Teaching and Teacher Education 50 (2015) 102e113
Contents lists available at ScienceDirect
Teaching and Teacher Education
journal homepage: www.elsevier.com/locate/tate
Evidence-based practices to stimulate emergent literacy skills in
kindergarten in France: A large-scale study
le
ne Labat a, b, f, Marion Le Cam c, Thierry Rocher c, Laurent Cros d,
Jean Ecalle a, f, *, He
Annie Magnan a, e, f
Laboratoire EMC Etude des M
ecanismes Cognitifs (EA 3082), Universit
e Lyon2, France
Laboratoire Paragraphe, Equipe Compr
ehension, Raisonnement et Acquisition des Connaissances (EA 349), Universit
e Paris8, France
c
DEPP Direction de l'Evaluation Prospective et de la Performance, Minist
ere de l'Education Nationale, France
d
Association Agir pour l'Ecole, Paris, France
e
IUF Institut Universitaire de France, France
f
LabEx Cortex Lyon ANR-11-LABX-0042, France
a
b
h i g h l i g h t s
Evidence-based (EB) literacy practices were proposed in a randomized controlled trial.
Analyses involving propensity scores were conducted to examine their impact.
These revealed a global effect on trained literacy skills in experimental group.
The impact was more robust in the children with the lowest scores.
The links between EB research, EB practices and EB policy are presented.
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 22 August 2014
Received in revised form
18 March 2015
Accepted 18 May 2015
Available online 22 May 2015
In a randomized controlled trial with 3569 kindergarten children, evidence-based literacy practices
(EBLP) were proposed by teachers to an experimental group (EG). A control group did not receive any
specific interventions during the same period. The EBLP related to the alphabetic code, phonological
awareness and oral comprehension. Analyses based on propensity scores showed significant gains in the
targeted domains and in pseudoword reading in EG after comparison between the two groups. The gains
were higher for children who had the lowest scores. However, no effect was observed in word reading
and vocabulary. EBLP could be a valuable pedagogical tool.
© 2015 Elsevier Ltd. All rights reserved.
Keywords:
Emergent literacy
Interventions
Word reading
Reading comprehension
Propensity scores analysis
1. Introduction
As in other countries, and in particular English-speaking parts of
the world where government funds are used to enhance school
readiness (Griffiths & Stuart, 2013), the French educational authorities tend to promote evidence-based practices to reduce difficulties in children during kindergarten and more specifically to
stimulate literacy skills. Indeed, a significant proportion of children
s* Corresponding author. Laboratoire EMC, University of Lyon2, 5, av Mende
dex, France.
France, 69676 Bron Ce
E-mail addresses: [email protected], [email protected] (J. Ecalle).
http://dx.doi.org/10.1016/j.tate.2015.05.002
0742-051X/© 2015 Elsevier Ltd. All rights reserved.
in France, and in particular those from lower socio-economic status
(see Fluss et al., 2008), experience considerable difficulties when
learning to read. The general purpose of this article is to present
how new pedagogical practices implemented by teachers trained in
evidence-based research could improve literacy skills in young
children.
For over three decades, evidence-based research has clearly
revealed that reading is underpinned by two components, namely
word recognition and reading comprehension. Indeed, developmental studies have provided evidence of distinctive and stable
r, Keenan, Olson, Byrne, &
predictors of these components (Elwe
Samuelsson, 2013; Kendeou, Van den Broek, White, & Lynch,
J. Ecalle et al. / Teaching and Teacher Education 50 (2015) 102e113
2009). Based on this scientific knowledge, it might be possible to
promote best practices to stimulate the emergent literacy skills
which are considered to be the foundation of reading (Greenwood,
Tapia, Abbott, & Walton, 2003). To this end, a large-scale study
conducted with more than three thousand children in kindergarten
was carried out and is reported here.
1.1. Learning to read and emergent literacy skills
According to the Simple View of Reading (Gough & Tunmer,
1986; but see also Aaron, Joshi, Gooden, & Bentum, 2008;
Kendeou, Savage, & van den Broeck, 2009), reading comprehension can be thought of as the product of word decoding, which is
specific to reading and is responsible for translating print into
language, and language comprehension skills that make sense of
this linguistic information. Numerous studies have indicated that
reading acquisition requires many important component skills,
namely phonological processing abilities, print knowledge, and oral
r et al.,
language, e.g., vocabulary, grammar, comprehension (Elwe
2013; Oakhill & Cain, 2012). The evidence indicates that these
skills are present during the preschool period. Thus, some emergent literacy skills are code-related, and other emergent literacy
skills are meaning-related.
1.1.1. Predictors of word recognition
Among the various candidate predictors (in alphabetic languages) that might explain improvements in word reading, it seems
that letter knowledge and phonological awareness are among the
best and most robust predictors of multiple reading outcomes in
Grades 1 and 2 (for a recent French longitudinal study, see: Costa
et al., 2013; in Finnish: Puolakahano et al., 2007; in English:
Schatschneider, Fletcher, Francis, Carlson, & Foorman, 2004; in
Hebrew: Levin, Shatil-Carmon, & Asif-Rave, 2006). Indeed, according to the self-teaching hypothesis, the development of word
reading is based on decoding procedure which requires the
involvement of letter knowledge (name and sound) and phonological awareness. Knowledge about letters (their shapes, their
names, and their linguistic functions) is known to play an important role in the development of reading and spelling ability (Foulin,
2005; Huang, Tortorelli, & Invernizzi, 2014; Treiman, 2006) provided that the children are proficient speakers of the language
being read. Children's ability to identify letter names and letter
sounds has been shown to be one of the best indicators of reading
achievement (Ecalle, Magnan, & Biot-Chevrier, 2008; Puranik,
Petscher, & Lonigan, 2013).
Moreover, a large body of evidence has emphasized the
important role of phonological awareness as a significant predictor
of the learning of word reading (Castles & Coltheart, 2004). More
precisely the pivotal role of phonemic awareness as a predictor of
individual differences in reading development has recently been
pointed out in a meta-analytic review (Melby-Lervåg, Lyster, &
Hulme, 2012). Researchers have suggested that the relationship
between phonological awareness and reading is bidirectional such
that phonological awareness facilitates reading abilities and
reading acquisition in turn improves phonological awareness
(Morais, 2003). These studies have often been limited to an investigation of the relations between explicit awareness of phonological
units and reading. They have not considered the early phonological
sensitivity children acquire through implicit learning before formal
instruction (however, see Ecalle & Magnan, 2002, 2007; Savage,
Blair, & Rvachew, 2006). Many studies suggest that phonological
awareness is a single, unified ability that is present during the
preschool and early elementary school years and that manifests
itself in a variety of skills throughout a child's development
(Anthony & Francis, 2005).
103
In summary, a number of different studies have shown that
kindergarten measures of phonological awareness and alphabet
knowledge are highly predictive of reading achievement in the
primary grades. As a result, the present study will place the
emphasis on children's phonological awareness and alphabet
knowledge.
1.1.2. Predictors of reading comprehension
Numerous studies have examined different predictors of
reading comprehension. The literature emphasizes four of these
predictors: grammatical skills (Muter, Hulme, Snowling, &
Stevenson, 2004; Nation & Snowling, 2000; Oakhill, Cain, &
Bryant, 2003), working memory (Cain, Oakhill, & Bryant, 2004),
inferencing (Kendeou, Bohn-Gettler, White, & Van Den Broeck,
2008), and more importantly, vocabulary and oral comprehension.
