Reading and Writing

, Volume 21, Issue 4, pp 349–369

Explicit instruction in phonemic awareness and phonemically based decoding skills as an intervention strategy for struggling readers in whole language classrooms

Authors

  • Janice F. Ryder
    • College of EducationMassey University
    • College of EducationMassey University
  • Keith T. Greaney
    • College of EducationMassey University
Article

DOI: 10.1007/s11145-007-9080-z

Cite this article as:
Ryder, J.F., Tunmer, W.E. & Greaney, K.T. Read Writ (2008) 21: 349. doi:10.1007/s11145-007-9080-z

Abstract

The aim of this study was to determine whether explicit instruction in phonemic awareness and phonemically based decoding skills would be an effective intervention strategy for children with early reading difficulties in a whole language instructional environment. Twenty-four 6- and 7-year-old struggling readers were randomly assigned to an intervention or control group, with the intervention group being divided into four groups of three children each. The intervention program was carried out over a period of 24 weeks and comprised 56 highly sequenced, semi-scripted lessons in phonemic awareness and alphabetic coding skills delivered by a teacher aide who received training and ongoing support from a remedial reading specialist. Posttests results showed that the intervention group significantly outperformed the control group on measures of phonemic awareness, pseudoword decoding, context free word recognition, and reading comprehension. Two-year follow-up data indicated that the positive effects of the intervention program were not only maintained but had generalized to word recognition accuracy in connected text.

Keywords

Phonemic awarenessDecoding skillsReading difficultiesReading intervention

Introduction

Shankweiler and Fowler (2004) recently reviewed research with the purpose of providing answers to recurring questions about the role of phonological processes in learning to read. They drew several conclusions, two of which are particularly relevant to the present study: first, “phoneme awareness is key to reading an alphabetic system” (p. 488), and second, “explicit, systematic instruction in the code relating spellings to pronunciations is necessary for most children” (p. 493). Similarly, from an examination of findings from a wide array of sources (e.g., studies of reading development, specific instructional practices, and effective teachers and schools), Snow and Juel (2005) concluded that explicit attention to alphabetic coding skills in early reading instruction is “helpful for all children, harmful for none, and crucial for some” (p. 518).

Snow and Juel further noted that “the development of detailed orthographic representations is vital to the automatization of word recognition” (p. 511). Research has established that making use of letter–sound relationships to identify unfamiliar words is the basic mechanism for acquiring sight word (i.e., word-specific) knowledge, including knowledge of irregularly spelled words (Adams, 1990; Adams & Bruck, 1993; Ehri, 1992, 1997, 2005; Gough & Walsh, 1991; Harm & Seidenberg, 1999; Hulme, Snowling, & Quinlan, 1991; Perfetti, 1992; Share, 1995; Tunmer & Chapman, 1998, 2006). The formation of sub-lexical, visuophonological connections between printed words and their spoken counterparts in memory provides the basis for fast, efficient recognition of words, which in turn frees up cognitive resources for allocation to sentence comprehension and text integration processes (Perfetti, 1985; Pressley, 2006; Share, 1995).

A key question that remains, however, is the extent to which explicit instruction in phonemic awareness and phonemically based decoding strategies is needed by students. It is not possible to teach students all, or even most, of the letter–sound relationships of English orthography. In their influential theory of reading acquisition, Gough and Hillinger (1980) pointed out that there are simply too many letter–sound relationships for children to acquire by direct instruction. Rather, as the reading attempts of beginning readers with a firm understanding of the alphabetic principle become more successful, they will begin making greater independent use of letter–sound information (possibly supplemented by sentence context cues) to identify unfamiliar words in text. Phonologically recoding specific words a few times ultimately cements the orthographic representations of the words in lexical memory from which additional spelling-sound relationships can then be induced without explicit instruction (Share, 1995). Snow and Juel (2005) claimed that “phonics may be useful to children not because of the specific letter–sound relations taught, but because a phonics approach gives children the chance to discover the alphabetic principle, and provides practice looking closely at word spelling” (p. 516).

Although children must rely on induction to acquire the spelling-sound relations necessary for learning to read, the amount of explicit phonics instruction required to “kick-start” the process of “self-teaching” (Share, 1995) may vary considerably across children. As Snow and Juel (2005) pointed out, “perhaps the key feature of classroom instruction ensuring that one child learns to read is systematic progress through a long list of letter–sound correspondence rules, whereas for another child the key feature is teaching just a few phonics rules but providing lots of interesting and increasingly difficult books for practice and self-teaching” (p. 517). For children encountering difficulty in developing the ability to intuitively perceive redundant patterns and connections between speech and print, explicit instruction in phonemic awareness and alphabetic coding skills is likely to be critical.

Consistent with this suggestion is a considerable amount of research indicating that children enter school with large individual differences in the experiences and competencies important in learning to read (Nicholson, 2003; Tunmer, Chapman, & Prochnow, 2003; Whitehurst & Lonigan, 2001). These reading-related skills include receptive vocabulary (promoted by shared reading and verbal interaction in the home), familiarity with “book” or “decontextualized” language and basic understanding of concepts and conventions of printed language (also promoted by adult storybook reading), knowledge of letter names and sounds (developed by exposure to ABC books and games), ability to produce preconventional spellings of words (developed by manipulating movable letters to form “invented” spellings; e.g., writing color as KLR, or fairy as FRE), and sensitivity to the subcomponents of spoken words (promoted by playing rhyming and sound analysis games and being read books that increase phonological sensitivity; e.g., pig latin, I spy, nursery rhymes, Dr. Seuss books). Children who possess higher levels of these skills at school entry profit more from reading instruction, learn to read sooner, and read better than children who do not (Whitehurst & Lonigan, 2001).

