average band of all childrenfor their grade level. At the same time, the meanText Reading level of the Comparison group fellbelow the average band at each grade level.

These data show that, as a group, discon-tinued Reading Recovery children continue toperform at a level consistent with the averageband for the appropriate grade level for two yearsfollowing their Reading Recovery year, withoutthe need of any further intervention,

Summary of Follow-up Studies --The follow-up studies involvek. the cohort

of first-grade children served by the Reading Re-covery program ;n the Columbus Public Schoolsin 1985-86, the initial yea. of the longitudinalstudy. In these studies, Reading Recovery chil-dren performed better than Comparison childrenat the end of the intervenfon period, and theycontinued to perform better one and two yearsafter withdrawal of the intervention. Higher pro-portions of Reading Recovery children scored ataverage or above-average levels for Text Reading.Reading Recovery children continued to makeprogress, while the Comparison group continuedto fall further behind each year. Successfullydiscontinued children, as a group, were perform-iag within the average range for their grade-levelpeers at the end of the intervention and continuedto perform withi:, the average range for their

grade-level peers through the second and thirdgrades.

Conclusion

The longitudinal study conducted in theColumbus Public Schools provides importantinformation for persons int-rested in implement-ing the Reading Recovery early intervention pro-gram.

First, the evidence indicates that the pro-gram benefits a large proportion of children hav-ing the most difficulty in leaning to read. In

general, these effects are sustained for at least twoyears following the release or children from theprogram. It is reasonable to expect, therefore,that if the program is implemented with care andintegrity, it will produce positive immediate andlong-term effects.

Reading Recovery will not solve all prob-lems related to school failure. It ..s evident fromour experience with several years of implemen-tation that teachers' perceptions of ;hildren anddefinitions of competency are hard to change.What this program will do is improve the readingability of a large proportion of at-risk children.We need continued opportunities within schoolsand classrooms for these young children, many ofwhom still live within an "at-risk" population, tomake the most of their potential abilities.

The Reading Recovery team, includingschool distr..1 officials, teachers, and Ohio StateUniversity personnel, are continuing to study theprogram carefully. Children from the 1985-86first-grade cohort will be followed, and research-ers will do additional in-depth study of the pro-cesses involved in the program. Procedures fortraining new Reading Recovery teachers and foridentifying and working with children, particu-larly those from highly mobile populations, arebeing further developed and n fined.

A r--t -.:


Highlights from Table 1:Although the Reading Recovery and Comparison groups had similar mean scores for the :;ix compo-nents assessed in September, 1985, the Reading Recovery gronp (including both discontinued and notdiscontinued students) scored higher on all seven measure at the end of the school year.

While the Reading Recovery children had been assessed as having reading skills that placed them amongthe bottom 20 percent of all first graders in September, their May scores approximated those of theRandom Sample of all first graders.

Reading Recovery scores were slightly lower than the Random Sample scores for text reading, but theReading Recovery children actually did better than the Random Samplechildren on several othermeasures.

Table 1. Means and Standard Deviations for All 1 jading Recovery Children,Comparison Children, and Random Sample of First Graders on Seven Measures,

September and May 1985-86

Reading Recovery ComparisonChildren' Children Random Sample

Measure Month Di Mean S.DJ 11 Mean S.DI N" Mean S.DJ

Text Sept 131 .85 51 .96 - -1,82 1.6$Reading May 133 9,9S 3.14 51 6.96 3.07 101 11.15 5.62(Max =26)

Writing Sept - . - - - -Sample May 131 2,94 1.17 46 2,$3 1.10 100 2,102 19(Max=6)

Letter Sept 131 3633 14.67 51 X93.29 14.96 - 'Ident. May 133 41.92 3.34 51 OA 8.33 102 $I.1 3.99(Max =54)

Word Sept 131 231 2.24 51 2.,04 2.09Test May 133 13.62 1.79 51 11.98 3.92 102 13,91 1.95(Max =15)

Concepts Sept 131 7,39 3.68 55 6,96 2.92 - ..... -About May 133 1640 2.82 51 13,98 3.31 102 100 3.25Print(Max=24)

..

Writing Sept 131 2.99 2.81 51 3.51 3.09 - .4. -Vcab. May 133 .34,611 12.82 51 2$,$7 14.33 102 3832 18.46(13 min)

Dictation Sept 131 5.51 5.75 51 6,22 5.72 - -(Max=37) May 133 $1.20 5.88 51 23.80 7.99 102 3044 6.96

iIncludes both successfully discontimcd children and not-discontinued children who received at least 60 Reading Recovery lessonsduring first-grade year. There were a total of 136 such children who were served at some time during the y-ar.

Less than 156 because students entered program utter September.Less than 136 because 3 students moved before May testing.

Differences in Ns reflect students who were absent duripit testint.

4J


Highlights from Table 2:

Mean scores of Successfully Dontinued Reading Recovery children (who started the school year indie bottom 20 percent of first graders) al:. higher than Random Sample mean scores on all seven ele-ments of skills tested, including text reading.

Although Not Dacontinued Reading Recovery children score lower than the Random Sample childrenon all seven measures, their scores are very similar to those of the Comparison Children who wereserved during the school year by a remedial reading program. Lower scores for the Not DiscontinuedReading Recovery children would have been expected since the Comparison Children group stillincludes a broader range of ability /skills than the less successful Reading Recovery children representedby the Not Disci itinued segment.

Table 2.-Means and Standard Deviations of Suct...fully Discontinued RR Children,Not-Discontinued RR Children, Comparison Children, and a Ztandom Sample of First-Grade Children

READING RECOVERY CHILDREN

1) Succ. Discontinued RR 2) Not - Discontinued RR COMPARISON CHILDREN

i

RANDOM

SAMPLE

IN* Mean S.D. 1 N* Mean S.D. N* Mean S.D. I1* Mean S.D.

Text Sept 95 1.96 I .85 36 L47 .77 51 1.65 .96Reading May 97 1132 2.45 36 6.28 1.30 51 6.96 3.07 101 11.15 5.62(Max=26)

WritingSample May 96 3.19 1.15 35 2.26 .92 46 233 1.10 100 2.92 1.39(Max)

Letter Sept 95 40.56 12.07 36 2164 13.97 51 33.29 14.96Ident. May 97 52.66 1.21 36 49.94 5.70 51 49.61 8.33 102 51.78 3.99(Max=54)

Word Sept 95 2.59 2.36 36 16 1.22 51 2.04 2.09Test May 97 14,32 .96 36 1L72 1.11 51 1L98 3.92 102 13.91 1.95(Max=15)

Concepts Sept t 4%01 3.44 36 512 2.83 55 6.96 2.92About May . 33 2.20 36 13.89 2.80 51 13.98 3.31 102 16.00 3.25Print(Max=24)

Writing Sept 95 3.68 2.94 36 1.17 1.18 51 3.51 3.09Vocab. May 97 39,08 11.21 36 22.81 8.76 51 2537 14.33 102 38.12 18.46(Max=10 min.)

Dictation Sept 95 6.65 5.79 36 1.94 2.07 51 6/2 5.72(Max=37) May 97 53.59 3.48 36 24.78 6.24 51 23.80 7.99 102 30.24 6.96

Ave. No.Lessons

67 100ll

area s m Ns reflect students who were absent during May testing

Longitudinal Study, Columbus Public Schools


When students were given the Comprehensive Test of Basic Skills, the Reading Recovery children as agroup (including both Discontinued and Not Discontinued children) gained ground relative to the level ofskills expected of them in the fall and again in May.

Comparison children, on the other hand, had a minus gain score. In other words, their skill level wasworse in relation to expectations for first-grade students in May than in the fall.

Table 3.Normal Curve Equivalent (NCE) Gain Scores for Reading Recoveryand Comparison Groups, Comprehensive Tests of Basic Skills, May 1986

Reding Vocabulary

Pretest(Fall 1985)

Posttest(May 1986)

NCEGain Score

Reading Recovery N=126* 30.9 38.3 7.4

Comparison N=26** 31.5 28.9 -2.6Reading Comprehension

Reading Recovery N=125 31.8 38.8 7.0

Comparison N=26 28.8 24.3 -4,5

TOTAL READING***

Reading Recovery N=124 2c.4 38.0 8.6

Comparison N=26 26.7 24.3 .4

* N<136 because of absent students when CTBS was given.** N<51 t ecause Comparison children analysis was comnuted for children who were both present when CTBS was given and

in school 80 percent of the time.*** Total Reading is a combined score consisting of the average of individual scale scores in vocabulary and oral cm reading

comprehension.

4?,

68 70.8% 1


Highlights from Table 4.

In the seven skills tested, between 70.8 percent and 99.0 percent of Successfully Discontinued ReadingRecovery children scored within or above die; average band of test scons of the Random Sample of first-grade students. If the skill levels of these Reading Recovery students had remained the same throughoutthe school year relative to other first graders, all of them would have been expected to fall in the bottom20 percent of students at the end of the year.

Almost 93 percent of the Successfully Discontinued Reading Recovery children fell within or above theaverage band of test scores (+ or - .5 standard deviation from the mean score of the Random Sample ofall first graders) for text reading. Even higher percentages fell within or above the average bands for theskills of letter identification, dictation, and basal word test.

Even on the writing sample segment of the test, the skill for which scores of the Successfully Dis-continued Reading Recovery children were lowest relative to the Random Sample group, seven out of tenof the children achieved scores that fell within or above the average band.

Table 4.Numbers and Percentages of Successfully DiscontinuedReading Recovery Children in End-of-Year Average Band, May 1986

Measure Band Number Percent

Text Reading

Writing Sample

Letter Identification

Basal Word Test

Concepts About Print

Writing Vocabulary

Dictation

Met or Exceeded AverageBelow Average

90 92.8%7 7.2%

IMet or Exceeded AverageBelow Average 28

IMet or Exceeded Average 96Below Average

L.L.4etor Exceeded Avera eBelow Average

1

29.2%

99.07Yal

L0%

Met or Exceeded Avera:e'Below Average

Met or Exceeded Avera eBelow Average

I Met or Exceeded AverageBelow Average

91 93.8%6 6.2%

86 88.7%11 11.3%

77 704%20 2U.6%

94 96.9%3 3.1%

NOTE: Average band is + or .5 standard deviation from the mean score of the Random Samplz of all first-grade students.

1

42 Longitudinal Study, Columbus Pu . Schools


A comparison of the text reading level scores of Successfully Discontinued Reading Recovery studentswith those of the Random Sample students demonstrates the long-term effectiveness of the Reading Re-covery program from the standpoint of helpmg students to incorporate successful strategies for readinginto their repertoires. At the end of third grade, the mean text reading level score for SuccessfullyDiscontinued Reading Recovery children is 23.99 and that of the Random Sample 23.50.

Mean scores of the total Reading Recovery (discontinued and not discontinued) group are also higherthan the mean scores of the Comparison group (both of which are composed of students who wereidentified in the beginning of their first grade year as falling within the bottom 20 percent of all firstgraders on the basis of reading skills) at each of the three administrations of the test.

Although mean scores for the Not Discontinued Reading Recovery students are lower than those of theComparison group, it should be remembered that the Comparison group still contains students from thewhole "bottom 20 percent" stratum while the Not Discontinued means represent only the scores of theleast successful Reading Recovery students.

Table 5.-Means and Standard Deviations on Text Reading Level for All of theYear I Cohort (1985-86) Children at Four Points in Time

Reading Recovery Comparison Random Sample*Date Discontinued Not Discontinued Total

N Mean S.D1 N Mean S.D. N Mean S.D. N Mean S.D. N Mean S.D.

Sept. 85 95 .85 36 .77 131 .85 51 .961.96 1.46 1.82 1.65 [---IViay 86 97 11,32 2.45 36 6.28 1.30 133 9,95 3.14 51 6,1 3.07 101 11.15 5.62

May 87 85 16.71 5.84 31 8.03 2.39 116 14.39 6 42 43 11.23 4.88 68 18.60 6.67

May 88 78 23.99 6.48 27 10 19 3 49 105 19.70 5.71 42 16.71 6.80 67 23 50 9.00

NOTE: Numbers of studen s vary due to iNence on day of testing or movement from the school system. Numbers in Reading Recovery andComparison groups include all children who remained in the school district - both those promoted each year and those retained in first orsecond grade.Each year a new Random Sample was drawn. The numbers in the table reflect Random Sample students presenton the day of testing.



