Emergent Literacy Intervention for Prekindergarteners at Risk for Reading Failure

Abstract

This research examined the effectiveness of an emergent literacy intervention for prekindergarten children at risk for reading failure, to replicate and improve on significant findings from Year 1 of the study. Data are reported for 266 children in 72 child care and preschool sites in Year 2 of the study and for 374 children at 102 sites in Year 3. The intervention consisted of eighteen 30-min lessons delivered twice weekly to small groups of children. Lessons targeted critical emergent literacy skills through explicit, developmentally appropriate activities for prekindergarteners. Hierarchical linear models were used to nest children within center and measure treatment effects on phonological awareness, alphabet knowledge, and vocabulary skills. Results indicated significant treatment effects on multiple measures in Years 2 and 3. This study replicated and strengthened findings from Year 1 in demonstrating a positive impact of this intervention for prekindergarteners at risk for reading failure.

Keywords

at risk prevention dyslexia preschool emergent literacy

The preschool years are critical for preparing children for future success in reading. It is well documented that children who enter kindergarten with a gap in their foundational reading skills tend to remain behind their typically achieving peers (Bierman et al., 2008; Snow, Burns, & Griffin, 1998). The costs of remedial intervention increase with each passing year, and long-term results tend to diminish when assistance begins later in students’ school careers (Fielding, Kerr, & Rosier, 2007; Shaywitz & Shaywitz, 2004; Torgesen, 2002). For these reasons, interest and efforts in defining, measuring, and teaching emergent literacy skills to preschoolers have dramatically increased over the past decade.

We conducted a multiyear assessment and intervention study targeting prekindergarten children at risk for reading failure in an effort to demonstrate an effective approach for reducing the gap in reading preparedness that already exists by the time such students enter kindergarten (West, Denton, & Reaney, 2000). For 3 consecutive years, we have implemented an experimental emergent literacy intervention with successive groups of young children identified through screening as being potentially at risk for future reading difficulties. The primary goals were to identify 4- to 5-year-old prekindergarteners who demonstrated delays in the acquisition of critical emergent literacy skills and to provide a targeted, experimental educational intervention to better prepare them for formal reading instruction in elementary school. The project had the following general structure: (a) universal screening at selected preschool and child care sites to identify all prekindergarten children at risk for reading failure using an established prekindergarten screening measure; (b) educational intervention with small groups of at-risk children, using an experimental curriculum; and (c) program delivery in the ecologically valid setting of child care centers and preschools.

This approach was selected to increase early identification, in prekindergarten, of children at risk for a high-incidence disorder (i.e., reading disability), provide focused intervention in a timely manner, and reduce barriers to access for intervention by providing all services in typical child care and preschool settings. It is consistent with the response-to-intervention (RTI) approaches that have been developed for use in the elementary school grades (e.g., D. Fuchs & Fuchs, 1998; L. Fuchs, 2003) and the comparable approaches currently under development or recently reported for preschool students (Buysse, Winton, & Zimmerman, 2007; Gettinger & Stoiber, 2008; Koutsoftas, Harmon, & Gray, 2009). The core purpose of the RTI approach is to monitor objectively the progress of all students on key academic and behavioral indicators in the regular classroom and provide more intensive, timely instruction for students demonstrating risk. Instruction is conceptualized in tiers, with Tier 1 referring to a strong general classroom curriculum and high-quality instruction. Tier 2 is a more intensive, targeted instructional approach for at-risk students who need more assistance than that provided through Tier 1 instruction. Children who fail to make adequate progress through Tier 2 instruction receive Tier 3 services, consisting of additional intensive intervention, special education evaluation, or both (D. Fuchs & Fuchs, 2006; Koutsoftas et al., 2009).

For the present project, a unique Tier 2 emergent literacy curriculum was developed to address the specific needs of prekindergarteners at risk for reading failure. This intervention was designed to deliver more intensive, explicit, and systematic emergent literacy instruction for at-risk prekindergarteners as a supplement to their Tier 1 classroom emergent literacy instruction. In addition, this curriculum was designed to accommodate the maturational differences in cognitive, linguistic, and behavioral characteristics of preschoolers versus older children (Lonigan, Burgess, Anthony, & Barker, 1998; Snow et al., 1998) and common differences in classroom structure and routines in preschool classrooms versus typical kindergarten and first grade classrooms. For example, instruction involved frequent repetition to enhance language and metalinguistic scaffolding and play-based, manipulative activities to maintain students’ attention and motivation. The overarching goal was to create an effective Tier 2 emergent literacy intervention that was developmentally appropriate for at-risk prekindergarteners and reflective of current research on critical content and instructional strategies proven to propel young children toward reading readiness.

Children’s Emergent Literacy Skills

The development of reading ability is preceded by and causally related to the development of several critical emergent or precursor literacy skills, which can be measured during the preschool years and are strongly predictive of future reading abilities (Lonigan et al., 2009; National Early Literacy Panel [NELP], 2008; Snow et al., 1998). English is an alphabetic language, such that its written symbols (e.g., letters or graphemes) represent speech sounds, or phonemes, as the primary sound unit for mapping speech into print. Phonemic awareness refers to the ability to detect, analyze, and blend phonemes accurately within spoken words, which then enables the association of those phonemes to their written symbols, that is, letters (Bailet, 2001; Lieberman & Shankweiler, 1985; Lonigan et al., 1998). Mastery of this fundamental relationship between spoken and written English, known as the alphabetic principle, is the single most important factor in the development of literacy in our language (Adams, 1990; Stanovich, 1986; Stubbs, 1980; Wagner, 1988; Wagner & Torgesen, 1987).

Of particular interest for the research described here are the precursor skills for phonemic awareness and the alphabetic principle: phonological awareness and alphabet knowledge. Phonological awareness refers to a hierarchical set of skills for becoming aware of speech sound units within words, culminating in full phonemic awareness (Anthony & Francis, 2005; Lonigan et al., 1998; Metsala & Walley, 1998). Detection of rhyming and alliteration are two of the earliest phonological awareness skills to emerge in preschoolers, whereas syllable and onset–rime awareness and segmenting and blending skills emerge somewhat later in the preschool years. These phonological awareness skills, for the most part, pertain to larger “chunks” of words and therefore are easier for preschoolers to grasp than smaller phonemic sound units. One exception is awareness of the onset of a word, which may be a single phoneme (i.e., /b/-/at/). Identifying phonemes in the initial position of a word, as distinct from the rime portion, is typically easier than identifying all phonemes, and thus the onset is often a starting point for developing phonemic awareness in prereaders (Slocum, O’Connor, & Jenkins, 1993; Treiman, 1992).

Equally important to and predictive of early reading success is adequate alphabet knowledge, including both letter names and letter sounds. Empirical evidence suggests that young children begin to learn letter–sound correspondences by first learning letter names, which subsequently may promote understanding of the alphabetic principle (Ehri & Roberts, 2006; Foorman, Anthony, Seals, & Mouzaki, 2002; McBride-Chang, 1999; Treiman, Tincoff, Rodriguez, Mouzaki, & Francis, 1998; Whitehurst & Lonigan, 2002). Both phonological awareness and alphabet knowledge contribute significant yet unique variance to the development of reading proficiency, such that instructional programs targeting both types of skills reflect best practices for preschool emergent literacy instruction (e.g., Justice, Chow, Capellini, Flanigan, & Colton, 2003; Landry, Anthony, Swank, & Monseque-Bailey, 2009; Lundberg, Frost, & Petersen, 1988; Piasta & Wagner, 2010; Schneider, Kuspert, Roth, Vise, & Marx, 1997; Schneider, Roth, & Ennemoser, 2000; for a review, see National Institute of Child Health and Human Development, 2000; NELP, 2008; NIFL, 2008). In a recent meta-analysis of the effects of preschool instruction on various alphabet knowledge outcomes, Piasta and Wagner (2010) found that effect sizes tended to be larger when multicomponent literacy instruction was provided. These findings lend further support to the current experimental instructional curriculum, which incorporated substantial emphasis on a broad array of alphabet knowledge and phonological awareness skills.

Emergent Literacy Instructional Intervention

A full rationale and detailed description of the emergent literacy intervention on which the present study is premised are reported in Bailet, Repper, Piasta, and Murphy (2009). The reader is also referred to that article for a review of prior studies pertinent to the current research. In brief, several studies have demonstrated the potential to improve prekindergarteners’ emergent literacy skills through either whole-class or small-group explicit instruction using code-focused activities (e.g., Connor, Morrison, & Slominski, 2006; Justice et al., 2003; Yeh, 2003), thus lending empirical support to the multiyear research we are pursuing. Koutsoftas et al. (2009) recently reported on a Tier 2 RTI study of phonemic awareness in at-risk preschoolers attending low-income classrooms with high-quality Tier 1 instruction. They completed a 6-week experimental intervention focusing on beginning sound awareness, delivered by trained teachers and speech-language pathologists. Results indicated that 71% of the children made significant progress in this skill on immediate retesting. In addition, most of the participants maintained their gains on further testing in the fall of their kindergarten year.

Additional research is needed with large sample sizes across a diverse array of typical preschool and child care settings, employing high-quality experimental design and statistical analysis features, to more comprehensively evaluate the effectiveness of emergent literacy intervention with at-risk prekindergarteners. This type of project has the potential to make a significant contribution to the science and knowledge base pertaining to identification and intervention for such children. There is growing awareness and increasing expectations that the preschool years afford an untapped opportunity to increase the proportion of children entering kindergarten truly “ready” for the level of reading instruction that is typical across the United States. However, significant gaps remain in our understanding of the precise skills that prekindergarteners should be taught, the most appropriate sequence, pace, and dosage, and best instructional methods for diverse learners. Filling that gap is important if the instructional opportunities are to be actualized in an impactful, yet developmentally appropriate manner.

The current research may help to bridge this gap in the following ways. First, this study provided universal screening of prekindergarteners across many typical preschool and child care sites, combined with differentiated instruction and reassessment. Although such work is now commonplace in the elementary school grades, little such research has been conducted with preschoolers. The current research also strikes an important balance between adherence to both rigorous experimental design features and implementation feasibility in the natural environment, which is essential to translate research into routine practice. Last, a unique curriculum was created for this project, providing an RTI Tier 2 intervention for at-risk prekindergarteners.