Of the various reading-related skills that develop in prereading children, phonemic awareness (the ability to reflect on and manipulate the phonemic elements of spoken words) is one of the most important (Shankweiler & Fowler, 2004). New school entrants who experience ongoing difficulties in detecting phonemic sequences in words will not be able to fully grasp the alphabetic principle and discover spelling-to-sound relationships. Without specific intervention, the development of word recognition skills in these children will be impeded, resulting in negative Matthew (or poor-get-poorer) effects in reading achievement (Stanovich, 1986).

The present study was carried out in New Zealand, which follows a predominantly constructivist, whole language approach to reading instruction and intervention in which literacy learning is largely seen as the natural by-product of active mental engagement (Wilkinson, Freebody, & Elkins, 2000). As Stanovich (1994) noted, this instructional approach assumes “that self-discovery is the most efficacious mode of learning, that most learning can be characterized as ‘natural’ and that cognitive components should never be isolated/fractionated during the learning process” (p. 264).

From the assumption that the ability to read evolves naturally and spontaneously out of children’s prereading experiences with “environmental print” (commonly occurring environmental labels accompanied by context or logos, such as the word stop appearing on an octagonally-shaped sign), whole language theorists concluded that literacy teaching should be modelled on first-language acquisition, where the focus is on meaning construction, not the abstract structural units that provide the basis for mapping print onto spoken language. If children are immersed in a print-rich environment in which the focus is on the meaning of print, they will readily acquire reading skills, according to this view. Children can be taught what they need to know to learn to read “as the need arises” through frequent encounters with absorbing reading materials. The focus of this approach, then, is on learning to read by reading, with minimal attention being given to the development of phonemically based, word-level skills and strategies. As Smith and Elley (1994), two leading proponents of the constructivist approach to reading education in New Zealand, claimed, “children learn to read themselves; direct teaching plays only a minor role” (p. 87). In Byrne’s (2005) terminology, this approach to reading instruction assumes that the processes of learning to read are highly learner dependent, with children relying on induction to a very large degree.

Another key aspect of the constructivist approach to literacy education is the assumption that reading acquisition is primarily a process in which children learn to use multiple cues (syntactic, semantic, visual, graphophonic) to predict the next words in text (Snow & Juel, 2005; Tracey & Morrow, 2006; Tunmer & Chapman, 2002). The latest handbook for beginning reading teachers in New Zealand, Effective Literacy Practice in Years 1–4 (Ministry of Education, 2003a), recommends teaching children to identify unfamiliar words in text by encouraging them to use all sources of information (knowledge and experience, semantic sources of information, syntactic sources of information, and visual and graphophonic sources of information) simultaneously in predicting, cross-checking, confirming, and self correcting as they read text (pp. 28–31, 130). The major shortcoming of this instructional philosophy, however, is that it stresses the importance of using information from many sources in identifying unfamiliar words without recognizing that skills and strategies involving phonological information are of primary importance in beginning literacy development (Perfetti, 1985; Pressley, 2006; Shankweiler & Fowler, 2004).

Although Effective Literacy Practice acknowledges the importance of phonemic awareness and phonics, it recommends that these skills should largely be taught through reading and writing activities, as described in Sound sense: Phonics and Phonological Awareness (Ministry of Education, 2003b), where it is stated that, “children are more likely to make connections between phonics and their reading and writing of texts if they are engaged and involved in making discoveries for themselves” (p. 7). Similarly, in another widely distributed Ministry of Education publication, Picking up the Pace: A Summary (Keith, 2002), it is stated that instruction in alphabetic coding skills should occur “overwhelmingly” in the context of real reading. According to this view, word analysis skills should only be taught incidentally as “mini-lessons” given in response to children’s oral reading errors during text reading because alphabetic coding skills acquired in isolation are of “limited versatility” (McNaughton, 2002).

In opposition to this view is a large body of research indicating that the degree of explicitness and detail with which word-level skills and strategies are taught is important for most beginning readers and especially so for struggling readers (Adams, 1990; Ehri, 2004; Oakhill & Beard, 1999; Perfetti, 1991; Pressley, 2006; Shankweiler & Fowler, 2004; Snow & Juel, 2005). There are several advantages of providing beginning and struggling readers with explicit and systematic instruction in orthographic patterns and word identification strategies outside the context of reading connected text rather than teaching word analysis skills only as the need arises during text reading. First, instruction in word analysis skills that is deliberately separated from meaningful context allows children to pay full attention to the letter–sound patterns that are taught (Morris, Tyner, & Perney, 2000). Second, this instructional approach helps children to learn word decoding skills that may be useful in reading all texts, not just a specific text. That is, acquiring detailed knowledge of letter–sound patterns is generative; it provides a way of identifying words not seen before. Third, including isolated word study in beginning and remedial reading programs helps struggling readers to overcome their tendency to rely on ineffective word identification strategies, such as using picture cues and sentence context cues to identify unfamiliar words in text rather than using these cues to supplement word-level information. Fourth, such instruction also helps to ensure that beginning readers see the importance of focusing on word-level cues as the most useful source of information in identifying unfamiliar words.

A largely unidimensional (i.e., “one-size-fits-all”), constructivist approach to reading instruction (in which word analysis activities, if any, arise primarily from the child’s responses during text reading) may be suitable for most children who possess high levels of reading-related knowledge, skills, and experiences (or literate cultural capital; see Tunmer et al., 2003; Tunmer, Chapman, & Prochnow, in press) at school entry. However, we contend that struggling and at-risk readers will almost always benefit more from reading instruction that involves explicit and systematic teaching in alphabetic coding skills both in isolation from reading connected text and in combination with plenty of opportunities to practice and receive feedback on applying their developing word identification skills and strategies during text reading (Greaney, Tunmer, & Chapman, 1997; Hatcher, Hulme, & Ellis, 1994; Iversen & Tunmer, 1993; Iversen, Tunmer, & Chapman, 2005; Tunmer et al., 2003). Byrne (2005) describes the processes of learning to read in children who require a fairly structured and teacher-supported introduction to reading as environment-dependent.