At the end of third grade, the percentages of Discontinued Reading Recovery children who performedwithin or above the average band on Text Reading looked very similar to the percentages of RandomSample children performing at these levels.

The percentages of Comparison children performing within or above the average band at the end of thirdgrade fell far below the percentages for Discontinued Reading Recovery children and the RandomSample.

The distribution of Not Discontinued Reading Recovery students was far lower than the other groups.However, it should be remembered that only the least successful Reading Recovery students arerepresented by the Not Discontinued group. When the Reading Recovery group is viewed as a whole(including both Discontinued and Not Discontinued students), the distribution is still marked:), higherthan that of the Comparison group.

Table 6.-Numbers and Percentages of 1985.86 First-Grade CohortReading Recovery and Comparison Children Who Fell Within the Average Band on Text Reading Level,

May 1987 and May 1988

May 1987

Band

Above Av.

Average*

Below Av.

May 1988

Band

Above Av.

Average*

Below Av.

Reading RecoveryDiscontinued Not Discontinued

N %1 RINIC1Total

15 0 1518.1 0.0 15.8

21 253 0 0.0 21 22.1

47 56.6 12 100.0 59 62.1

30 48.4 0 0.0 30 44.1

13 21.0 0 0.0 13 19.1

19 30.6 6 100.0 25 36.8

Random Sample

19

21

28

34

14

19

27.9

30.9

41.2

50.7

2Q.9

28.4

NOTE: Numbers of children in Table 6 include only those who were actually in the appropn.....-. ,,,ade level at the time of testing. and thereforediffer from thenumbers in Table 5, which include el 1985-86 first -gran :ohort children regardless of the grade they were in at the time of testing.

The average baud is defined u plus or minus .5 standout, deviation from the mean score of the Random Sample. Using the Random Sample meansandstandard deviations shown in Table 5, the average band for each testing date is:

Mean SD Average BandMay 1986 11.15 5.62 8.34-13.96May 1987 18.60 6.67 15.27-21.94May 1988 23.50 9.00 19.00-28.00

53

Chapter 4:Ohio ReadingRecovery Project

Studies of Reading Recovery in Ohio

1984-85 1985-86 1986-87 1987-88

Columbus Pilot Study

Columbus Longitudinal Study

Ohio State Study, 1st Year

Ohio State Study, 2nd Year

Ohio State Study, 3rd Year

]

Implementation at State Sites

Based on the positive results of the 1984-85pilot study in Columbus, tiu-. Ohio General As-sembly provided fmancial support to implement 2statewide Reading Recovery program. r:e OhioDepartment of Education developed a four-yearplan to implement an early reading interventionprogram. The program was designed to eventu-ally reach an estimated 15 percent of all firstgraders in Ohio.

1985-86In 1985-1986, the same year that the longi-

tudinal Columbus project was implemented, 28Teacher Leaders were enrolled in the trainingprogram at The Ohio State University 22 rep-resenting Reading Recovery sites distributed geo-graphically throughout the State of Ohio, and sixprospective Teacner Leaders from the university.

Teacher Leaders were selected for exper-ience and leadership. Each Teacher Leader taughtfour children each day in Reading Recoverylessons and completed various other responsibil-ities in his or her district. A total of 110 childrenwere served. One day a week throughout theyear, Teacher Leaders travelled to The Ohio StateUniversity for an all-day training session wherethey learned Reading Recovery diagnostic and

45

0-7

intervention procedures, participated in demon-stration lessons using a one-way observationglass, and acquired the theoretical and practicalknowledge needed to perform the Teacher Leaderrole the following year.

1986-87During the following year, 1986-1987,

trained Teacher Leaders taught 235 teachers at 23regional training sites in Ohio. An additional 22Teacher Leaders began training. Several of theregional training sites were consortia in whichschool districts and colleges or universities co-operatively implemented the program.

The state project in 1986-87 involved 167school districts. The student population repre-sented a wide geographic distribution and in-cluded students from urban, suburban, and ruraldistricts. A total population of 1,130 studentsstatewide had at least 60 Reading Recovery les-sons or were successfully discontinued. Thesechildren were the participants in the evaluationstudy in 1986-87.

1987-88During the 1987-88 school year, 2,648 chil-

dren received at least 60 days of Reading Recov-ery instruction. These first-grade students wereparticipants in the 1987-88 study. Twenty-three

46 Ohio Reading Recovery Project

regional training sites, 228 school districts, and416 teachers were involved.

Study of Statewide Results

As in the Columbus longitudinal study,children selected for Reading Recovery statewideranked in the lowest 20 percent in their first-gradeclassrooms in reading skills. The reading skilllevels of Reading Recovery children on ente.ingand leaving the program varied according to theaverage skill levels for first-grade children at eachsite. The numbers of children served at each sitevaried according to the number of teachers-in-trainirg and the hours available to work withReading Recovery children.

Unlike the Columbus study, however, thestate site project had no group of Comparisonchildren against which to measure the ReadingRecovery children's progress. Instead, th .::. effec-tiveness of Reading Recovery statewide wasmeasured in this study by comparing the ReadingRecovery children with the average bands of firstgraders at their respective school sites and byexamining their gains during the interventionyear.

Children who received Reading Recoveryinstruction in the first grade in 1985-86, 1986-87,and 1987-88 were tested at the end of their re-spective first-grade years to determine the effectsof Reading Recovery. The data for the state eval-uation study were analyzed to answer the fol-lowing research questions:

1) What proportion of children served by theReading Recovery program statewide weresuccessfully discontinued?

2) What proportion of discontinued ReadingRecovery children achieved end-of-yearscores equal to or exceeding the averageband of their respective school sites?

3) What was the rate of progress from entrythrough the end-of-the-year tasting for chil-dren who were successfully discontinued

from the program during the firs sixmonths?

4) What was the average Normal Curve Equi-valent (NCE) gain score on a nationallynormed test for the population of success-fully discontinued Reading Recovery chil-dren and the population of not-discontinuedReading Recovery children?

ProceduresAt the beginning of the year, at the time

w ..1 they /ere discontinued, and at the end ofthe school year, Reading Recovery children wereassessed on 10 dependent measures. (See Ap-pendix C for a description of each measure):

1) Text Reading Level;2) Letter Identification;3) Word Test;4) Concepts About Print;5) Writing Vocabulary;6) Dictation;7) Sight Vocabulary;8) Phoneme/Grapheme: Consonants;9) Phoneme/Grapheme: Vowels; and

10) Reading Comprehension.

Results and Analysis

Question #1 What proportion of childrenserved by the Reading Recovery program state-wide w' re successfully discontinued?

The percentage of program children whowere successfully discontinued either during theyear or at the eild of the school year was calcu-lated.

During 1985-1986, a total population of 110children were served by the 22 Teacher Leaders intraining. Of that population, 73 percent were dis-continued. The results were good considering thefact that, because of the all-day training sessionson Mondays, Teacher Leaders worked with chil-

Ohio Reading Recovery Project 47

dren only four days each week instead of therequisite five days.

During the 1986-1987 year, trained TeacherLeaders taught classes involving 235 teachers at23 sites in Ohio. Of the 1,130 children who re-ceived a full Reading Recovery program (at least60 days of instruction), 82 percent were success-fully discontinued. In contrast to the previousyear, all children, except those served by TeacherLeaders in training, received the requisite fivedays each week.

During the 1987-1988 year, of the 2,648children who received a full Reading Recoveryprogram, 86 percent were successfully released.The data for 1986-87 and 1987-88 are shown forindividual state sites in Figure 15 (page 49).

Question #2 What proporron of discon-tinued Reading Recovery children achieved end-of-year scores equal to or exceeding the averageband of their respective school sites?

To address this question, discontinuedReading Recovery children at each site werecompared with the average band of their respec-tive school. site. Figure 16 on page 50 lists thepercentages of students meeting or exceeding theaverage band for each test at the end of first gradein spring 1987 and spring 1988. The data showthat high percentages of the Reading Recoverychildren, ranging from 68.5 percent to 94.8 per-cent, achieved scores within the average band.Results are provided for two years. In May,1988, only three of the measures were adminis-tered because of the time involved in testing. Ex-perience had shown that it was not necessary toadminister all measures.

Question #3 What was the rate of progressfrom entry through end-of-year testing for chil-dren who were successfully discontinued fromthe rE2mdglrin the first six months?

To address this question, the continuedprogress of students who were released from theprogram during the first six months of the schoolyear was assessed for six measures. The purposeof this analysis was to determine whether thosestudents who were released from the program andgiven no further help in Reading Recovery wouldcontinue to make progress in the classroom in-structional program through the rest of the year.

In 1986-1987, a total of 231 children werediscontinued from the program during the first sixmonths of the school year These students madeconsistent gains across the school year, even afterthe withdrawal of extra help. In the 1987-1988school year, a total of 699 children were dis-continued from the program during the first sixmonths. These results replicated the trends shownin the previous year. Results are summarized inTable 7 on page 51 and are shown graphically inFigure 17, page 52.

Question #4 What was the average NormalCurve Equivalent (NCE) gain score on a na-tionally normed test for the population of suc-cessfully discontinued Reading Recovery childrenand the population of not-discontinued ReadingRecovery children?

All Reading Recovery children were as-sessed at the entry point of the program and at theend of the year on three scales of a nationallynormed test: 1) Reading Comprehension; 2) SightVocabulary; and 3) Phoneme/ Grapheme: Conso-nants.

The Normal Curve Equivalent (NCE) gainscores for the children served in each of the twoyears are displayed in Table 8 on page 53. Again,the data show that Reading Recovery childrenmade consistent NCE gains during the first twoyears of statewide evaluation. The Normal CurveEquivalent gain for children making average

48 Ohio Reading Recovery Project

progress in other words, those remaining inthe same place on the Normal Curve for theirgrade level would be zero. For the total groupof Reading Recovery children, the NCE gain inReading Comprehension was 9.6 points in 1986-87 and 7.0 points in 1987-88; in SightVocabulary, 8.1 points in 1986-87 and 4.6 pointsin 1987-88; and in Phoneme/Grapheme:Consonants, 8.5 points in 1986-87 and 7.3 pointsin 1987-88. In each case, the NCE gain coresfor successfully discontinued Reading Recoverychildren exceeded the gain scores for not-discontinued Reading Recovery children.

Summary

Evidence from the state evaluation studiesindicates that the program has been successful inits first three years of implementation.

The successfully discontinued rate for proj-ect sites was high. During the first year of imple-mentation (1985-86), the rate was above 73 per-cent for all sites statewide. During the secondyear, the rate was 82 percent for all sites, andduring the third year, the rate was 86 percent. Inaddition, test score data indicated that ReadingRecovery statewide produced gains in studentreading ability similar to those documented in theColumbus Public Schools during the first year ofthe longitudinal study.

Examination of a random sample of teacherrecords showed that ',he program was imple-mented with integrity. Teacher Leaders provided

the necessary amount of instruction and followedguidelines set by the university. Documented siteobservations and project results provided evi-dence that, in general, instruction in the teachercourses was of high quality.

Teacher Leaders' reports (end-of-year doc-uments from each site) of the responses of rele-vant other personnel (administrators, other teach-ers, children, and parents) indicate that the projecthas been positively received at the local level.The state Reading Recovery conference, whichdrew close to 900 participants, including manyfrom outside Ohio, was successful in providinginformation about the project. Project sites hostedmany visitors from Ohio and outside the state.

Evidence gathered during the state studyalso identified some concerns that should be ad-dressed. First, there is a continuing need to in-vestigate ways to integrate more children in to theprogram. Attendance and mobility will continueto lx. problems until coverage is increased.

Second, more communication with class-room teachers is needed. Classroom teachersneed high quality staff development in order tounderstand methods of assessing children's abili-ty and working with them, thus providing thekind of consistent service that will make the mostof Reading Recovery instruction.

Third, it is necessary to continue coopera-tive work among the various departments andfunding sources involved in the program so thatteachers will devote less time to administrativeduties, and more time to working with children.


FIGURE 15. Percent of Children Receiving Reading RecoveryInstruction Who Were Successfully Discontinued, Ohio State Sites

196647

191748

ttat

e

S

teS

41.1.