To summarize our methodological approach, we implemented a prototypical version of the experimental emergent literacy intervention with children deemed at risk for later reading difficulties. In all, 18 lessons were developed by project staff and delivered by project teachers to small groups of children over 9 weeks in participating child care and preschool sites. These sites were randomly assigned to two groups, matched on several demographic variables. One group of sites received the experimental intervention early in the school year (immediate intervention), and the other group of sites received the experimental intervention later in the year (delayed intervention). Lessons focused on emergent literacy skills identified through research as critical to future reading success, as previously described (e.g., Neuman & Dickinson, 2002; Stone, Silliman, Ehren, & Apel, 2004). Specific code-focused activities included letter names and letter sounds, syllable counting, blending and segmentation, rhyming, alliteration, and onset–rime. Several lessons included an emergent writing component, the exact number of which varied by year (see Table 1).

Table 1.

Intervention Lessons Overview.

Percentage of Lessons, Out of 18, Containing Specific Early Literacy Activities
	Year 1	Year 2	Year 3
	2005–2006	2006–2007	2007–2008
Syllable segmentation	39	22	50
Elision and blending	0	0	22
Name activities^a	27	100	100
Print awareness	100	100	100
Vocabulary	67	55	72
Letter name–sound	44	67	56
Beginning sound	44	67	56
Onset–rime	17	0	22
Rhyming	22	33	27
Alliteration	33	27	16
Emergent writing	22	50	55
Parent letter	17	50	100

Activities that are organized around participating children’s names

Results from Year 1 of the study indicated significant emergent literacy gains in response to the experimental intervention. Statistically significant gains in the immediate intervention group were obtained midyear, as compared to the delayed intervention group, on two of the screening measures, the Get It, Got It, Go! (GGG; Early Childhood Research Institute on Measuring Growth and Development, 1998) Rhyming and Alliteration subtests. These results indicated that children who received the intervention demonstrated greater phonological awareness learning in rhyming and alliteration skills than children who did not. Considered within the RTI context, such results demonstrated that differentiated and more intensive early literacy instruction for at-risk prekindergarteners was feasible in typical preschool and child care settings. However, this study had some limitations related to instrumentation and sample size that were addressed in Years 2 and 3.

The Present Study

This article reports results from the 2nd and 3rd years of a multiyear research project in which each successive year consisted of new students and slight changes in outcomes measures and intervention lessons. The basic cluster-randomized structure of the program remained the same across all 3 years of this project. Larger numbers of child care and preschool sites participated in Years 2 and 3 relative to Year 1. The assessment methods were strengthened in Years 2 and 3 by establishing a more stringent cutoff score on the screening measure for intervention eligibility and including different emergent literacy outcomes measures. The curriculum was strengthened iteratively in Years 2 and 3, in response to prior results and accumulating new research and study experiences. The intent of these changes was to increase statistical power, address psychometric limitations of outcomes measures noted in Year 1 (Bailet et al., 2009), and enhance the curriculum in response to initial outcomes and experiences.

We hoped to replicate the following hypotheses from Year 1 in Years 2 and 3: (a) at-risk children will show significant and meaningful gains in emergent reading skills following their participation in the twice weekly, 9-week intervention; (b) the two intervention groups will not differ when reassessed following the delayed group’s treatment period at the end of the school year (i.e., there will be no “time of year” effect); and (c) gains made by the immediate intervention group will be maintained over the latter half of their prekindergarten year following cessation of their intervention.

In Year 3, particular emphasis was placed on segmentation and blending of syllables and onset–rime, relative to Years 1 and 2 of the project (see Table 1). As a result of this increased instructional focus, we hypothesized that immediate intervention children would show greater gains in these phonological awareness skills, culminating in a significant treatment effect that would be evident in Year 3.

For Year 1, we also had hypothesized and obtained a significant dosage effect for all outcome measures. In Years 2 and 3, we emphasized to project teachers the importance of completing as many lessons as possible with each participating child because of the significant dosage effects identified in Year 1. As a result, most children completed nearly all the lessons, and the variance was too small to allow for statistical analyses of dosage effects. In Year 2, 91% of participants received 15 or more lessons out of a maximum possible 18 lessons; in Year 3, 86.8% received 15 or more lessons.

Method

Participants

Preschools

Year 2 of the study took place in 72 private preschool and child care sites, located in a large city in the southeastern United States. All of the 38 sites that participated in the study for Year 1 also participated in Year 2. Year 3 of this study took place in 102 sites located in the same city as the prior 2 years. Most of the sites (> 90%) that participated in Years 1 and 2 continued to participate in Year 3. In addition, numerous other sites were added to increase the number of schools and children served in all quadrants of the county for Year 3.

As was the case for Year 1, it remained a requirement for Years 2 and 3 that all participating sites have a developmentally appropriate Tier 1 early literacy curriculum in use daily in their prekindergarten classrooms. This was important because the experimental intervention was designed to provide instruction that was supplemental (i.e., Tier 2) to ongoing classroom instruction. Therefore, research sites were selected with this factor in mind. Examples of Tier 1 curricula used in study sites included the Early Language and Literacy Model (Florida Institute for Education and the University of North Florida, 2002) and Links to Literacy (Episcopal Children’s Services of Northeast Florida, 2006). All research sites also were participating in the state-funded prekindergarten program, which requires 540 hr of language and literacy instruction during the school year.

As in Year 1 of the study, for Years 2 and 3 pairs of child care centers were matched by zip code and percentage of children receiving a state-sponsored financial subsidy, where applicable. For most schools, randomization of treatment occurred at the school level, with the at-risk students who attended each individual school being assigned to either the immediate or delayed treatment group. In these schools, teachers provided intervention services to the at-risk children in either the fall or the spring semester. However, some sites were quite large and/or had many children who qualified (10 in Year 2 and 8 in Year 3). In these schools, teachers provided intervention in both the fall and the spring to serve all children who qualified. At-risk children in these “split” schools received intervention at only one of the two time points (as did their “nonsplit school” peers); however, the randomization to treatment group occurred at the child level. Thus, children within a split school were randomly assigned to be in either the immediate or delayed intervention group at that school. Each split school was then randomly chosen to contribute either its immediate or delayed group data to the subsequent analyses. Although this method resulted in a loss of a subset of the total number of student participants (59 in Year 2 and 36 in Year 3), it allowed each school to contribute data to the analyses and maintained the integrity of the random assignment as much as possible. Socioeconomic status remained comparable across the “split” and “nonsplit” sites.

In both split and nonsplit sites, the delayed intervention group served as an untreated control group with which to compare the immediate intervention group. Impact of the intervention was assessed through an “intent-to-treat” framework (Shadish, Cook, & Campbell, 2002). Children serving as controls during the immediate intervention phase received the same experimental intervention during the second half of the prekindergarten year. Gains of the delayed intervention group were compared to gains of the immediate intervention group to validate their results and assess potential time-of-year instructional effects.

Students

Years 2 and 3 of this multiyear study involved new groups of 4-year-old child participants each year. Screening was limited to those children eligible, by birth date, to enter kindergarten the following school year. A total of 2,004 children attending prekindergarten during the fall of the 2006–2007 school year were screened in Year 2 (screening measures and procedures described subsequently). Out of those screened, 349 children received intervention. Children were omitted from the analyses if they were outside of the age range of the screening tool (i.e., older or younger than 4 years of age; n = 4) or were lost to general attrition such as moving and/or changing preschools or withdrawing from the program (n = 20). An additional 59 children had both pre- and postintervention scores but were excluded from the analyses as they received either immediate or delayed intervention in a “split” school that was chosen at random not to contribute data for that particular time point. Thus, 266 children provided scores for the analyses in Year 2.

During the fall of the 2007–2008 school year (Year 3), a total of 2,777 children attending prekindergarten were screened. Out of those screened, 429 received intervention. Children were omitted from the analyses if they were outside of the age range of the screening tool (i.e., older than 4 years of age; n = 15) or were lost to general attrition such as moving and/or changing preschools or withdrawing from the program (n = 4). As in Year 2, the inclusion of split schools in the data analyses allowed only the data from one of their two time points to be used; this resulted in the exclusion of 36 children. Thus, 374 children provided scores for the analyses in Year 3. Demographic and descriptive information on the sample is provided in Table 2.

Table 2.

Years 2 and 3: Demographic Information and Descriptive Statistics for Immediate and Delayed Intervention Groups.

n and Percentage or Mean and Standard Deviation	Year 2 Intervention Groups				Year 3 Intervention Groups
	Immediate		Delayed		Immediate		Delayed
	n	%	n	%	n	%	n	%
Sex
Girls	57	44.5	47	34.3	82	40.4	66	38.6
Boys	68	53.1	81	59.1	121	59.6	105	61.4
Race/ethnicity
Caucasian	53	41.4	48	35.0	100	49.3	77	40.7
African American	56	43.8	50	36.5	56	27.6	64	33.9
Hispanic	7	5.5	8	5.8	21	10.3	21	11.1
Other	7	5.5	20	14.6	23	11.3	27	14.3
Risk category
Fall GRTR 0–4	16	12.5	20	14.6	23	11.3	24	14.0
Fall GRTR 5–6	42	32.8	38	27.7	62	30.5	50	29.2
Fall GRTR 7–8	70	54.7	79	57.7	118	58.1	97	56.7

	M	SD	M	SD	M	SD	M	SD

Age in months	52.44	3.49	52.67	3.28	53.38	3.29	53.14	3.21
Lessons completed	16.89	1.80	16.80	2.58	16.77	2.59	16.01	3.06

Note: Demographic data for the intervention groups were available for 95% and 99% for Years 2 and 3, respectively. GRTR = Get Ready to Read.

Children who were screened each year and did not qualify for the intervention (i.e., typically developing peers) were also rescreened in the spring of their prekindergarten year. This afforded an opportunity to compare end-of-year screening score gains for the intervention group versus their typical classmates, when all children had received the same Tier 1 classroom emergent literacy instruction. For Year 2, there were 1,575 children identified as typical on their fall screening; for Year 3, there were 2,281 such children.

