DOCUMENT RESUME

ED 405 554 CS 012 742

AUTHOR Stahl, Steven A.; And Others TITLE Fluency-Oriented Reading Instruction. Reading Research Report No. 79. INSTITUTION National Reading Research Center, Athens, GA.; National Reading Research Center, College Park, MD. SPONS AGENCY Office of Educational Research and Improvement (ED), Washington, DC. PUB DATE 97 CONTRACT 117A20007 NOTE 55p.; For an earlier report, see ED 379 610. PUB TYPE Reports Research/Technical (143)

EDRS PRICE MF01/PC03 Plus Postage. DESCRIPTORS *Basal Reading; Grade 2; *Instructional Effectiveness; Primary Education; *Reading Ability; *Reading Improvement; Reading Material Selection; Reading Research; Student Attitudes; Teacher Attitudes; *Word Recognition IDENTIFIERS Partner Reading; Reading Fluency; Repeated Readings

ABSTRACT A 2-year project was designed to reorganize basal reading instruction as to stress fluent reading and automatic word recognition. The reorganized reading program had three components: a redesigned basal reading lesson, stressing repeated reading and partner reading; a choice reading period during the day; and a home reading program. Over the 2 years of the program, students made significantly greater than expected growth in reading ability in all 14 classes. All but 2 children who entered second grade reading at primer level or higher (and half of those who did not) were reading at grade level or higher by the end of the year. Growth in fluency and accuracy appeared to be consistent over the whole year. Students' and teachers' attitudes toward the program were positive. In evaluating individual components, results indicated that self-selected partnerings seemed to work best and that children chose partners primarily out of friendship. Children tended to choose books that were about or slightly below their instructional level. In addition, children seemed to benefit instructionally from more difficult materials, with the greater amount of scaffolding provided in this program. (Contains 49 references, and 4 tables and 6 figures of data. An appendix presents interview questions.) (Author/RS)

*********************************************************************** Reproductions supplied by EDRS are the. best that can be made from the original document. *********************************************************************** Fluency-Oriented Reading Instruction

Steven A. Stahl University of Georgia

Kathleen Heubach Virginia Commonwealth University

Bonnie Cramond University of Georgia

U.S. DEPARTMENT OF EDUCATION Office of Educational Research and Improvement ED CATIONAL RESOURCES INFORMATION CENTER (ERIC) This document has been reproduced as received from the person or organization originating it. 0 Minor changes have been made to improve reproduction quality.

Points of view or opinions stated in this document do not necessarily represent official OERI position or policy.

National Reading Research Center

READING RESEARCH REPORT NO. 79 Winter 1997 2 BEST COPYAVAILABLE NRRC National Reading Research Center

Fluency-Oriented Reading Instruction

Steven A. Stahl University of Georgia

Kathleen Heubach Virginia Commonwealth University

Bonnie Cramond University of Georgia

READING RESEARCH REPORT NO. 79 Winter 1997

The work reported herein is a National Reading Research CenterProject of the University of Georgia and University of Maryland. It was supported under theEducational Research and Development Centers Program (PR/AWARD NO. 117A20007) asadministered by the Office of Educational Research and Improvement, U.S. Department ofEducation. The findings and opinions expressed here do not necessarily reflect the position orpolicies of the National Reading Research Center, the Office of Educational Research andImprovement, or the U.S. Department of Education.

3 NRRC National Reading Research Center

Executive Committee National Advisory Board Donna E. Alvermann, Co-Director Phyllis W. Aldrich University of Georgia Saratoga Warren Board of Cooperative Educational John T. Guthrie, Co-Director Services, Saratoga Springs, New York University of Maryland College Park Arthur N. Applebee James F. Baumann, Associate Director State University of New York, Albany University of Georgia Ronald S. Brandt Patricia S. Koskinen, Associate Director University of Maryland College Park Association for Supervision and Curriculum Jamie Lynn Metsala, Associate Director Development University of Maryland College Park Marsha T. DeLain Penny Oldfather Delaware Department of Public Instruction University of Georgia Carl A. Grant John F. O'Flahavan University of Wisconsin-Madison University of Maryland College Park Barbara McCombs James V. Hoffman Mid-Continent Regional Educational Laboratory (MCREL) University of Texas at Austin Luis C. Moll Cynthia R. Hynd University of Arizona University of Georgia Carol M. Santa Robert Serpell School District No. 5 University of Maryland Baltimore County Kalispell, Montana Betty Shockley-Bisplinghoff Anne P. Sweet Clarke County School District, Athens, Georgia Office of Educational Research and Improvement, Linda DeGroff U.S. Department of Education University of Georgia Louise Cherry Wilkinson Rutgers University Publications Editors Peter Winograd University of Kentucky Research Reports and Perspectives Production Editor Linda DeGroff, Editor Katherine P. Hutchison University of Georgia University of Georgia James V. Hoffman, Associate Editor University of Texas at Austin Assistant Production Editor Mariam Jean Dreher, Associate Editor Jennifer Moon University of Maryland College Park University of Georgia Instructional Resources Dissemination Coordinator Lee Galda, University of Georgia Jordana E. Rich Research Highlights University of Georgia William G. Holliday University of Maryland College Park Text Formatter Policy Briefs Angela R. Wilson James V. Hoffman University of Georgia University of Texas at Austin NRRC - University of Georgia Videos 318 Aderhold Shawn M. Glynn, University of Georgia University of Georgia Athens, Georgia 30602-7125 NRRC Staff (706) 542-3674 Fax: (706) 542-3678 Barbara F. Howard, Office Manager INTERNET: [email protected] Kathy B. Davis, Senior Secretary NRRC - University of Maryland College Park University of Georgia 3216 J. M. Patterson Building University of Maryland Barbara A. Neitzey, Administrative Assistant College Park, Maryland 20742 Valerie Tyra, Accountant (301) 405-8035 Fax: (301) 314-9625 University of Maryland College Park INTERNET: [email protected] About the National Reading Research Center

The National Reading Research Center (NRRC) is Dissemination is an important feature of NRRC funded by the Office of Educational Research and activities. Information on NRRC research appears in Improvement of the U.S. Department of Education to several formats. Research Reports communicate the conduct research on reading and reading instruction. results of original research or synthesize the findings of The NRRC is operated by a consortium of the Univer- several lines of inquiry. They are written primarily for sity of Georgia and the University of Maryland College researchers studying various areas of reading and Park in collaboration with researchers at several institu- reading instruction. The Perspective Series presents a tions nationwide. wide range of publications, from calls for research and The NRRC's mission is to discover and document commentary on research and practice to first-person those conditions in homes, schools, and communities accounts of experiencesinschools.Instructional that encourage children to become skilled, enthusiastic, Resources include curriculum materials, instructional lifelong readers. NRRC researchers are committed to guides, and materials for professional growth, designed advancing the development of instructional programs primarily for teachers. sensitive to the cognitive, sociocultural, and motiva- For more information about the NRRC's research tional factors that affect children's success in reading. projects and other activities, or to have your name NRRC researchers from a variety of disciplines conduct added to the mailing list, please contact: studies with teachers and students from widely diverse cultural and socioeconomic backgrounds in pre-kinder- Donna E. Alvermann, Co-Director garten through grade 12 classrooms. Research projects National Reading Research Center deal with the influence of family and family-school 318 Aderhold Hall interactions on the development of literacy; the interac- University of Georgia tion of sociocultural factors and motivation to read; the Athens, GA 30602-7125 impact of literature-based reading programs on reading (706) 542-3674 achievement; the effects of reading strategies instruction on comprehension and critical thinking in literature, John T. Guthrie, Co-Director science, and history; the influence of innovative group National Reading Research Center participation structures on motivation and learning; the 3216 J. M. Patterson Building potential of computer technology to enhance literacy; University of Maryland and the development of methods and standards for College Park, MD 20742 alternative literacy assessments. (301) 405-8035 The NRRC is further committed to the participation of teachers as full partners in its research. A better understanding of how teachers view the development of literacy, how they use knowledge from research, and how they approach change in the classroom is crucial to improving instruction. To further this understanding, the NRRC conducts school-based research in which teachers explore their own philosophical and pedagogi- cal orientations and trace their professional growth. NRRCEditorial Review Board

Peter Afflerbach Suzanne Clewell Mary Graham University of Maryland College Park Montgomery County Public Schools McLean, Virginia Rockville, Maryland Jane Agee Rachel Grant University of Georgia Joan Coley University of Maryland College Park Western Maryland College JoBeth Allen Barbara Guzzetti University of Georgia Michelle Commeyras Arizona State University University of Georgia Janice F. Almasi Frances Hancock Linda Cooper University of Buffalo-SUIVY Concordia College of Saint Paul, Shaker Heights City Schools Minnesota Shaker Heights, Ohio Patty Anders Kathleen Heubach University of Arizona Karen Costello Virginia Commonwealth University Connecticut Department of Education Harriette Arrington Hartford, Connecticut Sally Hudson-Ross University of Kentucky University of Georgia. Jim Cunningham Mania Banning Gibsonville, North Carolina Cynthia Hynd University of Utah University of Georgia Karin Dahl Jill Bartoli Ohio State University Gay Ivey Elizabethtown College University of Georgia Marcia Delany Eurydice Bauer Wilkes County Public Schools David Jardine University of Georgia Washington, Georgia University of Calgary

Janet Benton Lynne Diaz-Rico Robert Jimenez Bowling Green, Kentucky California State University-San University of Oregon Bernardino Irene Blum Michelle Kelly Pine Springs Elementary School Mark Dressman University of Utah Falls Church, Virginia New Mexico State University James King David Bloome Ann Duffy University of South Florida Vanderbilt University University of Georgia Kate Kirby John Borkowski Ann Egan-Robertson Georgia State University Notre Dame University Amherst College Linda Labbo Fenice Boyd Jim Flood University of Georgia University of Georgia San Diego State University Michael Law Karen Bromley Dana Fox University of Georgia Binghamton University University of Arizona Donald T. Leu Martha Carr Linda Gambrell Syracuse University University of Georgia University of Maryland College Park Susan Lytle Barbara M. Pahner Bernard Spodek University of Pennsylvania Mount Saint Mary's College University of Illinois

Bert Mangino Stephen Phelps Bettie St. Pierre Las Vegas, Nevada Buffalo State College University of Georgia

Susan Mazzoni Mike Pickle Steve Stahl Baltimore, Maryland Georgia Southern University University of Georgia

Ann Dacey McCann Amber T. Prince Roger Stewart University of Maryland College Park Berry College Boise State University