Indeed, measures of general oral language have repeatedly been
found to be strongly related to early reading achievement, specifically in the domain of reading comprehension. Results have shown
that early comprehension performance, assessed in 6-year-old
children kindergarten or in 4-year-old children preschool by means
of non-reading tasks, is highly predictive of later reading comprehension performance (Kendeou et al., 2008). Moreover, vocabulary
has also been identified as a skill that plays a critical role in reading
comprehension performance throughout the elementary period
(Verhoeven, van Leeuwe, & Vermeer, 2011). The lexical quality
hypothesis states that the degree of comprehension is influenced
by the size of the lexicon, as well as the quality and flexibility of
individual lexical representations (Perfetti & Stafura, 2014). In
addition, a strong relationship between vocabulary and comprehension has been found in school-age children (Cain & Oakhill,
2011; Lonigan, Burgess, & Anthony, 2000; Vellutino, Tunmer,
, &
Jaccard, & Chen, 2007), and young children (Florit, Roch, Altoe
Levorato, 2009; Roth, Speece, & Cooper, 2002). However, according to the lexical restructuring hypothesis, vocabulary can provide
the foundations for phonological sensitivity (Dickinson, McCabe,
Anastasopoulos, Peisner-Feinberg, & Poe, 2003).
Thus, different emergent literacy skills are differentially predictive of different components of reading. In this study, we focused
on letter knowledge, phonological skills, vocabulary and oral
comprehension. The present research mobilizes evidence-based
pedagogic tools developed within the framework of developmental cognitive psychology studies that have attempted to identify the main predictors of learning to read. Our aim is to test the
effectiveness of early interventions in boosting the literacy skills
described as predictive of reading and to examine their impact on
early reading performance.
1.2. Evidence-based emergent literacy practices
During preschool and kindergarten, children show considerable
variability in their levels of emergent literacy skills (Cabell, Justice,
Konold, & McGinty, 2011). Numerous studies suggest that the
majority of reading problems could be prevented by reducing the
number of children who enter school with low levels of emergent
literacy skills (Snow, Burns, & Griffin, 1998). Targeted interventions
could be proposed in the light of evidence-based research. “Evidence-based practices are instructional techniques with meaningful research supporting their effectiveness that represent critical
tools in bridging the research-to-practice gap and improving student outcomes” (Cook & Cook, 2011, p. 71). Because two types of
specific predictors of reading ability have been identified, one for
word reading and one for reading comprehension, two types of
early intervention could be used in a focused way to boost codefocused and/or meaning-focused skills.
104
J. Ecalle et al. / Teaching and Teacher Education 50 (2015) 102e113
1.2.1. Code-focused interventions
Letter knowledge and phonological awareness are related to the
code in so far as they are needed to decode new words on the basis
of grapheme-phoneme correspondences. Studies involving interventions in children with reading difficulties have shown that
letter-sound knowledge and phoneme awareness are two causal
influences on the development of children's early literacy skills
(Hulme, Bowyer-Crane, Carroll, Duff, & Snowling, 2012). A metaanalysis has provided evidence that phonological awareness
training in conjunction with instruction in letter-sound correspondences can facilitate the process of learning to read (Ehri et al.,
2001). It is therefore recommended that phonological awareness
training take place before the start of formal instruction. Numerous
studies of the effectiveness of interventions involving both
phonological awareness training and practice in the learning of
sound-symbol correspondences have been conducted and been
found to be efficient (Magnan & Ecalle, 2006; Byrne & FieldingBarnsley, 2000; de Graaff, Bosman, Hasselman, & Verhoeven,
2009; Torgesen et al., 1999).
1.2.2. Meaning-focused interventions
Comprehension skills refer, for example, to the detection of inconsistencies, resolving logical inferences, and understanding story
structure in narratives. The fact that comprehension skills in the
preschool years predict later reading comprehension suggests that
the development of these skills may benefit from separate and
targeted instruction as early as the preschool years. At this age,
children are not yet proficient at reading and this type of instruction would therefore need to take place in non-reading contexts. A
number of studies have suggested that comprehension strategies
generalize to a considerable degree across media and, therefore,
that comprehension interventions can be conducted using nonreading materials, for example in the form of aural or televised
stories (Fuchs, Fuchs, Mathes, & Simmons, 1997).
Several studies have demonstrated that comprehension skills
can be stimulated through explicit teaching. This type of instruction
has been shown to have a positive effect on comprehension performance (Trabasso & Bouchard, 2002). Different kinds of
comprehension instruction have been used. For instance,
knowledge-based inference was trained in 7e8-year-old poor
comprehenders (Yuill & Oakhill, 1988). In another study (de Corte,
Verschaffel, & Van De Ven, 2001) conducted among children in 5th
grade, text comprehension strategies (such as activating prior
knowledge, formulating the main idea, etc.) were trained using
interactional techniques. These studies, like many others, have
found positive effects on children's comprehension performance
(see also the meta-analyses of Berkeley, Scruggs, and Mastropieri
(2010), and Edmonds et al. (2009)).
1.2.3. Multiple interventions
Many children who are at risk for later reading difficulties are
often at risk for problems in the domains of both code and language
comprehension. Studies indicate that the effects of interventions
are specific to these domains. For example, in a large sample of
more than one thousand 4-year-old children monitored over a
period of 3 years (Bianco et al., 2010), phonological training
improved phonological awareness but not oral comprehension,
while comprehension training improved oral comprehension but
not phonological awareness. These results underline the independence of these two domains. Two comprehension programs were
proposed. These involved the analysis of story-books and were
designed to stimulate comprehension skills In addition, phonological awareness training in pre-kindergarten and kindergarten
has been found to have a positive effect on reading skills (words
and pseudowords) in Grade 1 while comprehension training has
been shown to have a positive effect on reading comprehension
(Bianco et al., 2012).
Furthermore, Bowyer-Crane et al. (2008) compared the effects
of a 20-week oral language comprehension intervention and a 20week code-focused intervention that taught kindergarten children
phonological awareness and reading skills. The results obtained at
post-test and in a follow-up assessment performed 6 months later
showed that the children who received the code-focused intervention performed better than those who received the oral language intervention on measures of letter knowledge, phonological
awareness, spelling, and reading. In contrast, the children who
received the oral language intervention performed better than
those in the code-focused intervention group on measures of vocabulary and grammar. In another study, an early intervention
targeting various oral language skills (code- and meaning-focused)
had a beneficial effect on later reading comprehension (Fricke,
Bowyer-Crane, Haley, Hulme, & Snowling, 2013). Similarly, after a
print referencing intervention during classroom-based storybook
reading sessions conducted over an academic year, Justice,
McGinty, Piasta, Kaderavek, and Fan (2010) reported that the positive effects of the intervention did not extend to measures of
children's language comprehension skills. Children whose teachers
used a print referencing style exhibited larger gains only on standard measures of print knowledge (i.e., print concept knowledge,
alphabet knowledge, and name writing). Recently, Lonigan,
Purpura, Wilson, Walker, and Clancy-Menchetti (2013) examined
the effectiveness of interventions designed to increase the development of emergent literacy skills with a sample of preschool
children who were at risk for later problems in reading, and to
evaluate the specific effects of different interventions. In this study,
324 preschoolers were randomly assigned to combinations of
meaning-focused (dialogic reading or shared reading) and codefocused (phonological awareness, letter knowledge) interventions, to both interventions or to a control group. First, the
results showed that the scores in the three interventions groups
were higher than those of the control group (children who received
only their classroom curriculum). Secondly, the impact was more
robust in the targeted domain. And thirdly, when interventions
were combined (for example, dialogic reading plus phonological
awareness or dialogic reading plus letter knowledge), no larger
effects were found, contrary to the authors' predictions.