Measures

As in Year 1, the Get Ready to Read (GRTR; Whitehurst, 2001) screening measure was administered to determine eligibility for intervention in Years 2 and 3. For those children identified as eligible, additional outcomes measures were used for each successive year of the study. The GGG test used in Year 1 was dropped and replaced by the newly available Test of Preschool Early Literacy (TOPEL; Lonigan, Wagner, & Torgesen, 2007) in Years 2 and 3. This decision reflected the TOPEL’s strong psychometric qualities and incorporation of syllable blending and elision and onset–rime tasks, which recent research has shown are strong predictors of subsequent reading ability (NELP, 2008). For Year 3, the Rhyme Knowledge subtest of the Assessment of Literacy and Language (ALL; Lombardino, Lieberman, & Brown, 2005) was added to the outcomes measures to document intervention participants’ progress in this early phonological awareness skill, which is not assessed by the TOPEL. The goal was to use the best available instruments from a psychometric standpoint and to incorporate measures reflecting the most current research that also were appropriate based on the participants’ learning characteristics and the content of the intervention. All measures included in Years 2 and 3 are described subsequently in greater detail.

Get Ready to Read

GRTR is a 20-item multiple-choice test that measures print awareness, knowledge of letter names and sounds, and beginning phonological awareness skills, such as syllable deletion and blending. It is designed for 4-year-olds. The technical report on GRTR (Whitehurst, 2001) describes scores of 6 to 9 in the fall of prekindergarten as “weak” and scores of 0 to 6 as “very weak.” The authors of GRTR report adequate reliability (Cronbach’s α = .78) and a classification accuracy of 78% (predictive validity of poor vs. good readers) when children were reassessed 4 years later with comprehensive reading measures (Whitehurst, 2001). For Years 2 and 3, “at risk” was operationally defined as a score of less than 9 on the GRTR emergent literacy assessment tool. This is in contrast to Year 1, for which the “at-risk” cutoff was set at less than 10 on GRTR. For Years 2 and 3 of this study, no children who scored low on GRTR were excluded from the experimental intervention. Thus, a small proportion of child participants had a known or suspected disability such as autism, global developmental delay, or behavioral or emotional impairment.

Test of Preschool Early Literacy

The TOPEL was included as an additional outcomes measure beginning in Year 2. The TOPEL consists of three subtests: Print Knowledge, Definitional Vocabulary, and Phonological Awareness. Raw scores for each subtest were used for data analyses in the present study. The TOPEL was added to the outcomes measures as soon as it became available (in Years 2 and 3), in part because its psychometric properties are stronger than those of GGG. In addition, the TOPEL test content reflects recent research indicating that performance on syllable blending and elision and onset–rime blending tasks during prekindergarten is strongly predictive of future reading ability (NELP, 2008). For the present study, the TOPEL Phonological Awareness subtest was divided into blending and elision item clusters for statistical analyses, as prior research has shown that these skills may develop at somewhat separate rates (Farver, Nakamoto, & Lonigan, 2007; NELP, 2008).

The TOPEL Examiner’s Manual (Lonigan, Wagner, Torgesen, & Rashotte, 2007) cites internal consistency reliability coefficients of .95, .94, and .87 and test–retest reliability coefficients of .89, .82, and .83 on the Print Knowledge, Definitional Vocabulary, and Phonological Awareness subtests, respectively. Numerous studies of criterion validity have indicated correlations of large to very large magnitude with other early literacy measures, including GRTR (.45, .51, and .57 for the TOPEL Print Knowledge, Definitional Vocabulary, and Phonological Awareness subtests, respectively), the Test of Early Reading Ability–Third Edition, Alphabet subtest (Reid, Hresko, & Hammill, 2001; .77 with TOPEL Print Knowledge subtest), and the Expressive One-Word Picture Vocabulary Test (Brownell, 2000; .71 with TOPEL Definitional Vocabulary subtest). For Year 3, the TOPEL Definitional Vocabulary subtest was dropped from the assessment battery because of time constraints for individual child assessment and the increased instructional focus on phonological awareness.

Assessment of Literacy and Language Rhyme Knowledge subtest

The ALL Rhyme Knowledge subtest was included as an additional outcome measure during Year 3. The ALL is designed to measure oral language and emergent literacy skills in prekindergarten, kindergarten, and first grade children. Test–retest reliability coefficients across all subtests and grades, cited in the Examiner’s Manual (Lombardino et al., 2005), are adequate to excellent. For prekindergarteners, the Rhyme Knowledge subtest test–retest coefficient is .95, whereas the internal consistency correlations range from .89 to .94. Reported correlations between the ALL Rhyme Knowledge subtest and similar subtests on established measures of oral language and early reading skills have demonstrated satisfactory correlations. For example, the ALL Rhyme Knowledge subtest demonstrated a correlation of .63 with the Rhyme Awareness subtest of the Pre-Reading Inventory of Phonological Awareness (Dodd, Crosbie, McIntosh, Teitzel, & Ozanne, 2003) and a correlation of .69 with the Rhyming subtest of the Early Reading Diagnostic Assessment (Harcourt Assessment, 2003). Overall, the psychometric properties and content validity of the ALL Rhyme Knowledge subtest were judged to be adequate for the purposes of the present study.

Recognizing that strong correlations exist among the multiple outcomes measures employed during Years 2 and 3 of this study, it remained important to include a larger number of items assessing specific alphabet knowledge and phonological awareness skills than are included in GRTR. With these longer, more skill-specific subtests from the TOPEL and ALL, the ability to document more precisely what improvements in emergent literacy skills resulted from the experimental intervention, if any, was substantially increased.

Procedure

Assessment protocol

Children initially were screened using GRTR, and children deemed eligible, based on the criteria previously described, were enrolled in the experimental intervention. Once the pool of children who were eligible for intervention was determined using GRTR, the additional outcomes measures were administered to them. Data were collected at three times during the school year: Time 1 assessment occurred in August and September, Time 2 assessment occurred in January and February (after the immediate group had completed intervention, but before the delayed group started intervention), and Time 3 assessment occurred in April and May. Time 1 and 2 assessments served as pretest and posttest for the immediate intervention study comparing treatment and control children; Time 2 and Time 3 assessments served as pretest and posttest for the replication study with the delayed intervention group. Time 3 assessments also served as a posttest to assess maintenance of gains for the immediate intervention group.

All measures were administered by experienced teachers hired for the project. Training was provided and teachers were subsequently supervised by a PhD school psychologist and PhD clinical psychologist with more than 25 years of collective experience in psychological and educational assessment. After formal training on each instrument, teachers were observed while practicing administration with each other. Corrective feedback and demonstration were given as needed, until all teachers displayed mastery. Ongoing supervision and oversight were provided throughout the screening process to ensure accuracy in administration and scoring.

Assessments were administered individually in relatively quiet environments at participating sites and required approximately 30 to 40 min per child.

Intervention

In both Years 2 and 3, the intervention consisted of eighteen 30-min lessons delivered over 9-week sessions. The sequence, instructions, materials, and activities for the entire curriculum were standardized across project teachers, with a systematic instructional approach embedded within and across lessons. Activities were designed to be developmentally appropriate and highly engaging for prekindergarteners. Each lesson included explicitly taught, specific emergent literacy skills (e.g., letter names and sounds, syllable counting, rhyming, etc.) and a literary element (e.g., story, poem, or song) reinforcing the specific skill focus or letters for that lesson. Teacher instruction was scripted to facilitate the children’s understanding of key concepts and to be explicit about literacy terms, such as “syllable” or “letter.” Lessons were designed to establish a predictable routine throughout the intervention and to provide ample repetition and practice of concepts within and across lessons. Thus, each lesson began with a consistent opening routine, followed by introduction and practice with one or more alphabet letters, a shared reading activity, phonological awareness instructional sequence, and lesson review. Pictures and objects were presented to facilitate language comprehension and child engagement for all core activities. A sample lesson is included in Appendix A.

Multisensory materials and instructional techniques played a prominent role in the experimental curriculum. Use of multisensory, simultaneous instruction with struggling readers, especially those with dyslexia, has been widely practiced for decades. Although specific empirical support has been elusive in prior research, these techniques appear, at a minimum, to enhance attention and motivation and theoretically may activate neural pathways that lead to greater learning and memory (Moats & Farrell, 2005). In the present study, examples of multisensory simultaneous instruction included (a) tracing alphabet letters with various tactile substances such as sand, corrugated paper, or rice, while simultaneously saying the sound of the letter; (b) seeing and feeling one’s mouth and throat while saying the corresponding letter sound; and (c) writing an alphabet letter in the air while saying its name or sound. These types of activities and others were sequenced within the lesson to provide multiple multisensory opportunities and ample repetition. Phonological awareness tasks were taught using manipulative materials that reinforced recognition and scaffolded mental manipulation of word parts, such as syllables and onsets.

The intervention was delivered by teachers hired to implement the project. In Year 2, there were 17 project teachers. In Year 3, there were again 17 project teachers, 9 of whom continued from Year 2. All intervention teachers had a minimum of a bachelor’s degree in elementary education, reading, or early childhood and prior teaching experience. They received approximately 20 hr of intensive professional development by project leaders on key aspects of language processing problems in young children, learning to read, and working with struggling learners. In addition, they received explicit training specific to the experimental curriculum for approximately 10 hr prior to the start of intervention.

Implementation fidelity was monitored through observing each teacher in the field and offering critical feedback about lesson adherence and general teaching style. Supervisors observed the following elements of lessons: consistency of opening and closing routines; ability to establish and maintain rapport, regulate child attention and behavior, and provide positive feedback; use of verbal instructions as prescribed; availability and appropriate use of required materials; and completion of all elements within the lesson. These practices exceeded those described as essential for “high standardization” in a recent review (Wanzek & Vaughn, 2007) and are consistent with procedural fidelity features described by Justice, Mashburn, Hamre, and Pianta (2008). All teachers met together weekly with their supervisors to discuss successes, challenges, and questions about lessons and prepare for the upcoming week of lessons. There were no indications of significant deviations from the lessons as designed. Project teachers demonstrated a high level of implementation fidelity and diligence to the intent and format of lessons.

The maximum number of children in an intervention group was four. Some groups consisted of two or three children, depending on the number of children participating in intervention in each site. Occasionally, teachers worked individually with participants who were struggling in the small-group situation or to make up a lesson that had been missed. Each child’s attendance was recorded for each lesson, as an additional indicator of implementation fidelity and dosage. Teachers made considerable effort to complete all 18 lessons with each child. The average number of completed lessons across the immediate and delayed intervention groups was 16.85 for Year 2 and 16.43 for Year 3. There was not a significant difference between the immediate and delayed groups for the mean number of lessons completed for either year.