In support of this claim about differences in learning processes (i.e., learner dependent versus environment-dependent) among beginning readers are the results of recent studies indicating an ability by treatment interaction in which children with limited literacy-related skills and knowledge at school entry benefit more from code-emphasis approaches to beginning reading instruction than from whole language/book experience approaches, whereas the opposite pattern occurs with children who have high levels of literacy-related abilities at the beginning of school (Connor, Morrison, & Katch, 2004; Juel & Minden-Cupp, 2000; Xue & Meisels, 2004). As Juel and Minden-Cupp (2000) noted from the findings of their study, “children who entered first grade with few literacy skills benefited from a heavy dose of phonics”, whereas children who entered first grade with comparatively high levels of literacy and literacy-related skills “did exceptionally well in a classroom that included a less structured phonics curriculum and more reading of trade books and writing of text” (p. 484).

Further evidence pointing to important differences in literacy learning processes among new school entrants comes from the results of three studies carried out in New Zealand. In the first study, Tunmer et al. (2003) found that incorporating into beginning literacy programs supplementary materials and procedures designed to help at risk children (i.e., those with limited literate cultural capital at school entry) develop awareness of the sound components of spoken words and make greater use of letter–sound patterns in reading unfamiliar words produced significantly greater gains in reading achievement for these children than the standard whole language approach to literacy instruction. In the second study, Greaney et al. (1997) found in a carefully designed project that even children with persistent reading problems (in Years 3 through 6), most of whom had been referred on from Reading Recovery (the nationally implemented, whole-language oriented, early intervention program in New Zealand; Clay, 1993), were able to derive long-term benefits from explicit instruction in alphabetic coding skills. Thirdly, Chapman, Tunmer and Prochnow (2001) found in a longitudinal study of beginning literacy development that struggling readers selected by their schools for Reading Recovery (after one year in school) were, without exception, experiencing major difficulties in detecting phonemic sequences in words (i.e., phonemic awareness) and in relating letters to sounds (i.e., alphabetic coding) during the year preceding entry into the program, that participation in Reading Recovery did not appreciably reduce these deficiencies, and that the failure to remedy these problems severely limited the immediate and long-term effectiveness of the program (see Tunmer & Chapman, 2003, for a critique of Reading Recovery). Given the negative outcomes for Reading Recovery and the implications of the distinction between learner dependent and environment-dependent literacy learning processes, the aim of this study was to determine whether explicit instruction in phonemic awareness and phonemically based decoding skills would be an effective intervention strategy for children with early reading difficulties in a whole language instructional environment.

Method

Sample and design

Children in New Zealand commence school on or around their fifth birthday, and formal reading instruction begins at that time. The children who participated in the study were selected from a larger pool of 64 6- and 7-year old, native English-speaking children from four separate Year 2 and Year 3 classrooms in a primary school with a total student population of 264, comprising 64% European, 34% Maori, 1% Pacific Islander and 1% Chinese. New Zealand schools are rated from a decile of 1 (low) to 10 (high) according to the socioeconomic community the school serves. The school participating in this study had a decile rating of 3, reflecting a mixture of middle and low income families in the catchment area of the school. Eight children who were receiving special education assistance or Reading Recovery at the school were excluded from the study.

Early in the school year the 64 children in the larger pool were given the Burt Word Reading Test, New Zealand Revision (Gilmore, Croft, & Reid, 1981). From the subgroup of 24 children whose Burt raw scores were the lowest, 12 closely matched pairs were formed, and the children in each pair were randomly assigned to an intervention or control group. The children in the intervention and control groups were distributed approximately equally across the four classrooms. This procedure resulted in three Maori and nine European children in each group. The low reading scores of the 24 target children were corroborated by poor performance on classroom reading tasks. All but four of these children had already been identified as struggling readers and placed in the lowest reading groups by their classroom teachers.

The children were then individually administered tests of phonemic awareness (Robertson & Salter, 1997), phonological decoding ability (Richardson & DiBenedetto, 1985), accuracy of recognizing words in connected text (Neale, 1988), and reading comprehension (Neale, 1988). Although the children in the intervention and control groups were initially matched on context free word recognition ability (as measured by the Burt Test, which yielded mean reading ages of less that 6 years for both groups), the means presented in Table 1 indicate that this procedure also resulted in close matches in reading accuracy (i.e., word recognition in connected text) and reading comprehension. As expected, t-tests of differences between the means of the two groups for each of these variables and age revealed no significant differences, t (22) < 1.0; p > 0.05.
Table 1

Pre- and posttest means and standard deviations for the control and intervention groups on all measures

Measure

Maximum score

Intervention group (n = 12)

Control group (n = 12)

Pretest

Posttest

Pretest

Posttest

M

SD

M

SD

M

SD

M

SD

Age (years; months)

6; 8

0; 4

6; 9

0; 3

Phoneme awareness total score

40

11.08

6.19

31.83

4.30

9.92

5.55

15.42

5.16

    Phoneme segmentation

10

1.83

1.99

8.33

1.07

1.42

1.31

1.75

1.36

    Phoneme deletion

10

5.08

2.68

8.25

1.76

5.00

2.83

6.50

2.11

    Phoneme blending

10

3.00

2.41

9.58

0.67

2.75

2.67

5.17

3.24

    Phoneme substitution

10

1.17

1.11

5.67

1.92

0.75

0.75

2.00

1.86

Pseudoword decoding

30

0.67

0.98

18.50

2.71

0.83

1.47

4.75

5.36

Burt raw score

110

15.83

7.91

33.58

7.39

16.17

7.69

26.17

8.00

Neale accuracy raw score

100

10.00

6.81

24.00

7.93

8.33

5.21

18.17

7.96

Neale comprehension raw score

44

4.67

2.10

10.58

2.35

4.17

3.66

7.67

2.90

The intervention program was carried out over a period of 24 weeks during the first three terms of a four-term school year. Following the intervention the children in the intervention and control groups were given the same tests that were administered prior to the intervention program. Two years after the intervention, follow-up data were obtained from 10 of the 12 matched pairs using the Burt Test and the Neale Accuracy subtest. The tester for the follow-up testing was blind to group assignment.