A 79

13

C

D

E

F

711

399

J

116

K

L

M 7395

79ss

7693

$490

0

Percent Successfully Discontinued

75 100

"s s :s

1..=iiss - s.'`..

ss,

Z.

zs

N

o rs

Q

R

T

-,'-''.."7".1.111111110111-1

U 1°391

V

W 91

Average 2

so Ohic Reading Recovery Project

Highlights from Figure 16:

Both the 1986-87 and 1987-88 Successfully Discontinued Reading Recovery students performed sig-nificantly better on a variety of measures of reading-related skills at the end of first grade than theirperformance at the beginning of first grade would. have predicted.

Of the 1986-87 state site group, mon. than two- thirds of the Successfully Discontinued Reading Recov-ery children achieved scorts in text reading equal to or exceeding the average band (+ or - .5 standarddeviation of the mean score) of scores of a randomly selected group of first graders. For the 1987-88group, the corresponding figure was 74.1 percent.

Figure 16. Percent of DiscIntinued Reading Recovery Children at OhioState Sites Achieving Scores Equal to or Exceeding

Average Band at the End of First Grade

0

Text Reading

Writing Vocabulary

Dictation gf

Letter Identi (cation

Ohio Word Test


Text Reading

Writing Vocabulary

Dictation

0

May 1987 100

68.5

May 1988*

86.8

94.2

94.6

92.6

86.2

100

74.1

82.1

kra.",: 94.8

*Note: All measures were not used in 1988 because experience had shown that it was unnecessary.

Ohio Readin Recovery Project 51


For Successfully Discontinued Reading Recovery students who completed the program within the firstsix months of the school year, the tests of reading skills were administered three times -- at the beginningof the program, at the point at which they were discontinued, and again at the end of the school year. Inall cases, the students showed continued progress in skill development even after they were no longerinvolved with Cie Reading Recovery program.

While scores for many of the subskills were, at the exit point, so close to the maximum possible scorethat little absolute improvement was possible, the students demonstrated substantial improvement in theareas of text reading (where exit scores had peen less than half of all possible points) and in writingvocabulary (where there was a time limitation rather than a specific score possible).

Table 7.-Statewide Progress of Children Who Were SuccessfullyDiscontinued from the Program During the First Six Months of the School Year

INSTRUMENT ENTRY EXIT END-OF-YEAR

11986.87 1(N=231)*

Text Reading

Writing Vocabulary

Dictation


Ohio Word Test

Concepts About °rim

11987-88 l(N=699)*

Text Reading

Writing Vocabulary

Dictation


Ohio Word Test


.78 10.61 20.05 (Max=26)

5.54 37.84 49.11 (10 min.)

8.80 33.88 35.38 (Max =37)

45.36 52.86 53.41 (Max=54)

.82 15.69 19.15 (Max=20)

9.43 18.87 20.44 (Max=24)

.13 11.63 19.41 (Max=26)

5.57 39.76 51.20 (10 min.)

9.01 34.17 35.28 (Max : :37)

44.39 52.97 53.37 (Max=54)

.76 15.58 19.17 (Max=20)

9.74 19.36 20.44 (Max=24)

*Includes only those children succe. sfully discontinued during first six months of school year. Many more children were successfullydiscontinued later in the year.

PA,

Ohio Reading Recovery Project

FIGURE 17. Statewide Progress of Children Who Were SuccessfullyDiscontinued Within the First Six Months of the School Year

IN Entry 3 Exit 0 End-Of-Year la-- Maximum Score .1

25

20

15

10

5

14

Text ReadingMax=26

60

40

20

Writing Vocabulary(10 minutes)

60

50

40

30

20

10

0

1986- 1 1987-88

Letter IdentificationMax=54

1986-87 1987-88

Ohio Word TestMax=20

1936-87 1987-88

25

Concepts About PrintMax=24

1986-87 1987-88


Highlights from Table 8:As with the Columbus longitudinal study, Ohio Reading Recov -Ty students at he state sites achievedsubstantially positive normal curve equivalent (NCE) gain scores. The mean gain scot is for SuccessfullyDiscontinued Reading Recovery students were substantially greater than the mean gain scores for Not-Discontinued Reading Recovery students.

Table 8.-Gain in Normal Curve Equivalent (NCE) Scores forReading Recovery Students Statewide, 1986-87 and 1987-88

Year One(1 9 8 6 - 8 7)

Year Two(1 9 8 7- 8 8)

READING COMPREHENSIONFALL SPRING**

Gain(N) NCE* (N) NCE NCE

TOTAL (908) 36.5 ',081) 46.1 9.6Discontinued (732) 37.7 086) 47.9 10.2

Not Discontinued (176) 31.5 (195) 37.7 6.2

TOTAL (1945) 37.4 (2,232) 44.4 7.0Discontinued (1674) 38.4 (1,929) 46.1 7.7

Not Discontinued (271) 31.5 (303) 33.9

SIGHT VOCABULARY

I FALL SPRING**

(N) NCE* (N) NCE

Year One TOTAL (908) 36.5 (1,082) 44.1

(198 6 - 8 7) Discontinued (732) 38.3 (886) 47.4

Not Discontinued (176) 29.0 (1%) 31.4

Year Two TOTAL (1,947) 42.2 (2,234) 46.8

(1 9 8 7- 8 8) Discontinued (1,676) 43.8 (1,929) 49.1

Not Discontinued (271) 32.4 (305) 32.2

NCE GainT9.12.4

4.65.3-0.2

PHONEME/GRAPHEME: CONSONANTSFALL

(N) NCE*Year One Total (909) 35.1

(1 9 8 6- 8 7) Discontinued (732) 37.1

Not Discontinued (177) 28.2

Year Two Total (1,947) 40.6

(1 9 8 7 -8 8) Dixontinued (1,676) 42.0

Not Discontinued (271) 32.4

SPRING**

(N) NCE NCE Gain(1,081) 43.6 8.5

(886) 46.8 9.7(195) 29.9 1.7 j

(2,228) 47.9

(1,924) 49.9 7.9(304) 35.5 3.1

*Mean scores were rounded to the closest whole number, and that converted from appropriate tables to percentiles wth,7h were thenconverted to NCE scores.**Spring has a larger N bec ause of students who entered the program after fall testing.

Ti::)

Chapter 5:Reading RecoveryStaffing andStaff DevelopmentProgram

Program developers and implementers agreethat the key to success in Reading Recovery is theyear-long staff development course for teachersand Teacher Leaders (Clay, 1982; Pinnell andWoolsey, 1985; Lyons et al., 1986). Teachers'participation in this special preparation ensureseffective implementation of the program proce-dures.

In this chapter, we describe the roles playedby Reading Recovery teachers and Teacher Lead-ers. We then discuss the staff development pro-gram used to prepare teachers for their difficultand challenging Reading Recovery assignments.

Importance of theStaff Development Component

The basis of good instruction is the teach-er's knowledge of learning and of the particularchild, along with the ability to use that knowledgein assessment and instruction. In the ReadingRecovery staff development program, teachersexamine and discuss their own assumptions aboutreading and learning to read. Teacher Leaderssupport teachers and challenge them so that theycan push their own boundaries towards newlearning.

Reading Recovery teachers are not just goodteachers who are "intuitive" or who make thechild feel good and comfortable. They engage in

55

6U

high-level analytical activity, an ongoing "peda-gogical -easoning process" (Rentel and Pinnell,1987). !This process is an integral part of theprogram. The eramework of procedures cannotbe effectively applied without the simultaneousdevelopment of a set of teacher understandings.

Teacher learning cannot be short-changed.It takes time, but once accomplished, it is an ir-reversible process. It may require resources, butit is not a consumable material, nor will it gatherdust on shelves. Knowledgeable teachers repre-sem a resource for the school system that willcontinue to produce results year after year.

Reading Recovery Teachers

Reading Recovery teachers are certificatedteaches who: volunteer for the program; have hadexperience in teaching young children (minimumthree years); and possess proven ability to estab-lish a rapport with children. In general, ReadingRecovery teachers should be selected becausethey have demonstrated good teaching at the pri-mary level.

Reading Recovery teachers' primary re-sponsibility is to work with at least four childreneach day in Reading Recovery lessons. Althougha teacher could work all day in the program,experience has si.own this to be undesirable be-cLuse of the intensity of the teacher-student inter-

56 Reading Recovery Staffing and Staff Dershcastlghram

action required io Reading Recovery lessons.Therefore, the Reading Recovery teacher shouldhave a half-day assignment in some other area.

Several implementation models are used inOhio. In one arrangement, a regular first-gradeteacher and a reading specialist, both of whomhave received Reading Recovery training, share afirst-grade classroom, each working half thedaywith Reading Recovery children and the otherhalf as a regular classroom teacher. Funding orthis model works out as two half-time specialreading teachers.

Classroomteacher

Chapter Oneteacher

A.M.

Teach 1st grade

Teach FR(4 - 5 children)

P.M.Teach RR

(4 - 5) children

Teach 1st grade

One advantage of this model is that knowl-edge gained in Reading Recovery can informclassroom teaching. By rotating teachers, thismodel could also contribute to the expertise offirst-grade teachers. It provides the additionaladvantage of helping special reading teachers un-derstand the exvectations of children in the class-room setting.

Some districts find this first model difficultto fund. Another model is to have a special read-ing teacher work one half day in ReadingRecov-ery and the other half day with groups of childrefrom a variety of grade levels. This model en-ables Reading Recovery teachers to transfer someof their new skills to working with groups ofchildren (although it is important to note that thespecific teaching procedures of Reading Recoveryare not suitable for group instruction).

The first two organizational patterns men-timed require reallocation of remedial reading re-sources, but they do not require hiring an addi-tional teacher. If a district is able to employ addi-tional personnel to implement the program, sev-eral more options open up. Two regular class-

room teachers might share a first-grade class,each working one-half day in the classroom andone-half day in Reading Recovery. Or, if kinder-garten is a half-day program, kindergarten teach-ers might work the other half of the day withReading Recovery students. Some districts havehired permanent tutors to provide Reading Re-covery on a half-day basis; however, tutorsshould be cel ified teachers, should work everyday, and should make a commitment to take theyear-long training and to work in the school foran entire year.

Reading Recovery teachers are responsiblefor communicating with parents about the pro-gram and for holding all necessary conferencesand contacts with parents of the individual chil-dren they teach. Reading Recovery teachers alsowork with classroom teachers in the building sothat they understand the program and can act aspartners in helping the children make acceleratedprogress. Reading Recovery works best whenthe Reading Recovery teacher and the regularclassroom teacher are working together to help thechild develop an independent reading system.

Reading Recovery Teacher Leaders

The key implementer of Reading Recoveryis the Teacher Leader. Called "tutor" in NewZealand, this role combines teaching of childrenwith the responsibilities of teacher educator andresearcher. The Leader's role is a complex onethat a soh( al district will want to inves.igatecare-fully before deciding to implement the program.

Teacher Leaders should have a master's de-gree or equivalent and experience in teachingyoung children. In addition, they should haveexperience in leadership roles and in providing in-service, inining for other teachers. The Leaderrole requires a high energy level, strong commit-ment, and a variety of skills for working withadults.

Reading Recovery Staffing and Staff Development Program 57

A lull -time Teacher Leader is necessary inorder to provide training and institute the self-renewing system of the program; however, aLeader can serve more than one district. Exceptfor very large districts, Leaders in Ohio work inregional sites, monitoring and supporting teachersin several different districts. Although a few uni-versity professors have undertaken the role inOhio, the Teacher Leader usually works for aschool district. The Leader is also qualified toserve in an adjunct role with the college or uni-versity providing cou. e credit for the teachercourse.

Teacher Leaders perform many duties.First, it is essential that they continue to workwith Wiar children eac:: day. Only by teachingchildren can Leaders refine and further developtheir understanding of the procedures, techniques,and strategies included in the program. Experi-ence has shown that when Leaders do not con-tinue to work with a range of children, the resu'tsof the program are not as good.

The Teacher Leader also conducts the year-long course for teachers, beginning with a sum-mer workshop so that teachers will be ready touse diagnostic procedures when school starts.The course meets each week during fall, winter,and spring. The Leader leads "behind the glass"sessions and discussions after lesson demonstra-tions. Additionally, the Leader provides lecturesand leads discussions to help teachers broadentheir theoretical base. The Leader also visitsteachers in training at their schools, workingindividually with them at least four times duringthe training year. The Leader prepares coursehandouts, syllabi, and other materials. He/sheobserves the teacher group closely, in both groupsessions and individual lessons, and provides theinstruction needed to develop essential knowledgeand skills.