Sarah McCarthey Gaoyin Qian Anne P. Sweet University of Texas at Austin Lehman College-CUNY Office of Educational Research and Improvement Veda McClain Tom Reeves University of Georgia University of Georgia Louise Tomlinson University of Georgia Lisa McFalls Lenore Ringler University of Georgia New York University Bruce VanSledright University of Maryland College Park Randy McGinnis Mary Roe University of Maryland University of Delaware Barbara Walker Eastern Montana University-Billings Mike McKenna Nadeen T. Ruiz Georgia Southern University California State University- Louise Waynant Sacramento Barbara Michalove Prince George's County Schools Upper Marlboro, Maryland Fourth Street Elementary School Olivia Saracho Athens, Georgia University of Maryland College Park Dera Weaver Elizabeth B. Moje Paula Schwanenflugel Athens Academy University of Utah University of Georgia Athens, Georgia

Lesley Morrow Robert Serpell Jane West Rutgers University University of Maryland Baltimore Agnes Scott College County Bruce Murray Renee Weisburg Auburn University Betty Shockley-Bisplinghoff Elkins Park, Pennsylvania Barnett Shoals Elementary School Susan Neuman Athens, Georgia Allan Wigfield Temple University University of Maryland College Park Wayne H. Slater John O'Flahavan University of Maryland College Park Shelley Wong University of Maryland College Park University of Maryland College Park Margaret Smith Marilyn OhEhausen-McKinney Las Vegas, Nevada Josephine Peyton Young University of Nevada University of Georgia Susan Sonnenschein Penny Oldfather University of Maryland Baltimore Hallie Yopp University of Georgia County California State University About the Authors

Steven A. Stahl is a Professor of Reading Bonnie Cramond is an Associate Professor in the Education at the University of Georgia and a Department of Educational Psychology at the principal investigator at the National Reading University of Georgia where she is the coordinator Research Center. He was formerly at the Center of the graduate program in Gifted and Creative for the Study of Reading at the University of Education. She taught elementary and middle Illinois.After teaching reading in elementary school for six years before returning to graduate schools in New York and Maine, he received his school for the Ph.D. in Gifted Education with an doctorate in reading from the Harvard Graduate emphasis on creativity. Her research interests School of Education. His research interests are include assessing and and nurturing creativity as centered around issues in reading instruction at the well as addressing the curriculum needs of bright primary grades and the relationship between students. vocabulary knowledge and reading comprehension.

Kathleen M. Heubach is an Assistant Professor atVirginia Commonwealth University.This research was conducted when she was a doctoral candidate in the Department of Reading Education at the University of Georgia, where she has also served as a Research Assistant at the National Reading Research Center. Her research interests include reading teacher education and the imple- mentation of literature-based instructional strate- gies within elementary content subject areas. She taught elementary and high school during a 12- year public school teaching career. National Reading Research Center Universities of Georgia and Maryland Reading Research Report No. 79 Winter 1997

Fluency-Oriented Reading Instruction

Steven A. Stahl University of Georgia

Kathleen Heubach Virginia Commonwealth University

Bonnie Cramond University of Georgia

Abstract. This paper reports the results of a two- good reader.Yet, according to Allington year project designed to reorganize basal reading (1983a), traditional conceptions of reading instruction as to stress fluent reading and automatic have ignored fluency as a goal. Instead, tradi- word recognition. The reorganized reading program tional classes have placed greater emphasis on had three components: a redesigned basal reading accurate reading of more and more difficult lesson, stressing repeated reading and partner reading; a choice reading period during the day; material rather than fluent reading. and a home reading program. Over the two years of This paper documents an attempt to re- the program, students made significantly greater organizesecond-grade classes around the goal than expected growth in reading ability in all 14 of fluency. We choose second grade because classes. All but two children who entered second we see this grade as a transition between the grade reading at a primer level or higher (and half simple and predictable material used in first of those who did not) were reading at grade level or grade to teach children to decode and the more higher by the end of the year. Growth in fluency and complex stories and expository text used in accuracy appeared to be consistent over the whole year. Students' and teachers' attitudes toward the third grade and higher. program were positive. In evaluating individual components, we found that self-selected partnerings Stages of Reading Development seemed to work best and that children chose part- ners primarily out of friendship. Children tended to Underlying our belief in the importance of choose books that were about or slightly below their fluency development in second grade is our instructional level. In addition, children seemed to view that reading is a series of stages, where benefit instructionally from more difficult materials than generally assumed, with the greater amount of development in one stage is dependent on scaffolding provided in this program. concepts learned in previous stages and prereq- Fluent and automatic word recognition has uisite for development in subsequent stages. traditionally been considered the hallmark of a The advantage of a stage model is that it pro-

1 2 Stahl, Heubach, & Cramond

vides a map describing what is to be expected the functions of print, such as directionality, at different levels of development. print conventions, and some knowledge of Stage models assume that reading is qualita- spelling patterns in the language; (3) letter tively different at different stages of develop- knowledge, or knowledge of the alphabet used; ment. That is, a child who is at one stage will (4) knowledge of the language (vocabulary and have different skills, knowledge, and beliefs syntax) that one is learning to read. about reading than a child at a higher or a Decoding. In this stage, the student begins lower stage. At each stage, the knowledge and to learn about sound-symbol correspondences. skills needed for the next stage are being devel- The student's reading performance here is oped. There have been a number of stage "glued to the text," in that she or he is trying models of reading including those of Doerhing to carefully reproduce what the text says. It and Aulls (1979), Downing (1979), and Mc- often sounds like "grunting and groaning," Cormick and Mason (1986). We will limit our because the child is not yet fluent. discussion to Chall's (1983) model because this Confirmation and Fluency. In this stage, model essentially contains the basic features of the student learns both to decode words fluent- the others, with greater elaboration. ly and accurately, as well as to orchestrate the Chall's model. Chall (1983) described the use of syntactic and semantic information in development of reading ability in six stages, text to confirm word recognition. In this stage, ranging from pre-reading to the advanced the child moves from the short, simple, and reading typical of graduate students. Her ap- possibly predictable texts of the Decoding stage proach is a global one, encompassing the to more complex texts with complex plots. At development of decoding, comprehension, and the end of this stage, children are viewed as critical evaluation. Because it is global, Chall's able to decode much of what is in their knowl- model describes broad trends in children's edge base, limited mainly by vocabulary knowl- development as readers. Her stages are as edge and world knowledge. follows: Learning the New (Single Viewpoint). In Emergent Literacy. .' In this stage, the child this stage, children learn to use their reading develops concepts about the forms and func- skill to extract information from text. At this tions of literacy. Recent research has suggested point, children are expected to learn from that four areas are most important for success content area textbooks, with increasingly less in initial reading: (1) phoneme awareness, or teacher guidance. the ability to manipulate sounds in spoken Multiple Viewpoints. In this stage, the child words; (2) print concepts, or the awareness of synthesizes information from different texts, acknowledging multiple viewpoints, but still keeping them separate from one's own. 'Chall uses the term "Readiness," but we have A World View. In this stage, adults develop adopted the term "Emergent Literacy" as more consistent the selectivity to weigh information and to add with the field. information from text to their world view.

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79 Fluency-Oriented Reading Instruction 3

Chall's model is useful for examining how children's word recognition and comprehension literacy develops over time and has important abilities are intertwined. Children learn to implications for instruction. For example, recognize words quickly and automatically in beginning reading instruction for children who the process of reading them in connected text lack phoneme awareness is likely to result in for the purpose of comprehension. reading difficulty. Juel (1988) found that no child who ranked in the lowest 25 % in pho- Transition from Decoding to Automaticity neme awareness at the beginning of first grade ranked higher than the lowest 25 % in reading The transitions in this model are extremely achievement by fourth grade. This finding has important. The transition between Emergent been replicated a number of times (see Adams, Literacy and Decoding is effected usually with 1990). instruction, although there are a number of Although each stage builds upon concepts self-taught readers who make the transition on developed in the previous stages, holding their own (see Durkin, 1974). The transition children at a stage for too long can also be between initial decoding and automaticity may detrimental to their growth. Holding children come only with practice. Samuels (1985; to a standard of word perfect oral reading, Samuels, Schermer, & Reinking, 1992) argues which might be appropriate for a child in the that automaticity comes out of children's Decoding stage, may retard their use of context practice through wide reading of different texts cues typical of the next stage. For example, if and/or repeated readings of the same texts. students are corrected for each deviation from Samuels reviews the work done on repeated the text whether it makes sense or not, they reading and concludes that ample evidence may not develop the risk-taking skills needed to exists that such practice does improve automa- use context and may concentrate on saying the ticity words "right" and not on the construction of Although Samuels (1985) allows that wide meaning (Allington, 1984). reading can also improve automaticity, there is A literal reading of Chall's model may some evidence that wide reading is not enough. slightly distort the actual development of read- Carver and Leibert (1995) failed to find that ing at various stages. By concentrating on the reading library books improved the reading of development of automatic word recognition children in grades 3, 4, and 5. Taylor, Frye, during the early stages, Chall may appear to and Maruyama (1990) did find an effect for the slight the comprehension that also occurs amount of reading fifth-grade children did at during the early grades. Although the develop- school (but not at home). However, this effect ment of automatic word recognition is the was small, about 1% of the total variance (see hallmark of these years, children's basic com- Carver & Leibert, 1995). These studies and prehension abilities also are growing at this others (e.g., Anderson, Wilson, & Fielding, time. As Adams (1990) points out, given the 1988) were done with intermediate grade interactive nature of the reading process, children, using reading logs to measure time

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO.79

11 4 Stahl, Heubach, & Cramond

spent on reading. Chall's theory suggests that Teachers read each text aloud, students and reading practice is especially important at this teachers read the texts chorally, and students transition point, so it is possible that wide practiced reading in pairs. Because of the short reading might have a greater effect in second texts, teachers were able to do all parts of the grade than with older children. Intervention lesson in a 15- min session each day. The only studies such as those of Hoffman (1987), gains attributable to the treatment were in Morris and Nelson (1992), and Rasinski, reading rate. There were no significant differ- Padak, Linek, and Sturtevant (1994) have ences between the experimental treatment and found that fluency-oriented instruction has the control in overall reading level as measured positive effects on second-grade children's by an informal reading inventory. reading .3 Hoffman (1987) describes an oral recitation Goals for our Fluency-Based Reading lesson format to substitute for a traditional Program basal reader lesson. In this format, the teacher begins by reading the story from the reader Using the stage model of reading, the pur- aloud and discussing its content. In this way pose of our fluency-oriented reading instruc- comprehension is dealt with prior to practice in tion was to help children move from the accu- oral reading. The teacher then re-reads the racy-driven decoding typical of the Decoding story, paragraph by paragraph, with the chil- stage to the fluency and automaticity needed to dren following along and echoing back each take advantage of reading to learn. We hypoth- paragraph. The students then choose or are esized that children move through this fluency assigned a portion of the text to master. They stage largely through practice in reading con- practice this text and read it to the group. They nected text for comprehension, using both then go on to the next story. On their own, repeated readings of the same text and wide children are to practice the story until they can readings of different texts. Therefore, we read it at an adequate rate with no errors. developed five goals for our Fluency-Based Hoffman reports that the lessons were success- Reading Program. They were: ful, but does not present statistical data. Morris Lessons would be comprehension orient- and Nelson (1992) found that a program based ed, even when smooth and fluent oral read- on Hoffman's, but including partner reading ing was being emphasized. This was impor- rather than small-group work, helped children tant because we wanted the students to be in one class develop word recognition skills. aware that the purpose of reading is getting However, they did not use a control group and meaning, and that the practice they were un- did not report statistical tests. dertaking would make them better comprehen- Rasinski, Padak, Linek, and Sturtevant ders, not simply better word callers. Anderson, (1994) used a similar format in their fluency Wilkinson, and Mason (1990), in their analysis development lesson; but instead of using basal of oral reading lessons, found that maintaining reader stories, they used 50- to 150-word texts. a focus on comprehension during reading