The reading programs of the four classrooms from which the children in the intervention and control groups were drawn strongly adhered to the whole language philosophy of teaching reading (for detailed descriptions of the New Zealand version of whole language, see Connelly, Johnston, & Thompson, 2001; Nicholson, 2000; Smith & Elley, 1994; Tunmer & Chapman, 2002). New Zealand has a unified national education system in which almost everything relating to literacy education is controlled centrally by the Ministry of Education, including the setting and monitoring of the national curriculum, the production of beginning reading materials and instructional guides for beginning reading teachers, and the development and implementation of nationwide professional development programs for literacy teachers. Consequently, compared with other countries like the United States, there is considerably less variation in the reading methods and instructional strategies used in New Zealand classrooms.

The relatively uniform approach to literacy instruction used across the four classrooms was confirmed in two ways; first, by informal observations of literacy instruction in each classroom at the beginning, middle, and end of the intervention, and second, by semi-structured interviews with the teachers that focused on their experience and training, general approaches to reading instruction, and specific strategies for teaching word identification skills. The observations and interviews revealed that for struggling readers (which included the children in the control group), all four teachers used a one-to-one, remedial procedure for teaching word identification strategies known as Pause, Prompt and Praise (Glynn, 1994). The basic technique used in this approach is to wait (pause) at least five seconds if the child gives no response to a difficult word in text or makes an error while reading in order to give the child a chance to self-correct before providing them with a prompt. If the child offers no response to an unfamiliar word, they are asked (prompted) to read to the end of the sentence, start again, and put in a word that makes sense. If the child makes an error that does not make sense, they are prompted with clues about the meaning of the story. Attending to the first letter or blend of an unknown word is encouraged only as a last resort during reading instruction. The strategy of “sounding out” an unknown word in text is not encouraged.

In addition to the classroom observations and teacher interviews, the reading scores of the intervention and control group children in each of the four classrooms were examined for possible differential effects due to classroom teachers. Although no significant differences were found, the size of the samples were too small to draw any firm conclusions regarding teacher effects. However, because the children in the intervention and control groups were distributed approximately equally across the four classrooms, any such differential effects would likely have been minimal.

Tests

The tests selected for this study included measures of phonemic awareness, phonological decoding ability, context free word recognition ability, accuracy of recognizing words in connected text, and reading comprehension. All five measures were used as pre- and posttests, and the measures of context-free word recognition and word recognition accuracy in text were used as the two-year follow-up measures. Due to limited resources, reading comprehension was not assessed two years after the intervention. However, Blaiklock (1997) reported that the Burt Test (our measure of context free word recognition) correlated highly with reading comprehension ability (r = 0.85), as assessed by means of the reading comprehension subtest of the Neale Analysis of Reading Ability, Revised (Neale, 1988).

Phonemic awareness

Four subtests of the Phonological Awareness Test (Robertson & Salter, 1997) were used to measure phoneme awareness: phoneme segmentation, phoneme blending, phoneme deletion, and phoneme substitution. The rhyme discrimination and rhyme production subtests were also given to the children. However, because the children tended to perform at ceiling levels, these two subtests were dropped from the test battery. In the phoneme segmentation subtest, the child was presented with a spoken word and asked to break it up into individual phonemes (i.e., “Tell me each sound in cat”). In the phoneme blending subtest, the child was asked to blend individual phonemes to make a real word that was presented as isolated sounds by the tester (i.e., “What word is this? /p-o-p/”). In the phoneme deletion subtest, the child was asked to repeat a word presented orally by the tester and then say it again deleting one of its phonemes (i.e., “Say cat. Now say it again but don’t say /k/”). In the phoneme substitution task the child was asked to isolate a sound in a given spoken word and then change the given sound to another, forming a new word (i.e., “Say paint. Change /p/ to /f/. What’s the new word?”). Scoring was based on the number of correct responses for each of the subtests.

Phonological decoding

An adapted version of a pseudoword decoding task developed by Richardson and DiBenedetto (1985) was used to measure phonological decoding ability. Thirty monosyllabic pseudowords from ‘Results’ of their Decoding Skills Test were presented in the form of a game in which the children were asked to try to read the “funny sounding names of children who live in far away lands.” The pseudowords were presented in order of increasing difficulty, ranging from simple consonant-vowel-consonant patterns (e.g., jit, med, dut) to blends, digraphs, and vowel variations (e.g., prew, thrain, fruice). Two practice items with corrective feedback were given followed by the 30 test items with no corrective feedback. Scoring was based on the number of items pronounced correctly.

Context free word recognition

The Burt Word Reading Test, New Zealand Revision (Gilmore et al., 1981) provided a measure of context free word recognition ability. The Burt Word Reading Test is a standardized test in which children are presented with a list of 110 words of increasing difficulty and asked to look at each word carefully and read it aloud. Testing continued until 10 successive words were read incorrectly or not attempted. Scoring was based on the number of words read correctly.