In addition to teaching the class, TeacherLeaders continue to work with and monitor pre-

viously trained teachers. They visit trained teach-ers at sites to provide technical assistance and toconfirm that the quality of the prc, ;ram is beingmaintained after the training year ends. The Lead-ers monitor children's progress to make sure thatindividual sites do not "let down" in the press foraccelerated progress and high achievement.

Teacher Leaders also serve a general in-service role for the school districts. They conductawareness sessions for parents, administrators,and classroom teachers concerning the ReadingRecovery -grogram. In addition, Leaders can helpclassroom teachers work more effectively withyoung children in reading and writing. In somedistricts, the Leader trains primary classroomteachers to use the diagnostic tools of ReadingRecovery, such as the running record, to improvetheir classroom work.

Teacher Leaders work with district admin-istrators to set up the arrangements for imple-menting Reading Recovery in district schools. InOhio, an administrator usually is designated asSite Coordinator. The Site Coordinator providesfor administration of the program, funding, andteacher assignment.

Finally, Teacher Leaders coordinate the re-cording and collection of data, and use these dataduring the year to monitor children's progress.Data are sent in a specified format to The OhioState University fcr analysis and feedback for thesites. At the end of the year, each Leader pre-pares a site report detailing progress of children inthe program that year. This close monitoring ofprogress provides ongoing evaluation of theinstruction and the training program.

Goals of StaffDevelopment Program

The staff development program is designed tohelp teachers learn very specific skills and devel-op the necessary base of knowledge needed in

62

58 Reading Recovery Staffing r d Staff Development Program

order to teach young e.hildren in one-to-one situ-ations. This program of staff development is notintended to improve classroom teaching (althoughparticiponts often claim that the growth in knowl-edge they experienc-- is valuable and can beapplied in other areas). Teachers who have diffi-culty in classroom teaching should not be placedin Reading Recovery for improvement. The de-manding program involves development beyondthat required for good classroom teaching. Goalsof teachers in the in-service program are to:

develop a deeper understanding of the natureof the reading and writing processes;become sensitive observers of the reading andwriting behavior of children;become competent in using specific teachingprocedures with iadividual children;support children so that they make acceleratedprogress in reading;become comfortable and in articulatinghow they are responding to the child duri.ginstruction;be able to critically evaluate their own as-sumptions and practices as well as those ofothers.

Origin of the StaffDevelopment Program

The staff development program grew out ofthe activities of Clay's original research team, agroup of teachers who were -i.edying children'sreading behavior.

During the research project, good readingteachers worked with individual children behind aone-way glass while the research team watchedfrom the other side. The observers could heareverything said by the teacher and child. Stand-ing close to the glass, they could observe beha-vior in detail. Over a period of time, theygathered precise information about children's be-havior and about teacher responses that seemed to

help children. Researchers selected the mostpowerful responses and produced a set of guide-lines that teachers could use to make effectivedecisions while working with individual children.Those guideline' represented a repertoire of responses. They were subsequently tested withgood results.

While planning for wider dissemination,Clay's research team did something unique inresearch and development efforts. Consistent withtheir theoretical view of learning as a process ofconstructing meaning, the researchers hypothe-sized that teachers must participate in constructingtheir own understandings, just as the researchershad constructed their theories. The original teamvalued the "behind-the-glass" activities so highlythat they recommended that all future traineescontinue the inquiry process, thus, in effect,engaging in theory-building activities forthemselves.

Since that time, the Reading Recovery staffdevelopment model has been refined and mademore explicit, but it has not lost the most criticalelement teachers developing their own knowl-edge by obsering actual Reading Recovery in-struction and by talking together.

Teacher Training

During the first year of program involve-ment, Leachers work daily in one-to-one sessionswith at least four first-grade children. A mini-mum of four children is essential so that teachersnew to the program can experience the range offlexibility that a Reading Recovery teacher mustposses to adjust the instruction to individuals.Once each week, the teachers participate in ananalysis of teaching from two live demonstrationlessons. Then, they discuss the lesson and ir -sights gained with the demonstration teacherspresent. In addition to the weekly lesson andseminars, teachers rec6ve visits at their school

60

IV

Reading Recovery Staffing and Staff Development Program 59

sites from colleagues and experienced TeacherLeaders. Their work with children is monitoredclosely, and professional colleagueship is estab-lished.

In the first four to six weeks, the coursefocuses on assessment training in which teachersbecome skilled in observing and recording exactlywhat children are doing as they work at readingand writing. Teachers learn to administer theDiagnostic Survey, to take running records ofchildren's oral text reading, and then to make theirobservations more explicit in a Diagnostic Sum-mary Report that describes a child's behavior andsome conclusions for analysis of that behavior.

On all observations, teachers are encouragedto look beyond "stores" or "competencies" tointerpreting the shades of meaning in each child'sapproach to the reading and writing tasks. Theinformation collected helps teachers to select thestudents most in need of interventions and todetermine the particular instructional programneeded by each child.

During the rest of the training year, teachersconcentrate on learning more about the readingprocess as they use the procedures with children.They observe and interpret children's behavior,and they note teacher-child interactions that sup-port c' ';:en's progress as well as those that mayinterfem. As teachers move into the instructionalprogram with their children, the focus of theweekly seminar shifts to teaching demonstrationsand discussions with extensive use of the one-way glass.

After their initial training year, Reading Re-covery teachers come together for continuingcontact sessions at least four times each schoolyear. These sessions may take place after school;ideally they are half-day sessions with teachersreleased from duties. Without these sessions,teachers may begin to c aft and to start resorting tot.1d practices. They need to continue to grow intheir understanding of the reading and writing

processes. At these continuing contact sessions,trained teachers may view a lesson demonstration,now working at a higher level to refine theirobservational skills. They may also choose toanalyze a taped session or to make an in-depthstudy of some aspect of the program. Often, theydiscuss research articles.

Teacher Leader Training

Training for Teacher Leaders now takesplace only at The Ohio crate University, althoughother states may soon establish such trainingsites.

Teacher Leaders participate in the sametraining provided for teachers. In addition, theymust learn how to teach others and to -;stablish thecommunication system for a Reading Recoveryproject. For one academic year, Leaders take aclinical class, similar to the teacher class, .iid atheoretical seminar to establish an understanding

the research base for the program. One dayeach week, they spend the entire day in class atthe university. During the other days of the week,they teach four children each day in ReadingRecovery lessons and participate in a variety ofintr.-ship activities at a Reading Recovery site.

Each Teacher Leader observes and graduallytakes a leadership role in an existing teacher class.This provides the opportunity to practice leading"behind the glass" sessions and discussions. Italso gives the Leader a chance t. observe thechanges in the teacher group over time and to seean experienced Leader at work. Internship : .ivi-ties also include visiting individual teachers inschools and participating in continuing contactsessions.

Talking While Observing

The Teacher Leader works with the class of10 to 12 teachers for the entire year. During

60 Reading Recovery Staffing and Staff Development Program

weekly "behind the glass" sessions, the Leaderstimulates the discussion and challenges the ob-servers to deal with difficult ideas, to revise theirthinking, and to speculate about the learning thatmay be taking place.

Through this behind-the-glass experience,the teachers learn to be sensitive observers ofchildren's reading and writing bellNior and todevelop skill in making those moment-to-momentdiagnoses that inform instruction. The goodteacher makes decisions "on the run" while teach-ing and quickly tailors his/her own responses tosupport the child's learning. The lesson is fastpaced; observers must test diagnostic and deci-sion-making skills as the lesson unfolds.

In "talking whilt, observing," teachers arerequired to state their observations aloud and torisk making inferences and reaching on-the-spotconclusions. Th ey hone their conceptual skills asthey speculate about the possible decisions. Theybecome aware that the way they respond to chil-dren has a powerful influence.

In behind-the-glass discussion, teachersmay challenge each other. Participants do not al-ways agree on the best possible action. TeacherLeaders say that the most poweiful learning takesplace when teachers in the group start to argueand defend their statements with evidence fromthe child's behavior or from external sources.

After the observation, the group memberstalk with each other and the demonstrating teach-er. During this time, the participants can reflecton the lesson and check their observations withthe teacher's views and with other sources. Al-though some critical feedback may be given to thedemonstrating teacher, the purpose of the obser-vation is not simply to critique the teachers' be-havior or style of teaching. It is not to comparethe teacher to a checklist or any model of teach-ing. The observation session is mainly for thebenefit of the group of observers.

The talking-while-observing process mightbe compared to a group of doctors observing acomplicated operation. Freed from the responsi-bility of action, the observers have the chance todevelop their ability to predict and to make deci-sions. Teachers can become "noticing teachers."In the process, they begin to think and talk in newways about Cr:. process of becoming literate andabout children's potential for learning.

In Reading Recovery staff development,teachers reflect on their teaching decisions; theyengage in discussion with their peers, they haveon-the-spot assistance from a specially trainedTeacher Leader. In addition, through the processof making their ideas explicit and backing them upwith evidence, they build a knowledge base fromwhich they can act.

In a sense, teachers develop their own "selfgenerating" systems a set of understandingsand beliefs that undergird their actions with chil-dren. They build their own theories and simulta-neously put those theories into practice.

Research on Reading RecoveryStaff Development

Research on teachers who have been in-volved in Reading Recovery shows that the train-ing program has a powerful impact or participants(Geeke, 1988; Pinnell & Woolsey, 1985; andLyons et al. Reading Recovery Technical Re-ports, 1985-1988). Individuals generally experi-ence a shift in theoretical orientation, movingfrom a "skills-oriented" view of reading, vihichfocuses on materials and sequential leaming ofspecific aspects of reading, towards an orientationwhich suggests that children "orchestrate" a rangeof skills and knowledge when they learn to readand write.

Administrators in districts where ReadingRecovery has been implemented comment thatReading Recovery teachers seem to change theirvows of children, that they are better observers of

Reading Recovery Staffing and Stpq Develo)ment Program 61

children's progress, and that they have more con-fidence in their own ability to teach children toread. All of those changes are the result of con-necting theoretical knowledge with closework with children and with the development ofobservational skill that allows teachers to noticeand document progress, even in very small steps.When these slowest children begin to make accel-erated progress, teachers' confidence in them-selves and in the children grows.

A year-long qualitative study of one groupof teachers revealed continuous shifts in teachers'focus of attention throughout the training period(Pinnell and Woolsey, 1985). The data for thestudy consisted of transcribed group discussionsthat were held every two weeks following trainingsessions. This study or teacher language in aninformal setting revealed teacher' concerns, theirgrowth of concepts, and the complexity of ideaswith which they were able to cope at variouspoints in the in-service process.

The shift of focus over the period of threeacademic quarters is illustrated in Figure 18, page62. At the beginning of the training, teacher lan-guage centered on the logistics of implementing anew program. They wanted to know "how to doit" and to be told the "right" way. Teachers wereconcerned about management, supplies, and theinevitable conflict, questions, and stresses ^ar-rounding any innovation. Whatever the topic, lo-gistics crent in and consumed attention. It is nowo 'r that college professors find it so difficultto pt. suade teachers to deal with underlying the-ory in a two-week course or workshop.

This surface-level focus continued for severalmonths. Then, as teachers appeared to becomecomfortable with the new activities, they began tofocus on their own discoveries and insights. Thetopic of children received more attention as theyreported detailed observations of children, createdand shared metaphors, learned about each others'students, and told stories about their work.

Finally, toward the end of their trainingyear, another shift could be observed in teachers'focus of attention. In their discussions, they be-gan to generalize and to make theoretical state-ments and to hypothesize. Specific descriptionsand stories were still evident but were fewer, andteachers began to link their ideas into more co-hesive statements. The increase in theorizing wasnot evident in teachers' discussions until they hadbeen immersed for a long time in the process ofobserving, talking, and decision making in theReadiag Recovery seminar. Their use of knowl-edge at this later time illustrates that they had be-gun to internalize the behaviors they were invitedto engage in during class sessions.

I general, the study supports the idea oflong-term training to help teachers develop theirown theoretical ideas. The teachers grew in theirunderstanding of children's learning and of theirown teaching, but they also learned about Seingeducational innovators. The educational systemwas influenced by the changes these teacherswere making, and this spirit of change began tohave impact on policy decision making at localand state levels. Through interviews and discus-sions with teachers, the researchers came to thefollowing conclusions:

1. In the initial stages of an edu ational changeeffort, innovators may encounter skepticismon the part of teachers involved.