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

1.22. Fluency-Oriented Reading Instruction 5

lessons not only improves comprehension, but students develop automaticity through repeated also improves children's word recognition exposures to words in context. Repeated read- skills beyond that of an emphasis on accuracy. ings have been found to effectively improve the Children would read material at their oral reading and comprehension of normally instructional level. Traditionally it is thought achieving students (e.g., Martinez, & Roser, that reading material that is too difficult or too 1985; Taylor, Wade, & Yekovich, 1985) and easy does not improve children's reading as of disabled and developmental readers of efficiently as reading material that is well various ages (e.g., Dowhower, 1989; Rasinski, matched to the child's ability (Allington, 1989). 1984). As will be discussed, our findings ques- Children will engage in partner reading. tion this assumption because children read Partner reading provides an opportunity for material that was well-above their instructional students to read connected text within a social- level, with a great deal of scaffolding, and ly supportive context. This context should both appeared to benefit greatly. We originally motivate children to read well and provide a defined instructional level as the level at which supportive environment to aid the development they could read with 95-98% accuracy (Wix- of reading skill. For these reasons, partner son & Lipson, 1991). Previous research (e.g., reading is used by both traditional educators Gambrell, Wilson, & Gantt, 1981) suggested and those who adhere to a more holistic per- that children do very little reading of connected spective (Routman, 1991; Vacca & Rasinski, text at an appropriate instructional level, as 1992). little as 2 to 3 min per day. Our initial goal was Partner reading was used for two primary to increase the amount of material that children reasons. First, it offered an effective alter- read at this level. However, as will be dis- native to round robin reading for increasing cussed later, district constraints forced us to the amount of time that children spend modify this goal so that we also increased the reading orally. In round robin reading, amount of reading children did above conven- children are spending only a small portion tional instructional levels. of the reading period actually reading text Children will be supported in their read- (Gambrell, Wilson, & Gantt, 1981). In ing through repeated readings. This was the partnerreading,children are spending key aspect of the reading program. Children considerably more time engaged in text. A read each story numerous timesthrough echo number of studies (e.g., Topping, 1987) reading, at home with their parents, with have found that such approaches can in- partners, and by themselves. The repeated crease the amount of engaged time spent in reading component of the program was intend- reading as well as encourage children to ed to provide practice so that children would read more difficult material. Second, part- develop fluent and automatic reading. Samuels, ner reading would allow teachers to monitor Schermer, and Reinking (1992) and Rasinski children's reading progress by going around (1991), among many others, suggested that the room and listening to children read. In

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO.79

13 6 Stahl, Heubach, & Cramond

the lower grades, teachers often organize reading of the same text. Most evaluations of repeated readings as a paired reading activ- programs that involved repeated reading were ity. either short-term or did repeated reading for Children will increase the amount of only a portion of the day. In our program, we reading that they do at home as well as in were worried that repeated reading both at school. Since the school day is limited in home and in school would bore students and length, we thought that children would gain teachers alike. Another problem is that of significantly in reading proficiency with some dealing with diverse reading abilities. In our practice at home. Anderson, Wilson, and classrooms, for example, children ranged from Fielding (1988) found that even small differ- virtual non-readers to children who could ences in home-reading practice could make handle fourth-grade level material comfortably. large differences in children's reading ability. These classes were in schools representing Because the home circumstances of our chil- mixed to lower socioeconomic status children dren differed dramatically from school to and were probably representative of similar school and from child to child, we tried a populations.Providing both materialand number of approaches. Several teachers con- instruction that is appropriate to the different nected the home-reading program with Book levels requires new organizational modes. It', a reading-incentive program. Other teach- ers included reading as part of the child's Developing a Reading Program homework. One school was involved with a Reading Millionaires project (N. Baumann, During the summer of 1992, two university- 1996; O'Masta & Wolf, 1991). In this project, based researchers (Stahl and Heubach) met the number of minutes read by students in the with four elementary classroom teachers,2 two whole school were tabulated, with the goal of based in Clarke County, Georgia (Gwen Black- reading one million minutes school-wide. The well and Alice Kay Copeland) and two based in time spent in our project reading at home was Greene County, Georgia (JoAnn Hayes and added to the number of minutes that the school Nancy Gutherie) to discuss how these princi- as a whole read. In addition, as will be dis- ples could be instantiated into a reading pro- cussed below, children were given structured gram. Our goal was to develop a plan for assignments to read portions of their basal teaching reading throughout the year that reading book at home as part of the lesson would be flexible, adaptable to different classes structure. and different stories, and focused on fluency. These five components have all been studied The plan also needed to have enough variety individually, but not as part of a total reading program. Implementing a total fluency-based program over a full school year creates a 2In this report, pseudonyms are not used for either unique set of problems. One problem is main- teachers or schools. Their names are given because they taining interest in a program that involves re- co-created much of the program.

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

14 Fluency-Oriented Reading Instruction 7

for both teacher and student so that it would one at their grade placement, and many were not become tedious. Because of the need to reading significantly below grade level, we make this instruction practical, we relied heavi- used repeated reading of the same material to ly on the teachers' experience in developing help children be successful with more difficult this program. Certain aspects, most notably material. We followed the logic of an Oral monitoring children's reading using running Recitation Lesson format, which has been records, were dropped or heavily modified effective in supporting children with reading based on teacher input. There were other difficulties (Hoffman, 1987), but made signifi- givens or elements that had to be part of the cant modifications. program. For example,allteachers were Each story is different and requires a slight- committed to using basal reading programs, ly different approach. Also, teachers and through both district policy and personal students need variation to maintain interest. We choice. Therefore, we had to design lessons around the basal material. Also, in Clarke did not want to have a formula lesson; instead, we provided many options for the teacher to County a new superintendent mandated whole class reading instruction. Therefore, we had to use. The basic structure of the lesson is shown deliver lessons to the whole class. In Greene in Figure 1. County, the classes were organized homoge- Story introduction. In Hoffman's oral recita- neously; one of our classes was a high-achiev- tion format, the teacher begins by reading the ing class and one was a low-achieving class. story aloud and discussing it, using a story These different levels had to be taken into map. In this way, the teacher deals with com- account. prehension prior to the fluency practice, keep- Our meetings stressed one principle per ing the lesson focused on comprehension. We week. We read descriptions of other fluency- followed a similar procedure with the teacher based programs, such as Hoffman (1987), and reading the story aloud to begin the lesson. discussed how those ideas would fit into the Following this read-aloud, we used a variety of teachers' classrooms and into our overall goals. procedures to discuss the story, including story At the end of the summer we had a general maps, traditional questions, student-generated plan for reading instruction. questions, and other graphic organizers, in- The general plan had three componentsa cluding various types of story maps, plot redesigned basal reading lesson, a home read- charts, Venn diagrams, and so on. This usually ing program, and a daily free-choice reading comprised the first day's lesson. period. These will be discussed, in turn, be- Children who needed some extra help with low. the story were pulled aside for echo reading. If A Redesigned Basal Reading Lesson the story was particularly challenging, echo reading was done with the whole class. In echo Since all of the students in one school were reading, the teacher read a paragraph at a time, required to read the same basal reading lesson, with the students echoing it back. This was

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO.79

15 8 Stahl, Heubach, & Cramond

Fluency-Based Classroom Reading

Read'Story to Class

Discuss story

Option, Reading;

Option: Children learnone section of text.

....Option: . Children .Read Story at :Mine two Cir:MOre times.

Partner Reading of Story

Option: :Children *read atoll/ as a play Children do Journals / Worksheets impairs or as a class

Figure 1.

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORTNO.79

BEST COPY AVAILABLE 16 Fluency-Oriented Reading Instruction 9

done to scaffold students' recognition of words of reviewing the story content. Students who and to help them successfully read the story. were having difficulty were assigned to read Partner reading. The next component was the story one more time at home. In addition, partner reading. Children were to read the teachers sometimes had students re-read por- story in pairs, with one member of the pair tions of the story for performance or made the reading a portion of the story aloud and the story into a play to provide more practice in other monitoring and providing assistance if reading. needed. How large the portion was to be was agreed upon by the partners, but most often Home Reading one child would read one page and the other child would read the next. Students did two types of reading at home. We tried a number of variations in how Students read the basal reading selection at partners were assigned. Because of our forma- home at least one or two days a week. The tive study on partner reading, discussed below, story was sent home the first day with the teachers used self-selection for partner reading. instruction for the student to read it to a parent One table at a time would get to choose their or other person in the household. More able partners. The pairings varied throughout the readers read the story to themselves, but most year, depending on who was getting along with students read it aloud. We met with the parents whom. We continued to observe cooperation, before school started and talked about reading especially as the year progressed. Students also at home. Often parents sat with the child and began to work in pairs during free-choice followed along. But sometimes the child read reading time, on their own. while the parent was doing something else, Two more points need to be made about such as making dinner. In many families, time partner reading. First, it was difficult to set up is often short, so many alternatives needed to in the beginning. It took several weeks of be provided. Parents were not able to read with practice before the partner reading jelled and their children every day, as evidenced by the students knew what their roles were. Second, responses to sheets that we sent home with the odd numbers were handled in different ways. children, but there seems to have been a gener- Sometimes a group of three was formed. Some- al effort to read at home. We also gave parents times the teacher read with the odd child. some guidance in how to correct errors. Be- Generally, however, teachers avoided reading cause we met during Parent-Teacher Organiza- with children because this impaired their ability tion meetings, our time was limited. Conse- to monitor the reading throughout the class. quently, we could not provide as extensive Additional instruction. The following day, parent training as provided by Mudre and the teacher worked with the journals that came McCormick (1989), whose training procedure with the basal reading program. Teachers seemed to improve parents' response to their varied in how they did this. Usually, journal children's oral reading. Also, the percentage of pages were discussed as a whole class as a way parents who were able to attend this meeting