Word recognition accuracy in text and reading comprehension

The accuracy and comprehension subtests of the Neale Analysis of Reading Ability, Revised (Neale, 1988) were used to assess word recognition accuracy in text and reading comprehension ability. The children were asked to read aloud a series of short passages that were graded in difficulty. As the child read a passage, the tester recorded all reading errors on a record sheet, which provided the basis for deriving the reading accuracy score. After completing each passage the child was presented with a series of questions relating to the passage. The reading comprehension score was based on the total number of correct responses to these questions.

Intervention programme

The intervention programme comprised 56 highly sequenced, semi-scripted lessons in phonemic awareness and phonemically based decoding strategies delivered to the 12 intervention group children by a teacher aide over a period of 24 weeks during the first three terms of a four-term school year. The children in the intervention group were divided into four groups of three. Each group received four lessons per week that varied between 20 and 30 min in duration, but typically lasted about 25 min. Each child in the intervention group therefore received a total of 1 h and 20 min of teacher aide-directed small group instruction per week, which was in addition to their regular classroom literacy instruction. Of the 56 lessons included in the intervention program, most were completed in a single session, but some lessons were extended over two days. During the intervention the matched group of 12 poor readers who were not given explicit training in phonemic awareness and alphabetic coding skills continued to receive the standard whole language instruction and supplementary activities for struggling readers provided by their classroom teacher.

Teacher aide

The employment of teacher aides in New Zealand has increased as a growing number of schools are struggling to cope with the demands of the full-inclusion special education policy adopted by the government (Ministry of Education, 1996). While there is at least one teacher aide working at every primary school in New Zealand, many of the larger urban schools employ several teacher aides to work in classrooms alongside the teachers. The main role of the teacher aide is to assist the teacher with administrative and non-instructional duties within the regular classroom setting. Although they do not have teacher training qualifications, most teacher aides also undertake teaching tasks, usually with small groups of students or individual students. This situation is similar to what has occurred in the United States, where many teacher aides are assigned instructional responsibilities for which they are ill-prepared (Gerber, Finn, Achilles, & Boyd-Zaharias, 2001). Non-mandatory professional training opportunities such as the Teacher Aide Certificate are available through some tertiary institutions in New Zealand, and many teacher aides complete this qualification during their employment.

The teacher aide who delivered the intervention lessons had worked in the school for three years prior to becoming involved in the study. Her tasks included helping with photocopying and miscellaneous support duties within the classrooms. Although she was familiar with most of the children participating in the study, she was not aware of their academic abilities. At the time of her involvement in the intervention she had no teaching qualifications but had completed one module of the Teacher Aide Certificate. The module was on reading and included material that stressed the discussion of stories prior to reading and the use of the Pause, Pause and Praise instructional strategy.

Before the intervention began, the teacher aide received some initial training that included background information on phonemic awareness and its importance in early literacy development, demonstrations on ways to introduce specific letter sounds, and practice with feedback to ensure that the teacher aide fully grasped the concept of phonemic awareness. Particular attention was given to making sure that the teacher aide was producing the sounds correctly (e.g., saying /m/ instead of “muh” by “putting the brakes on” before adding the extra vowel to the sound /m/, as was demonstrated to the teacher aide by not moving the mouth or lips from their initial position as the sound was produced). The teacher aide was told that she would be following daily, semi-scripted lesson plans in which everything that she needed to say (presented in bold type in the lesson plans) and do was written down for her, and that she was not to deviate from the lesson plans without first consulting her supervisor (the first author), a remedial reading specialist called a Resource Teacher of Literacy (RTLit). RTLits are specially trained professionals in New Zealand who provide individual instruction to children with persistent literacy learning difficulties. Meetings between the RTLit and teacher aide were held throughout the intervention to clarify lesson plans, discuss the use of the intervention materials, and discuss any specific problems regarding the participating children. The close monitoring and supervision of the teacher aide by the RTLit throughout the study helped to ensure that the fidelity of the intervention program was maintained.

Intervention materials

The overall aim of the intervention program was to provide struggling readers with direct instruction in making connections between their speech sounds and the corresponding letters and letter patterns of English orthography. The lessons and materials used in the intervention were based on activities, theories, and research drawn from a variety of sources (Adams, Foorman, Lundberg, & Beeler, 1998; Cunningham & Hall, 1998; Eldredge, 1995; Herman, 1993; McGuinness, 1997; O’Connor, Notari-Syverson, & Vadasy, 1998) and were adapted for use in New Zealand schools (e.g., by taking into consideration dialect differences in the pronunciation of some speech sounds). Each of the 56 semi-scripted lesson plans was presented in a set format, which included the following components:
  1. (1)

    Materials required

     
  2. (2)

    Recap (a review of previously learned material)-1 or 2 min

     
  3. (3)

    Phonemic awareness exercises-5 min

     
  4. (4)

    Main lesson focusing on teaching letter–sound correspondences (introduced in a fixed order across lessons)-10 to 15 min

     
  5. (5)

    Activity reinforcing the learning of the new material introduced in the main lesson-5 min

     
  6. (6)

    Reading of phonetically controlled storybooks (i.e., decodable texts) introduced in a sequential manner after Lesson 28-storybooks read at least two times a week between lessons

     

The materials used in the intervention included an alphabet chart with 29 accompanying grapheme/picture flash cards, 11 syllable picture cards, 68 rhyme picture cards, 41 phoneme counting picture cards, three picture cards for each of the first 25 phonemes (used when introducing new letter–sound patterns), sound mats and letter tiles, 72 picture cards representing CVC (consonant, short vowel, consonant) words, word lists and word slides, grapheme bingo cards, worksheets, phonetic storybooks, hand-held mirrors, individual white boards and pens, toy microphones, and stretch toys. The letters on the alphabet chart were displayed after each new sound was taught. The chart included individual pictures and the corresponding letter or letter pattern (i.e., grapheme) for the sound. Around 35 phonemes were represented on the chart, including both long and short vowel sounds (e.g., the picture of the sun had an s beside it, the picture of an apple appeared with the letter a, the picture of oats appeared with the grapheme oa, the picture of a hole appeared with the o_e letter pattern). Accompanying the alphabet chart was a smaller matching set of 29 grapheme/picture flash cards with the picture on one side and the corresponding grapheme on the back. The flash cards were used in every lesson to teach, reinforce, and review the graphemes and sounds.