2. In the initial stages, participants are likely toconcentrate on the practical arrangementsnecessary for implementation.

3. As logistic concerns diminish and teacherscontinue to talk together and observe chil-dren, the focus can shift to acconiplishing thegoals of the project and to more substanti.c.issues.

4 Change is facilitated by creating a sharedlanguage among those who participate in theeffort.

62 Reading Recovery Staffing and Staff Development Program

FIGURE 18.Change in Teacher Focus Over a Period of Reading Recovery Training:Percentage of Teacher Comments Falling into Each Category During Group Discussions

FALL

WINTER

SPRING

45

40

35

30

25

20

15

10

5

0

45

40

35

30

25

20

15

10

5

0

45

40

35

30

25

20

15

10

5

0

14.2

40.2

17.5

9.8

15.6

2.9

Children Imple- Practice System Theory Homesmentation

30.7

210

4.0

Children Imp le- Practice System Theory Homesmentation

Children Imp le- Practice System Theory Homesmentation

Readin Recover Staffin and Staff Develo ment Pro ram 6 3

5. Change initiators must take into account thepowerful demands of the system in whichthe new processes must be implemented.

6. Involving teachers in observation and peercritiquing helps teachers develop the newunderstandings and confidence that providethe basi° for change.

7. Change in understandings and attitudes pre-cedes observable change in classroom prac-tice.

8. Change requires time, hard work, and un-usual effort on the part of everyone in-volved.

9. Change is often accompanied by pain andsense of discontinuity with existing beliefs.

10. Change is facilitated by a support groupwho can talk honestly, sh..-re concerns, andpro vide feedback.

11. Change is made possible when teachers re-flect on their work, realize their success,and feel the power of their own teaching.People who know they are making a dif-ference have the motivation to do what theyneed to do.

A study of the Reading Recovery programin Australia confirmed the value of the programfor helping teachers learn (Geeke, 1988). Geekeclaims that the Reading Recovery staff develop-ment program seemed to achieve genuine changein teachers. He suggests a number of possiblereasons for this, including: the fact that the Read-ing Recovery in-service extended over a full year,the observation of Reading Recovery sessionsthrough a one-way window; and the follow-upvisits by Teacher Leaders to the schools. But thecomments of the teachers on what they hadlearned from Reading Recovery point to yet an-other factor contributing to the teacher changeachieved by the program: the reconstruction ofbeliefs about learning and teaching.

The '-iterview data show that most of the participat-ing teachers had their existing beliefs shaken during

the early in-service sessions. They were quicklypersuaded that their current methods of teachingreading and writing were based on false asvmp-dons about teaching and learning. Subsequently,on the basis cf their observations of children andtheir experiences during in-service sessions, theydeveloped new beliefs about teaching and learning.This set of beliefs then acted as a framework intowhich the specific teaching practices of ReadingRecovery could be placed.

. . . the teaching procedures were not given to theteachers as 1, set of "ideas" for teaching literacy.Instead, the teachers were expected to use the !go-cedures in a way that reflected the set of basic be-liefs which were being develope4 at the same time.The ultimate aim of the training program seems tohave been the development of a dynamic relation-ship between belief an practice, with belief a',.:ingas an individualizing influence on instruction(Geeke, p. 144).

There woillt; seem to be a lesson here for the devel-opers of vther in-service courses. Despite the com-mon cynicism found among teachers about "theory"as opposed to "practice," we have noted that it wasa combination of specific practice relotld to generalprinciple which made the tuition offered by ReadingRecovery teachers so effective. It seems that realteacher change is unlikely to be achieve.; by simplyintroducing a "new method ;-,f instruction" in somecurriculum area. The new "method" will only bereally effective if teachers have thoroughly acceptedthe underlying principles of the program as well asits teaching practices. The techniques employed byReading Recovery to achieve this result deserveclose examination, especially as it appears to havebeen much more successful than usual in achievingteacher change in the group immediately involved(Geeice, p. 145).

Through interviews with ReadIng Recoveryteachers in training, Geeke identified six beliefsthat teachers said they had developed from theirinvolvement in Reading Recovery. He states thesebeliefs as the following:

1. Effective learning depends on the child as-saming responsibility for learning.

2. Effective learning is built on the child's cur-rent knowledge and skills, and depends on

v

4

4

S 4 Reading Recovery Staffing and Staff Development Program

the child understanding what is expected ofhim or her.

3. Effective learning leads to an awareness ofone's mental processes, self- monitoring ofthe cognitive strategies being employed, andthe development of a self-correcting system.

4. Effective teaching depends on accurate ob-servation and sensitive response, within aframework of coherent beliefs and e ..ctivepractice.

5. Effective teaching depends on the quality ofinteraction with the child. In particular, itdepends on astute questioning which showsthe child how to solve his own learningproblems.

6. Effective reaching depends on the teacher'sunderstanding of the learning process,checked against the actuality of children'sobservable learning behaviors. Only if theteacher really knows 11,-)w children learn willhe/she be able to adapt teaching methodsappropriately in response to the children'sdemonstrated needs (Geeke, p. 145).

Concl'ision

These studies provide evidence that theReading Recovery staff development program hasa powerful impact on tea..:iers and Teacher Lead-ers who participate. These findings also haveimplications for the creation of staff developmentprograms in general.

Attention is being given now to the impor-tance of developing teachers as professional deci-sion makers. Administrators and staff developersare turning away from the idea of "materials ap-proaches" that provide scripts and step-by-stepactivities for teachers to fork .v. They are startingto concentrete instead on the expansion of theteachers' "knowledge base." In order for teachersto change their views and develop their profes-sional knowledge, staff developers and teachersmust be prepared to invest time and unusual effortin long-term learr'ng experiences. Re.ding Re-covery presents a promising new direction forstaff develop dent in literacy education.

Finally, it must be recognized that theTeacher Leaders in Reading Recovery are takingon new professional roles. This task is furthercomplicated because there is no comparable rolewithin our educational system. The Leader is si-multaneously a teacher educLtor, a teacher ofchildren, a researcher, a supervisor/evaluator, anda university liaison. The Teacher Leader mustdemand excellent performance from teachers andmust be prepared to give critical feedback. Read-ing Recovery staff development is not consideredsuccessful unless it has observable outcomes interms of student learning; this situation i3 uniqueamong in-se_vice programs and causes somepi-c.:sure for teachers and Teacher Leaders. Thesepositive outcomes have been observed at all sites.Continuing to improve outcomes for children de-pends on continuing to offer a quality staff devel-opment program.

Chapter 6:Implementing aReading RecoveryProgram

Researchthe topic H .....ilTake fact

finding visit

\li

Generatelocal support

'6" Successful ReadingTrain teacher I Recovery Program

leaders

------,,,i Select teacherleaders

Establishing a Reading Recovery programmay at first seem like a complicated and difficultoperation. Educators involVed with Reading Le-covery sometimes say that "what makes it hard,also makes it easy." That is, getting started re-quires effort and commitment, but maintenanceand continuing improvement are built in. In thischapter, we will expand on that idea and providesome concrete suggestions for implementing ahigh quality program.

Need for a Quality,Long-Term Effort

The program described here is radicallydifferent from the usual innovations that come andgo in school districts. Most experienced schooladministrators can point to dozen innovations,all with promise, that have been adopted and dis-carded within a three-year period. Each time aninnovation is superficially implemented and thenquickly phased out in favor of the next "hot" item,personnel become a little more cynical.

We new have a teaching population thattends to stay for an extended career. Those teach-ers need a chance to engage in long-term changeand to see the results of their efforts. No newapproach or program lasts forever, but hig:i qual-ity implementation, with maximum chance of

success, will encourage staff members to putmore effort into growth and change.

A System Approach

Reading Rec :very is ~.such more than adefined curriculum and instructional program forchildren. It is a complex set of interlocking sys-tems, all aces dry to implement the program. Infact, it is a system intervention. This systemicaspect gives the program the chance to be estab-lished, to be implemented in an effective way, andto last long enough for rigorous evaluation to takeplace.

Initiating the Reading Recovery system re-quires commitment, resources, time, and effort.Once in place, however, it operates as a self-maintaining and self-renewing system. Peopleinvolved continue to become more knowledgeableand skillful uther than out of date or routinized.

The Ohio Model

Since Reading Recovery is a state-spon-sored program in Ohio, it is appropriate to de-scribe the state network that exists for implement-irg the program in that state.

In Ohio, various institutions have estab-lished working relationships to provide the ad-ministrative and technical )ort needed. Within

6570

66 Implementing a Reading Recovery Program

each school district and at the state level, ChapterOne staff and curriculum area staff work togetherto sponsor the program. In most districts inOhio, Reading Recovery is a Chapter One inno-vation, although several other funding sources arealso used. Some districts support Reading Re-covery througii the general fund. At the statelevel, departments related to federal programs,staff development, and curriculum cooperate toprovide resources, evaluate the program, andperform necessary administrative tasks. The uni-versity, school districts, and the state agencywork closely together. The state agency adminis-ters funds; selects Reading Recovery sites,teachers, and Teacher Leaders; and collects booksand materials. The Ohio State University pro-vides training of Teacher Leaders and teachers,monitors collection and analysis of data, and pre-pares a final report each year. The university alsoprovides continuing support to Teacher LeadersthiJughout the year.

The effectiveness of the program dependson the knowledge base cf teachers and TeacherLeaders. Thus, it is necessary to continually ex-pand and update their knowledge. Their trainingmakes it possible for teachers to apply newknowledge to their work with children, refiningtheir decisions and skills as they do so. Eachyear, data on children at the regional sites pro-vides a feedback system for refining the program,while research in the field provides new insights.

The State Agency RoleFor Reading Recovery to be implemented

on a statewide basis, the state education agencymust provide strong leadership. In the Ohioproject, the state agency helped in securing thegrants for the pilot study and devised the plan forimplementing Reading Recovery as a statewideprogram the next year. The state education de-partment selects sites, admiiiisters funding, andensures that the program meetF regulations. Col-laboration between the Division of In-service Ed-

ucation and the Division of Federal Programs hasensured that Reading Recovery projects fundedby Chapter One meet the guidelines for that pro-gran,. The State Department of Education alsoprovides the small books used by the childre:i inReading Recovery and the instructional materialsfor the teacher classes; centralized ordering savesboth cost and time. The state agency also spon-sored an outside evaluation of the prograin, con-ducted by a nationally selected panel of experts i 1reading education.

The Univers4y RoleThe Ohio State University serves as a sys-

tem coordinator for instruction of Teacher Lead-ers, teachers, and children. University personneldevelop guidelines for program implementationand design supporting materials for trainingclasses. The university assumes responsibilityfor training Teacher Leaders and providing con-tinuing contact sessions for previously trainedLeaders.

The Teacher Leader course, as described inChapter 5, is a one-year experience that includesclinical training, a beoretical seminar, and a vari-ety of internship experiences to help the prc ;pee-tive Leader learn all facets of the role. Continuingcontact sessions include four one-day conferencesduring which Teacher Leaders have a chance toshare ideas and update their knowledge. Twomajor events are the Reading Recovery confer-ence, sponsored yearly by the Ohio Department ofEducation, and th,:, Sui imer Institute, an intensivefour-day session in June of each year. The Sum-mer Institute provide:, a time for further profes-sional development, for reviewing the success ofthe project in the previous year, and for workingto improve program quality. During the 1988Summer Institute, for example, Teacher Leaderscreak 7uidelines and suggestions for workingmore effectively with parents, re-evaluated thebook list, planned for colleagues visits, and re-fined the course syllabus for the teacher classes.

Inplementing a Reading Recovery Program 67

They also discussed their work in relation to the-oretical work by severe authors and explored theliterature related to metacognition.

University personnel also supervise thecourses by approving syllabi, visiting classes,and updating instruction when necessary. In ad-dition, university personnel serve a role in datacollection and management. Children's test scoresand other information are collected, sent to theuniversity, and analyzed Each year, the univer-sity prepares a state report which is provided to allsites and to the State Department of Education.

Getting Started

For a school system or a state agency, start-ing a ReaCng Recovery project will probablyinvolve a two-year process, including awareness,groundwork, decision making, Teacher Leadertraining, and teacher training. This process isoutlined in the five steps below.