To summarize, multiple interventions seem to be effective in the
targeted domains, on the one hand, and, on the other, have a more
global effect on reading comprehension by stimulating both codefocused and meaning-focused skills.
1.3. Purpose of the current study
The purpose of this study was 1/to propose interventions
combining both code-focused and meaning-focused skills and 2/to
cover a large number of participants in different regions. Two target
domains related to reading were stimulated, with code-focused
interventions designed to develop the low-level processes
involved in word reading being used, on the one hand, and
meaning-focused interventions intended to develop high-level
comprehension processes, on the other. We expected that codefocused interventions relating to the alphabetic code and phonological skills would impact letter knowledge, phonological skills
and, more specially, word and pseudoword reading. Moreover,
stimulating the various processes involved in comprehension
should impact oral comprehension. We also expected that
comprehension-related interventions would impact vocabulary
since improved comprehension could help promote vocabulary
acquisition during other language activities such as shared storybook reading.
J. Ecalle et al. / Teaching and Teacher Education 50 (2015) 102e113
2. Method
A randomized controlled trial with two groups, one experimental and one control, was designed to examine the impact of
evidence-based literacy practices during a school year in kindergarten. First, all the teachers were recommended to assess the literacy of the children in their classes using booklets which were
provided to them (see below). Then, in order to implement the
innovation, the teachers in the experimental group were trained for
a day in the use of the new pedagogical practices recommended by
evidence-based research (see below).
2.1. Participants
Before the beginning of the study, a large number of teachers in
different urban areas of France were contacted by academic authorities and invited to participate in this study. Ninety-eight
teachers agreed. They then received a letter from the Minister of
National Education giving them more information about the
research and randomly assigning their class to one of the groups.
Their classes were assigned to the experimental group (N ¼ 2067;
48 kindergartens) and seventy-two classes were selected (from a
nationally representative sample) to form the control group
(N ¼ 1502; 32 kindergartens).1
A large number of children (N ¼ 3569; mean age: 5.9 y-o;
sd ¼ .32; range: 59e71 months) participated to the study and were
evaluated at the beginning (t1) and at the end (t2) of kindergarten
in 118 schools. The gender ratio was 50.8% boys, 45.6% girls (3.6%
were not specified). Their SES varied from low to high and they
were all French speaking (but not all had French as L1), and had no
specific problems identified.2
The teachers scheduled the assessment sessions during normal
classroom hours. They received two booklets. One was intended for
them and contained all the instructions for the assessments that is,
the explicit orders for each task which was administered in small
groups. The other booklet was for each of the children, who wrote
their answers in it using a pen. The responses were collected and
recorded by one of the authors (with the help of assistants who
were trained to record the data). In the control group, conventional
classroom teaching included non-formal learning of emergent literacy, namely an item forming part of the kindergarten curriculum.
2.2. Measures and procedure
The construction of the tasks proved to be difficult due to time
constraints. This is because the teachers did not have much time to
conduct their assessment sessions. As a result, only four domains
related to reading acquisition were investigated during the two
sessions (t1 and t2). Two of these were code-focused, i.e. letter
knowledge and phonological skills, and two were meaningfocused, i.e. vocabulary and oral comprehension. At t2, the number of items was reduced (the easiest items were discarded) and
two reading tasks were added, one using words and the other
pseudowords. The tasks were administered to the children in small
groups (5e8). The children were given a booklet in which they were
told to circle the correct answer in all the tasks. In each task, in
order to capture the children's attention, they were told to look at
an image in front of each item. For example they were told: “You see
a white star. Put your finger on. You must circle the picture …”. The
1
It was more difficult to enroll teachers in the control group.
The data of children who had significant difficulties identified by a school
psychologist were not taken into account in this study. We also have no specific
information about these children's mother tongue.
2
105
tasks were administered in the same order across groups and took
place over three sessions.
2.2.1. Letter knowledge
This task related to letter-name knowledge only. Uppercase
letters were successively named by the teacher. The children were
asked to circle the named letter which was presented in a set of 7
letters. Because the scores were high at t1, only the letters with the
lowest scores at t1 were presented at t2. The number of correct
answers was recorded (max ¼ 26 at t1; max ¼ 15 at t2).
2.2.2. Phonological skills
The same two tasks were proposed at t1 and t2. In the oddity
task, the teacher named three pictures and children had to find and
circle the picture corresponding to the word which did not sound
like the others (cheveux (hair), pigeon (pigeon), chemin (path)). The
words shared either a syllable (6 items) or a phoneme (6 items) at
initial or final position. The children completed two practice trials
(one with a common syllable and the other with a common
phoneme).
The deletion task required the children to delete a phonological
unit in a word. They had to retrieve the initial syllable of a bisyllabic
word, delete it, and then circle the new word (ex: pinceau (brush) e
seau (bucket), crayon (pen), arc (bow), landau (pram)). One practice
trial was completed followed by six test items. The number of
correct answers in the two tasks was recorded (max ¼ 18).
2.2.3. Vocabulary
The children's receptive vocabulary was assessed using a classical procedure similar to that used in the well-known Peabody
riault-Whalen, & Dunn, 1993). Four
vocabulary test (Dunn, The
pictures were presented and the children had to circle the one
corresponding to the named word (example: for the target word
luge (toboggan), the other word-pictures were fraise (strawberry);
ski; cygne (swan)). Twenty-three words, 12 high-frequency and 11
low-frequency, were selected from a French lexical database (New,
Brysbaert, Veronis, & Pallier, 2007). These words were presented
after one practice trial. Because high scores were observed at t1, the
items with the lowest scores were retained at t2 (max ¼ 13).
2.2.4. Oral comprehension
Listening comprehension was assessed using the same short
narrative (11 sentences, 136 words) at t1 and t2. The story was read
aloud twice by the teacher (no pictorial aids were used). The children were instructed to listen closely so that they could answer
questions after the story was over. Comprehension was evaluated
using a forced-choice task. For each question, the children had to
circle the picture named by the teacher out of a selection of four
pictures. Twelve questions were asked: 3 based on information
explicitly stated in the text (literal comprehension) and 9 requiring
the generation of inferences (text-connecting, knowledge-based
and elaborative inferences; see Cain & Oakhill, 1999). This means
that the children were required to use both textual information and
other processes when responding. The number of correct answers
was recorded (max ¼ 12).
2.2.5. Reading
2.2.5.1. Word reading. A silent, forced-choice reading task was
administered at t2. The teacher named a word and the children had
to recognize and circle the written word in a selection of five test
items. The target word was present in a list of 5 items consisting of
the orthographically correct word (e.g., lapin, rabbit), and 4 pseudowords, namely a homophone (lapain), a visually similar item
(lapiu), an item sharing the same initial letters (lapon) and an item
containing an illegal letter sequence (lpina). After a practice trial, 10
106
J. Ecalle et al. / Teaching and Teacher Education 50 (2015) 102e113
words were named. The words were selected from a lexical database based on their frequency of occurrence in books for 3 to 5-y-o
children (Arabia-Guidet, Chevrie-Muller, & Louis, 2000). The
number of correct responses was recorded (max ¼ 10).