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

17 10 Stahl, Heubach, & Cramond

varied considerably, from two or three parents Does the program lead to gains in oral per class in one school to three-quarters of the reading with comprehension? parents in another school. Because we are stressing oral reading in the Students also were supposed to read a book program, we wanted to assess the program's of their choice. Children brought books from effects on oral reading. But we did not want to the school library, the class library, the public, produce word callers who decoded the text or read books that they owned. The object of without comprehension, so we also assessed the home reading program was to extend the comprehension. amount of practice that students do by adding What happens during partner reading? to the school day. Partner reading is a prominent feature of this program. The teachers in the program Free-Choice Reading Period wanted to know what were the most effective pairings of students; we wanted to know more The teachers encouraged students to read a about why children chose each other, what variety of books on their own. The purpose of were the dynamics of partner reading, and so this reading time was to increase interest in on. reading as well as to promote reading at the What are the effects of reading at home? students' own level. The teachers provided We wanted to know whether differences in periods of time (15-20 min) for independent parental support of home reading would have reading and students were also encouraged to an effect on children's gains in reading, but read as they completed assignments throughout the day. were unable to collect this data in a manner that was reliable. Evaluation of the Study What types of books do children choose during choice reading? We used the following questions about this If children are to benefit from reading intervention to guide our examination of this during choice reading, they need to choose program. books that are at or near their instructional Can a fluency-oriented reading program level. (At least this is the conventional wisdom, be sustained over a full year? but see Carver and Leibert, 1995.) We wanted Although most of the approaches used in to examine the relative difficulty of books that our re-organization have been tested before, children choose. We also wanted to see why these tests have generally been of short dura- children chose the books they did. tions, usually one or two months at the most. What are the effects of the program on Because lesson structure involved repeated struggling readers? readings of the same text, some observers Given Stanovich's (1986) notion of Matthew thought that either the children or the teachers Effects and Allington's (1983b) observations would tire of the procedures, leading to nega- about the differences in the amount of reading tive attitudes. done by struggling and normally achieving

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

18 Fluency-Oriented Reading Instruction 11

readers, we felt that radically increasing the pared with their achievement in May. (The amount of reading done would have an espe- first year we also included an interim measure cially large effect on struggling readers. in February.) Traditionally,programevaluationsare Results conducted with either an experimental or quasi- experimental design (Campbell & Stanley, To assess the effectiveness of the program, 1966). In such a design, there is a treatment we conducted a series of evaluations. Because group and a control group. In our original plan this program is complex and was undertaken for this study, we had planned to use the first over the course of two years, the evaluation year to develop the program, conducting only procedures are complex as well. Some evalua- formativestudies and one pretest-posttest tions used the entire population of children evaluation. The second year was intended to be participating; others used only a sample of that an experimental test of the program we devel- population. Because the samples differ from oped during the first year. However, the re- substudy to substudy, they will be reported as sults of the first year were so unexpectedly traditional studies, with a description of the strong that we felt that denying treatment to a sample, methods, results and discussion. control set of classes was unethical. Therefore, The studies reported below come from we decided to use all of our classes as treat- questions that we had about the program, ment classes, and we developed a pretest- beginning with whether it could be sustained posttest design to evaluate the program. and whether it affected children's growth in The logic for the analysis is that if the reading to more specific questions about com- program is more successful than conventional ponents of the program. Some of these ques- instruction, children then will make greater tions were generated by the researchers; others progress on a standard measure of reading than came from concerns of the teachers participat- the one-year growth expected in one year's ing. The first studies reported deal with ques- time. If such growth occurs in a substantial tionsconcerning overall program effects, proportion of the classrooms that we have followed by questions about specific compo- worked with, we then can argue that this nents. growth is due not to chance variations but to the effects of the program itself. Study 1Overall Program Evaluation Participants. To assess the overall program effects, we used the entire population of stu- Because of the nature of the program and dents for both the first and second year. The our theoretical orientation discussed earlier, we student population during the first year consist- used a measure of oral reading with compre- ed of 84 students, 49 in Oglethorpe Avenue hension to evaluate the program. The basic School in Clarke County (Blackwell, Cope- design used was a pretest-posttest design in land) and the remainder in Greensboro Primary which children's scores in August were corn- School in Greene County (Gutherie, Hayes).

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

19 12 Stahl, Heubach, & Cramond

The students at Oglethorpe Avenue were of inventory, during the first month of school and mixed socioeconomic status. Approximately the last month of school. During the first year, 60% were African-American, the remainder the QRI was administered in February as an European-American. In Greene County, 85 % interim measure. The QRI was chosen because of the students were African-American and this measure gave equal emphasis to oral were predominantly from homes with a lower reading and to comprehension, matching our socioeconomic status. program objectives. According to the material The student population during the second in the manual, alternate form reliability was year was similar in characteristics to that used high. This was assessed by calculating the the first year, except it was considerably larg- reliability of decision making using the individ- er. We added an additional teacher at Ogle- ual passages of the test. For the eight levels of thorpe Avenue School (Martha Cartwright) and the test, Leslie and Caldwell (1988) report that added three teachers at Barnett Shoals School all reliabilities were above 80% and three- in Clarke County (Martha Cornish, Ric Os- quarters were above 90 %. In addition, Leslie walt, Mary Todd). We also added two more and Caldwell report the concurrent validity of teachers in Greene County (Barbara Dean, the QRI, as measured by the correlations Stephanie Hart). Of the six new teachers, one between instructional level on the QRI and teacher was male, the remainder female, two performance on an unnamed standardized teachers were African-American, the remainder achievement test, ranged between .44 and .72 European-American. Two of the new teachers with the majority of correlations above .70. had fewer than five years experience; the Year One. Figure 2 shows the QRI results remainder had more than ten years experience. for the first year. As shown on that figure, The second-grade students who participated students made an average gain of 1.88 grade from Oglethorpe Avenue and Greene County levels in their instructional level over the were demographically similar to those who course of the year. This gain was uniform for participated thefirstyear. The additional all four classes. The ordinary assumption is students from Barnett Shoals contained a wider that students will average about one year's variety of parental backgrounds. Approximate- growth in one year's time. We compared the ly 40% of these students were African-Ameri- actual growth to this assumed growth through can. We had an exceptionally high rate of a series of t-tests. For each class, we took the mobility during the second year. We included, mean growth and tested whether it was signifi- at some point, 180 different students, but only cantly different from one. In all four classes, 125 were present from the beginning to the the growth over the year was significantly end. greater than one year, all p < .01. Procedure. All students participating in the Furthermore, as shown in Figure 3, gains project were given the Qualitative Reading were made by students entering at different Inventory (QRI, Leslie & Caldwell, 1988), an reading levels. That is, the average child en- individually administered informal reading tering second grade reading below the primer

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

20 Fluency-Oriented Reading Instruction

C WM= N

U) a. C 0 =Il C)

ca ca

CO

Figure 2.

C\1 NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79 14 Stahl, Heubach, & Cramond - CN)

-co

LO

c\I

15 ,1). 43' 1:3 17 "0 E E i I I 22 .c CC CC EC a.",_ CL 2 g c2 0 F1 E u.18 LE corD lanaireuo! joru 6uipeau 6uiaa1u3 Figure 3.

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79 Fluency-Oriented Reading Instruction 15

level made an average of two years progress children who would ordinarily have a great during the course of the year. The average deal of difficulty learning to read. Of the child reading at the third-reader level in the remaining 105 students who had pretest and beginning of the year made a gain of three posttest data, only 2 failed to read at the second years during the school year. Of the 85 stu- grade level or higher by the end of the study. dents in the four classes, only 3 were still Both of these students began reading at the unable to read the second-grade passage by the primer level and were able to progress only to end of the year. the first-reader level. Year Two. The second year pretest-posttest Thus, for all 14 classes over the first two evaluations are shown in Figure 4. As can be years of the project, students made significant- seen in this figure, the yearly gains were nearly ly more progress than one year's growth in one as high, averaging 1.77 years' growth in in- school year. By the logic discussed earlier, we structional level. Again, for each class, we maintain that this indicates that fluency-orient- took the mean growth and tested whether it was ed reading instruction is more effective than significantly different from one. In 8 of the 9 conventional instruction. classes,3 the growth over the year was signifi- cantly greater than one. All but one tested at p Study 2Growth of Rate and Accuracy < .01, and the remaining class made approxi- mately one year's gain in one year's time. To examine the development of fluency As shown in Figure 5, children at all enter- over the course of the year, we initiated a ing reading levels made gains in the second series of fluency checks during the second year as they did in the first year. Again, these year. We operationalized fluent reading as gains were relatively uniform. Children who reading that is both rapid and accurate. There- entered second grade reading below the primer fore, to examine fluency, we looked at both level had ended with an average instructional accuracy and rate. The purpose of these checks level of 2.25, somewhat below the second- was to examine the effects of each lesson on children's accuracy and rate of reading the grade level (which we would have coded 2.5). basal reading selections. We also wanted to see Of the 20 students who could not read a primer passage at the beginning of the school year, 9 how readers of different entering abilities developed over the course of the year. were reading at a second-grade level or higher Participants. The participantsin these by the end of the year, and all but one could sampling studies were the students in the six read at a primer or higher level. This sug- classes in Barnett Shoals and Oglethorpe Ave- gests that this program was successful even for nue Schools. Because there were different numbers of students in these classes during the year, the numbers varied. There were 91 3Guthrie and Hayes, two of our original teachers, team-taught a combined, larger class. Their children are students sampled in November, 87 in January, reported together. and 89 in May.

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

25 16 Stahl, Heubach, & Cramond

C\1 0 82 pa

IJeH

map

apati saAeH

IBMS()

Pcol

LIS11.1100

sq6pmImo

puelodoo

I le m>1 oe 1E1 1111111111 In U) nr cn cv o aT f7 cv Figure 4.

