How Text Matters

In this quasi-experimental study, which is part of a series of investigations on supplemental reading tutoring variations, the relative effectiveness of more intense decoding instruction or text reading practice was examined. Fifty-seven first-grade students scoring in the lowest quartile for reading skills received either classroom reading instruction or one of two treatments: tutoring in word study with text reading practice, or word study tutoring alone. Individual instruction was provided by trained paraprofessional tutors. At the end of first grade, treatment students significantly outperformed their nontutored peers on measures of reading accuracy, reading comprehension, reading efficiency, passage reading fluency, and spelling. Differential treatment effects on passage reading fluency are examined, taking into consideration pretest skill levels and text reading practice characteristics. Word identification is the major obstacle for students with reading problems, and the ability to accurately and quickly identify words is essential for reading success. One critical condition for word identification is phonological awareness, or knowing that words are composed of individual sounds or phonemes. A second prerequisite is alphabetic understanding, or knowing that individual letters in words correspond to sounds and being able to map print onto speech. Early interventions in phonological and explicit phonics instruction have been shown to help students at risk for reading disabilities (RD) and are widely recommended components of effective classroom reading instruction (Adams, 1990; Armbruster, Lehr, & Osborne, 2001; National Reading Panel, 2000). Systematic phonics instruction is also a feature of many supplemental interventions for at-risk beginning readers (Berninger & Traweek, 1991; Blachman, Tangel, Ball, Black, & McGraw, 1999; Foorman, Francis, Fletcher, Schatschneider, & Mehta, 1998; D. Fuchs et al., 2001; Iversen & Tunmer, 1993; O’Connor, 2000; Torgesen, Alexander, et al., 2001; Torgesen, Wagner, Rashotte, Alexander, & Conway, 1997; Vellutino et al., 1996), including one-toone tutoring (see Elbaum, Vaughn, Hughes, & Moody, 2000). Students who struggle with word identification often have difficulty becoming fluent readers. Reading practice becomes increasingly important as children acquire complex phoneme awareness and high-quality word representations (Stahl & Murray, 1998). Most early reading interventions include some type of reading practice, including word list and oral storybook reading (Baker, Gersten, & Keating, 2000; Elbaum et al., 2000; Foorman et al., 1998). In fact, many literacy programs schedule the majority of tutoring time for book reading (Pinnell & Fountas, 1997). Tutors often select from lists of leveled books for practice with their students (Blachman, Schatschneider, Fletcher, & Clonan, 2003; Johnston, Invernizzi, & Juel, 1998). Text reading practice is included in supplementary interventions to foster a love of reading and to develop vocabulary and comprehension (National Reading Panel, 2000). Reading practice reinforces decoding and word-level reading skills in authentic connected text, allowing students to develop the fluency required to construct meaning from texts (National Research Council, 1998). The focus of this article—the nature and level of phonics instruction and oral reading practice provided in supplementary tutoring interventions— is important for several reasons. First, reading practice is the central activity in most tutoring programs (Elbaum et al., 2000). Second, opportunities for classroom oral reading practice are often limited, especially for the lowest skilled readers (Chard & Kame’enui, 2000; Haynes & Jenkins, 1986). Supplementary tutoring may offer the only significant period of oral reading practice and intensive phonics instruction for many at-risk students. OpportuniVOLUME 38, NUMBER 4, JULY/AUGUST 2005 365 ties for supplementary tutoring are difficult to create in schools, and therefore tutoring activities must be carefully selected for ease and reliable use by tutors and evidence of treatment effectiveness. This article describes a study that compared two forms of phonicsbased tutoring designed to promote transfer of learning by children at risk for RD. Our particular concern was the transfer of skill gained from wordlevel or text-level reading practice scaffolded by a tutor. Role of Reading Practice As students begin to acquire decoding skills, they typically engage in two kinds of reading practice: reading individual words and reading connected text. Word study is the repeated reading of word lists, often combined with instruction in phonological and phonics skills. For example, the tutor may introduce letter–sound correspondences and have students practice reading words that feature the new correspondences. The tutor may introduce new sight words, have students practice reading and spelling these new words, and review previously taught sight words. Training in individual words improves the quality of lexical representations, leading to automatic activation (Perfetti, 1992) and improved accuracy and speed in reading connected text. Research shows that practice in reading single words increases speed and accuracy in reading the trained words, with inconclusive results for transfer to comprehension (Fleisher, Jenkins, & Pany, 1979; Levy, Abello, & Lysynchuk, 1997; Tan & Nicholson, 1997). Text-level training transfer to new materials appears to be mediated by reader skill and text type and difficulty. Faulkner and Levy’s (1994, 1999) research identified factors that influence the benefits of reading connected text. Good and average readers showed fluency and accuracy transfer to a new text when the two texts shared content, but not when there was a high word overlap only. Poor readers, however, showed the most transfer in speed and accuracy of processing in a second text when there was high word overlap only, and they appeared to benefit from repeated exposure to individual words. Context and content appear most beneficial for readers who already have fluent and accurate word reading skills and who can focus on content rather than word identification. Repeated reading appears to help poor readers to learn more words, and readers lacking automatic word recognition skills— whether due to reader skill or text difficulty—benefit from rereading word lists. Prior exposure to words outside of a meaningful context produced transfer to reading the same words presented in an intact text for poor readers but not for good readers (Faulkner & Levy, 1999). For nonfluent readers, transfer was observed at the single-word level, and text difficulty interacted with reader skill in reading transfer. Studies by Rashotte and Torgesen (1985) and