The 11 syllable cards were used in the first few lessons. They included pictures of five three-syllable words (e.g., butterfly), five four-syllable words, and one challenging five-syllable word (i.e., hippopotamus). The 68 rhyme picture cards included 34 pairs of pictured words that rhymed (e.g., clock/lock, box/fox, sun/run). The 41 phoneme counting picture cards were designed for identifying and counting phonemes in words. Cards one through 12 included pictures of two-phoneme words (e.g., pea, bow, hay). Cards 13 through 21 included pictures of three-phoneme words (e.g. leaf, bat, fish). Cards 22 through 30 included words with beginning or ending consonant clusters (e.g. lamp, snail, mask). Cards 31 through 41 were the most challenging and included four- and five-phoneme words (e.g. tiger, peanut). The sets of three picture cards were used to introduce each of the first 25 phonemes. For example, pictures of a needle, nose, and necklace (along with the picture of a net depicted on the grapheme/picture flash card for the letter n) were used to introduce the /n/ phoneme.

The concept of sound mats and letter tiles was adapted from Elkonin (1973). The mats had three or four squares drawn in a row. They were primarily used with the letter tiles in exercises in which the teacher aide dictated three- or four-phoneme words and assisted the children in spelling the words with their letter tiles. The 72 picture cards representing CVC words were used in exercises in which the children independently attempted to spell the CVC words with their letter tiles.

Twenty-four CVC, CCVC, and CVCC word lists, each of which contained words with the same vowel sound, and a word slide for each vowel sound were used to teach children to decode familiar words quickly. For example, the teacher aide would insert a list with short e words into the slide with the picture of an egg, the key word for the letter e. She would then tell the children that she didn’t want to hear them sounding out the words. Instead, she wanted them to do that in their heads quietly before saying what the word was. The primary goal of using the word lists and slides was to increase speed and accuracy when decoding.

Playing bingo with grapheme bingo cards was used as a reinforcement activity and was played in several different ways. Each child was given a bingo card with graphemes on it. The teacher aide called out a sound or showed a picture from the numerous picture cards that were available. She then identified whether she wanted the focus to be on the beginning or ending sound. The child named the picture, and from their letter choices on the bingo card, determined what the grapheme representation of the sound was, placing a token on the grapheme.

Silly sentence worksheets were introduced as an individual activity to assess decoding and comprehension. The four worksheets contained a “yes” or “no” answer to a silly question (e.g., “Can a bug kick a duck?”). There were eight to 10 questions on each worksheet. A further 13 worksheets were available and designed for use with the long vowel lessons (48 through 56). These worksheets contained four rows of words with different regular spelling patterns for the same long vowel (e.g., a_e, ai, ay, eigh). After individually reading the words on the worksheet, the student was then asked to sort and write down the words in the worksheet according to their long vowel spelling pattern.

The phonetically controlled storybooks were used when the children were able to decode simple words, which occurred upon completion of Lesson 28. These books consisted of individual stories written in highly decodable text, and were introduced in order of the vowel sound mastered (i.e., the children began reading the short a books first, the short o books next, and so on). The books, which were commercially available (Educators Publishing Service, 1995; Modern Curriculum Press, 1986; Steck-Vaughn Company, 1991), were read individually by the children to give them practice in both decoding and reading connected text, and as a way of encouraging them to use word-level reading strategies instead of relying primarily on picture and sentence context cues.

The hand-held mirrors enabled the children to see what their mouth was doing when they pronounced different sounds. The individual white boards and pens were used in writing and spelling activities. The toy microphones were used to “sing” the sounds of words together in blending exercises. The stretch toys were used to help children grasp the concept of “stretching” out individual sounds.

Intervention lessons

The lessons presented by the teacher aide focused on the development of phonemic awareness and phonemically based decoding strategies. Particular emphasis was placed on helping the struggling readers to understand that their own speech was represented by individual letters and letter combinations. The lessons were highly sequenced, beginning with initial phonemes that could be easily segmented and “stretched”. The first six graphemes introduced in the lessons were represented by continuant phonemes (e.g., m, s, r, f, a, n). As soon as the children had learned a few grapheme–phoneme correspondences, they immediately began reading and spelling words containing those correspondences. As the children moved through the lessons, they received constant reinforcement of previously learned material, and practice with reading connected text.

Each lesson presented by the teacher aide had three major components: the phonemic awareness exercises, the main lesson focusing on teaching grapheme–phoneme correspondences that were introduced in a fixed order across lessons, and an activity that reinforced the learning of the new material introduced in the main lesson. The phonemic awareness exercises were primarily oral activities that included tasks of rhyme identification and production, syllable counting, and phoneme isolation, segmentation, blending, and substitution. For example, in the phonemic awareness exercise for Lesson 1, in which the concept of sounds in words was introduced, the children were asked, “What happens to your mouth when you say words?” After responding to this question, each child was asked to say their name slowly into the mirror and describe what their mouth was doing. The teacher aide then said the child’s name slowly and exaggerated the sounds in the child’s name to demonstrate how the lips and tongue articulate. She compared the similarities and differences in the way each child’s name sounded, and encouraged the children to exaggerate the sounds in their name by saying their names very slowly. She told them to pretend that they were aliens from space and that everything they did was in “slow motion.” She then discussed what their mouth was doing as they began to say their name, and had each child demonstrate the sound at the beginning of their name by talking in slow motion like aliens, but this time “putting the brakes on” (i.e., stretching the beginning sound only). The teachers wrote each child’s name on the white board and pointed out that the first thing the child’s mouth does when they say their name represents the first letter in their name. In later lessons the children would be asked questions like, “Listen to this word, man. Is there a /m/ in man? Is it at the beginning? Middle? End? Do the words man and sun begin with the same sound? Can you say sun without the /s/? Who can say man without the /m/?”