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79 May Instructional Second Year of Study Level, by Entering Level, 0 Fourth trade or Higher Third Grade -J SecondFirst GradeReader Below Primer Primer 0 1 2 Mean Instructional 3 Level,4 May 5 6 7 28 29 18 Stahl, Heubach, & Cramond

Method. These fluency checks were con- unread and reread stories), and are making ducted over a two-week period. At the end of progress over time (comparing the unread the first week, after the teacher had finished a stories in November, January, and May). This story, children read orally two selections of progress is most pronounced from November between 150 and 200 words. The first selection to January, suggesting that the bulk of the is taken from the story just completed; the children's reading growth occurred during that second selection is from the story not yet read time period. This is similar to the results from but to be begun the following day. The second the informal reading inventory given during the week students re-read the selection from the first year. We found that students made a gain story they had just finished. For each story of a full year in the four months between segment, we noted both accuracy of word September and January, and somewhat less identification and rate of reading. These checks than that between January and May. were given in November, February, and May. Looking at the growth over time at each These checks allowed us to compare each level, as presented on Tables 1 and 2, it seems child's reading of an unread story with one that that there were different patterns of growth in was just completed and with their reading of rate and accuracy among children with differ- the same story after a week of treatment. In ent entering abilities. In terms of accuracy, addition, we could compare children's reading children reading initially at a second-grade of the previously unread story, which could be level or higher generally made little improve- considered a baseline, with their reading later ment in rate over the year as suggested by their on in the year, allowing us to assess progress reading of the unread selection. These students in both accuracy and rate. were generally reading the material at or above All deviations from the text were considered an instructional level of 95 % accuracy (or a errors for the purpose of this study. We did not 5 % or lower error rate). There simply was not distinguish between meaning changing and much room for them to grow. For students non-meaning changing miscues in our analysis who began the year reading below an instruc- for higher interrater reliability.' tional level of second grade, there were differ- Results. The results from the checks (see ent patterns of growth. Again concentrating on Figure 6) suggest that students made significant the error rates for the unread selection, those progress in both rate and accuracy because of children initially reading at the first-reader the practice (comparing the read story with the level dropped their average error rate from 9% in October to 6% in February and May. This improvement suggests that they raised their 'In the instruction, however, teachers did make those instructional level to that expected at their distinctions. In our summer program, we made the distinction between miscues that change the meaning and grade level. The error rate of children reading those that do not, and our observations of the teachers at or below the primer levelin October indicated that they generally did not correct non-meaning dropped significantly on the unread selec- changing miscues. tion, but remained considerably above the 95 %

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

30 Percent of Miscues Previously bIV Ab0 'oCI)0 0CO O 0 Previously No. of Words Per Minute0IV 0A 0CO 0CO 0 0N.) Read,Oct. Unread, Cct. ° Read, Oct . Unread, ° Re-read, Oct. Re-read, Oct.Oct. 0A) CD 30:A) CD Previously to PreviouslyRead, Jan 0CD Read, Jan Unread, Jan. 9)11. Re-read,Unread, Jan. Re-read, Jan. CD3 Jan. 00. Previously 0.to PreviouslyRead, May Unread, Read, May Unread, May Re-read, MayMay Re-read, May ErrorTable Rate,1 by Condition and Entering Level October January May Below Primer Previously (.09)Read.16 Unread (.10) .25 Re-Read (.09) .15 Previously (.08)Read.13 Unread(.13) .20 Re-Read (.12) .12 Previously (.05)Read.11 Unread(.09) .19 Re-Read (.09) .12 First Reader Primer (.04)(.04) .04.08 (.06) .09.18 (.03) .04.06 (.02)(.04) .03.06 (.03)(.04) .06.11 (.01)(.02) .01.04 (.02)(.04) .03.07 (.04)(.04) .06.12 (.02)(.03) .04.06 Second GradeThird Grade (.02)(.03) .02.03 (.02)(.03) .04.05 (.02) .02 (.01)(.02) .01.02 (.03)(.02) .04.05 (.01)(.01) .01.02 (.01)(.03) .01.03 (.02)(.03) .03.04 (.01)(.02) .01.03 32 Fourth Grade or Higher (.001) .01 (.02) .02 (.01) .01 (.001) .01 (.02) .02 (.01) .01 (.001).01 (.01) .02 (.01) .01 33 z z0 Rate,Table by 2 Condition and Entering Level rriz Previously Read October Unread Re-Read Previously Read January Unread Re-Read Previously Read UnreadMay Re-Read Below Primer Primer (14.12)(17.34)60.4238.34 (13.22)(12.96)46.6028.69 (17.48)(20.75)42.3663.35 (19.18)(19.42)44.6570.62 (11.74)(23.29)62.1737.21 (23.11)(21.30)52.2576.61 (21.48)(14.25)69.4639.57 (21.93)(12.25)55.9432.00 (19.16)(16.63)70.4344.39 SecondFirst Grade Reader (23.88)(32.88)92.0876.70 (14.53)(27.86)69.9255.12 (16.98)(36.58)91.3074.85 (23.71)(30.89)103.9697.98 (15.78)(31.10)79.7274.81 (23.34)(28.75)114.37111.22 (28.60)108.92(34.92)95.37 (28.80)(22.49)84.6862.16 (30.02)(25.60)108.6478.89 to FourthThird Grade Gradeor Higher (34.44)143.38(22.09)111.02 (32.16)(16.65)104.1878.29 (35.47)(21.01)128.36106.21 (26.38)(21.64)138.15114.05 (37.78)(26.76)124.7893.67 (28.04)(22.68)153.86131.27 (38.31)(20.93)149.24122.88 (36.48)(21.51)119.0690.95 (41.05)(21.89)144.37120.50 Po 34 35 22 Stahl, Heubach, & Cramond

accuracy level. With re-reading, though, chil- readers, those initially reading at a fourth- dren who began the year reading at a primer grade level or higher, grew from104words level were able to bring their error rate to near per minute in October to119words per minute the95%level, as evidenced by their perfor- in May, suggesting that even these able stu- mance on both the previously read and re-read dents were making palpable gains. stories. But the error rate of children who began reading below the primer level remained Study 3Student Attitudes very low. The results on the growth of accuracy We were also concerned with the attitudes mirror the pretest-posttest results. This pro- that students might have about the program. gram seems to be highly successful for children Because this program involved a great deal of who begin the second-grade year with a read- repeated reading, we were concerned that ing level at or above the primer level; that is, students might consider it drudgery. We were for children who can recognize a simple corp especially concerned about gifted children's of words. In Chall's(1983)stage model, responses to this program. We worried that discussed above, these would be children who gifted children may feel like tutors in coopera- are at the Decoding stage or higher. tive learning situations, especially where they Although children reading initially at a were reading material below their own instruc- second-grade level or higher did not make tional level (see Martin,1984).We wanted to gains in accuracy, they did make gains in rate, question gifted and non-gifted children in our especially between the October and February sample to assess their attitude toward the sampling. Children at all levels at or above the program. We also wanted to assess the atti- primer made average gains of at level 10 words tudes of children toward the various compo- per minute from October to February. Between nents of the programpartner reading, home February and May, gains were inconsistent. reading, choice reading. Some groups of students read at somewhat To assess student interest in reading lessons, lower rates in May than in February. The fall- we individually interviewed a sample of stu- off of those children who initially read at a dents about their attitudes toward the program first-reader level was dramatic, from75to 62 and toward reading in general. The interviews words per minute from February to May. An were conducted in April of the second year so even larger fall-off occurred on their reading of that students had experienced nearly a whole the re-read selection. This may be nothing school year of fluency-oriented instruction. more than a problem with a particular passage Participants. The sample consisted of44 or a somewhat different selection of students at students from the three classes at Barnett different times due to absences. This was the Shoals Elementary School. They were the group with the smallest number(9or 10) and entire group of students present that day, most susceptible to attrition effects. However, excluding children who were receiving Chapter the average reading rate of even our most able 1 services or special education services.

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO.79

36 Fluency-Oriented Reading Instruction 23

Procedure. Each child was interviewed home reading, the reading of the story by the individually by one of the researchers. The teacher, and especially partner reading. interview consists of nine questions, some of We categorized the responses to the ques- which were multipart. (These questions are tions of "What do you like about the reading included in Appendix A.) The questions were program" and "What don't you like about the designed to examine students' attitudes toward reading program" into broad categories of the reading program and how they perceived responses. The largest number, 19 children, components of the instruction. We transcribed liked something about the stories or the materi- the responses to the questions and rated them in als they could read (e.g., "read a lot of books, gross categories. For questions that asked for chapter books," "finding out what's in the an overall evaluation of the program or a story," "get to read the good"). The next component, answers were categorized as posi- largest number, 10 children, mentioned some- tive ("Good" "I like it a lot") or negative ("I thing that we categorized as having to do with don't like it") or neutral ("It's OK"). For other self-efficacy, or the reader's sense of compe- questions, two of the researchers categorized tence as a reader. In this category were men- the responses by attempting to group together tions of practice and growing skill. Examples responses that contained similar wording or include "learning how to read good and prac- ideas. We agreed on both the gross categories tice," "learning new words," and "you can and which responses fit under each category. find out stuff." The only program component Results. Table 3 records the number of mentioned was partner reading, which was positive, negative, and neutral responses to the mentioned by seven children, and an additional evaluation questions. As can be seen on this five mentioned external motivation, including table, attitudes toward the program have been participation in the Book It program and the overwhelmingly positive from children of all Reading Millionaires project (N. Baumann, ability levels. 1996). Overall program. The main impression is When asked what they did not like, children that children of all ability levels and of both tended to mention factors involving length or genders felt positive about their reading class. difficulty of the stories (e.g., "takes some All students felt they had learned to read better people a long time to read," "some books have this year. All but one student gave an overall too many words"). The next largest group of positive rating to the program. (The one child children could not come up with things they who did not like the program was extremely did not like. Ten children mentioned various gifted; he was reading at a seventh-grade level aspects of the program that they did not like. at the beginning of the year.) Three students There was no particular component, however, were lukewarm in their responses ("It's OK"); that was disliked by more than one individual. the remainder were enthusiastic. When we Home reading. According to the interview, looked at the various components of the pro- the median number of times a basal story was gram, students were enthusiastic about the read at home was twice with a range from none

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

37 24 Stahl, Heubach, & Cramond

Table 3 Answers to Interview Questions Overall Program Do you think you have learned to read better this year?

Yes No

Girls 13 0

Boys 20 0 Home Reading What do you think about (reading the story at home)?

Positive Negative

Girls 13 0

Boys 15 1 Do you practice reading the story at home?

Yes No Sometimes

Girls 19 3 2

Boys 16 4 1 How many times do you practice at home?

1 5 Girls 4 Boys

1 or 2 1 Girl 1 Boy

2 6 Girls 2 Boys

2-3 1 Girl 1 Boy

3 3 Girls .1 Boy

4 1 Girl 1 Boy

5 2 Girls 3 Boys Choice Reading What do you think about DEAR time?

Positive Negative Sort-Of

Girls 13 0 0

Boys 19 0 1

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

as Fluency-Oriented Reading Instruction 25

Partner Reading What do you think about partner reading?