2.2.5.2. Pseudoword reading. This task was based on the same
principle: The children were required to circle the pseudword
named by the teacher from among five test items (e.g., pseudoword
named: mida; test items: mida-nida-mipa-mdia-ufno). After a
practice trial, 10 pseudowords with different structure were proposed, Consonant-Vowel (2), VC (2), CVCV (4), CCVCV (2). The
number of correct responses was recorded (max ¼ 10).
problem by means of different strategies. Twenty-one sessions
were planned and different modules focused on various processes
involved in comprehension (Bianco, Coda, Gourge, & Robert, 2002).
Some example exercises were given for each module. In the
“detection of inconsistencies” module, the children had to find
oddities in pictures depicting a situation and explain why this was
an inappropriate representation of the situation. In the “situation
model” module, they had to represent the situation described in a
narrative. In the “causality” module, the children had to identify the
causes and effects of actions. In the case of the “anaphora” module,
they had to understand and interpret the anaphoric iterations inside a text, and in the “deduction” module, the teacher asked the
children to eliminate the clues that were not relevant in a text.
2.3. Interventions
The children in the experimental group were taught in special
pedagogical sessions as described below. Two aspects of literacy
skills were targeted: on the one hand, code-focused processes
involved in word reading, and on the other, meaning-focused processes involved in reading comprehension. These sequences were
explicit, systematic, gradual and intensive. The children in the control group continued with their conventional classroom teaching.
2.3.1. Stimulating code-focused processes (alphabetic code and
phonological awareness training)
The aim of the alphabetic code training was a/to identify and
recognize the letters of the alphabet (letter-name, letter-sound and
three letter forms, i.e., uppercase, script, cursive) and b/connect
orthographic and phonological representations in order to read
simple syllables (Mirgalet & Zorman, 2011).
The training consisted of 4 sets of sessions. In each set, session 1
focused on simple and regular letters (A, L, I, R, T, O, P, M, U, B),
session 2 on other regular letters (F, E, N, D, V, Q, J) and complex
letters (C, G, S) in which each letter could be matched with several
sounds and inversely. The aim of session 3 was to learn inconsistent
letters (K, H, Z, X, Y, W). Finally, session 4 focused on learning how
to read the CV syllable. During these exercises, the children saw a
card with a letter (e.g., the letter “l”), a graphical depiction which
represented a key-word with the target letter in the initial position
(e.g., the graphical depiction of loup (wolf) and the corresponding
letter-sound (/l/)).
The aim of the phonological awareness training was to stimulate
word listening, segmental analysis, verbal memory, and articulatory cues (Zorman & Jacquier-Roux, 2002). During this training,
different linguistic units (rhyme, syllable and phoneme) were
involved in two levels of phonological skills: implicit or epilinguistic processing and explicit or metalinguisitic processing (for
this distinction, see Ecalle & Magnan, 2002; Savage et al., 2006).
Overall, 36 exercises were performed. On the one hand, the epilinguistic skills (11 exercises) referred to the discrimination of oral
units such as syllable and rhyme (e.g., finding words which shared
the same units, moving away from the meaning to focus on the
phonological cues, breaking the word down into syllables or
merging the syllables to form a word). On the other, the metalinguistic skills (25 exercises) focused on phonemic awareness and
letter-sound mapping (e.g., removing and merging phonemes,
segmenting syllables into phonemes and matching them with the
orthographic units).
2.3.2. Stimulating meaning-focused processes (comprehension
training)
Listening comprehension was trained in sessions during which
the teachers stimulated comprehension skills by administering
explicit exercises. The aim was to encourage the elaboration of a
situation model which required the children to solve a situational
2.3.3. Implementation training
The children were trained by their teachers in their normal
classrooms. The teachers were instructed by educational advisors
to teach the children in small groups (4e7 per group) and in split
classes comprising children of similar academic levels in terms of
letter knowledge, phonological awareness and comprehension.
During these special sessions, the other children in the same
classroom continued to perform their normal classroom activities.3
The experimental protocol required the teachers to train each
domain intensively from January to June. The alphabetic code
training was administered twice a week for each group (two split
classes), the phonological training was also administered twice a
week for each group (4 groups per class, around 6 children per
group) and the oral comprehension training was administered once
a week for each group (4 groups per class). Each training session
lasted 30 min. To summarize, each child should have received
approximately 9 h of oral comprehension training, 18 h of alphabetic code training and 18 h of phonological training.
3. Results
After a psychometric analysis of the items in each task and the
presentation of descriptive data, the potential impact of the intervention was examined by comparing the scores of the two groups
after the intervention, i.e. at t2, while taking account of the groups'
initial levels. Following the analysis of the global effect, the subsequent analyses investigated the differential effects of intervention on the basis of the children's initial and final levels in each
domain.
3.1. Psychometric analysis
All the tasks required forced-choice responses. Performance on
all the items was therefore first controlled for: None of these performances was at chance level. Biserial coefficients were then
calculated for each item in all tasks. When this coefficient was less
than .20, the item was excluded from the following analyses.
Cronbach's alphas were computed with the retained items in order
to determine internal consistency (Table 1).
3.2. Data in literacy skills
Descriptive data and skewness are presented for each group at
t1 and t2 in Table 2. It can be noted that, contrary to expectations,
the results obtained for the experimental group at t1 were poorer
3
In fact, teachers in France are not always trained to teach specific instructional
designs to small groups of children. Consequently, training sessions administered in
experimental groups were organized for the teachers who were also familiarized
with new pedagogical material promoted by researchers in educational psychology.
J. Ecalle et al. / Teaching and Teacher Education 50 (2015) 102e113
107
Table 1
Item analysis with internal consistency (Cronbach's alpha) in each domain at session 1 and 2.
t1
Letter K
Phono S
Voc
Oral Comp
Word R
Pword R
t2
nb items
nb items retained
a
a (equiv)a
nb items
nb items retained
a
a (equiv)a
26
18
23
12
e
e
26
17
23
12
e
e
.93
.77
.82
.72
e
e
.93
.84
.84
.85
15
18
13
12
10
10
15
17
13
12
8
10
.86
.80
.72
.73
.58
.77
.87
.80
.77
.79
.75
.85
Notes: Letter K: letter knowledge; Phono S: phonological skills; Voc: vocabulary; Oral Comp: oral comprehension; Word R: word reading; Pword R: pseudoword reading.
a
This new Cronbach alpha was calculated for an equivalent number of items in each domain (as if all tasks had 26 items as in letter knowledge at t1 and 17 items as in
phonological skills at t2).
Table 2
Descriptive data in experimental and control group with Manovas for comparison between groups at t1 and t2.