Changes in the experimental intervention across study years

Within each study year, no changes were made to the experimental intervention, such that each participant that year received the same exact lessons. Each summer between study years, the experimental curriculum underwent minor modifications in response to the most recent child outcomes, teachers’ reported experiences with the lessons, and new research that became available. The overall structure, intent, and focus of the curriculum remained consistent over time. Any changes reflected shifts in the proportions of activities devoted to a particular emergent literacy skill and enhancement of instructional support techniques. When this multiyear project was conceived, the plan was to enhance the curriculum each year in this manner, recognizing that inevitably there would be aspects that could be improved.

Table 1 displays the percentage of lessons, out of a possible 18, that contained specific emergent literacy skill foci across Years 1, 2, and 3, for comparison purposes. For Year 2, the major changes as compared to Year 1 involved the following: (a) reduced emphasis on syllable segmentation, alliteration, and onset–rime; (b) additional emphasis on activities organized around participating children’s names, rhyming, alphabet letter names and sounds, and emergent writing; and (c) additional parent letters, to foster greater school–home communication. The rationale for the reduced areas of emphasis reflected the difficulty participants had with these skills during Year 1 and recognition that participants in Year 2, as a group, would be slightly lower in their baseline skills because of the reduction in the GRTR cutoff to qualify for intervention. Hence, the concern was that the selected child participants would not be ready developmentally to comprehend these tasks. The increase in emergent writing reflected growing research emphasizing the importance of these activities in building both phonological awareness and alphabet knowledge (Welsch, Sullivan, & Justice, 2003; Whitehurst & Lonigan, 1998) and project teachers’ observations in Year 1 that students enjoyed and benefited from these writing activities.

Similarly, in response to Year 2 experiences and results, additional curriculum adjustments were made for Year 3. One major change involved a reversal of the decision for Year 2 to reduce syllable segmentation, elision and blending, and onset–rime activities. The rationale for increasing the focus again on syllable segmentation, elision and blending, and onset–rime in Year 3 reflected the increasing published research on the strongly predictive relationship of these skills with later reading success (NELP, 2008). Multisensory components, as previously described, were increased to help participating struggling learners grasp these challenging concepts. We also believed it was important to teach at-risk participants those phonological elision and blending skills measured by the outcomes tasks, for two reasons. First, these are some of the skills proving to be most strongly predictive of future reading ability in the early elementary school grades. Second, we recognized the at-risk participants’ probable difficulty in generalizing newly acquired skills across differing phonological awareness constructs and task formats. Therefore, explicit instruction in these elision and blending skills was deemed both beneficial and appropriate.

Curricular emphasis on rhyming and alliteration was decreased in Year 3, reflecting newer research suggesting that these skills are less reliable in predicting subsequent development of phonemic awareness and reading ability (Anthony & Lonigan, 2004; Lonigan et al., 2009; NELP, 2008). Parent letters were sent home after every lesson in Year 3. These letters were simple descriptions of a key early literacy activity included in that day’s intervention lesson, such as the letter “Bb” or clapping syllables in words, and a simple suggestion for parents to reinforce that concept at home. A sample parent letter is included in Appendix B.

Results

Preliminary Analyses

For both years, there were no significant differences between the immediate and delayed intervention groups for gender, ethnicity, or chronological age. Means, standard deviations, and ranges for all Time 1, Time 2, and Time 3 dependent variables can be found in Tables 3 (Year 2) and 4 (Year 3), respectively. In both Years 2 and 3, the immediate and the delayed intervention groups’ scores on all dependent variables in the fall were not statistically different.

Table 3.

Year 2 Outcomes: Time 1, Time 2, and Time 3 Descriptive Statistics for Outcome Variables by Intervention Group.

Student Outcome	Time 1				Time 2				Time 3
Student Outcome	n	M	SD	Range	n	M	SD	Range	n	M	SD	Range
Immediate intervention
GRTR	128	6.31	1.67	0–8	128	11.73	3.53	3–19	120	13.67	3.33	4–20
TOPEL PK	123	6.96	5.85	0–29	128	16.45	9.40	0–34	120	20.73	9.74	1–41
TOPEL VOC	123	35.07	12.77	0–63	128	41.55	12.11	1–65	120	46.18	9.96	0–62
TOPEL EL	123	4.72	2.35	0–12	128	5.21	2.27	0–12	120	5.92	2.35	1–12
TOPEL BL	123	5.40	3.16	0–18	128	7.30	3.49	0–15	120	7.44	3.29	0–15
Delayed intervention
GRTR	138	6.41	1.57	0–8	138	10.72	3.61	3–18	125	13.43	3.85	4–20
TOPEL PK	132	7.69	6.18	0–30	124	12.08	8.64	0–36	112	18.12	9.86	0–36
TOPEL VOC	132	36.09	14.03	1–62	124	41.94	12.76	2–62	112	45.83	12.85	2–64
TOPEL EL	132	4.27	2.55	0–11	124	5.27	2.63	0–12	112	6.11	2.97	0–12
TOPEL BL	132	4.97	3.09	0–11	124	6.48	3.55	0–15	112	7.64	3.59	1–15

Note: GRTR = Get Ready to Read; TOPEL = Test of Preschool Early Literacy; PK = Print Knowledge; VOC = Vocabulary; EL = Elision; BL = Blending.

Table 4.

Year 3 Outcomes: Time 1, Time 2, and Time 3 Descriptive Statistics for Outcome Variables by Intervention Group.

Student Outcome	Time 1				Time 2				Time 3
Student Outcome	n	M	SD	Range	n	M	SD	Range	n	M	SD	Range
Immediate intervention
GRTR	203	6.55	1.51	2–8	203	11.97	3.89	3–20	164	13.82	4.12	2–20
TOPEL PK	200	6.19	4.34	0–21	201	16.18	10.13	0–36	160	20.54	10.06	1–36
TOPEL EL	200	4.12	2.18	0–9	201	5.32	2.44	0–12	160	6.10	2.42	0–15
TOPEL BL	200	5.18	2.98	0–15	201	7.09	3.16	0–15	160	8.08	3.32	0–15
ALL RHY	199	3.18	2.04	0–15	200	4.77	3.52	0–19	159	6.25	4.61	0–19
Delayed intervention
GRTR	171	6.48	1.55	1–8	171	11.14	3.61	1–19	149	14.45	3.31	5–20
TOPEL PK	170	6.46	4.71	0–25	171	15.08	9.25	0–36	145	21.01	9.45	1–36
TOPEL EL	170	3.75	2.08	0–11	170	5.11	2.09	0–11	145	5.91	2.65	0–12
TOPEL BL	170	5.38	3.92	0–15	170	6.02	2.98	0–15	145	7.46	3.52	0–15
ALL RHY	170	3.05	1.68	0–9	171	3.95	3.20	0–17	146	6.12	4.86	0–24

Note: GRTR = Get Ready to Read; TOPEL = Test of Preschool Early Literacy; PK = Print Knowledge; EL = Elision; BL = Blending; ALL RHY = Assessment of Literacy and Language, Rhyme Knowledge.

In the Year 2 sample, 21 children (8 in the immediate and 13 in the delayed group) withdrew from their child care centers prior to Time 3 assessment. Attrition rates differed somewhat by intervention group (6.25% attrition for immediate group vs. 9.42% attrition for delayed group at Time 3 data collection). However, there were no statistically significant differences on initial emergent literacy scores between the children who withdrew from their child care sites prior to completion of the intervention and the children who completed the intervention.

In the Year 3 sample, 61 children (39 in the immediate and 22 in the delayed group) withdrew from their child care centers prior to Time 3 assessment. Attrition rates differed by intervention group (19.2% attrition for immediate group vs. 12.9% attrition for delayed group at Time 3 data collection). As in Year 2, there were no statistically significant differences on initial emergent literacy scores between children who withdrew from their child care sites prior to completion of the intervention and children who completed the intervention in Year 3.

The data were carefully entered, and missing data were assessed to determine if patterns of results related to missing data emerged or might be considered missing at random (Shadish et al., 2002). Missing data at the student level were only the result of attrition. Because of the use of residualized gain in analyses, pretest and posttest data were required for each set of analyses (i.e., Time 1 to Time 2 for the treatment effect). Children missing either pretest or posttest data for a given timeframe were excluded from the relevant analysis.

Fall treatment outcomes: Immediate versus delayed (control) treatment

To test the overall treatment effect across both study years, hierarchical linear modeling (HLM; Raudenbush & Bryk, 2002) was used to account for the shared variance among children attending the same center (two-level model) as the participants were nested within center. Although children participated in small groups within center, we did not use these small groups as a level of analysis because these groups were often fluid. That is, at their discretion, teachers moved children from one group to another if they felt this better facilitated each child’s ability to participate in the intervention. The analytic strategy employed, however, correctly accounted for the shared variance of children within centers when estimating the effects of the treatment.

Because of the differences in randomization across the larger “split” and relatively smaller “nonsplit” sites, an analytic method that took differences in the level of randomization into account was necessary. To address this issue, HLM analyses were run and effect sizes were calculated for the split and nonsplit schools separately that were then merged in a meta-analytic method per Shadish and Haddock (2009). This method ensured that the treatment variable was entered at the appropriate level of analysis for each cohort of sites (i.e., at the school level for the nonsplit sites and at the child level for the split sites), while maintaining high statistical rigor in the data analyses (L. Hedges, personal communication, July 28, 2010).

Nonsplit sites

All analyses considered student gains in the outcomes of interest by examining Time 2 outcomes controlling for Time 1 scores (i.e., residualized gain). All models were fit using full maximum likelihood estimation in the HLM software (version 6.02a; Raudenbush, Bryk, & Congdon, 2005).

The HLM models estimated the overall treatment effect on each of the outcomes variables, with separate models built for each outcome. Time 1 student scores were entered at Level 1 of the model, with Time 2 scores serving as the dependent variable. These scores were centered at the grand mean, thus allowing interpretation of the results for a child who scored at the Time 1 sample mean (i.e., the mean of the group of children meeting initial selection criteria). Treatment was assigned at the center level and therefore was dummy coded and entered at Level 2 (center level) of these models. A significant treatment effect signified a reliable difference in emergent literacy skill gain for children in the immediate versus delayed intervention group. Effects using robust standard errors were interpreted given the sufficiently large number of schools. These estimates are generally more conservative (Raudenbush & Bryk, 2002). Random effects and interaction terms were considered and removed if nonsignificant to create the most parsimonious models. Results of these HLM analyses are shown in Table 5 for Year 2 and Table 7 for Year 3.