Positive Negative

Girls 13 0

Boys 19 1 Do you enjoy reading with (your partner)?

Yes No Sometimes

Girls 11 0 2

Boys 18 2 to five days a week. When asked how this reading time. When asked how choice reading makes you a better reader, 22 provided some helps one read an approximately equal number answer related to self-efficacy, 10 mentioned of responses mentioned the importance of their family, and 4 mentioned the atmosphere. additional time for practice and the atmosphere Again, we see the mentions of family as moti- conducive to reading. As carried out in these vational, suggesting that these children largely schools, choice reading could be individual or see this component as providing both practice could involve a small group. Only two people in reading skill and motivation for that prac- mentioned working with a partner; two others tice. When asked how home reading improves mentioned some other aspect of social interac- tion. When asked why choice reading made interest, 9 mentioned self-efficacy, reflecting one more interested in reading, seven re- what appears to be a belief that greater compe- sponses mentioned the importance of an oppor- tence in reading leads to greater interest. Ex- tunity to sample (e.g., "yes, if you read you amples include "if you read at home you can can see another book you want to read," "yes, read books like S .If Mrs. C can't we read different books after you get done with pronounce a word, S knows it," "yes, the first one," "yes, can read more books and it's fun, when you get the hang of it, it is fun," know about them and want to read more") and "yes, because I can be better at reading," and two mentioned interest in a particular type of "yes, because when we read a book we know book. Six responses mentioned reading skill. the words so we can make it more interesting As discussed above in regards to the home- and fun." reading component, these six responses may Choice reading. As noted on Table 2, reflect a belief that greater skill in reading children were uniformly positive about choice leads to more interest.

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

39 26 Stahl, Heubach, & Cramond

Partner reading. As noted on Table 3, basal reading lesson and during the free-choice nearly all students were positive about partner reading period, we wanted to find out what reading. When asked how partner reading helps went on during partner reading. Our interest you read, two categories predominated. Eigh- began with the teachers' questions about how teen students mentioned some sort of assistance best to organize partner reading, whether with words and 15 mentioned the social inter- teachers should assign children to work in actions. The social interaction statements seem heterogeneous groups, as had been done in one to suggest that working with a friend is motiva- class, or whether students should select their tional ("because A is my friend, she reads a little better than me and it helps me"). own partners, as had been done in another These results suggest that the students see class. However, our interests were somewhat partner reading much as we did when planning deeper. We wanted to capture, qualitatively, the program. some aspects of what made partnerships func- When asked how partner reading makes one tion in our reading lessons. more interested in reading, 10 also mentioned To understand these questions, we conduct- social interactions with another group mention- ed a qualitative and a quantitative analysis of ing more general self-efficacy. The largest data collected in two econd-grade classrooms group, however, was unable to provide reasons where partner reading is an integral part of why it improved interest and four children said reading instruction. We were interested in (a) that it did not. the relative efficiency of different pairings of The majority of children mentioned friend- students, (b) the types of interaction taking ship as the primary reason they chose their partners. Reading ability was mentioned by place while children read in pairs, (c) the somewhat fewer than half of the children. factors that influence decision-making within When asked what level of reading ability they the pairs, and (d) the factors that influence would prefer their partner to have, there was a smooth and fluent reading. wide variety of responses. Most children did We view partner reading as an example of not care about the level, but some preferred a a closed social circle embedded within a larger child at the same level, some preferred to work classroom context. We examined the functions with poorer readers, and some preferred to of literacy within this smaller context and how work with better readers. Those who preferred these functions related to the goals of the to work with better readers were nearly always classroom at large. In this study, literacy children who were having reading problems. learning and paired reading are examined in These responses confirm a fuller study of relationship to the larger social fabric of the partner reading done in the first year and classroom. discussed below. Subjects. This substudy was conducted the Study 4Partner Reading first year, using the two second-grade class- rooms from Oglethorpe Avenue School, with Because partner reading was an important a total of 42 children. The children are lafgely aspect of our program, both in the redesigned from middle-class families of diverse ethnicity

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

40 Fluency-Oriented Reading Instruction 27

and represent a range of reading abilities. In running records of samples of material read both classrooms a newly adopted basal is the using the partner reading sessions. We collect- foundation for reading instruction. ed data from October to March, revisiting each Method. During the first data-collection class every four weeks. There were a total of cycle, children were assigned to partners in six data-collection cycles, 1 three-week cycle each of three ways(1) pairs heterogenous in in which both the quantitative and qualitative ability assigned by the teacher,(2)pairs homo- data were collected, and5one-day observa- geneous in ability assigned by the teacher, and tions in which only qualitative information was (3) student-chosen pairs. This cycle lasted three gathered. weeks. Each child participated in each of the Analysis. The qualitative data analysis was three selection conditions and thus served as his done simultaneously with its collection using or her own control. Following each partner the constant comparative method (Glaser & reading session, each child was given a seg- Strauss,1967).The analysis began with the ment of the basal reader text that was read collection of data through observation and during that session and asked to read it orally. taping of paired reading sessions. Immediately A running record (Clay,1985)was taken of following the paired reading, one researcher this oral reading. The error rate following each took the students out individually and gave reading was used to examine the relative effi- them a structured interview relating to their ciency of the different types of pairings. perceptions of partner reading and the reasons The remainder of the time students chose for their choices in partners. The interviews their own partners. The students read in pairs and tapes were transcribed and reviewed. After after the basal story had been introduced and transcribing, that researcher looked for patterns read orally to the class or individually by each as initial categories and relationships emerged. student. The children also read in pairs during Subsequent collection cycles were used to DEAR (Drop Everything And Read) time, confirm or disconfirm initial assertions about which provides an opportunity for reading of the social interactions taking place during self-selected books. In these classes there was paired reading. an average of15-20min per day allotted to The running records were used as a dynam- DEAR time. Children can choose to read alone ic measure of reading level to note whether stu- or with partners during this time. dents were working on the same level or not. During the remainder of the observations, They were also used to compare the effective- data were collected from multiple sources, ness of the three different partnering situations. including (a) audio recordings of6target Quantitative Results. The running records students (3 from each class) as they read with were analyzed to look directly at the effects of a partner, (b) video recordings of pairs of different partnerings on student reading. We students, (c) field notes taken as the students did not find a significant difference between participate in paired readings, (d) interviews types of pairings. (The results are presented in with the students and the teachers, and (e) Table 4.) However, the level of performance

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO.79

41_ 28 Stahl, Heubach, & Cramond

Table 4 Partner Reading, Grouping Study

Self-Selected Heterogeneous Homogeneous Grouping Grouping Grouping

Mean .94 .93 .91 Standard Deviation (.09) (.11) (.11)

was so high that a ceiling effect might have The most important relationship for the obscured real differences. All children who students was that of friendship. When asked initially read at a level of primer or higher how a certain person became a partner, the could read second-grade material at an instruc- majority of students responded with remarks tional level (approximately 95 % or higher) or that were categorized as "friendship." For close to it. Of those who read below the primer example, one child, when asked how Peter level, half could read second-grade material at became his partner, responded "I play kickball an instructional level. This suggests that the with him everyday." With only one exception, support discussed above is useful in helping children accepted the partner who had chosen nearly every child read successfully in grade- them even if that person is not someone they level materials. Confirming the qualitative would have chosen themselves. Data from the analysis, selecting one's partner seems to teacher interviews confirmed that children produce better results, especially with lower tended to pick partners with whom they are achieving readers. getting along at the moment. Qualitative Results. The major assertion Although friendship was the main property generated from the data analysis was that the of the relationship category, other factors were relationship the children shared before they involved as well. Students were likely to work paired for partner reading not only helped to with others who had the same working style as their own. For example, a no-nonsense type determine their choice of a partner but guided their actions during reading as well. This reader who wanted to immediately get started would choose another no-nonsense type. Gen- relationship seems to be the most important der did not play much of a role in determining factor in determining how effectively students who was chosen for partner reading. Although work together in completing the paired reading same-gender pairs were the norm, it was not task. Analysis of this data indicates that a uncommon to find boys and girls working positive and established relationship between together by choice. the partners is important for effective partner Each new pair of students had to work out reading. procedures for reading the story. Decisions

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

4 2 Fluency-Oriented Reading Instruction 29

were necessary about where to go in the room time devoted to reading material at their in- for reading, whether to read sitting or lying on structional level. To develop fluency,itis the floor, who would go first, and how turns important that children read material at or near would be taken. This decision making was their instructional level, which we defined greatly affected by the nature of the relation- initially as roughly 95 % accuracy. Because we ship already established. If there were dis- were required to provide whole class instruc- agreements about procedure, the self-selected tion using the basal reader, most children were pairs worked these out without including the reading material at or above their instructional teacher or wasting real time. Often there was level during that time. Relatively few second- not a need to discuss procedures. For example, graders were actually reading at second-grade one pair was so in tune that when one child level. For example, only 42 of the 152 students rolled over on his stomach, the other followed. assessed at the beginning of the second year One of the key features of paired reading is actually had a score at second-grade instruc- the assistance that one child gives another when tional level. Our assumption was that during fluent reading breaks down. The most frequent the period of choice reading, children, both form of assistance took place when a child gifted and struggling, would be able to read could not read a word. The reader would stop, material at their level. This assumption has wait for the partner to provide the word, and not, to our knowledge, been tested. then continue reading. Assistance appears to be One purpose of this substudy is to check connected to the relationship that is established whether students actually chose books that are before a paired reading session begins. When instructionally appropriate. A second purpose children already have a working relationship, is to develop a theory of why children choose they are more likely to ask for help when it is the books that they do. needed and assistance is given in a more effi- Method. We asked children in two classes to cient manner. fill out logs of the books they were reading When off-task behaviors were noted, they during SSR time for two weeks. Subjects in often did not interrupt the reading when the this study were 43 students in the two Greene pairs were self-selected. This is because the County classes during the first year of the pairs had already established routines and ways program. After two weeks, we interviewed of relating to one another. These behaviors each child individually about why they chose became more frequent and were more likely to these particular books and what criteria they hinder smooth sailing when the partners were use for choosing books in general. We also not self-chosen. took one of the books that the child had read during the preceding week and did a running Study 5Choice Reading record on a small section of the book to find out its relative level of difficulty for the child. One of the assumptions in developing this We used oral reading error rate as a measure program is that children should have ample of relative difficulty.