Experimental group
N
Manovaa
Control group
Mean (SD)
Skewness
N
Mean (SD)
Skewness
F; p; (h2)
34.84
.0001
(.01)
84.74
.0001
(.02)
35.62
.0001
(.01)
2.69
ns
t1
Letter K (/26)
2028
21 (6.26)
1.37
1439
22.27 (5.27)
1.82
Phono S (/17)
1978
7.02 (3.51)
.32
1357
8.23 (3.63)
.09
Voc (/23)
2021
1.07
1457
18.6 (3.59)
1.24
Oral comp (/12)
2011
6.87 (2.75)
.10
1415
7 (2.74)
.19
Letter K (/15)
1971
13.87 (2.24)
3.04
1449
13.77 (2.32)
2.86
Phono S (/17)
1932
10.62 (3.37)
.71
1395
10.41 (3.45)
.61
Voc (/13)
2008
10.46 (2.58)
1.43
1447
10.93 (2.14)
Oral comp (/12)
1944
8.5 (2.53)
.64
1431
8.52 (2.65)
.59
1907
1.87 (1.68)
1.21
1422
2 (1.72)
1.06
1913
5.7 (2.7)
.14
1412
17.81 (4.1)
t2
Word R
(/8)
Psword R (/10)
1.26
ns
2.42
ns
16.67
.0001
(.01)
1.06
ns
3.84
.05
(.00)
51.01
.0001
(.02)
1.4
.23
5.01 (2.7)
Notes: Letter K: letter knowledge; Phono S: phonological skills; Voc: vocabulary; Oral Comp: oral comprehension; Word R: word reading; Psword R: pseudoword reading.
a
One-way MANOVA for comparison between groups at t1 and t2.
than those observed in the control group (see results of Manovas in
Table 2). This observation underlines the need to use specific statistical analyses to examine the impact of interventions. High
negative skewness for the letter knowledge and vocabulary of both
groups at t1 and t2 means that their scores were very high. This was
unexpected given the precautions taken during experimental
design. Table 3 shows the correlations between scores in each
domain. All of these were significantly positive. Among the highest
coefficients, those between t1 and t2 for vocabulary (.61), letter
knowledge (.57), oral comprehension (.57) and phonological skills
(.53) indicated some stability in the individual scores for each
domain. Moreover, other high coefficients should be noted, namely
between oral comprehension and vocabulary (.56 at t1 and .52 at
t2) and between pseudoword reading and phonological skills at t2
(.54).
3.3. Global effect of the interventions
Generally, in observational studies, random assignment is not
feasible. In this study, the fact that there was a difference, in favor of
the control group, between the two groups before the beginning of
the interventions meant that we needed to use a specific technique
Table 3
Correlationsa matrix between scores at t1 and t2 (N ¼ 3056).
Let1
Co1
Ph1
Voc2
Let2
Co2
Ph2
WR2
pWR2
Voc1
Let1
OC1
Ph1
Voc2
Let2
OC2
Ph2
WR2
.42
.56
.53
.61
.28
.51
.47
.22
.32
.27
.40
.32
.57
.28
.38
.24
.43
.55
.43
.17
.57
.45
.21
.29
.44
.25
.47
.53
.33
.41
.30
.52
.46
.20
.33
.26
.37
.18
.41
.54
.24
.34
.33
.54
.47
pWR2
Notes: Voc: vocabulary; Let: letter knowledge; OC: oral comprehension; Ph:
phonological skills; WR: word reading; pWR: pseudoword reading; 1: session 1; 2:
session 2.
a
All coefficients are significant at .05.
108
J. Ecalle et al. / Teaching and Teacher Education 50 (2015) 102e113
to achieve the best match between the experimental and control
groups. Propensity score-based approaches (matching, stratification, and weighting) may be used “to approximate the randomization process by creating treated and untreated groups that are
equivalent (in expectation) on measured confounding variables”
(Sullivan & Field, 2013, p. 245). For an individual, the propensity
score is the conditional probability of being treated given the individual covariates. In our analysis, we decided to use genetic
matching (for a description of the technique, see Diamond &
Sekhon, 2013). In general terms, this technique is based on
matching the data of the treatment and control groups in order to
estimate causal treatment effects. Using the matching technique,
propensity scores are calculated in such a way that each treated
individual is matched with the untreated individual with the
closest score provided that the difference between the two scores is
not too large.
Recently used in the education field (Sullivan & Field, 2013;
Wen, Leow, Hahs-Vaughn, Korfmacher, & Marcus, 2012), genetic
matching was used in our study to measure the mean effect of the
interventions on letter knowledge, phonological abilities, comprehension in terms of the scores obtained in literacy skills assessed at
t2, i.e., letter knowledge, phonological skills, vocabulary, comprehension and word and pseudoword reading, while taking account
of the scores at t1 for letter knowledge, phonological skills, vocabulary and comprehension. Missing data at t1 have been eliminated (N ¼ 338). The statistical package R was used to run analyses
(R Core Team, 2013).
A significant effect of the interventions (Table 4) was observed
on letter knowledge (15%), phonological ability (25%), comprehension (15%) and pseudoword reading (41%). The absence of a
significant gain in vocabulary is not very surprising given that the
children did not receive any specific vocabulary-related intervention. At the same time, no significant gain was observed in word
reading even though a more substantial improvement might have
been expected due to the code-focused intervention which targeted letter knowledge and phonological abilities.
3.4. Differential effects of intervention
Two types of analyses were used to examine the specific impact
of the interventions based on the inter-individual differences
identified first at t1 and then at t2. The first of these were conducted on four (when this was possible; see below) clusters as a
function of their initial levels at t1. The children were subdivided
into four clusters: cluster A with the highest standardized scores,
i.e., more than one standard deviation above the mean, cluster B
with scores between þ1 sd and 0, cluster C with scores between 1
sd and 0 and cluster D with the lowest scores, i.e., more than one
standard deviation below the mean. A regression analysis was run
to obtain an estimate for interventions in letter knowledge,
phonological skills, vocabulary and oral comprehension after
chronological age, gender and educational level had been
Table 4
Results of genetic matching.
Letter knowledge
Phonological skills
Vocabulary
Oral comprehension
Word reading
Pseudoword reading
ATE
s.e.
T
p-value
Std control
Std effect
.31
.82
.00
.37
.03
1.1
.08
.12
.09
.10
.07
.10
4.05
6.65
.05
3.63
.43
10.8
.00
.00
.96
.00
.66
.00
1.99
3.33
2.16
2.52
1.72
2.68
.16
.25
.00
.15
.02
.41
Notes. ATE: average treatment effect; s.e.: standard error; T ¼ ATE/s.e.; std: standard
deviation.
Table 5
Standardized data at t1 in groups and estimates of interventions.
Clusters Experimental group
N
Mean (SD)
Letter knowledge
A
1217
.55
B
404
.41
C
432
2.23
Phonological skills
A
286
1.33
B
374
.49
C
812
.45
D
581
1.42
Vocabulary
a
A
152
1.23
B
917
.54
C
588
.52
D
396
2.10
Comprehension
A
390
1.39
B
712
.36
C
511
.52
D
440
1.48
(.19)
(.33)
(.78)
Control group
Effects of interventions
N
Estimate
Mean (SD)
1009
.55 (.20)
263
.39 (.30)
218 2.10 (.74)
SE
p-value
.05
.25
.57
.02
.07
.16
.05
.00
.00
(.27)
(.22)
(.30)
(.30)
309
1.38
375
.52
537
.41
269 1.48
(.29)
(.22)
(.30)
(.34)
.15
.23
.22
.35
.07
.08
.09
.10
.04
.00
.01
.00
(.00)
(.30)
(.29)
(.79)
122
1.23
763
.57
411
.50
194 1.95
(.00)
(.31)
.00
(.30) .08
(.75) .25
.07
.12
.18
.99
.55
.17
(.28)
(.30)
(.18)
(.41)
302
1.38
540
.35
378
.53
270 1.50
(.28)
(.30)
(.18)
(.44)
.07
.07
.09
.09
.41
.28
.65
.00
.06
.08
.04
.28
a
These two clusters had a maximum mean score and the standard deviation was
therefore zero. Consequently, no effect of interventions could be calculated.
controlled for (Table 5).