Table 5.

Year 2 Outcomes: Hierarchical Linear Modeling Results for Treatment Effect on Student Fall to Winter Gains, Nonsplit Sites.

Fixed Effect	Coeff.		SE	Approx. df	p Value
Winter TOPEL PK intercept (γ₀₀)	11.89		0.88	59	.000
Child level
Fall TOPEL PK score (γ₁₀)	0.78		0.08	185	.000
Preschool level
Fall treatment (γ₀₁)	5.16		1.17	59	.000
Winter TOPEL VOC intercept (γ₀₀)	41.12		1.10	59	.000
Child level
Fall TOPEL VOC score (γ₁₀)	0.65		0.05	185	.000
Preschool level
Fall treatment (γ₀₁)	−0.12		1.69	59	.943
Winter TOPEL EL intercept (γ₀₀)	5.25		0.22	59	.000
Child level
Fall TOPEL EL score (γ₁₀)	0.42		0.08	185	.000
Preschool level
Fall treatment (γ₀₁)	−0.41		0.33	59	.215
Winter TOPEL BL intercept (γ₀₀)	6.53		0.42	59	.000
Child level
Fall TOPEL BL score (γ₁₀)	0.41		0.09	185	.000
Preschool level
Fall treatment (γ₀₁)	0.88		0.65	59	.182
Winter GRTR intercept (γ₀₀)	10.38		0.44	60	.000
Child level
Fall GRTR score (γ₁₀)	0.42		0.16	195	.008
Preschool level
Fall treatment (γ₀₁)	1.38		0.61	60	.026

	Final Estimation of Variance Components
Random Effects	SD	Variance	df	χ²	p Value

Winter TOPEL PK
Preschool site level (U₀)	1.06	1.12	59	71.40	.129
Child level (R)	7.88	60.47
Winter TOPEL VOC
Preschool site level (U₀)	5.06	25.61	59	172.39	.000
Child level (R)	6.56	43.03
Winter TOPEL EL
Preschool site level (U₀)	0.21	0.04	59	72.16	.117
Child level (R)	2.19	4.80
Winter TOPEL BL
Preschool site level (U₀)	1.93	3.72	59	170.32	.000
Child level (R)	2.58	6.65
Winter GRTR
Preschool site level (U₀)	1.62	2.62	60	116.96	.000
Child level (R)	3.00	9.02

Note: Results with robust standard errors are reported. Random effects are reported where significant. TOPEL = Test of Preschool Early Literacy; PK = Print Knowledge; VOC = Vocabulary; EL = Elision; BL = Blending; GRTR = Get Ready to Read.

Effect sizes for each of the dependent variables were estimated by computing Cohen’s d for the Time 2 treatment versus control comparisons (Cohen, 1988). The gamma coefficient corresponding to the treatment effect was entered as the numerator, and this was divided by the pooled SD for each dependent variable. Computing the effect size in this manner makes the resulting effect size analogous to the HLM results, controlling for the other covariates entered into the model (e.g., Time 1 scores).

Split sites

HLM analyses mirrored the method used with nonsplit sites, with some exceptions. In the split sites, treatment was assigned at the child level and therefore was dummy coded and entered at Level 1 (child level) of the model. As is true for the nonsplit schools analysis, a significant treatment effect signified a reliable difference in emergent literacy skill gain for children in the immediate versus delayed intervention group.

Unlike the nonsplit school analysis, robust standard errors could not be interpreted because of the relatively smaller sample size of the split site cohort. Thus, fixed effects were interpreted without the robust standard errors for these models. Likewise, random effects were not retained in the model for the split school analyses for this same reason, as it is less likely that results from these smaller analyses can (or should be) reliably generalized. Results of these split school HLM analyses are shown in Table 6 for Year 2 and Table 8 for Year 3.

Table 6.

Year 2 Outcomes: Hierarchical Linear Modeling Results for Treatment Effect on Student Fall to Winter Gains, Split Sites.

Fixed Effect	Coeff.		SE	Approx. df	p Value
Winter TOPEL PK intercept (γ₀₀)	13.34		1.33	9	.000
Child level
Fall TOPEL PK score (γ₁₀)	1.01		0.15	50	.000
Fall treatment (γ₂₀)	4.36		1.74	50	.016
Winter TOPEL VOC intercept (γ₀₀)	46.83		1.45	9	.000
Child level
Fall TOPEL VOC score (γ₁₀)	0.71		0.08	50	.000
Fall treatment (γ₂₀)	–2.97		1.86	50	.117
Winter TOPEL EL intercept (γ₀₀)	6.72		0.48	9	.000
Child level
Fall TOPEL EL score (γ₁₀)	0.50		0.14	50	.001
Fall treatment (γ₂₀)	–1.23		0.63	50	.055
Winter TOPEL BL intercept (γ₀₀)	8.24		0.64	9	.000
Child level
Fall TOPEL BL score (γ₁₀)	0.58		0.13	50	.000
Fall treatment (γ₂₀)	–1.43		0.83	50	.090
Winter GRTR intercept (γ₀₀)	11.97		0.51	9	.000
Child level
Fall GRTR score (γ₁₀)	1.38		0.27	65	.000
Fall treatment (γ₂₀)	0.07		0.75	65	.929

	Final Estimation of Variance Components
Random Effects	SD	Variance	df	χ²	p Value

Winter TOPEL PK
Preschool site level (U₀)	0.16	0.03	9	9.11	.427
Child level (R)	6.06	36.69
Winter TOPEL VOC
Preschool site level (U₀)	0.10	0.01	9	5.40	> .500
Child level (R)	6.64	44.03
Winter TOPEL EL
Preschool site level (U₀)	0.29	0.09	9	11.67	.232
Child level (R)	2.05	4.19
Winter TOPEL BL
Preschool site level (U₀)	0.46	0.21	9	11.24	.259
Child level (R)	2.74	7.51
Winter GRTR
Preschool site level (U₀)	0.04	0.002	9	4.67	> .500
Child level (R)	3.06	9.35

Note: TOPEL = Test of Preschool Early Literacy; PK = Print Knowledge; VOC = Vocabulary; EL = Elision; BL = Blending; GRTR = Get Ready to Read.

Table 7.

Year 3 Outcomes: Hierarchical Linear Modeling Results for Treatment Effect on Student Fall to Winter Gains, Nonsplit Sites.

Fixed Effect	Coeff.		SE	Approx. df	p Value
Winter TOPEL PK intercept (γ₀₀)	14.61		0.90	92	.000
Child level
Fall TOPEL PK score (γ₁₀)	0.81		0.09	92	.000
Preschool level
Fall treatment (γ₀₁)	1.69		1.26	322	.183
Winter TOPEL EL intercept (γ₀₀)	5.15		0.18	92	.000
Child level
Fall TOPEL EL score (γ₁₀)	0.45		0.06	322	.000
Preschool level
Fall treatment (γ₀₁)	0.19		0.26	92	.458
Winter TOPEL BL intercept (γ₀₀)	5.89		0.31	92	.000
Child level
Fall TOPEL BL score (γ₁₀)	0.27		0.05	322	.000
Preschool level
Fall treatment (γ₀₁)	1.12		0.44	92	.013
Winter ALL RHY intercept (γ₀₀)	3.99		0.29	91	.000
Child level
Fall ALL RHY score (γ₁₀)	0.53		0.15	320	.001
Preschool level
Fall treatment (γ₀₁)	0.60		0.40	91	.135
Winter GRTR intercept (γ₀₀)	11.07		0.38	92	.000
Child level
Fall GRTR score (γ₁₀)	0.78		0.13	327	.000
Preschool level
Fall treatment (γ₀₁)	0.74		0.48	92	.125

	Final Estimation of Variance Components
Random Effects	SD	Variance	df	χ²	p Value

Winter TOPEL PK
Preschool site level (U₀)	4.21	17.71	92	196.47	.000
Child level (R)	7.52	56.51
Winter TOPEL EL
Preschool site level (U₀)	0.71	0.50	92	145.42	.000
Child level (R)	1.92	3.67
Winter TOPEL BL
Preschool site level (U₀)	1.62	2.62	92	242.51	.000
Child level (R)	2.34	5.47
Winter ALL RHY
Preschool site level (U₀)	0.87	0.76	91	116.05	.039
Child level (R)	3.09	9.53
Winter GRTR
Preschool site level (U₀)	1.32	1.74	92	148.31	.000
Child level (R)	3.24	10.50

Note: Results with robust standard errors are reported. Random effects are reported where significant. TOPEL = Test of Preschool Early Literacy. PK = Print Knowledge; EL = Elision; BL = Blending; ALL RHY = Assessment of Literacy and Language, Rhyme Knowledge; GRTR = Get Ready to Read.

Table 8.

Year 3 Outcomes: Hierarchical Linear Modeling Results for Treatment Effect on Student Fall to Winter Gains, Split Sites.

Fixed effect	Coeff.		SE	Approx. df	p Value
Winter TOPEL PK intercept (γ₀₀)	16.71		1.88	7	.000
Child level
Fall TOPEL PK score (γ₁₀)	1.14		0.33	40	.002
Fall treatment (γ₂₀)	3.87		2.99	40	.203
Winter TOPEL EL intercept (γ₀₀)	5.30		0.38	7	.000
Child level
Fall TOPEL EL score (γ₁₀)	0.70		0.13	39	.000
Fall treatment (γ₂₀)	0.17		0.56	39	.756
Winter TOPEL BL intercept (γ₀₀)	6.03		0.98	7	.000
Child level
Fall TOPEL BL score (γ₁₀)	0.40		0.15	39	.013
Fall treatment (γ₂₀)	2.73		1.39	39	.057
Winter ALL RHY intercept (γ₀₀)	3.77		0.54	7	.000
Child level
Fall ALL RHY score (γ₁₀)	0.16		0.21	40	.439
Fall treatment (γ₂₀)	2.21		0.81	40	.010
Winter GRTR intercept (γ₀₀)	10.97		0.75	7	.000
Child level
Fall GRTR score (γ₁₀)	0.89		0.42	41	.041
Fall treatment (γ₂₀)	2.87		1.11	41	.014

	Final Estimation of Variance Components
Random Effects	SD	Variance	df	χ²	p Value

Winter TOPEL PK
Preschool site level (U₀)	0.33	0.11	7	7.91	.340
Child level (R)	8.78	77.03
Winter TOPEL EL
Preschool site level (U₀)	0.18	0.03	7	8.63	.279
Child level (R)	1.74	3.04
Winter TOPEL BL
Preschool site level (U₀)	1.45	2.11	7	18.37	.011
Child level (R)	2.92	8.54
Winter ALL RHY
Preschool site level (U₀)	0.03	0.00	7	2.20	> .500
Child level (R)	2.64	6.95
Winter GRTR
Preschool site level (U₀)	0.08	0.01	7	7.65	.364
Child level (R)	3.66	13.41

Note: TOPEL = Test of Preschool Early Literacy; PK = Print Knowledge; EL = Elision; BL = Blending; ALL RHY = Assessment of Literacy and Language, Rhyme Knowledge; GRTR = Get Ready to Read.