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

43 30 Stahl, Heubach, & Cramond

Quantitative results. In the running records, Discussion all children except one had chosen books that were at or near their instructional level. With As noted earlier, the studies used to evalu- one exception, students were able to read their ate this program were driven by a series of chosen book with 92% accuracy or higher; the questions we had about the program. We had average was 95.5 % accuracy. This rate sug- some of the questions prior to planning the gests that students are choosing material near program; other questions arose during the their instructional level but considerably more implementation and came either from the difficult than their independent level, which has researchers or from the participating teachers. traditionally been thought of as 98 % accuracy (Wixson & Lipson, 1991). As noted below, we Can a Fluency-Oriented Reading Program Be have reason to reassess this traditional notion. Sustained Over a Full Year? The one student whose accuracy was consid- erably below this level (62 %) was a child who Teachers' viewpoints. As noted above, for was placed in a homogeneous above average the first year of this study there were four reading class for reasons unrelated to his read- teachers who developed the program. Three of ing, but read significantly below the class the four teachers were highly experienced, average. He chose books that looked like those each with more than ten-years teaching experi- his classmates were reading, even though they ence, mostly at second grade. The fourth were too difficult for him. We feel, bolstered teacher was in her fourth year of teaching. She by his interview, that he had chosen books for had spent the past year teaching in a supple- social reasons, to look as if he were competi- tive with his peers. If these results can be mental program aimed at low-income children. replicated, it suggests that SSR is a valid way This year she was returning to second grade. of increasing children's fluency, becuase they One of the teachers was African-American; the will choose books that are instructionally remainder were European-American. All are appropriate. It also suggests that social pres- female. All of these teachers would consider sures need to be taken into account in imple- themselves traditional. They all had experience menting SSR. using basal reading programs and preferred to The students were also interviewed about use such programs. their reasons for their choices. The teachers The second year we expanded our group to were a major influence. Books that the teachers ten teachers. Our intention was to see whether had read to the class were chosen often. In the success of this program could be replicated addition, one teacher encouraged her students with teachers who did not participate in the to read chapter books that challenged their creation of it. All of the teachers participating ability in reading. These exhortations were during the first year also continued to partici- mentioned often by her students. Students did pate during the second year. not mention their peers as influences on their At the end of the first year, the four teach- reading choice. ers reported that they were very happy with the

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

44 Fluency-Oriented Reading Instruction 31

procedures involved and would enthusiastically was also mentioned. When asked about inter- continue them into the second year. Of the ten est, a proportion of students mentioned aspects teachers who participated the second year, all of self-efficacy, suggesting that some students have reported that they are using the proce- felt that improvement in reading ability was dures this current year, even though we are no related to reading interest (see Nell, 1988). longer providing direct support. These results Other students mentioned the importance of suggest that the program is sustainable over the being able to sample new books, of choice course of the school year, and that teachers reading, and of atmosphere, such as a quiet maintain their enthusiasm about the instruction. room. Although nearly all students felt that Student effects. From the interviews, it also home reading, partner reading, and choice seems that students had positive attitudes to- reading improved interest, many did not articu- ward the program. We found uniformly that late why they felt so. students were positive about the program, both overall and about each individual component. Does the Program Lead to Gains in Oral Contrary to our fears, students did not seem to Reading With Comprehension? get bored with repeated readings. Instead, they said the repeated readings led to greater mas- The results of our two-year study of fluen- tery of the material. When asked what they cy-oriented reading instruction suggests that liked about the program, the majority of the reorganizing instruction so as to stress fluency comments concerned the stories they were seems to have had positive effects on second- reading. This suggests that students' focus was grade children's growth as readers. These on the stories and not on the procedures used to effects were most pronounced on children teach the stories. Where they had complaints entering the second- grade year reading at a about the program, the majority dealt with the primer level or higher. Over the two years of length or difficulty of the material that was the program, all such children but two were read. There were some complaints about indi- reading at grade level or higher by the end of vidual components, but no component was the year. As might be expected, this program mentioned by more than one child. had its largest effects on measures of rate and When asked about individual components of accuracy in reading. Its effects on comprehen- the program, students felt positively about sion were significant because we used a mea- reading at home, about choice reading, and sure of oral reading with comprehension as a about partner reading. The same themes pretest and posttest. We do not report results seemed to emerge from questions about these from standardized comprehension measures various components. When asked about how all because of the difficulty of accessing such three components helped to improve their results. However, according to the teachers, reading, students tended to mention the impor- the effects on standardized reading comprehen- tance of practice to improve one's reading. sion tests were less pronounced and did not Social interaction, both with peers and parents, seem to differ from those of previous years.

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

45 32 Stahl, Heubach, & Cramond

What Happens During Partner Reading? process of reading at home positively, we felt that reading log data was not sufficiently reli- Partner reading was the only aspect of the able from which to draw conclusions. Students program mentioned by the children interviewed reported reading the story from the basal, with when asked what they liked about the reading a median of one reading and a range of 0 to 5. program, and it was the most positively per- We did find that students enjoyed reading at ceived. We found that having children choose home and believed that it made them better their partners rather than having them assigned on the basis of reading ability was the most readers. Some students enjoyed the opportunity efficacious approach. In our study, such an to read in quiet; others enjoyed the interaction arrangement did not lead to significantly higher with their family. When asked whether reading achievement because the oral reading levels at home made them more interested in reading, after repeated readings were very high for all nearly all children said "Yes," but generally arrangements.However,the management did not articulate why. aspects of partner reading seemed to go easier when there was a friendship between partners. What Types of Books Do Children Choose There were fewer disturbances, and off-task During Choice Reading? behaviors were more easily handled. We found, in two separate sets of inter- We had two concerns about the books views, that friendship was the primary reason children chose. First, we wanted students to that partners chose each other. Help with choose books at an appropriate level. For a reading was the secondary reason and, as might fluency-oriented program to work, children be expected, this reason was given by strug- gling readers. The social aspect of partner need to be practicing material at an appropriate reading appears to have mitigated the effects of level. Initially, we defined instructional level in children reading at levels well above or well the conventional manner, 95 % accuracy with below their instructional level. As noted above, acceptable comprehension (Wixson & Lipson, only about a quarter of the students in the 1991). We did find that children chose books second year of the program were reading at a that they could read at an appropriate level, second-grade instructional level at the begin- although our view of what such a level might ning of the year. Those who were reading be in a program like this has changed. Children above grade level enjoyed partner reading as a were choosing books that they could read with way of sharing interesting stories with a friend. an average accuracy level of 95 %, ranging, For those reading below grade level, the social with one exception, from 92 % to 100%.This aspect of partner reading made it easier to ask level of material seems to be appropriate for for and receive help. them to gain practice in reading connected text. What Are the Effects of Reading at Home? The one exception, as discussed above, chose his books for social reasons, because he was This question could not be answered with reading at a considerably lower level than the our data. Although students perceived the rest of his class.

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

46 Fluency-Oriented Reading Instruction 33

We also wanted to know what influenced Instructional Level? children's book choice. Here we found that the predominant influence was the teacher, not the These results suggest that children can other students. A book was more likely to be benefit from reading material well below the chosen if the teacher had read it aloud, if the 95 % accuracy rate traditionally recommended teacher had specifically recommended it, or if for instruction (Wixson & Lipson, 1991). In the teacher had stressed the importance of fact, students appeared to benefit from reading reading more difficult material, as was the case stories in the first sampling even though they in the two classes we studied. Where the teach- were reading them with an average accuracy er does not make such a recommendation, as rate of 85%, which would be considered the in the classes studied by Ivey and Heubach Frustration level. The reason why students (1994), children tend to read easier material. were able to benefit from reading material at The influence of the teacher may be more pro- these lower levels of accuracy was the higher nounced at the second-grade level and similar support they were given for the reading results may not be found in older grades. Also, through the routines of the program. In this these were all teacher-centered classes, where program, students were supported in their children have not been explicitly prepared for reading by having multiple exposures to the making choices. Where children are more same material, by having the stories read to accustomed to making choices, peers may have them, by exposure to the vocabulary prior to more influence. their own reading, by reading the story at home once or more, possibly by echo reading, What Level of Material Should Children Be Reading? and by partner reading. This high level of support is considerably greater than was pro- Because students generally choose books at vided in a traditional Directed Reading Activ- a 92% accuracy rate or higher, rather than the ity. traditional 98 %, we feel that this somewhat We would like to suggest that the instruc- more difficult rate should be thought of as the tional level for a given child is inversely relat- child's instructional level, at least in a program ed to the degree of support given to the reader. similar to this. This somewhat more difficult That is, the more support given, the lower the level has also been suggested by Clay (1985) accuracy level needed for a child to benefit and Powell (cited in Wixson & Lipson, 1991) from instruction. In classroom organizations and adopted by Wixson and Lipson as well. such as our fluency-oriented instruction, stu- We also have some evidence that children are dents can benefit from reading material at able to gain instructionally from somewhat greater relative difficulty because they are more difficult material than is traditionally given greater amounts of support for that assumed. This evidence comes from our obser- reading. vations of the effects of repeated reading on Another source of support for word recog- oral reading accuracy and rate. nition is pictures. Pictures in texts can improve

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79 47 34 Stahl, Heubach, & Cramond

children's word recognition (Denburg, 1976- first graders. The approach taken here is easy 77), at least while the picture is present (al- to implement, involves only classroom teach- though pictures can retard the development of ers, and works with second-grade children. context-free word recognition, since readers The effects of this program on children who may rely on illustrations for cues [e.g., Singer, initially read below the primer level were Samuels, & Spiroff, 1973-74]). Pictures can mixed. About half of these children made also aid in comprehension (see Schallert, 1980, adequate progress, the remainder did not. For for review). The use of picture books in early these children, the teachers made special adap- grades may support the ability of children to tations, including books with reduced vocabu- read material with lower numbers of words lary, providing extra time for reading, and so correctly recognized. This may account for the on. A program based on repeated readings of lower criteria for instructional level observed grade-level material requires a certain initial by Clay (1985) and Powell (cited in Wixson & level of competence. For those without such Lipson, 1991) for primary grade readers. Such competence, more intensive remediation is readers tend to read more heavily illustrated required. material, which indicates they are less reliant Since our struggling readers had more on knowing the words to read the text compe- exposure to the materials, through additional tently. readings at home and through some additional work in class, they were able to read materials What Are the Effects of the Program on of much greater than expected difficulty. In Struggling Readers? turn, the reading of more difficult material aided their growth as readers, allowing them to The most pronounced effects of this pro- read the second-grade material with more ease. gram were on children who were struggling; This seems to be the opposite process .to that that is, those reading above the primer level but involved in Matthew Effects (Stanovich, 1986). not at second-grade level. As noted earlier (see Stanovich suggests that struggling readers, Figures 3 and 5), all children reading at a because they read relatively easy material and primer level or higher at the beginning of the read less of it than proficient readers, fall year were reading at the second-grade level by further and further behind their better achiev- the end of the year. In ordinary classroom ing peers. Instead, we suggest that our class- situations, these children will fall further and room organization provides a mechanism for at further behind the average for their grade (Juel, least some children to catch up with their 1988; Stanovich, 1986). Programs that have peers. successfully accelerated the growth of these readers have been either fairly expensive and What Have We Learned? difficult to implement, like Reading Recovery (Clay, 1985) or Success for All (Madden, This paper has presented a complex evalua- Slavin et al., 1997), or have been directed to tion of a complex program, an attempt to re-