When ceiling scores were obtained for letter knowledge (30% of
the children achieved the maximum score), only three clusters
were distinguished. We observed that the interventions were statistically significant for clusters A (5%), B (25%), and C (57%). In the
case of phonological skills, the interventions were again statistically
significant in all four clusters, A (15%), B (23%), C (22%), and D (35%).
In the case of vocabulary, none of the clusters exhibited a significant
gain in favor of the experimental condition. Finally, in the case of
comprehension, the interventions brought about a significant gain
in the experimental group in cluster D only (28%).
The second set of analyses was based on the deciles of the scores
obtained at t2. In this case, a quantile regression analysis was carried out for each domain, except in the case of letter knowledge for
which ceiling scores were achieved (thus making it difficult to
discriminate deciles). The quantile regression analyses on scores at
t2 were conducted as a function of the scores observed at t1 (with
imputation4 of missing data at t2), as well as of gender and age. The
regressions were weighted by taking account of the propensity
scores explained by the same variables as in genetic matching. Fig. 1
(phonological skills), 2 (vocabulary), 3 (comprehension), 4 (word
reading), and 5 (pseudoword reading) present the coefficients
representing the effect of the intervention, calculated based on the
difference between the mean value of the experimental group
minus the mean value of the control group for each quantile.
The estimates of the intervention effects showed that, in the
case of phonological skills, comprehension and pseudoword
reading, the interventions had a greater impact in the children with
the lowest scores. The impact of intervention was particularly high
(around 50%) in quantiles 30 to 50 on pseudoword reading scores
(Fig. 5). Similarly, in the case of phonological skills, the greatest
impact was observed in quantiles 20 to 40 (from 30% to around
23%; Fig. 1). On the other hand, no significant effect (nearly null;
from 0 to 5%) was observed on vocabulary (Fig. 2) or word reading
(Fig. 4) irrespective of quantile. Finally, in the case of comprehension, gains of 15%e18% were observed for quantiles 20 to 30 (Fig. 3).
4
Missing data were replaced by a value computed using the Markov Chain
Monte Carlo method (based on Bayesian inference).
J. Ecalle et al. / Teaching and Teacher Education 50 (2015) 102e113
109
Fig. 1. Coefficients related to the impact of the intervention on phonological skills as a
function of the quantile of the propensity score at t2. Notes. Ph: Phonological skills;
Voc: vocabulary; OC: oral comprehension; WR: word reading; PWR: pseudoword
reading.
Fig. 3. Coefficients related to the impact of the intervention on oral comprehension as
a function of the quantile of the propensity score at t2. Notes. Ph: Phonological skills;
Voc: vocabulary; OC: oral comprehension; WR: word reading; PWR: pseudoword
reading.
Fig. 2. Coefficients related to the impact of the intervention on vocabulary as a function
of the quantile of the propensity score at t2. Notes. Ph: Phonological skills; Voc: vocabulary; OC: oral comprehension; WR: word reading; PWR: pseudoword reading.
Fig. 4. Coefficients related to the impact of the intervention on word reading as a
function of the quantile of the propensity score at t2. Notes. Ph: Phonological skills;
Voc: vocabulary; OC: oral comprehension; WR: word reading; PWR: pseudoword
reading.
To summarize: To examine the potential effect of the interventions, we first conducted a rigorous analysis of the items used
in each task. Very few items were rejected after the psychometric
analysis. We then ran specific data analyses to examine the impact
of evidence-based literacy practices scheduled during teaching
periods. Two approaches (global and differential) were successively
implemented to evaluate the short-term effect of interventions
during kindergarten.
4. Discussion
In response to evidence-based research, the academic authorities and the French Minister of Education decided to promote the
use of evidence-based literacy practices conducted by teachers in a
large number of kindergarten classes (i.e. before formal reading
instruction) which constituted an experimental group which was
then compared with a control group with which the teachers
continued to work as normal. This study is one of the most
important to have been undertaken in France (and in the literature)
given the very large sample of children involved (more than three
thousand) and the precautions taken during the different analyses
(for example, several items making a low contribution to the global
scores in each task were ultimately excluded). To examine the potential effect of the interventions, we first conducted a rigorous
analysis of the items used in each task. Very few items were
rejected after the psychometric analysis. We then ran specific data
analyses to examine the impact of evidence-based literacy practices
scheduled during teaching periods. Two approaches (global and
differential) were successively implemented to evaluate the shortterm effect of interventions during kindergarten.
In the case of letter knowledge, the interventions in the experimental group resulted in a significant overall effect of 15%
compared with the control group, with the effects decreasing from
110
J. Ecalle et al. / Teaching and Teacher Education 50 (2015) 102e113
Fig. 5. Coefficients related to the impact of the intervention on pseudoword reading as a
function of the quantile of the propensity score at t2. Notes. Ph: Phonological skills; Voc:
vocabulary; OC: oral comprehension; WR: word reading; PWR: pseudoword reading.
57% in cluster C, which had the poorest scores at t1, to 5% in cluster
A, which obtained the highest scores at t1. However, because of
ceiling effects at t2, no other analysis was possible. In the case of
phonological skills, the global effect amounted to 25% in the
experimental group, with significant gains being observed in the
four clusters and decreasing from cluster D to cluster A (35%, 22%,
23%, 15%). The quantile regression analysis performed on the scores
at t2 confirmed that the impact of the interventions was greater in
those children who obtained the lowest scores (around 25% in the
low deciles).
No global or differential effects were revealed in the case of
vocabulary, whereas a significant global effect of 15% was found for
oral comprehension, with the interventions having a more marked
effect in cluster D, which had the lowest scores at t1 (28%), and in
the low quantile (20) observed at t2 for which the effect was
approximately 20%. Finally, in word reading, no global and differential effect was found. In contrast, a major effect was observed for
pseudoword reading (41%). The differential analysis revealed a
noteworthy effect of around 50% in the low quantiles (from 30 to
50).
The results observed in this study were not obvious given that in
the past, this type of pedagogical intervention has not always systematically revealed a significant effect (Gentaz et al., 2013). Secondly, we did not observe any significant effect in domains which
were not directly targeted during the interventions, such as word
reading and vocabulary. In the case of word reading, this is not
particularly surprising given that it is first necessary for children to
make intensive use of phonological decoding before progressively
constructing an orthographic lexicon that makes it possible to read
words. In the case of vocabulary, we expected an indirect effect of
oral comprehension given that the training in the various types of
processing involved in comprehension might contribute to a better
understanding of narrative during shared book-reading and
consequently stimulate vocabulary in children. However, this was
not the case. In addition, the ceiling effects observed in connection
with vocabulary could mask the potential impact of oral comprehension interventions on the acquisition of new words, in particular in children with the lowest vocabulary levels. Moreover, the
short period of the oral comprehension intervention (9 h)
compared to other studies (more than 30 h in Bower-Crane et al.'s
study (2008) or in Clarke, Snowling, Truelove, & Hulme's study
(2010)) could explain the absence of an impact of oral comprehension training on vocabulary.