Split and nonsplit analysis effect size merge

To address the significant differences in sample size between the two sets of analyses, an inverse-variance sample weighting approach was used for fixed effects models per Shadish and Haddock (2009). Using this method, individual variances and weights based on sample size were calculated for the dependent variables for each “study” that were then statistically merged to create a weighted average effect size, variance, standard error, and confidence interval across studies. The 95% confidence interval was used to determine whether the weighted average effect size differed from zero and thus whether or not the null hypothesis should be rejected. Average weighted effect size estimates that exceeded the standard two-tailed critical value of 1.96 resulted in rejection of the null hypothesis. Additional statistics were calculated to determine whether the independent studies shared a common population effect size through measuring homogeneity of the variance (Q statistic) and an additional descriptor, I², that addresses the heterogeneity of the variance as recommended by Shadish and Haddock (2009). Weighted average effect size estimates and their associated statistics are presented in Table 9 for Year 2 and Table 10 for Year 3.

Table 9.

Merged Treatment Effect Sizes (d_o), Variances (v_o), Standard Errors (SE), Confidence Intervals (CI), and Associated Homogeneity and Heterogeneity Indices (Q, I², respectively) for Each Dependent Variable for Year 2.

Measure	n	d_o	v_o	SE	CI	Q	I ²
GRTR^a	266	.29	.015	.124	2.38	1.73	42.20
PK^a	252	.54	.017	.129	4.18	0.002	0.00
VOC	252	−.07	.016	.127	0.514	0.721	0.00
EL	252	−.24	.016	.127	1.89	1.31	23.61
BL	252	.11	.016	.127	0.856	4.49	77.73

Note: GRTR = Get Ready to Read; PK = Test of Preschool Early Literacy (TOPEL) Print Knowledge; VOC = TOPEL Vocabulary; EL = TOPEL Elision; BL = TOPEL Blending.

Indicates significant treatment effect at the 1.96 CI, two-tailed. Q values are significant when χ²(1) < 3.84, critical value when α = .05.

Table 10.

Measure	n	d_o	v_o	SE	CI	Q	I ²
GRTR^a	374	.25	.011	.105	2.39	2.16	0.54
PK	372	.20	.011	.105	1.92	0.29	0.00
EL	371	.09	.011	.105	0.814	0.001	0.00
BL^a	371	.41	.011	.105	3.88	1.13	11.35
ALL RHY^a	371	.24	.011	.105	2.26	3.09	67.64

Note: GRTR = Get Ready to Read; PK = Test of Preschool Early Literacy (TOPEL) Print Knowledge; EL = TOPEL Elision; BL = TOPEL Blending; ALL RHY = Assessment of Literacy and Language, Rhyme Knowledge.

Indicates significant treatment effect at the 1.96 CI, two-tailed. Q values are significant when χ²(1) < 3.84, critical value when α = .05.

Treatment Main Effects

GRTR

In Years 2 and 3, there was a significant effect of intervention on students’ fall (Time 1) to winter (Time 2) gains on GRTR. In Year 2, the weighted average effect size (d_o) for the impact of intervention on GRTR was .29, a small to medium effect size. In Year 3, the weighted average effect size for the impact of intervention on GRTR was .25, a small effect size.

TOPEL

In Year 2, there was a significant effect of intervention on students’ Time 1 to Time 2 gains on the TOPEL Print Knowledge subtest (TOPEL PK). The weighted average effect size for the impact of intervention on TOPEL PK was .54, a medium effect size. In Year 3, there was not a significant effect of intervention on students’ fall to winter gains on the TOPEL PK. In Year 2, there was not a significant effect of intervention on students’ fall to winter gains on the TOPEL Definitional Vocabulary subtest; this subtest was not administered in Year 3.

On the TOPEL Blending and Elision item clusters, there was not a significant effect of intervention in Year 2. However, in Year 3, HLM analyses indicated a significant effect of intervention for Blending (weighted average effect size = .41). Unlike Blending, the Elision task did not evidence a significant treatment effect in Year 3. Figure 1 displays mean raw scores on the TOPEL Blending subtest at each time of assessment by intervention group for Year 3. These results partially support the hypothesis related to specific gains in phonological awareness in Year 3.

Figure 1.

Year 3: Mean Test of Preschool Early Literacy Blending scores for immediate intervention group (solid line) and delayed intervention group (dashed line)

ALL Rhyme Knowledge subtest

In Year 3, HLM analyses indicated a significant effect of intervention on students’ fall to winter gains on the ALL Rhyme Knowledge subtest (ALL RHY). The weighted average effect size for the impact of intervention on ALL RHY was .24, a small effect size.

Replication of Treatment Gains and Maintenance of Effects

To address our second hypothesis, we assessed whether there was a benefit associated with receiving the experimental intervention in the fall versus the spring. These analyses also provided a means of addressing our third hypothesis, pertaining to the immediate group’s maintenance of treatment gains from Time 2 to Time 3, after their intervention had stopped.

Specific t tests were conducted to determine if the immediate and delayed intervention groups’ Time 3 scores differed significantly from one another within year. Because the treatment level did not affect this Time 3 analysis, both split and nonsplit data were combined within each year. In Year 2, there were no differences between the groups’ Time 3 scores on GRTR, TOPEL Elision, or TOPEL Blending item clusters (p > .05). The TOPEL PK subtest did evidence a significant group difference in spring scores across the two groups in Year 2, t(230) = -2.033, p = .04. It is unclear why the immediate intervention group scores (M = 20.73, SD = 9.74) remained higher than the delayed intervention group scores on this measure at Time 3 (M = 18.12, SD = 9.86). However, it is likely that this is a spurious finding given the large sample size, large standard deviations, and multiple comparisons. In Year 3, there were no differences between the groups’ Time 3 scores on any of the measures. Across both study years, the delayed intervention groups made similar progress to the immediate intervention groups on all of the early literacy outcome variables after receiving intervention. The lack of group differences at Time 3 also suggests that, although children in the immediate intervention groups in Years 2 and 3 did not make additional significant gains in emergent literacy skills following the completion of the intervention, they maintained gains observed at the Time 2 assessment.

End-of-Year Gains for Intervention Participants Versus Their Typical Peers

The administration of GRTR in both the fall and spring to both intervention and nonintervention children in participating preschools allowed the exploration of how intervention children compared to the nonintervention children at the end of the school year on this screening measure. Children who qualified for and received the intervention (both immediate and delayed) posted an average fall to spring gain on GRTR of about 7.5 points across both Years 2 and 3; by comparison, children who did not qualify for the intervention posted a gain of about 3 points, a figure that is consistent with previous research in typical populations (Whitehurst, 2003).

Discussion

We have carried out a multiyear cluster-randomized controlled trial of the impact of an experimental emergent literacy intervention with prekindergarteners at risk for future reading disability, in their typical preschool settings. The major structural underpinnings of the intervention have remained constant across all 3 years of research. Thus, for example, each year the intervention has consisted of eighteen 30-min standardized lessons, each with a specific emergent literacy skill focus and literary element. Instructional group size was four or fewer, and teachers were those hired and trained for the study. Project supervisors systematically observed intervention teachers hired for this study and gave corrective feedback as appropriate. No concerns about violations of implementation fidelity were raised in any year of the study.

Results from Years 2 and 3 replicated initial results from Year 1, with outcomes from all 3 years indicating a significant, favorable impact of this experimental intervention on emergent literacy skills status. For both Years 2 and 3, group differences on GRTR were statistically significant, whereas they were not in Year 1. GRTR scores attained by the immediate and delayed intervention groups for Years 2 and 3 were similar to those scores from Year 1, suggesting that the larger number of study sites in Years 2 and 3 improved statistical power to detect treatment-based differences. Cohen’s d effect sizes for GRTR across Years 1, 2, and 3 (weighted average in Years 2 and 3) were .24, .29, and .25, respectively, consistent with effect sizes reported by other emergent literacy researchers (Piasta & Wagner, 2010).

This data trend bears closer examination, however, because as a group, the at-risk children identified for the experimental intervention in Years 2 and 3 had a lower baseline mean score on GRTR (i.e., lower emergent literacy skills) than those in Year 1. This reflected the change in the cutoff score to less than 9 in Years 2 and 3 as compared to less than 10 in Year 1. Nonetheless, study participants achieved similar postintervention GRTR scores at the end of Year 2 and slightly higher GRTR scores at the end of Year 3 as children at the end of Year 1 (13.8, 13.55, and 14.12, for Years 1, 2, and 3, respectively). Their absolute degree of progress most likely reflected curriculum improvements relative to Year 1 as well as increased teacher experience with this type of at-risk prekindergarten population.

It should be noted that this approximate 7.5-point raw score gain on GRTR from fall to spring of the prekindergarten year represents substantial, meaningful progress in emergent literacy skills, as the reported typical gain on GRTR across the prekindergarten year is 3 points (Whitehurst, 2003). Within the context of our own research in Years 2 and 3, virtually identical gains of about 3 points were documented among typically developing children screened in participating sites. This suggests that the quality of Tier 1 instruction in participating research sites was comparable to that of prekindergarten classrooms participating in other research using GRTR as a screening measure. More importantly, the intervention children more than doubled the rate of gain on this measure as compared to their typical classmates, documenting an accelerated learning pace that reduced the emergent literacy achievement gap.