1. Examine descriptive materials.School officials at either the local or the state

level who are interested in the possibility of ini-tiating a Reading Recovery project should firstgather written niaterials, such as descriptive ar-ticles and research reports, and distribute thosematerials (or a summary of them) to appropriatedecision makers in the district or state. A plan-ning group might be established to include thefolio .ving: teachers, principals, the superinten-dent, the Chapter One director of personnel, rep-resentatives of the state agency, and person lelfrom a local college or university.

2. Make a fact-finding visit.After an initial meeting, the decision makers

may gerwrate questions which could be answeredby nersonnc: at The Ohio State University or atlocal Ohio sites. If interest is strong, the planninggroup may wish to invite a knowledgeable person

to come and provide more information or to makea fact-Fading visit.

There is no substitute for first-hand viewingof teachers and children at work. Ohio is nowthe only place where the complete program maybe viewed. A visit will provide information about:1) the in-service courses for teachers and TeacherLeaders; 2) administration of the project; 3) ma-terials; 4) instruction of children; 5) problems thatmight be encounterzd; 6) funding sources; and 7)evaluation. The best use of time during a visit isto examine the Teacher Leader training, the teach-er training, and the implementation in schools.Teacher Leader training and an overall view of thestate program should be the focus of the visit toThe Ohio State University; for the regional sitevisit, an area should be selected that is similar insize and characteristics to the visitors' home dis-trict. A typical schedule for a fact-fitiing visitmight look like this:

Monday: The Ohio State UniversityA. M. Orientation

Observation of Clinical CourseLunch with Teacher Leaders in Training

P.M. Observation of Theoretical Seminar, orMeeting With School District OfficialsQuestions

Tuesday: Regional SiteA.M. Observation in Schools

Lunch with District OfficialsP.M. Observation of Teacher Class

Questions

After the fact-finding visit, the planningteam will have a good idea about whether theywant to implement a Reading Recovery programia their district.

3. Generate local support.If the planning team has been well selected,

support will already have a broad base. The plan-ning team can conduct a series of presentations tohelp convince key derision makers to make a

7

t8 Implementing a Reading Recovery Program

commitment to the program. This commitmentmust be long-term supporting initial imple-mentation, teacher training, and continual supportof the Teacher ader in his/her role.

4. Select Teacher Leade;To begin Reading Recovery, the local dis-

trict must take care in selecting the key imple-menter, the Teacher Leader. A person or personsshould be sent to The Ohio State University forone academic year (end of September to first ofJune) to participate in the Teacher Leader courseand related internship activities. For a large schoolsystem, two Teacher Leaders are recommended.

Teacher Leaders should have an M.A. de-gree, documented experience in good teaching atthe primary level, documented experience in lead-ership of other adults, and the quality that denotesleadership. Leaders must be flexible; they mustbe able to analyze situations and find solutions toproblems; they must be hard workPrs. The personto select for this important role is one whom no-body can possibly do ithout for a year. TeacherLeaders must be voicnteers and willing to come toOhio for an extended time. In the future, Leadertraining may be, available in other geographicareas, but at present only The Ohio State Univer-sity offers the year-long course.

5. Apply for information about TeacherLeader training.

Districts that wis.: to obtain more informa-tion about having a trained Teacher Lead- shouldcontact The Ohio State University, Reading Re-covery Program, 200 Ramseyer Hall, 29 WestWoodruff Avenue, Columbus, Ohio 43210. Ap-plication materials for sending a Teacher Leadertrainee to the University are available.

Local Implementation

The Teacher Leader course begins in the falland continues through spring of an academic

year. During this year, the district or state shouldconduct an awareness campaign to ensure thatindividuals who will be affected by implementingReading Recovery are knowledgeable and in-volved. The Leader in training can return to con-duct such sessions beginning in December of thetraining year. Frcm the beginning, the imple-mentation of the program should he a shared ven-ture. It is a good idea to bring key adminis.rators,including principals, to an existing site for a visitonducted by the Leader in training. There should

also be sessions for classroom teachers, parents,the school board, and prospective Reading Recovery teachers. Often, districts send key peopleto the Reading Recovery conference, held near theend of January each year.

Local implementation begins in August dur-ing the year after the Teacher Leader completestraining. In spring and summer before the train-ing begins, the Leader will order books and ma-terials for teachers and will supervise the con-struction of a one-way glass and facility for theteacher training class. Training begins with aworkshop on diagnostic procedures, to be held inAugust prior to the start of school. As a result ofthis workshop, Reading Recovery teachers intraining will be able to test and select studentsimmediately when school begins. The Leaderwill supervise the selection of students and willcollect necessary data to be sent to The Ohio StateUniversity.

The Teacher Leader will conduct trainingsessions and provide individual assistance toteachers in training throughout the academic year.In addition, the Leader will work with children,provide in-service for other district personnel, andestablish administrative arrangements necessaryfor conducting the program arid collecting data.Final testing of children will be accomplished inlate May, leaving as much instructional time aspossible during the school year. The Leader willalso return to The Ohio State University for a

Implementing a Reading Recovery Program , 69

summer institute each June for continued learn-ing.

When a local district implements the Read-ing Recovery program, it puts in place a provensystem of operation. Initial training of the Teach-er Leader takes time and resources; however, thedistrict gains a knowledgeable person who canperform a valuable role in improving programsfor children at risk of failure in reading. Asinterest in the program spreads, Teacher Leadertraining may be available in local areas on a drive-in basis, thus reducing the costs considerably.Within a geographic area of 100-150 miles,teachers could work with children in their owndistricts and complete Teacher Leader require-ments by driving to a university campus anddemonstration sites once or twice each week.

The National Diffusion Network

Adopting Reading Recovery has been madeeasier since the program's approval and fundingby the National Diffusion Network (NDN),which was established in 1974 as a vehicle forassisting educators in a search of innovative so-lutions to practical problems. Operating with theU.S. Department of Education, the Network'sbasic goal is to identify exemplary programs andmake them available to private and public schools,colleges, and other educational institutions. With-in each state, a state facilitator serves as a link be-tween local educational agencies and NDN-desig-nated exemplary programs.

In January, 1987, Reading Recovery wasinvited to submit a proposal to NDN to become aDeveloper Demonsu ator project. In addition to acomplete description of the program, the proposaladdressed: 1) specific claims and evidence tosupport each claim reported; 2) procedures, dataanalysis, and results for each claim; 3) the rela-tinnship between effect and treatn. nt; 4) controlfor rival hypotheses; and 5) significance of theresults compared with other programs designed to

treat similar populations. The U.S. Department ofEducation's Program Effectiveness Panel agreedthat the proposal provided convincing evidence ofthe effectiveness of the Reading Recovery pro-gram, and was therefore eligible to compete fordissemination funds from NDN.

A second proposal and Reading Recoverytraining materials were reviewed by the U.S. Of-fice of Education's Program Significance Panel.This panel was responsible for weighing theneed, content, program design, evidence of ef-fectiveness and educational significance. ReadingRecovery was one of six new programs thatreceived funds for dissemination in 1987-1988.Funding will continue for three mc.%ie. years.

Reading Recovery is an unusual project forthe sponsorship of NDN because it requires along and complex training program and is noteasily disseminated. Furthermore, no consumablematerials are available. The program depends onthe knowledge of the teaches; there are sometraining materials, but they do not stand alone.

In 1987-88, NDN funding assisted thedevelopment of two Reading Recovery sites out-side Ohio: Richardson, Texas, and Summerville,South Carolina. Reading Recovery will also ex-pan, .o Tucson, Arizona, and to Scarsborough,Ontario. Teacher Leaders in this group of dis-tricts constitute the Reading Recovery nationalnetwork of Teacher Leaders. Results of theseprojects are collated through a collaborative re-search effort involving all projects. These dataare available through The Ohio State Universityand NDN as evidence of program effectiveness.

The Reading Recovery project in Ohio hasbeen visited by representatives from 37 states.Currently in training in the 1988-1989 TeacherLeader class are representatives from Halifax,Nova Scotia; Louisville, Kentucky; Madison,Wisconsin; Carrollton, Texas; Plano, Texas;Mercer County, West Virginia; Coeur d'Alene,Idaho; Chicago, Illinoi.., University of Illinois(Champaign, Illinois); Portland State University

70 Implementing a Reading_ Recovery Program

(Portland, Oregon); and New York University(New York, New York). Each year, two sites areselected to receive modest NDN grants. Recipi-ents also have tuition waived by The Ohio StateUniversity. Other participants may also be desig-nated as officially approved NDN sites, even ifthey do not receive the actual grant assistance.Advantages to the school district coalition are theassistance of OSU staff in the first year of oper-ation and the communication network offered bythe Reading Recovery project.

Recognition by NDN is not only a recogni-tion of the program's effectiveness. It provides a% .o ensure high quality implementation inthose sites that are sponsored by NDN a.ld togather and disseminate other research results.Sites that do not publish data provide no evidenceof the program's success other than the opinionsof people involved. Although we believe thatperceptions of professional educators are impor-tant, we also are aware that perceptions can bealtered by those involver., Through the NDN datacollection network, participants in Reading Re-covery can provide their own monitoring of theirefforts.

Cost Effectiveness

In the implementation of any new program,costs must be considered. According to Levin,Glass, and Meister (19;;6) cost effectiveness an-alyb's must satisfy two requirements.

First, the educational interventions must bereadily implemented. That means that the inter-ventions must have; been applied in conventionasettings, established for a reasonable time, anchave characteristics that make them transferable toother settings. Reading Recovery has been im-plemented nationally in New Zealand, at four sitesin Australia, and at 23 sites in Ohio. In Ohio, theprogram has been implemented in hundreds ofschool districts across the state. The 5,408 chil-dren served from 1984-1988 were from popula-

tions of low-achieving first graders from a varietyof urban, suburban, and rural elementary schools.The program was so highly regarded that it wasrecommended by the National Diffusion Net-work. Numerous replications of the programprovide evidence that the program can be trans-ferred to new settings with minimum adaptations.

Second, the methods used to evaluate costsand effectiveness must be acceptable (Levin et al.,1986). Reading Recovery has been subject toongoing evaluation in New Zealand and for fouryears in Ohio. In addition to the evaluations al-ready described in this monograph, an outsideevaluation team, headed by Richard Anderson,Director of the Center for the Study of Reading,University of Illinois, critically examined quali-tative and quantitative data from the first fouryears of the Ohio operation. This team of outsideevaluators verified the evaluation results, affirmedthe potential of Reading Reco iery for helping at-risk children, and suggested further research(Anderson et al., 1988). These evaluations war-rant further investigation of Reading Recovery todetermine its cost effectiveness.

It will take a longer period of investigationto determine whether Reading Recovery providesan acceptable cost benefit. First, full coverage isneeded so that students who are highly mobilewill have a better chance of staying in the programlong enough to be successfully discontinued.Second, an adequate sample of students must befollowed over a number of years to determinelong-term benefits. In the final analysis, it is al-ways difficult to determine benefits which arisefrom preventive programs because we can onlyspeculate about costs that might be saved in spe-cial education assignments, dropout problems, ordisciplinary problems. Although the project is inthe development phase, we can estimr.te cost ben-efits by considering savings related to length ofservice and retention. In most remedial programs,children are served on a continuing basis; in

Implementing a Reading Recovery Program 7 1

Reading Recovery, they are discontinue afterabout one-half of the academic year.

In deciding to implement the program, manyschool districts have projected tneir own cost-benefit estimates. A superintendent of a suburbanOhio school district serving approximately 5,000elementary students reported that by February ofthe second year of implementation, the programhad paid for itself in savings related to reductionof services and retention. A rural school districtin Ohio projected in January that 95 percent of thechildren selected for Reading Recovery would beretained. In June, only 10 percent were actuallyretained. Even if more of the children are retainedin subsequent grades, the savings will be sub-stantial enough to pay for the program.

Obviously, these projections need to be fol-lowed up with statewide studies of further serviceand retention. During 1988-1989, The Ohio StateUniversity will conduct surveys about childrenserved during the first and second years of state-wide implementation. In addition, a NationalEvaluation Panel will be conducting a cost benefitanalysis on a large sample of children.

Until these data are available, we encourageschool districts to analyze what they now spend tohelp at-risk children and what the results are interms of long-term help for these children. Ef-fective implementation of Reding Recovery doesrequire a major investment in tams of teachertraining, and the instructional program itself de-pends on labor-intensive one-to-one tutoring.However, Reading Recovery also offers distinct

cost advantages. It is a short-term, intensive pro-gram which can reduce children's '.;ed for ex-pens; ve remediation programs ove. any years ofschooling. In addition, Reading Recovery doesnot depend on custly consumable materials, asmany commercial remedial programs do. Con-sidering all these factors, each school district mustproject what it might be able to invest in ReadingRecovery.