In this study, word recognition was not trained directly. However, by stimulating code-focused domains related to word recognition (phonological skills, letter knowledge and alphabetic code)
we were able to reveal performance improvements in these three
domains and in particular in the children with the lowest scores, i.e.
at risk of reading failure.
One of the most important results observed here was the
remarkable gain in the experimental group observed in the pseudoword reading task as a result of the training of the alphabetic
code and phonological skills. Melby-Lervag et al. (2012) underline
the need to provide phonemic awareness and letter-sound training
to boost early reading and spelling skills. Stimulating decoding at
an early age could help trigger an initial procedure in young children which is very important for the acquisition of word reading.
We expect that the interventions proposed in kindergarten will
have an effect on word recognition in first grade as in Bianco et al.
(2012). At the same time, reading comprehension, as a form of
high-level processing, is underpinned by oral comprehension
which involves various skills (inferencing, mental model building,
anaphora processing, etc.). Longitudinal studies have clearly shown
that early listening comprehension and vocabulary are closely
related to reading comprehension (Kendeou et al., 2008; Kendeou,
Savage, et al., 2009, Kendeou, Van Den Broek, et al., 2009;
Verhoeven et al., 2011). If oral comprehension is to be stimulated
(perhaps for more than the 9 h during the school year used in our
study), then vocabulary-related interventions could be proposed to
children with poor lexical knowledge. In future evidence-based
practices conducted in France (for English-speaking children, see
Coyne et al., 2010; Loftus & Coyne, 2013), it would seem to be of
value to promote such interventions because vocabulary is a core
component in reading and acts both as a predictor of phonological
sensitivity and as a predictor of comprehension (see Section 1.1.2).
Some limitations should be pointed out. First, in the control
group, no specific information was collected about literacy practices and the time devoted to this. Similarly, we did not obtain any
information about the fidelity of adherence to instructions (that is,
implementation fidelity) in the experimental group. Secondly, the
class teachers administered the pre- and post-tests and this could
potentially pose a problem in terms of the neutrality of assessment.
However, this would not impact the effect of the intervention
because the same was true of both groups. Thirdly, the observed
ceiling effects in letter knowledge were unexpected. The task was
constructed on the basis of data obtained some years ago and
proved to be suitable for use with kindergarten children (Ecalle et
al., 2008). Since 2008, new instructions from the French Minister
of Education have highlighted the need to teach the alphabet and
letter names to children in kindergarten and this could explain the
high scores observed in this domain. In the case of vocabulary, high
scores were also observed despite the selection of low-frequency
words from the recent Lexique database (New et al., 2007). This
means that a new task should be constructed to assess, in particular, the quality of conceptual knowledge relating to words (see
Kearns & Biemiller, 2010). Finally, the data for the two groups,
experimental and control were not approximately equivalent at t1.
However, the recent data analyses borrowed from the field of
economics (Diamond & Sekhon, 2013) and applied recently to education (Wen et al., 2012) make it possible to evaluate the impact of
interventions in two groups that differ in their initial levels.
To summarize, in this large-scale study, the interventions had an
impact on the targeted domains (phonological skills, alphabetic
code and oral comprehension). When children receive stimulation
regarding the alphabetic code and its use (i.e. translating letters
into sounds, i.e. by using a decoding procedure), pseudoword
J. Ecalle et al. / Teaching and Teacher Education 50 (2015) 102e113
reading is boosted, in particular in low-scoring children.
In this sample, children will be followed during first and second
grade to examine the long-term effects of early interventions on
reading and spelling. Moreover, in the ongoing analyses, differential analyses will allow us to go beyond an expected global effect
and examine the impact of training as a function of inter-individual
differences. In fact, in the future, such specific pedagogical sessions
should be organized more specifically to help children at risk of
failure (see Lonigan et al., 2013). However, in order to achieve better
external validity, assessment sessions should be scheduled by
neutral experimenters and the faithful adherence to the proposed
intervention model will have to be monitored more precisely.
The experimental literature has highlighted the two independent components of reading, namely word recognition and
comprehension, and their separate and specific predictors (see
Section 1). The many scientific studies conducted on this topic
indicate that preventive (in the sense of proactive) actions, i.e. actions conducted before the start of formal reading instruction,
could prove to be a valuable pedagogical tool in preparing children
for reading acquisition. Reading difficulties in word recognition
and/or comprehension could be identified at an early age. It seems
that the best way to prevent these difficulties is to use multiple
interventions using evidence-based literacy practices. Indeed, these
have already shown their effectiveness on the specific skills which
were trained (Bianco et al., 2010; Bower-Crane et al., 2008; Fricke
et al., 2013; Justice et al., 2010; Lonigan et al., 2013). Our study
confirms these results by additionally highlighting the fact that the
interventions had the greatest effect in the children with the lowest
scores.
More generally, it is possible to distinguish between two types
of training study. The aim of ecological studies is to give teachers
specific pedagogical tools whose use can be explained by researchers over multiple sessions. In this case, teachers are themselves the “experimenters” examining ways of using the new tools
designed by researchers. It is therefore important that they remain
faithful to the intervention model if it is to act efficiently. In this
case, it is hard to verify the quality of implementation fidelity. On
the other hand, in experimental training studies, sessions are
scheduled directly and are strictly monitored by one (or more)
experimenter(s) during school hours after, it goes without saying,
obtaining the consent of the teachers (see for example, Ecalle,
Kleinsz, & Magnan, 2013; Labat, Ecalle, Baldy, & Magnan, 2014).
Here, the implementation of the innovative design is fully
controlled by the researchers and their partners, who are involved
directly during the training in schools. Finally, to summarize, these
two types of training study, ecological and experimental, both
constitute evidence-based (EB) practices; however implementation
fidelity could vary between the types and, in particular, cannot be
directly controlled during “ecological” experiments.
111
Fig. 6. Links between evidence-based levels accounting for evidence-based educational model.
In the field of learning and academic achievement, EB research
feeds EB practices and, in turn, by using specifically designed
training approaches, new EB practices can strengthen fundamental
research conducted in order to discover how processes and/or
knowledge are involved in learning. EB research could also inspire
educational policies relating to EB practice by highlighting how the
fundamental processes involved in learning can contribute to the
teaching guidelines set out by policy makers. Similarly, EB practices
can help mold educational policy and help stimulate the use of new
classroom practices, in particular in order to assist children at risk.
If educational policy makers observe a disjunction between
fundamental evidence-based research and the practices applied in
the field, they can encourage, or indeed commission, new ecological or experimental research which can be analyzed and taken into
account to drive the development of new pedagogic guidelines.
This would make it possible for educational policies to be underpinned by reliable empirical information (Jonson-Reid, 2012). Such
an approach could result in the promotion of “evidence-based
education” (Cook & Cook, 2011). The ecological training study reported in this paper could help contribute to educational policy. To
our knowledge, it is to date the only study of such a scope to have
been conducted in France.