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

48 Fluency-Oriented Reading Instruction 35

organizesecond-gradereadinginstruction Anderson, R. C., Wilkinson, I. A. G., & Mason, around a set of theory-derived principles. For J. M. (1990). A microanalysis of the small-group the most part, this reorganization was success- guided reading lesson: Effects of an emphasis on ful in achieving its goals. The program was global story meaning. Unpublished paper, Uni- sustainable over two years, teachers and chil- versity of Illinois at Urbana Champaign. dren perceived it and its various components Anderson, R. C., Wilson, P. T., & Fielding, L. G. positively, and it lead to overall gains in (1988). Growth in reading and how children spend their time outside of school. Reading achievement. These gains were found for all Research Quarterly, 23, 285-303. children reading at a primer level or higher Campbell, D. T., & Stanley, J. C. (1966). Experi- initially and for about half of those who could mental and quasi-experimental designs for not initially read a primer passage. research. Chicago: Rand-McNally. We also learned about the reciprocal nature Carver, R. P., & Leibert, R. E.(1995).The of instruction and text difficulty. The tradition- effect of reading library books at differing levels al notion of instructional level, based as it was of difficulty upon gain in reading ability. Read- on a traditional notion of instruction, seems not ing Research Quarterly, 30, 26-48. to be relevant to this type of classroom setting. Chall, J. S. (1983). Stages of reading development. Instead, with the greater support given to New York: McGraw-Hill. readers through repeated readings of the in- Chall, J. S., Jacobs, V., & Baldwin, L. (1990). The structional text in various venues and with reading crisis. Cambridge, MA: Harvard Uni- various procedures, children were able to learn versity Press. from material that they initially read with Chomsky, C. (1978). When you still can't read in third grade? After decoding, what? In S. J. greater difficulty than expected. This program Samuels (Ed.), What research has to say about provides that structure in a form easily usable reading instruction (pp. 13-30). Newark, DE: by teachers and responded to by students. International Reading Association. Clay, M. M. (1985). The early detection of reading References difficulties. Portsmouth, NH: Heinemann. Denburg, S. D. (1976-77). The interaction of Adams, M. J. (1990). Beginning to read: Thinking pictures and print in reading instruction. Read- and learning about print. Cambridge, MA: ing Research Quarterly, 12, 176-189. M.I.T. Press. Doehring, D. G., & Aulls, M. W. (1979). The Allington, R. L. (1983a). Fluency: The neglected interactive nature of reading acquisition. Jour- reading goal. The Reading Teacher, 37, 556- nal of Reading Behavior, 11, 27-40. 561. Dowhower, S. L. (1989). Repeated reading: Re- Allington, R. L. (1983b). The reading instruction search into practice. The Reading Teacher, 42, provided readers of differing reading abilities. 582-587. The Elementary School Journal, 83, 549-559. Downing, J. (1979). Reading and reasoning. Edin- Allington, R. L. (1984). Oral reading. In P. D. burgh, UK: W. & C. Black. Pearson, R. Barr, M. L. Kamil, & P. Mosenthal Durkin, D. (1974). A six-year study of children (Eds.), Handbook of reading research (pp. who learned to read in school at the age of four. 829-864). White Plains, NY: Longman. Reading Research Quarterly, 10, 9-61.

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

4 9 36 Stahl, Heubach, & Cramond

Gambrell, L. B., Wilson, R. M., & Gantt, W. N. Mudre, L. H., & McCormick, S. (1989). Effects of (1981). Classroom observations of task-attending meaning-focused cues on underachieving read- behaviors of good and poor readers. Journal of ers' context use, self-corrections, and literal Educational Research, 74, 400-404. comprehension. Reading Research Quarterly, Glaser, B. G., & Strauss. A. L. (1967). The dis- 24, 89-113. covery of grounded theory: Strategies for quali- Nell, V. (1988). Lost in a book: The psychology of tative research. Chicago: Aldine. reading for pleasure. New Haven, CT: Yale Heath, S. B. (1983). Ways with words. Cambridge, University Press. UK: Cambridge University Press. O'Masta, G., & Wolf, J. (1991). Encouraging Hoffman, J. (1987). Rethinking the role of oral independent reading through the reading million- reading. Elementary School Journal, 87, 367- aires club. The Reading Teacher, 44, 656-662. 373. Ivey, M. G. & Heubach, L. (1994, December). Rasinski, T.V. (1989). The effects of repeated Self-selection in free reading: Examining book reading and repeated listening while reading on difficulty and factors influencing choice in Grade reading fluency. ERIC Document Reproduction 2. Paper presented at annual meeting, National Service ED 384 666. Reading Conference, San Diego, CA. Rasinski, T. V. (1991). Fluency for everyone: Juel, C. (1988). Learning to read and write: A Incorporating fluency instruction in the class- longitudinal study of fifty-four children from room. The Reading Teacher, 43, 690-692. first through fourth grade. Journal of Education- Rasinski, T.V., Padak, N., Linek, W., & Sturte- al Psychology, 80, 437-447. vant, E. (1994). Effects of fluency development Leslie, L., & Caldwell, J. (1988). Qualitative on urban second grade readers.Journal of reading inventory. New York: HarperCollins. Educational Research, 87, 158-165. Madden, N., Slavin, R. E., Karweit, N. L., Do- Routman, R. (1991). Transitions. Portsmouth, NH: lan, L. J & Wasik, B. A. (in press). Reading, Heinemann. writing, and tutoring in Success for All, in S. A. Samuels, S. J. (1985). Automaticity and repeated Stahl & D. A. Hayes (Eds.), Instructional reading. In J. Osbom, P. T. Wilson, & R. C. models in reading. Hillsdale, NJ: Erlbaum. Anderson (Eds.), Reading education: Founda- Martin, C. E. (1984). Why some gifted children do tions for a literate America (pp. 215-230). not like to read. Roeper Review, 7,72-75. Lexington, MA: Lexington Books. Martinez, M., & Roser, N. (1985). Read it again: Samuels, S. J., Schermer, N., & Reinking, D. The value of repeated readings during storytime. (1992). Reading fluency: Techniques for making The Reading Teacher, 38, 782-786. decoding automatic. In S. J. Samuels & A. E. McCormick, C. E., & Mason, J. M. (1986). Inter- vention procedures for increasing preschool Farstrup (Eds.), What research says about children's interestin and knowledge about reading instruction, 2nd edition (pp. 124-144). reading. In W. H. Teale & E. Sulzby (Eds.), Newark, DE: International Reading Association. Emergent literacy: Writing and reading (pp. Schallert, D. L. (1980). The role of illustrations in 90-115). Norwood, NJ: Ablex. reading comprehension. In R. J. Spiro, B. C. Morris, D., & Nelson, L. (1992). Supported oral Bruce, & W. F. Brewer (Eds.), Theoretical reading with low achieving second graders. issues in reading comprehension (pp. 503-524). Reading Research and Instruction, 32, 49-63. Hillsdale, NJ: Erlbaum.

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

50 Fluency-Oriented Reading Instruction 37

Singer, H., Samuels, S. J., & Spiroff, J. (1973- 74). The effects of pictures and contextual conditions on learning responses to printed words. Reading Research Quarterly, 9, 555- 567. Stanovich, K. E. (1980). Toward an interactive- compensatory model of individual differences in the development of reading fluency. Reading Research Quarterly, 16, 32-71. Stanovich, K. E. (1986). Matthew effects in read- ing: Some consequences of individual differ- ences in the acquisition of literacy. Reading Re- search Quarterly, 21, 360-407. Taylor, B. M., Frye, B. J., & Maruyama, G. M. (1990). Time spent reading and reading growth. American Educational Research Journal, 27, 351-362. Taylor, N. E., Wade, M. R., & Yekovich, F. R. (1985).The effects of text manipulation and multiple reading strategies on the reading per- formance of good and poor readers. Reading Research Quarterly, 20, 566-585. Topping, K. (1987). Paired reading: A powerful technique for parent use. The Reading Teacher, 40, 608-614. Vacca, R., & Rasinski, T. (1992). Case studies in whole language. New York: Harper Collins. Wixson, K. K., & Lipson, M. Y. (1991). Reading disability. New York: Harper Collins.

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

51 38 Stahl, Heubach, & Cramond

APPENDIX

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 79

52 Interview Questions

Name Class

I want to talk to you about the reading program in your class this year. We need to know about what you think about your reading program and the parts of it.

1. How do you like reading in your class?

2. What do you like best about reading in your class?

3. What do you like least about reading in your class?

4. Do you think that you have learned to read better this year?

5. What do you think about when the teacher reads to you from your reading book?

a. Do you think that this helps you read?

b. Do you think that this makes you more interested in reading?

6. What does the teacher usually do afterwards?

a. What do you think about this?

b. Do you think that this helps you read?

c. Do you think that this makes you more interested in reading?

7. Do you practice reading the story at home?

53 a. About how many times? (If yes, above)

b. What do you think about this?

c. Do you think that this helps you read?

d. Do you think that this makes you more interested in reading?

8. What do you think about partner reading?

a. Do you think that this helps you read?

b. Do you think that this makes you more interested in reading?

c. Who do you read with most often?

d. Why did you choose that person?

e. Do you enjoy reading with

f. How well does read?

9. How often do you have DEAR time?

a. What do you think about this?

b. Do you think that this helps you read?

c. Do you think that this makes you more interested in reading?

54 MCNational rS Reading Research Center 318 Aderhold, University of Georgia, Athens, Georgia 30602-7125 3216J. M. Patterson Building, University of Maryland, College Park, MD 20742

55 U.S. DEPARTMENT OF EDUCATION Office of Educational Research and Improvement (OERI) Educational Resources Information Center (ERIC) ERIC

NOTICE

REPRODUCTION BASIS

This document is covered by a signed "ReproductionRelease (Blanket)" form (on file within the ERIC system), encompassingall or classes of documents from its source organization and, therefore, does not require a "Specific Document" Release form.

This document is Federally-funded, or carries itsown permission to Er reproduce, or is otherwise in the public domain and, therefore,may be reproduced by ERIC without a signed Reproduction Release form (either "Specific Document" or "Blanket").

(9/92)