Making a Difference

Changing the educational prospects for at-risk children will require enormous resources inthe coming years. That investment of resources isnecessary because we must increase the educa-tional level, quality of life, and productivity of anincreasing proportion of children at risk of schoolfailure. We have already invested much in reme-dial programs. We know that many of them donot make the fundamental changes that areneeded.

Reading Recovery has demonstrated its po-tential for improving both the reading success ofindividual children and the productivity of educa-tional systems. The program is designed for suc-cessful long-term implementation. The factors thatmake it di:ficult to initiate a Reading Rec weryprogram are the same factors that ultimately makeit easy to sustain and improve the program, there-by making a profound difference in the learningand lives of children.

70

BibliographyAnderson, R.A., Hiebert, E.H., Scoit, J.A., and Wilkinson, with contributions from members of

the Commission on Reading (1985). Becoming a Nation of Readers: The Report of theCommission on Reading. Washington, DC: The National Institute of Education.

Clay, M.M. (1979-a, 1988). The Early Detection of Reading Difficulties. Auckland, NewZealand: Heinemann Educational Books.

Clay, M.M. (1979-b). Reading: The Patterning of Complex Behavior. Auckland, New Zealand:Heinemann Educational Books.

Clay, M.M. (1982). "Reading Recovery: A Follow-up Study." In Observit g Young Readers:Selected Papers, by Marie M. Clay. Exeter, NH: Heinemann Educational Books.

Clay, M.M. (in press). Implementation of Reading Recovery in Three Countries.

DeFord, D.E. (1985). "Validating the Construct of Theoretical Orientation in Reading Instruc-tion." Reading Research Quarterly. Volume XX, No. 3, pp. 351-367.

DeFord, D.E. (1987). "Reading Strategy Development Growing Through New and FamiliarMaterials." Proceedings cf the National Reading Conference. Austin, Texas.

DeFord, D.E. (1987). "Teachers Learning Through Teaching and Peer Interaction." Proceedingsof the National Reading Conference. St. Petersburg lorida.

DeFord, D.E., Pinnell, G.S., Lyons, C.A., and Young, P. (1987). Report of the Follow-UpStudies Columbus, Ohio 'Reading Recovery Project 1985-1986, 1986-1987.(Technical Report). Columbus, OH: The Ohio State University.

DeFord, D.E., Pinnell, G.S., Lyons, C.A., and Young, P. (1988). Report of the Follow-UpStudy, Reading Recovery Project, Vol. IX. Columbus, OH: The Ohio State University.

Geeke, P. (1988). Evaluation Report on the Reading Recovery Field Trial in Centrai Victoria,1984. Australia: Centre for Studies in Literacy, University of Wollongong.

Goodman, Wa.son, and Burke (1987). Reading Miscue Inventory, 2nd Edition. New York:Richard C. Owen.

Holland, K. (1987). The Impact of Reading Recovery Program on Parents and Home LiteracyContexts. Unpublished doctoral dissertation, The Ohio State University.

Huck, D. and Pinnell, G. (1985). The Reading Recovery Project in Columbas, Ohio: F ;lot Year1984-t 985. (Technical Report). Columbus: The Ohio State University.

Levin, H.M., Glass, G.V., and Meister, G. (1986). "The Political Arithmetic of Cost-Effective-ness Analysis." Phi Delta Kappan V. 68 6 no. 1 (September 1986), pp. 69-72.Bloomington, IN: Phi Delta Kappa.

Lyons, C.A., Pinnell, G., Short, K., and Young, P. (1986). The Ohio Reading RecoveryProject: Volume IV Pilot Year. 1985-1986. (Technical Report). Columbus: The OhioState University.

73 7; -, 1

74

Lyons, C., Pinnell, G., Young P., and De Ford, D.E. (1987). Report of the Ohio ReadingRecovery Project, State c. Ohio Year 1 1986-1987. Volume VIII (Technical Report).Columbus: The Ohio State University.

Lyons, C.A., Pinnell, G.S., McCarrier, A, Young, P., DeFord, D. (1988). The Ohio ReadingRecovery Project: Volume X and State of Ohio Year 1987-88. Columbus, OH: The OhioState University.

Lyons, C.A. (1987). "Helcing Readers Make Accelerated Progress: Teacher Responses andKnowledge." Proceedings of the National Reading Corif4.-ence. St. Petersburg, Flofda.

Lyons, C.A. (1987). Reading Recovery: An Effective Intervention Program for LearningDisabled First Graders. Paper presented at the annual meeting of the American EducationResearch Association, Washington, DC.

Lyons, C. (1987). Rtport to the National Diffnsion Network.

Peterson, B. (1987). The Characteristics of Text Which Support Beginning Readers. Unpub-lished doctoral dissertation. The Ohio State Universtiy, Columbus.

Pinnell, G., Short, K., Lyons, C.A., and Young, P. (1986). The ReaLtng Recovery Project inC alumbur Ohio: Volume III. Year One, 1985-1986. (Technical Report). Columbus: TheOhio State University.

Pinnell, G., Short, K., Lyons, C.A., and Young, P. (1986). The Ohio Reading RecoveryProject: Follow-up Study in Columbus, Ohio: Volume II. 1985-1986. (Technical Report).Columbus: The Ohio State University.

Pinnell, G.S. (1967). "A Two-Year Study of Students Involved in the Ohio Reading RecoveryProject" Proceedings of the National li%:ading Conference. Aur:in, Texas.

Pinnell, G.S. (1987). "Helping Teachers Sce How Readers Read: Staff Development ThroughObservation." Theory Into Practice, 26 (1).

Pinnell, G.S. (i986). 'Helping Teachers Help Children at Risk: Insights from the ReadingRecovery Program." Peabody Journal, 62 (3), 70-85.

Pinnell, G.S., and Green, J. (April, 19E6). "Learning from Language Research to Talk AboutEducation." Language Arts, 63, 385-389.

Pinnell, G.S. and Woolsey, D. (1985). A Study of Teacher Change During a Research-OrientedProject. Report to the NCI E Reseal :11 Foundation.

Rent &, V.W. and Pinnell, G.S. (1987). "Study of Teachcr Reasoning Processed." Proceedingsof National Reading Confer we. St. Petersburg, Florida.

Short, K.G. (1987). Reading Recovery: Developing Strategies for Reading. Paper presented atthe annual meeting of the American Education .esearch Association, Washington, DC.

Woolsey, D.P. First Grade Children's Responses to Teacher Change in Literacy Contexts.Unpublished doctoral dissertation, The Ohio State University.

Young, P. and Pinneh, G.S. (1987). Reading Recovery: A Program Evaluation. Paper presentedat the annual meeting of the American Education Research Association, Washingotn, DC.

Appendix ASome Small Books Us :d in Reading Recovery

List prepares oy: Barbara Peterson

Children read four or more books during every Reading Recovery lesson. First, they readsevt,ial familiar books to build fluency and confidence. Each day a new book is introduced, pro-viding opportunity to use reading strategies on unfamiliar and more challenging texts. ReadingRecovery teachers carefully select each new book according to the strengths and needs of the child.

During their year-long inservice program, teachers learn how to select and introduce booksappropriate for each child. Books used in Reac ling Recovery are selected from a number ofsources. No single series can adequatelar serve the program.

Reading Recovery teachers select books from a list of hvndreds of titles. This list is revisedfrequently as new books are published and others are found inadequate. Consequently, booklistsare considered training materials and are not made available.

Here are EXAMPLES of the kinds of books we use in Reading Recovery. Books whichprovide support for beginning readers have familiar language patterns within the framework of apredictable story. Books for more independent readers challenge them in many ways, always inthe context of a complete story or message.

The purpose of this list is to highlight some of the small books available on a Reading Re-covery booklist. Most could be obtained through bookstores and school or public libraries. At theend of the book list, we have inserted a list of sources.

Level 1

Jonas, Ann NOW WE CAN GO GreenwillowMaris, Ron MY BOOK PuffinTafuri, Nancy HAVE YOU SEEN MY DUCKLING? PuffinWildsmith, Brian CAT ON THE MAT Oxford

Level 2Ziefert, Harriet WHERE IS MY DWNER? Grosset & DunlapZeifert, Harriet WHERE IS MY FRIEND? Grosset & Dunlap

Level 3Hutchins, Pat 1 HUNTER MorrowWildsmith, Brian ALL FALL DOWN OxfordWildsmith, Brian TOOT, TOOT Oxford

Martin, Bill

Peek, Merie

Level 4BROWN BEAR, BROWN BEAR Holt, Rinehart,

and WinstonROLL OVER Clarion

757E

76 Append:. 71 A

Peppe, RodneyStobbs, William

Burningham, JohnGinsburg, MirraLindgren, BarbroLindgren, BarbroLindgren, BarbroLind,fen, BarbroLindgren, BarbroPeek, Merle

Shaw, Charles

Burningham, JohnCampbell, RodHill, EricKraus, RobertLangstaff, JohnRoffey, Maureen

Asch, FrankCampbell, RodHutchins, PatLloyd, DavidMaris, RonMaris, RonStobbs, WilliamWest, ColinWest, Colin

Brown, RuthDe Regniers, BeatriceGerstein, MordecaiRockwell, AnneRockwell, HarlowStadler, JohnWard, CindyWatanabe, ShigeoWheeler, CindyWheeler, CindyWheeler, Cindy

Level 5

HUMPTY DUMMYONE, TWO, BUCKLE MY SHOW

Level 6

THE SCHOOLTHE CHICK AND THE DUCKLINGSAMS BALLSAM'S COOKIESAM'S LAMPSAM'S TEDDY BEARSAM'S WAGONMARY WORE HER RED DRESS

Level 7

IT LOOKED LIKE SPILT MILK

Level 8

THE BLANKETHENRY'S BUSY DAYWHERE'S SPOT?HERMAN THE HELPEROH, A-HUNTING WE WILL GOHOME SWEET HOME

Level 9

JUST LIKE DADDYDEAR ZOOROSIE'S WALKGRANDMA AND THE PIRATEARE YOU THERE, BEAR?IS ANYONE HOME?GREGORY'S GARDENHAVE YOU SEEN THE CROCODILE?"PARDON ?" SAID THE GT.RAI.FE

Level 10

A DARK DARK TALEGOING FOR A WALKWILLIAM, WHERE ARE YOU?CARSMY KITCHENHOORAY FOR SNAIL!COOKIE'S WEEKI'M THE KING OF THE CASTLE!MARMALADE'S NAPMARMALADE'S SNOWY DAYROSE

SO

VikingBodley

CrowellMacmillanMorrowMorrowMorrowMorrowMorrowClarion

Harper & Row

CrowellVikingPutnamWindmillAtheneumBodley

Prentice-HallFour WindsMacmillanCrownGreen willowGreenwillowOxfordHarper & RowHarper & Row

DialHarper & RowCrownCuttonGreenwillowHarper & RowPubnamPhilomelKnopfKnopfKnopf

Appendix A 77

Ahlverg, Janel and AllanKraus, RobertMack, StanRockwell, AnneStadler, john

Barton, ByronBurningham, JohnBurningham, JohnBurningham, JohnBurningham, JohnBurningham, JohnBumingham, JohnGinsburg, MirraHutchins, PatKrauss, RuthLong, ErleneShulevitz, UriStadler, JohnTaylor, JudyWatson, WendyWescott, Nadine

Alexander, MarthaJonas, AnnJonas, AnnKovakski, MaryannRockwell, AnneRockwell, Anne and HarlowTolstoy, Alexei

Adams, Pan

Baron, ByronBrown, Margaret WiseCarle, EricHutchins, Pat

Kraus, Rovert

Kraus, Rovert

Robart, RoseTaylor, Judy

Level 11

EACH PEACH PEAR PLUMWHOSE MOUSE ARE YOU?10 BEARS IN MY BEDBOATSSNAIL SAVES THE DAY

Level 12

BUZZ BUZZ BUZZTHE BABYTHE CUPBOARDTHE DOGTH. FRIENDTHE RABBITTHE SNOWTHREE KITTENSTITCHTHE CARROT SEEDGONE FISHINGONE MONDAY MORNINGTHREE CHEERS FOR HIPPOMY DOGLOLLIPOPPEANUT BUTTER AND JELLY