A new phoneme and its letter representation were typically introduced in the main lesson. For example, in Lesson 5 the first vowel sound (“sticky letter”) was introduced. The teacher aide stretched the /a/ sound as she said the following words and presented the associated picture cards: ax, ant, astronaut, apple. She held up the a letter card and said, “This letter represents the /a/ sound. The key word for this letter is apple. Can you hear the /a/ at the beginning of these sounds?” She then pointed to the picture cards and said, “Today we are going to learn how to say the /a/ sound. Who would like to try saying this sound? Before you try, first think about the keyword. To help you, begin saying the key word in your alien voices and then ‘put the brakes on’.” The teacher aide passed out the mirrors and listened to each child produce the sound correctly. She then discussed what their mouths were doing: “Is your mouth open or closed when you say /a/? Where is your tongue? Can you see a big smile when you say this speech sound? Sometimes this is called the ‘smiley’ sound.”

The main lesson also included reinforcement activities such as letter writing and/or simple word dictation or “chaining” exercises using the letter tiles. In the chaining exercises the children used their sound mats and tiles to spell the word dictated, followed by: “If that is _______, then show me ______.” This statement was then repeated with the new word produced and the next word in the chain. An example chain is the words pat, pot, not, nat, nap, map, tap, top, map.

During the activity portion of the lesson, the children usually made their own individual flash cards or made use of the numerous rhyming or phoneme picture cards in a group or individual game. For example, in the activity for Lesson 3, the teacher aide placed the m, r, s, and f letter cards on the table or floor and had each child take turns matching the picture cards for m, r, s, and f with the corresponding letter cards.

Results

The pre- and posttest means of the intervention and control groups on all measures are shown in Table 1. The phoneme awareness total score is the sum of the four phoneme awareness subtest scores. Although gains were made by both groups, the intervention group outperformed the control group at posttest on all measures. The effect sizes were 1.71 for the phonological awareness total score, 1.69 for pseudoword decoding, .88 for the Burt raw score, .70 for the Neale accuracy raw score, and .98 for the Neale comprehension raw score. A 2 (group) × 2 (time) ANOVA was performed on the data for each of these measures. Of particular interest were the significant Group × Time interaction effects found for phonological awareness (total score), F (1,22) = 60.76, p < .001; pseudoword decoding, F (1,22) = 64.79, p < .001; and the Burt Test, F (1,22) = 13.95, p < .001. In each case, the posttest mean of the intervention group was higher than that of the control group. These results suggest that the intervention program was successful in achieving its primary goal of significantly improving the phonological awareness skills, decoding ability, and context-free word recognition skills of struggling readers. An examination of the age norms for the Burt posttest scores revealed that the intervention group children performed, on average, only two months below age-appropriate levels, whereas the control group children performed ten months below age-appropriate levels.

Although the Neale comprehension posttest mean of the intervention group was higher than the control group mean, the Group × Time interaction effect for Neale comprehension approached, but did not reach, statistical significance, F (1,22) = 3.54, p = .073. This suggests that there was a tendency for the positive effects of the intervention program to generalize to reading comprehension, which is consistent with the well-established relationship between the development of word recognition skills and reading comprehension ability (Pressley, 2006). The Group × Time interaction effect for Neale accuracy, however, did not approach significance. A possible explanation for this result is that the children in the control group were able to use compensatory strategies when reading words in connected text, a finding similar to that reported by Greaney et al. (1997).

Two years after the completion of the intervention program, follow-up data were obtained from ten of the twelve matched pairs using the Burt Test and the Neale Accuracy Subtest. Tests of significant differences between the means of the intervention and control groups for these measures are presented in Table 2. For both the Burt Test and Neale Accuracy Subtest, the children in the intervention group significantly outperformed the children in the control group, indicating that the positive effects of the intervention program were not only maintained but had generalized to word recognition accuracy in text. The effect sizes were .72 for the Burt raw score and .81 for the Neale Accuracy raw score. The finding of a significant difference between the groups on the Neale Accuracy Subtest is particularly noteworthy, given that the Group × Time interaction effect for the Neale accuracy pre- and posttest scores failed to reach significance. A possible interpretation of this finding is that the substantial progress achieved by the intervention group children in their word recognition skills during the intervention was producing positive Matthew (rich-get-richer) effects in reading achievement (Stanovich, 1986).
Table 2

Tests of significant differences between the means of the intervention and control groups for the two-year follow-up measures

Measure

Intervention group (n = 10)

Control group (n = 10)

t (18)

M

SD

M

SD

Burt raw score

52.70

13.02

43.70

9.45

1.77*

Neale accuracy raw score

47.40

15.69

35.60

11.24

1.93*

*p < .05

The mean reading ages of the intervention and control groups for the follow-up reading measures are presented in Table 3. The difference in mean reading age between the intervention group children and their matched controls was 9 months for the Burt Test and 14 months for Neale accuracy. The control group was reading 17 months below their chronological age on the Burt Test and 16 months below their chronological age on the Neale Accuracy Subtest, whereas the intervention group children averaged only 7 months below their chronological age on the Burt Test and 1 month below age-appropriate levels on the Neale Accuracy Subtest. Although the intervention group children performed somewhat below average in reading, their scores were clearly within the normal range after two years following the completion of the intervention program.
Table 3