In Year 2, students participating in immediate intervention also made impressive gains on the TOPEL PK as compared with the delayed intervention control children. This particular subtest provided a nice complement to GRTR, with more items and broader assessment of print awareness and alphabet knowledge that appeared to capture a major strength of Year 2’s experimental curriculum. Contrary to expectations, however, for Year 3 we did not obtain statistically significant results between intervention groups on the TOPEL PK. We speculate that this result occurred, at least in part, because of increased instruction in print awareness and letter recognition in participating preschool and child care settings in general, favorably affecting all children in child care sites and enhancing print knowledge skills for both the immediate and delayed intervention groups. Piasta and Wagner (2010) noted that young children are frequently exposed to alphabet knowledge through teachers and parents, which can dilute the apparent effect of specific instructional interventions in this skill. In the community where this research was conducted, there has been an increasing citywide focus in the past few years on quality early literacy instruction in child care centers. Thus, it is feasible that general literacy instruction in the classroom was becoming more effective with each successive year of the study, especially for print and alphabet knowledge.

Results on the TOPEL Blending and Elision item clusters in Year 2 reflected no significant gains in response to the experimental intervention. In contrast, the experimental intervention results for Year 3 indicated a significant increase in scores on the TOPEL Syllable and Onset–Rime Blending item clusters within that subtest. Statistically significant gains in rhyming ability also were documented in Year 3.

Looking across all 3 years of our study with successive groups of children, it is clear that the experimental intervention, in general, resulted in gains in phonological awareness skills. The critical link required to demonstrate that effect was alignment of phonological awareness skill foci in the intervention lessons, with outcomes measures assessing the same skill that had been taught to the at-risk participants. Thus, when rhyming and alliteration were emphasized in Year 1 intervention and were also assessed directly with outcomes measures, a significant positive effect was documented. Similarly, in Year 3, when blending and segmenting of syllables and onset–rime were emphasized in intervention and assessed directly as outcomes, a significant positive effect was documented. Rhyming remained a focus for Years 2 and 3 but was specifically assessed again only in Year 3, when we obtained a significant effect of intervention.

This phonological awareness results pattern across our multiyear study is consistent with prior developmental research. The emergence of various phonological awareness skills occurs gradually from early to middle childhood and is affected by numerous child, linguistic, cognitive, and experiential factors (Anthony & Francis, 2005; Lonigan et al., 1998; Metsala & Walley, 1998). Existing studies indicate that detection of rhyme and alliteration is one of the earliest phonological awareness abilities to emerge, followed by blending of syllables and onset–rime, with segmentation skills emerging still later (Anthony & Francis, 2005; Anthony, Lonigan, Driscoll, Phillips, & Burgess, 2003; Lonigan et al., 1998; NELP, 2008). In the early childhood years, proficiency in these skills may vary significantly within and across individual children. Instruction and skill in one aspect of phonological awareness may not generalize to other aspects, particularly in at-risk learners. This would explain our lack of significant gains on the TOPEL Blending and Elision item clusters in Year 2, when the intervention lessons did not emphasize the specific skills assessed by those tasks. Future research may enable us to demonstrate the most effective sequence of phonological awareness skills, instructional strategies, and dosage needed for at-risk prekindergarteners to grasp all of these essential concepts.

Most children completed the majority of lessons in Years 2 and 3, which most likely strengthened study results, as a significant dosage effect had been documented in Year 1. This could not be measured specifically in Years 2 and 3 because of a lack of variance in lesson completion rate. Clearly the length of the experimental intervention, that is, 30 min twice per week for 9 weeks, should be manageable in many settings. Although additional lessons would likely be beneficial for most participants, fewer lessons might lessen the impact for many and is not recommended.

Across all 3 years of successive study, we did not find a time-of-year effect, with the exception of the TOPEL PK in Year 2, for which the immediate intervention group outperformed the delayed intervention group on spring rescreening. This may reflect a Type I error, given the number of statistical comparisons made. Overall, there was minimal apparent benefit to having the intervention earlier or later during the prekindergarten year. This is good news, as it suggests that this type of program can be adapted to meet the scheduling demands of a particular school or child care center, without adversely affecting its potential instructional impact. In addition, it is not uncommon for children to enter prekindergarten midyear, and these study results are encouraging for such children.

Years 2 and 3 of this multiyear research addressed several of the limitations identified within the Year 1 study. Specifically, the research outcomes measures were strengthened through the addition of the TOPEL beginning in Year 2 and then the ALL RHY in Year 3. The psychometric properties of these tests as compared to those of GGG resulted in more readily interpretable and robust results. Inclusion of a larger number of child care and preschool sites and participating children strengthened statistical power and increased confidence that these results are replicable and generalizable across many types of prekindergarten instructional settings. The continued use of an immediate versus delayed intervention group design, coupled with the rigorous HLM analyses, provided a strong experimental design that could be carried out in an ecologically valid, natural environment.

Implications of These Studies

The current research shows that an achievement gap in emergent literacy exists and can be measured at least as early as the beginning of the prekindergarten year. More importantly, our research indicates that this gap can be reduced substantially for many children who receive differentiated and more intensive instruction (i.e., RTI Tier 2). The majority of participants made meaningful progress toward reading readiness, with two thirds of participants moving from the below average to the average range on GRTR during the study year. Even children who scored in the most at-risk range on their baseline GRTR screening (raw scores of 0 to 4) averaged gains of approximately 8 points on their GRTR posttest for Years 2 and 3. Thus, study children not only made significant gains in reading preparedness but also actually reduced the gap with their typical peers by demonstrating an accelerated learning pace during intervention.

Not surprisingly, however, there was variability in the range of outcomes, as was the case in Year 1. Ample existing research has demonstrated that, regardless of the nature, length, and intensity of a remedial intervention, some children will fail to progress as hoped. Such findings are not without benefit, however. In the future, such findings for individual prekindergarteners could be used to initiate more comprehensive evaluation and intervention from an earlier age within the RTI framework (i.e., a Tier 3 component). From a research perspective, follow-up studies to identify critical differences in RTI would be helpful in designing future differentiated intervention.

Limitations and Future Research Directions

Although a strength of this study was its implementation in many typical preschool and child care settings rather than a laboratory or clinical setting, a limitation was the delivery of the experimental intervention by project teachers rather than classroom teachers. Although this model was necessary to ensure adequate instructional quality and consistency for the present research purposes, such a model does not allow speculation as to whether the intervention would be equally successful in the hands of typical early childhood professionals. A future research goal is to study the effectiveness of such intervention when delivered by typical child care workers and preschool teachers who receive various levels of training and support, for which initial projects are underway. Demonstrating this capability will be critical in encouraging efforts to implement similar programs in multiple communities (NELP, 2008). The 2008 NELP report specifically calls for translational research for this very reason.

We believe that the finely tuned activities within and across lessons within our experimental curriculum worked synergistically to produce the encouraging results that were obtained across 3 years of study. Subsequent research will help to clarify the most powerful components of these lessons and whether teaching them in different formats (e.g., whole-class or center-based activities) and for longer time periods will result in similar or even greater benefits. For example, emergent writing activities and parent letters both increased significantly from Year 1 to Year 3. Additional research to tease apart whether either of these components makes a quantifiable contribution to outcomes would be valuable. Emergent literacy assessment instruments have progressed substantially from a theoretical and psychometric standpoint since this research was begun. Nonetheless, longitudinal research is needed to identify the predictive validity of these various measures as well as the potential long-term impact of this early intervention relative to future reading success. Additional research addressing individual child, family, and school characteristics that relate to both immediate and longer term RTI also will be essential in developing targeted programs that reduce the risk of reading failure in children.

Footnotes

Appendix A

Appendix B

Author’s note

This research was made possible through funding by Nemours.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Adams

M. J.

(1990). Beginning to read: Thinking and learning about print. Cambridge, MA: MIT Press.

Anthony

J. L.

Francis

(2005). Development of phonological awareness. Current Directions in Psychological Science, 14, 255–259.

Anthony

J. L.

Lonigan

C. J.

(2004). The nature of phonological awareness: Converging evidence from four studies of preschool and early grade school children. Journal of Educational Psychology, 96, 43–55.

Anthony

J. L.

Lonigan

C. J.

Driscoll

Phillips

B. M.

Burgess

S. R.

(2003). Phonological sensitivity: A quasi-parallel progression of word structure units and cognitive operations. Reading Research Quarterly, 38, 470–487.

Bailet

L. L.

(2001). Development and disorders of spelling in the beginning school years. In Bain

A. M.

Bailet

L. L.

Moats

L. C.

(Eds.), Written language disorders: Theory into practice (2nd ed., pp. 1–41). Austin, TX: Pro-Ed.

Bailet

L. L.

Repper

K. K.

Piasta

S. B.

Murphy

S. P.

(2009). Emergent literacy intervention for prekindergarteners at risk for reading failure. Journal of Learning Disabilities, 42, 336–255.

Bierman

K. L.

Domitrovich

C. E.

Nix

R. L.

Gest

S. D.

Welsh

J. A.

Greenberg

M. T.

Gill

(2008). Promoting academic and social-emotional school readiness: The Head Start REDI program. Child Development, 6, 1802–1817.

Brownell

(2000). Expressive One-Word Picture Vocabulary Test (3rd ed.). Novato, CA: Academic Therapy.

Buysse

Winton

Zimmerman

(2007). RTI goes to pre-k: An early intervening system called recognition and response. Early Developments, 11, 6–10.

10.

Cohen

(1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.

11.

Connor

C. M.

Morrison

F. J.

Slominski

(2006). Preschool instruction and children’s emergent literacy growth. Journal of Educational Psychology, 98, 665–689.

12.

Dodd

Crosbie

McIntosh

Teitzel

Ozanne

(2003). Pre-reading Inventory of Phonological Awareness. San Antonio, TX: Psychological Corporation.

13.

Early Childhood Research Institute on Measuring Growth and Development. (1998). Research and development of individual growth and development indicators for children between birth and age eight (Tech. Rep. No. 4). Minneapolis: University of Minnesota, College of Education and Human Development. Retrieved from http://cehd.umn.edu/ceed/projects/ecri/default.html

14.

Ehri

L. C.

Roberts

(2006). The roots of learning to read and write: Acquisition of letters and phonemic awareness. In Dickinson

D. K.

Neuman

S. B.

(Eds.), Handbook of early literacy research (Vol. 2, pp. 113–131). New York, NY: Guilford.