Dohower (1987) have also shown that poor readers benefit from repeated readings with a high word overlap across passages, with transfer to new texts due to improved identification of individual words. Repeated reading of text is recommended as an effective method to develop accuracy and fluency as well as comprehension (National Reading Panel, 2000). Although many first-grade phonics-based classroom and tutoring interventions include text reading practice (see reviews by Bus & van IJzendoorn, 1999; Ehri, Nunes, Stahl, & Willows, 2001; Elbaum et al., 2000), oral reading fluency outcomes for first graders are often overlooked (L. S. Fuchs, Fuchs, Hosp, & Jenkins, 2001). Bus and van IJzendoorn (1999) reported that for the subset of nine randomized preschool– primary studies of phonological awareness training that included reading activities, the reading outcome effect (aggregating measures for reading words and sentences) was somewhat lower (d = .88) than for purely phonetic training programs (d = .96). However, text reading fluency was not explicitly examined or disaggregated from the combined reading outcome variable category. In Ehri et al.’s (2001) review of systematic first-grade classroom phonics instruction (most of which included text reading practice), effects for reading text orally (aggregating both accuracy and fluency outcomes) averaged d = .23, whereas the effects for decoding regular words averaged d = .98. Finally, in their meta-analysis of one-to-one tutoring interventions that often featured phonics instruction and text reading (28 of the 42 study samples included first graders only), Elbaum et al. (2000) reported an average effect size of d = .48 for oral passage reading. These findings underscore the challenge of closing the fluency gap for children with RD, which is often widened by growing differences in reading practice (Torgesen, Rashotte, & Alexander, 2001). Text Type and Difficulty Two considerations in choosing texts for reading practice are text type and difficulty. In their review of fluency practices, Kuhn and Stahl (2000) concluded that more difficult texts were associated with stronger reading effects. Faulkner and Levy (1994) found that poor readers who read difficult text seemed to focus more on individual words rather than on text content and showed transfer when texts shared words rather than content. One widely used measure of text difficulty, in particular for beginning readers, is decodability. Decodable text can be defined as text with “1) a proportion of words with phonically regular relationships between letters and sounds, and 2) a degree of match between the letter/ sound relationships represented in text and those the reader has been taught” (Mesmer, 2001, p. 122). Decodable text is implicitly recommended in systematic phonics instruction to encourage students to practice taught decoding strategies (Armbruster et al., 2001), and its use is explicitly required in many recent state reading adoptions (Allington JOURNAL OF LEARNING DISABILITIES 366 & Woodside-Jiron, 1998). Decodability includes features such as word regularity, frequency, complexity, and lesson-to-text match, and its definition continues to be a matter of discussion (Hiebert, 2002; Mesmer, 2001). Although decodable text is often recommended for beginning readers and students with reading problems (Adams, 1990; Anderson, Hiebert, Scott, & Wilkinson, 1985; Hiebert, 1998; Mesmer, 1999), there is limited research to support its use. One of the few studies of decodable text was conducted by Juel and Roper-Schneider (1985). Using a quasi-experimental design, they followed two groups of first graders who received the same scripted, synthetic phonics instruction. One group read from a phonics-oriented basal series emphasizing decodable text, and the other group read from a basal series emphasizing high-frequency words. Text type appeared to influence students’ word reading strategies in the first two thirds of the year, at which testing point the decodable group scored significantly higher than the high-frequency group on decoding skills. Although at the end of the year there were no significant differences between the groups on tests of decoding, reading comprehension, or reading vocabulary, the decodable group scored significantly higher on a switch list of words that were not common to both basal programs and was most influenced in word reading ability by word regularity, whereas the highfrequency group appeared most influenced by word repetition. The authors concluded that strategy use during reading was strongly influenced by the type of text used. In this regard, decodable text has been recommended to reinforce the application of word-level decoding skills. Oral reading practice in decodable text may help students to develop a selfteaching strategy (Jorm & Share, 1983; Share & Jorm, 1987); students’ successful decoding encounters in text provide opportunities to form orthographic representations that underlie skilled reading. As Ehri’s (Ehri & Sweet, 1991; Ehri & Wilce, 1985, 1987) work has demonstrated, even beginning readers with limited letter-sound knowledge and phonemic awareness possess the selfteaching skills to learn words. Text considerations like decodability appear to be more important for less skilled students who require more scaffolded literacy experiences. Two studies have suggested that decodable text is most useful for readers who do not yet apply letter–sound relationships to recognize words. In Juel and Roper-Schneider’s (1985) study, text type influenced strategy use during the first two units of the year, when it is most important for students to attend to word features. In a brief, 14day, first-grade phonics intervention comparing the contribution of reading practice for highly decodable and less decodable text, Mesmer (2005) also found that text decodability supported the application of letter–sound relationships for first graders at the partial alphabetic phase of word recognition and enabled students’ progress to the full alphabetic phase. Other reading researchers and practitioners (Brown, 1999/2000; Mesmer, 1999) have documented in teacher case studies that the use of decodable text (a) offers students practice in letter–sound correspondences they have been taught, (b) enables students to respond to these letter patterns automatically, and (c) helps students move into full alphabetic phase reading (Ehri & McCormick, 1998). Decodable text may operate as a temporary support, allowing students to rapidly associate letters with sounds and to generalize knowledge of taught word features to novel words. Text and instruction that are well integrated may be most effective in scaffolding the development of beginning word reading skills (Gaskins et al., 1997; Mesmer, 1999).