Mean reading ages (years, months) for the two-year follow-up measures as a function of group

Variable

Intervention group (n = 10)

Control group (n = 10)

Chronological age

9; 4

9; 5

Burt reading age

8; 9

8; 0

Neale accuracy reading age

9; 3

8; 1

Discussion

Recent research suggests that the processes of learning to read vary considerably across children (Snow & Juel, 2005). For children who possess high levels of reading-related knowledge, skills and experiences at school entry, the processes involved in becoming literate are typically learner dependent, with children relying on induction to a very large degree. A largely constructivist, whole language approach to beginning reading instruction (with a major emphasis on reading of trade books and writing of text) is likely to be more effective for these children than code-emphasis approaches. In contrast, for children who possess low levels of essential reading-related skills and experiences at the outset of formal reading instruction, the learning processes are typically highly environment-dependent, with the children requiring a fairly structured and teacher-supported introduction to reading. These children almost always benefit more from reading instruction that involves explicit and systematic instruction in orthographic patterns and word identification strategies.

Given these considerations, and the fact that the constructivist-orientated, Reading Recovery program implemented in New Zealand is largely ineffective (Chapman et al., 2001), the aim of this study was to determine whether explicit instruction in phonemic awareness and phonemically based decoding skills would be an effective intervention strategy for children with early reading difficulties in a whole language instructional environment. Twenty-four 6- and 7-year-old struggling readers were drawn from a larger pool and divided into 12 closely matched pairs of poor readers on the basis of context free word recognition ability. The children in each pair were then randomly assigned to an intervention or control group, and the children in the intervention group were divided into four groups of three. The intervention program was carried out over a period of 24 weeks during the first three terms of a four-term school year. The program comprised 56 highly sequenced, semi-scripted lessons in phonemic awareness and alphabetic coding skills delivered to the children in the intervention group by a teacher aide who received training and ongoing support from a remedial reading specialist. Posttests results showed that the intervention group significantly outperformed the control group on measures of phonemic awareness, pseudoword decoding, and context free word recognition ability, and that the Group × Time interaction effect for reading comprehension approached statistical significance. Two-year follow-up data showed that the positive effects of the intervention program were not only maintained but had generalized to word recognition accuracy in text. The mean reading age of the intervention group was 9 months ahead of the control group on the Burt Test and 14 months ahead on the Neale Accuracy Subtest. In addition, the scores of the intervention group children on both measures were within the normal range.

The results of the study have implications for educational practice. The finding that a teacher aide-delivered intervention program had a positive impact on the children’s reading achievement conflicts with a recent review of the effectiveness of teacher aides by Gerber et al. (2001), who concluded from their analyses that “teacher aides have little, if any, positive effect on students’ academic achievement” (p. 123). A likely explanation for this difference in educational outcomes is that the teacher aide in the present study was carefully trained and monitored, received ongoing supervision and support from a remedial reading specialist, and followed daily, semi-scripted lesson plans in which everything that she needed to say and do was written down. Also, the content of the lessons themselves (i.e., explicit and systematic instruction in phonemic awareness and alphabetic coding skills) directly addressed what previous research in New Zealand and elsewhere has shown are the most important skill deficiencies of struggling readers (Chapman et al., 2001).

Our findings also have implications for educational practice in New Zealand. For almost two decades New Zealand has consistently shown comparatively high levels of variability on the test scores from international surveys of literacy achievement, despite having a national education system with a relatively uniform approach to literacy instruction and intervention. We have argued that persistent literacy learning difficulties in New Zealand are largely triggered by a predominantly constructivist, whole language approach to reading instruction and intervention that fails to respond adequately to differences in essential reading-related skills and knowledge at school entry (Tunmer et al., 2003, 2004; Tunmer, Prochnow, Greaney, & Chapman, in press). As Adams and Bruck (1993) pointed out some years ago, “wherever children who cannot discover the alphabetic principle independently are denied explicit instruction on the regularities and conventions of letter strings, reading disability may well be the eventual consequence” (p. 90). The resulting high rate of early reading difficulties would, in turn, account for New Zealand’s unacceptably large literacy achievement gap in the later grades.

According to this view of literacy development, children who do not possess sufficient levels of essential literacy-related skills and experiences at the outset of formal reading instruction (and who are not provided with explicit instruction to overcome their weakness in these areas, especially phonological awareness), will be forced to rely increasingly on ineffective word identification strategies (such as using picture cues, partial word-level cues, and contextual guessing). If these ineffective strategies are left uncorrected, negative Matthew (poor-get-poorer) effects in reading and persistent literacy learning difficulties will be the likely result (Tunmer et al., 2003, 2004). Reliance on such counterproductive strategies would almost certainly be exacerbated by an instructional approach like whole language in which beginning readers are specifically encouraged to use text-based strategies over more effective word based strategies, such as those incorporated into the early intervention program of the present study.

The critical question that needs to be examined empirically is this: What is the most effective strategy that education systems can use to reduce the influence of differences in literate cultural capital at school entry on future reading achievement? Put more simply, we need an approach in which the new entrant with a limited amount of literate cultural capital has approximately the same probability of future success in learning to read and write as the child with an abundance of literate cultural capital at school entry. The findings of our study suggest that the most promising strategy for reducing the reading achievement gap is to place much greater emphasis on differential instruction, where beginning reading teachers and remedial specialists maximize the effectiveness of their literacy teaching by using research-based assessment procedures and instructional strategies to cater to the differing skill needs of beginning readers from the outset of schooling, with particular attention focused on ensuring the development of phonemically based word-level skills and strategies during the early stages of reading acquisition.

Copyright information

© Springer Science+Business Media B.V. 2007