15.

Episcopal Children’s Services of Northeast Florida. (2006). Links to Literacy. Lewisville, NC: Kaplan Early Learning.

16.

Farver

J. M.

Nakamoto

Lonigan

C. J.

(2007). Assessing preschoolers’ emergent literacy skills in English and Spanish with the Get Ready to Read! screening tool. Annals of Dyslexia, 57, 161–178.

17.

Fielding

Kerr

Rosier

(2007). Annual growth for all students: Catch-up growth for those who are behind. Kennewick, WA: New Foundation Press.

18.

Florida Institute for Education and the University of North Florida. (2002). Early Language and Literacy Model. Jacksonville: Author.

19.

Foorman

B. R.

Anthony

Seals

Mouzaki

(2002). Language development and emergent literacy in preschool. Seminars in Pediatric Neurology, 9, 173–184.

20.

Fuchs

L. S.

(1998). Researchers and teachers working together to adapt instruction for diverse learners. Learning Disabilities Research and Practice, 13, 126–137.

21.

Fuchs

L. S.

(2006). Introduction to response to intervention: What, why, and how valid is it? Reading Research Quarterly, 27, 216–227.

22.

Fuchs

(2003). Assessing intervention responsiveness: Conceptual and technical issues. Learning Disabilities Research and Practice, 18, 172–186.

23.

Gettinger

Stoiber

K. C.

(2008). Applying a response-to-intervention model for early literacy development in low-income children. Topics in Early Childhood Special Education, 27, 198–213.

24.

Harcourt Assessment. (2003). Early Reading Diagnostic Assessment (2nd ed.). San Antonio, TX: Author.

25.

Justice

L. M.

Chow

S. M.

Capellini

Flanigan

Colton

(2003). Emergent literacy intervention for vulnerable preschoolers: Relative effects of two approaches. American Journal of Speech-Language Pathology, 12, 320–332.

26.

Justice

L. M.

Mashburn

A. J.

Hamre

B. K.

Pianta

R. C.

(2008). Quality of language and literacy instruction in preschool classrooms serving at-risk pupils. Early Childhood Research Quarterly, 23, 51–68.

27.

Koutsoftas

A. D.

Harmon

M. T.

Gray

(2009). The effect of tier 2 intervention for phonemic awareness in a response-to-intervention model in low-income preschool classrooms. Language, Speech & Hearing Services in Schools, 40, 116–130.

28.

Landry

S. H.

Anthony

J. L.

Swank

P. R.

Monseque-Bailey

(2009). Effectiveness of comprehensive professional development for teachers of at-risk preschoolers. Journal of Educational Psychology, 101, 448–465.

29.

Lieberman

I. Y.

Shankweiler

(1985). Phonology and problems of learning to read and write. Remedial and Special Education, 6, 8–17.

30.

Lombardino

L. J.

Lieberman

R. J.

Brown

J. C.

(2005). Assessment of Literacy and Language. San Antonio, TX: Harcourt Assessment.

31.

Lonigan

C. J.

Anthony

J. L.

Phillips

B. M.

Purpura

D. J.

Wilson

S. B.

McQueen

J. D.

(2009). The nature of preschool phonological processing abilities and their relations to vocabulary, general cognitive abilities, and print knowledge. Journal of Educational Psychology, 101, 345–358.

32.

Lonigan

C. J.

Burgess

S. R.

Anthony

J. L.

Barker

T. A.

(1998). Development of phonological sensitivity in 2- to 5-year old children. Journal of Educational Psychology, 90, 294–311.

33.

Lonigan

C. J.

Wagner

R. K.

Torgesen

J. K.

(2007). Test of Preschool Early Literacy. Austin, TX: Pro-Ed.

34.

Lonigan

C. J.

Wagner

R. K.

Torgesen

J. K.

Rashotte

C. A.

(2007). Test of Preschool Early Literacy: Examiner’s manual. Austin, TX: Pro-Ed.

35.

Lundberg

Frost

Petersen

O. P.

(1988). Effects of an extensive program for stimulating phonological awareness in preschool children. Reading Research Quarterly, 23, 263–284.

36.

McBride-Chang

(1999). The ABC’s of the ABC’s: The development of letter-name and letter-sound knowledge. Merrill-Palmer Quarterly, 45, 278–301.

37.

Metsala

J. L.

Walley

A. C.

(1998). Spoken vocabulary growth and the segmental restructuring of lexical representations: Precursors to phonemic awareness and early reading ability. In Metsala

J. L.

Ehri

L. C.

(Eds.), Word recognition in beginning literacy (pp. 89–120). Mahwah, NJ: Lawrence Erlbaum.

38.

Moats

L. C.

Farrell

M. L.

(2005). Multisensory structured language education. In Birsh

J. R.

(Ed.), Multisensory teaching of basic language skills (2nd ed., pp. 23–42). Baltimore, MD: Brookes.

39.

National Early Literacy Panel. (2008). Developing early literacy: Report of the National Early Literacy Panel. Washington, DC: National Institute for Literacy.

40.

National Institute of Child Health and Human Development. (2000). Report of the National Reading Panel. Teaching children to read: An evidence-based assessment of the scientific research literature on reading and its implications for reading instruction (NIH Pub. No. 00-4769). Washington, DC: Government Printing Office.

41.

Neuman

S. B.

Dickinson

D. K.

(Eds.). (2002). Handbook of early literacy research. New York, NY: Guilford.

42.

Piasta

S. B.

Wagner

R. K.

(2010). Developing early literacy skills: A meta-analysis of alphabet learning and instruction. Reading Research Quarterly, 45, 8–38.

43.

Raudenbush

S. W.

Bryk

A. S.

(2002). Hierarchical linear models: Applications and data analysis method (3rd ed.). Thousand Oaks, CA: Sage.

44.

Raudenbush

S. W.

Bryk

A. S.

Congdon

(2005). HLM: Hierarchical linear and nonlinear modeling (Version 6.02a). Lincolnwood, IL: Scientific Software International.

45.

Reid

D. K.

Hresko

W. P.

Hammill

D. D.

(2001). Test of Early Reading Ability. Austin, TX: Pro-Ed.

46.

Schneider

Kuspert

Roth

Vise

Marx

(1997). Short and long term effects of training phonological awareness in kindergarten: Evidence from two German studies. Journal of Experimental Child Psychology, 66, 311–340.

47.

Schneider

Roth

Ennemoser

(2000). Training phonological skills and letter knowledge in children at risk for dyslexia: A comparison of three kindergarten intervention programs. Journal of Educational Psychology, 92, 284–295.

48.

Shadish

W. R.

Cook

T. D.

Campbell

D. T.

(2002). Experimental and quasi-experimental designs for generalized causal inference. Boston, MA: Houghton Mifflin.

49.

Shadish

W. R.

Haddock

C. K.

(2009). Combining estimates of effect size. In Cooper

Hedges

Valentine

(Eds.), The handbook of research synthesis and meta-analysis (2nd ed., pp. 257–277). New York, NY: Russell Sage.

50.

Shaywitz

S. E.

Shaywitz

B. A.

(2004). Neurobiologic basis for reading and reading disability. In McCardle

Chhabra

(Eds.), The voice of evidence in reading research (pp. 417–442). Baltimore, MD: Brookes.

51.

Slocum

T. A.

O’Connor

R. E.

Jenkins

J. R.

(1993). Transfer among phonological manipulation skills. Journal of Educational Psychology, 85, 618–630.

52.

Snow

C. E.

Burns

M. S.

Griffin

(Eds.). (1998). Preventing reading difficulties in young children. Washington, DC: National Academy Press.

53.

Stanovich

K. E.

(1986). Matthew effects in reading: Some consequences of individual differences in the acquisition of literacy. Reading Research Quarterly, 21, 360–407.

54.

Stone

C. A.

Silliman

E. R.

Ehren

B. J.

Apel

(Eds.). (2004). Handbook of language and literacy: Development and disorders. New York, NY: Guilford.

55.

Stubbs

(1980). Language and literacy: The sociolinguistics of reading and writing. Boston, MA: Routledge Kegan Paul.

56.

Torgesen

J. K.

(2002). The prevention of reading difficulties. Journal of School Psychology, 40, 7–26.

57.

Treiman

(1992). The role of intrasyllabic units in learning to read and spell. In Gough

P. B.

Ehri

L. C.

Treiman

(Eds.), Reading acquisition (pp. 65–106). Hillsdale, NJ: Lawrence Erlbaum.

58.

Treiman

Tincoff

Rodriguez

Mouzaki

Francis

D. J.

(1998). The foundations of literacy: Learning the sounds of letters. Child Development, 69, 1524–1540.

59.

Wagner

R. K.

(1988). Causal relations between the development of phonological processing abilities and the acquisition of reading skills: A meta-analysis. Merrill-Palmer Quarterly, 34, 261–279.

60.

Wagner

R. K.

Torgesen

J. K.

(1987). The nature of phonological processing and its causal role in the acquisition of reading skills. Psychological Bulletin, 101, 192–212.

61.

Wanzek

Vaughn

(2007). Research-based implications from extensive early reading interventions. School Psychology Review, 36, 541–561.

62.

Welsch

J. G.

Sullivan

Justice

L. M.

(2003). That’s my letter! What preschoolers’ name writing representations tell us about emergent literacy knowledge. Journal of Literacy Research, 35, 757–776.

63.

West

Denton

Reaney

L. M.

(2000). The kindergarten year: Findings from the early childhood longitudinal study, kindergarten class of 1998–99. Washington, DC: National Center for Education Statistics.

64.

Whitehurst

G. J.

(2001). Get Ready to Read. New York, NY: Pearson Early Learning.

65.

Whitehurst

G. J.

(2003). The NCLD Get Ready to Read screening tool technical report. Retrieved from http://www.getreadytoread.org

66.

Whitehurst

G. J.

Lonigan

C. J.

(1998). Child development and emergent literacy. Child Development, 69, 848–872.

67.

Whitehurst

G. J.

Lonigan

C. J.

(2002). Emergent literacy: Development from prereaders to readers. In Neuman

S. B.

Dickinson

D. K.

(Eds.), Handbook of early literacy research (pp. 11–29). New York, NY: Guilford.

68.

Yeh

S. S.

(2003). An evaluation of two approaches for teaching phonemic awareness to children in Head Start. Early Childhood Research Quarterly, 18, 513–529.