Level 13

BLACKBOARD BEARTWO BEAR CUBSWHEN YOU WERE A BABYTHE WHEELS ON THE BUSTHE AWFUL MESSTHE TOOL BOXTHE GREAT BIG ENORMOUS TURNIP

Level 14

THERE WAS AN OLD LADYWHO SWALLOWED A FLY

BUILDING A HOUSEGOODNIGHT MOONTHE VERY BUSY SPIDERYOU'LL SOON GROW

INTO THEM, TITCHCOME OUT AND PLAY,

LITTLE MOUSEWHERE ARE YOU GOING,

LITTLE MOUSE?THE CAKE THAT MACK ATEMY CAT

81

VikingMacmillanPantheonDuttonHarps. & Row

PuffinCrowellCrowellCrowellCrowellCrowellCrowellCrownMacmillanScholasticHoughtonScribner'sCrowellMacmillanPuffinDutton

DialGreenwillowPuffinLittle, BrownFour WindsCollierPan Picolo

Scholastic

PuffinHarper & RowPhilomelGreenwillow

Greenwillow

Greenwillow

Little BrownMacmillan

78 Appendix A

Fox, MemGuilfoile, ElizabethKline, SuzyMcPhail, DavidNodset, JoanWood, Audrey

Alexander, Martha

Bennett, JillBonsall, CrosbyHutchins, PatHutchins, PatKent, JackKuskin, KarlaKraus, RovertMayer, Mercer

McLeod, EmilieMinarick, ElseRuddell, ChrisSeuling, BarbaraWells, Rosemary

Bornstein, RuthGaldone, PaulGalsone, PaulHurd, Edith ThacherHutchins, PatJohnson, CrockettLobel, ArnoldLobel, ArnoldMayer, Mercer

Mayer, MercerNicol, HelenPeppe, RodneyUdry, Janice MayVipont, Elfrida

Carle, EricDabcovicn, Lydia

Ember ley, EdKrasilovsky, Phyllis

Lionni, Leo

Level 15

HA I nr. AND THE FOXNOBODY LISTENS TO ANDREWDONT TOUCH!FIX-ITWHO TOOK THE FARMER'S HATTHE NAPPING HOUSE

Level 16

WE'RE IN BIG TROUBLE,BLACKBOARD BEAR

TEENY TINYAND I MEAN IT STANLEYGOODNIGHT OWLHAPPY BIRTHDAY SAMTHE FAT CATJUST LIKE EVERYONE ELSELEO THE LATE BLOOMERTHERE'S A NIGHTMARE

IN MY CLOSETTHE BEAR'S BICYCLEA KISS FOR LITTLE BEARBEN AND THE BEARTHE TEENY TINY WOMANNOISY NORA

Level 17

LITTLE GORILLATHE LITTLE RED HENTHE THREE BEARSJOHNNY LION'S BOOKTHE DOORBELL RANGHAROLD AND THE PURPLE CRAYONMOUSE SOUPMOUSE '''ALESTHERE'S AN ALLIGATOR

UNDER MY BEDTHERE'S SOMETHING IN MY ATTICMEG AND NIOGTHE HOUSE THAT JACK BUILTLET'S BE ENEMIESTHE ELEPHANT AND THE BAD BABY

Level 18

THE VERY HUNGRY CATERPILLARMRS. HUGGINS AND

HER HEN HANNAHDRUMMER HOFFTHE MAN WHO DIDN'T

DO HIS DISHESLITTLE BLUE AND LITTLE YELLOW

BradburyModern CurriculumPuffinDuttonScholasticHBJ

Dial

PutnamHarperPuffinPuffinPuffinHarperWindmillDial

PuffinHarperHarper & RowPuffinDial

ClarionScholasticScholasticHarperGreenwillowHarperHarperHarperDial

DialPuffinDelacorteScholasticCoward

PuffinDutton

Prentice-HallDoubleday

Astor

Aunendix A 79

Lo xl, ArnoldMir ailc, ElseSendak, Maurice

Burningham, JohnBurn;ngham, JohnHe :chins, PatLobel, ArnoldLobel, t moldMurphy, JillRice, Eve

Allen, PamCrowe, R.Galdone, PaulHutchins, PatLobel, amidSendak, MauriceSlobodkina, EsphyrZoleto' , Charlotte

OWL AT HOMELITTLE BEARWHERE . HE WILD THIr IS ARE

Level 19

MR. GUMBY'S MOTORCARMR. GUMBY'S OUTINGTHE SURPRISE PARTYFROG AND TOAD ARE FRIENDSFROG r.ND TOAD TOGETHERWHt T NEXT BABY BEA.1?SAM . '1-10 NEVER FORGETS

Level 20

WHO SANK filE BOAT?TYLER TOAD AND THUNDERTilE THREE LITTLE PIGSThE WIND BLEWUNCLE ELEPHANTCHICKEN SOUP WITH RICECAPS FOR SALEI KNOW A LADY

SOURCES

Trade books rre available through bookstores orare given below. Some books may be out of 1,,i

Cypress Publishing Corporation1753 Gardena AvenueGlendale, CA 91204213/244-86j1

DLM Teaching ResourcesP.O. Box 4000One DLM ParkAllen, TX 75002800/527-4747800/442-4711 (in Texas)

Longman Inc.95 Church StreetWhite Plains, NY 10601-1505914/993-000

HarperHarperHarper

PuffinPuffinPuffinHarperHarperDialPuffin

CowardDutton3cholasticPuffinHarperScholasticHarper & RowPuffin

book jobbers. Addresses of text book publishersnt.

Richard C. Owen PublishersRockefeller CenterBox 819New York, NY 10185212/864-784;`

Rigby Educatioil454 S. Virginia StreetCrystal Lake, IL 60014815/455-7220

Scott Foresman and Co.1900 East Lake AvenueGlenview, IL 60C25312/729-3000

The Wright Group10949 Technology PlaceP.O. Bc t 27780San Diego, CA 92127800/523-2371

R3

Appendix B Description of Alternative .aterventionProgram, Columbus Longitudinal Study 1985-86

For the program evaluation study of the first full year of Reading Recovery implemen-tation, desct bed in Chapter 3, low-achieving suojects from the same schools we*: randomlyassigned either to Reading Recovery or to an alternative compensatory program. Subjects parti-cipated in one program or ene other, not both.

I-. both prog-ams, children received daily instruction. In Reading Recovery, the instructionwas individual for one half how daily. In the alternative corr.pensntory program, the instructiontook I:ace in small groups of two to four children although some instruction was individualand varied in time from approximately 30 to 45 minutes. (The latter description is taken frominformal observations by U.S.U. researchers; policy documents do not specify the amount of timespent with students except that 3.5 hours were spent in the classroom.)

The alternative compensatory program grew out of concern that previous use of aides hadnot demonstrated significant impact on student achievement. The purpose of this program was totrain aides to work efficiently and effectively, using deveoped methods and materials, witb nder-achieving students in the area of reading and language a:.-ts. The program was specifically designeeto provide direct instructional assistance to children rather than to assist teachers in non-instruc-tional tasks. Under *1-e program, paraprofessionals received extensive staffdevelopment. Theaides were given materials and were trained to implement a sper;fic program that reinforced skillsin ways different from those used in the regular classroom leading groups. The duties of hara-professionals were very clearly described, and supervision was provided *o make sure they ful-%lled their specific responsibilities rather than tasks not directly related to instruction of children.Paraprofessionals performed tasks such as: taking dictation from students; reading and tellingstories; reviewing and reinforcing lessons and skills taught by the teacher; conducting drills onpractice work; and assisting students with written work and manuscript writing.

Becaus of prior agreements with th, schocl district, the alternative compensatory programhas not bee:, described or identified by name in the Pleading Recovery research reports. Neitherthe school district nor the research team wanted to cast doubt on the effectiveness of a highly val-ued program.

Interesting follow-up information is available concerning the aide program. The programdirector and his associates studied Re:, 'ing Recovery very carefully in the year following thestudy, and a supervisor received training. Based on insights gained from this involvement, theprogram implementers redesigned and refinci4 the assessment and instructional approaches used bythis p of paraprofessionals. They are not doing Reading Recovery instruction, but they havecreated a new and dynamic paraprofessional program based on similar principles. The program isbeing tested now.

Appendix CDependent vleasures Used to AssessChildren in Reading Recovery Program

1) Text Reading Level: Measures of text reading level were obtained by constructing a gra-dient of Jifficulty of text, then testing for the highest level zead with accurac:, of 90 percent orbetter. Por the pilot year, levels were drawn from a basal reading system. When the stateproject began, a different series of text reading items was drawn in the same way from a set ofbooks not currently in use in any school district in Ohio. Levels of oifficulty were indicatedby the textbook publishing company; these books contained language similar to that in typicalreading texts. None of the text items used for testing were used in Reading Recoveryinstruction.

The first five levels were not drawn from basal texts but were specified easy books. For leveltwo, for example, the teacher read Where's Spot2 (written by Erie Hill and publishes by G. P.Putnum's Sons, 1980) to the child and on one page asked the child to point to the pattern ofwords that had been read on several previous pages. A child unable to respond to print in thisvery simple way was designated as "level one," and a child able to point to the weds and readthem was designated as having passed "level two." There were 30 levels .1 all, the highest:evel indicating approximately eighth-grade reading ability according to the basal series used.These materials were not used in Reading Recovery instruction but represented the kind oftexts chilci:en would be expected to perform on if they were to survive in the regular class-rooms without extra remedial help.

Individually, children were asked to read stories while the tester recorded a running record ofre?ding behavior and calculated an accuracy level. Children continued reading at higher levelsuntil they reached two levels below 90 percent accuracy. The score on text level is the highestlevel read with above 90 percent accuracy.

2) Let:er Identification. Children were asked to identify 54 different characters, includingupper-case and lower-case letters and conventiona, print for "a" and "g."

3) Word Test: Children were asked to read down a list of 15 words drawn from the mostfrequent words from the pre-primers in use in the district. A different list was used at eachadministration.

4) Concepts About Print: Children were asked to perform a variety of tasks during a bookreading. The tasks represented a standard situation to check on significant concepts aboutprinted language, such as directionality and concept of word. Two forms of the test wereused.

(continued)

84 Appendix C

5) Writing Vocabulary: Within a 10-minute peiiod, children were asked to write all thewords they knew. The Writing Vocabulary score was the number of words spelled accur-ately.

6) Dicvition Test: Testers read a sentence to the children who werc to write the words, indi-catirg their ability to analyze the word for sounds. A different sentence was use at eachadministration.

7) Standardized Tests of Reading:

a) The Stanford Achievement Test. The following sections were administered in a groupsetting: 1) Word Reading; 2) Reading Comprehension; 3) Word Study Skills. Scores forReading Subtotal and Total Reading were calculated. The Stanford Achievement test wasused only for :esearch children in the Pilot Cohort, 19M-1985.

b) The Metropolitan Achievement Test: Sad? Edition (MAT 6) Reading Diagnostic Tests.The MAT 6 Reading Diagnostic Test is a standardized test which consists of 11 subtestsEubtests of the MAT 6 Reading Diagnostic Test were administered to Reading Recoverychildren both in the fall and the spring of the academic year. In September, the PrimerForm was administered. Subtests included: 1) Sight Voc. bulary; 2) Phoneme/ Gra-pheme: Consonants; and 3) Reading Comprehension. In May, Primary 1 For,n L wasadministered to Reading Recovery children and a group of random sample, children.Subtests inch d d: 1) Sight Vocabulary; 2) Phoneme/Grapheme: Consonants; 3) Phu-neme/Grapheme: Vowels; and 4) Reading Comprehension. Raw scores were convertedto percentile ranks and Normal Curve Equivalent Scores. In the 1985-86 year, the MAT6 test was used for all Reading Recovery .ites eAcept Columbus.

c) The Comprehensive Test of Basic Skills was administered to Columbus subjects as partof the required evaluation of compensatory programs. Two sabtests, 1) Reading Voca-bulary and 2) Reading Comprehension were used for th.:. 1985-1986 study.

_1

E4711,1"'

Documents

DOCUMENT RESUME ED 303 790 AUTHOR Pinnell, Gay Su; And … · 2013. 8. 2. · DOCUMENT RESUME. CS 009 535. Pinnell, Gay Su; And Others Reading Recovery: Early Intervention for At-Risk