Acknowledgments
This work was supported by the French Ministry of Education
(Department for Youth, Community Education, and Life Skills and
the Evaluation, Forecasting and Performance Department (DEPP),
the Acting For School Association (Agir pour l'Ecole), the LabEx
Cortex (“Construction, Function and Cognitive Function and Rehabilitation of the Cortex”, ANR-10-LABX-0042) at the University of
Lyon (France), within the program “Investissements d'Avenir”
(ANR-11-IDEX-0007) organized by the French National Research
Agency (ANR).
5. Conclusions
References
A final point, more generally oriented towards educational
policy, should also be considered. Indeed, in the light of the
empirical information resulting from EB research and EB practices,
policy makers could call on a variety of studies to drive new
educational policy, in particular in order to combat potential
learning difficulties which are known to be more present in children from low socio-economic status. Consequently, it is possible to
distinguish between three EB approaches which are potentially
interlinked, as illustrated, in Fig. 6 in order to lay the foundations
for an evidence-based educational model. The third of these is an
EB policy based on empirical studies which have been strictly
selected as a function of their validity and significance in a given
domain (David, 2002).
Aaron, P. J., Joshi, R. M., Gooden, R., & Bentum, K. E. (2008). Diagnosis and
treatment of reading disabilities based on the component model of reading.
Journal of Learning Disabilities, 41(1), 67e84. http://dx.doi.org/10.1177/
0022219407310838.
Anthony, J. L., & Francis, D. (2005). Development of phonological awareness. Current
Directions in Psychological Science, 14, 255e259. http://dx.doi.org/10.1111/
j.0963-7214.2005.00376.x.
quence d'occurrence
Arabia-Guidet, C., Chevrie-Muller, C., & Louis, M. (2000). Fre
des mots dans les livres d'enfants de 3 a 5 ans [Frequency of words in books for
3e5 y-o children]. Revue Europ
eenne de Psychologie Appliqu
ee, 50, 3e16.
Berkeley, S., Scruggs, T. E., & Mastropieri, M. A. (2010). Reading comprehension
instruction for students with learning disabilities, 1995e2006: a meta-analysis.
Remedial and Special Education, 36, 423e436. http://dx.doi.org/10.1177/
0741932509355988.
Bianco, M., Bressoux, P., Doyen, A. L., Lambert, E., Lima, L., Pellenq, C., et al. (2010).
Early training of oral comprehension and phonological skills at preschool: the
results of a 3 years longitudinal study. Scientific Studies of Reading, 14, 211e246.
112
J. Ecalle et al. / Teaching and Teacher Education 50 (2015) 102e113
http://dx.doi.org/10.1080/10888430903117518.
Bianco, M., Coda, M., Gourge, D., & Robert, E. (2002). Compr
ehension GS-CP-CE
soutien [Support to comprehension in kindergarten, first and second grade].
Grenoble, France: Editions de la Cigale.
Bianco, M., Pellenq, C., Lambert, E., Bressoux, P., Lima, L., & Doyen, A. L. (2012).
Impact of early code-skills and oral-comprehension training on reading
achievement in first grade. Journal of Research in Reading, 35, 427e455. http://
dx.doi.org/10.1111/j.1467-9817.2010.01479.x.
Bowyer-Crane, C., Snowling, M. J., Duff, F. J., Fieldsend, E., Carroll, J. M., Miles, J., et al.
(2008). Improving early language and literacy skills: differential effects of an
oral language versus a phonology with reading intervention. Journal of Child
Psychology and Psychiatry, 49, 422e432. http://dx.doi.org/10.1111/j.14697610.2007.01849.x.
Byrne, B., & Fielding-Barnsley, R. (2000). Effects of preschool phoneme identity
training after six years: outcome level distinguished from rate of response.
Journal of Educational Psychology, 92, 659e667. http://dx.doi.org/10.1037/00220663.92.4.659.
Cabell, S. Q., Justice, L. M., Konold, T. R., & McGinty, A. S. (2011). Profiles of emergent
skills among preschool children who are at risk for academic difficulties. Early
Childhood
Research
Quarterly,
26,
1e14.
http://dx.doi.org/10.1016/
j.ecresq.2010.05.003.
Cain, K., & Oakhill, J. (1999). Inference making ability and its relation to comprehension failure. Reading and Writing, 11, 489e503. http://dx.doi.org/10.1023/A:
1008084120205.
Cain, K., & Oakhill, J. (2011). Matthew effects in young readers: reading comprehension and reading experience and vocabulary development. Journal of
Learning Disabilities, 44, 431e443. http://dx.doi.org/10.1177/0022219411410042.
Cain, K., Oakhill, J., & Bryant, P. E. (2004). Children's reading comprehension ability:
concurrent prediction by working memory, verbal ability, and component skills.
Journal of Educational Psychology, 96, 31e42. http://dx.doi.org/10.1037/00220663.96.1.31.
Castles, A., & Coltheart, M. (2004). Is there a causal link from phonological awareness to success in learning to read? Cognition, 91, 77e111. http://dx.doi.org/
10.1016/S0010-0277(03)00164-1.
Clarke, P. J., Snowling, M. J., Truelove, E., & Hulme, C. (2010). Ameliorating children's
reading-comprehension difficulties: a randomized controlled trial. Psychological
Science, 21(8), 1106e1116. http://dx.doi.org/10.1177/0956797610375449.
Cook, B. G., & Cook, S. C. (2011). Unraveling evidence-based practices in special
education. The Journal of Special Education, 47, 71e82. http://dx.doi.org/10.1177/
0022466911420877.
de Corte, E., Verschaffel, L., & Van De Ven, A. (2001). Improving text comprehension
strategies in upper primary school children: a design experiment. British
Journal of Educational Psychology, 71, 531e559. http://dx.doi.org/10.1348/
000709901158668.
Costa, H. C., Perdry, H., Soria, C., Pulgar, S., Cusin, F., & Dellatolas, G. (2013). Emergent
literacy skills, behavior problems and familial antecedents of reading difficulties: a follow-up study of reading achievement from kindergarten to fifth
grade. Research in Developmental Disabilities, 34, 1018e1035. http://dx.doi.org/
10.1016/j.ridd.2012.11.029.
Coyne, M. D., McCoach, D. B., Loftus, M. S., Zipoli, R., Ruby, M., Crevecoeur, Y. C., et al.
(2010). Direct and extended vocabulary instruction in kindergarten: investigating transfer effects. Journal of Research on Educational Effectiveness, 3,
93e120. http://dx.doi.org/10.1080/19345741003592410.
David, M. E. (2002). Introduction: themed section on evidence-based policy as a
concept for modernising governance and social science research. Social Policy
and Society, 1, 213e214. http://dx.doi.org/10.1017/S1474746402003056.
Diamond, A., & Sekhon, J. S. (2013). Genetic matching for estimating causal effects: a
general multivariate matching method for achieving balance in observational
studies. Review of Economics and Statistics, 95, 932e945.
Dickinson, D. K., McCabe, A., Anastasopoulos, L., Peisner-Feinberg, E. S., & Poe, M. D.
(2003). The comprehensive language approach to early literacy: the interrelationships among vocabulary, phonological sensitivity, and print knowledge among preschool-aged children. Journal of Educational Psychology, 95,
465e481. http://dx.doi.org/10.1037/0022-0663.95.3.465.
riault-Whalen, C. M., & Dunn, L. M. (1993). Echelle de vocabulaire en
Dunn, L. M., The
images Peabody [French version of peabody picture vocabulary test revised]. Toronto: PsyCan.
Ecalle, J., Kleinsz, N., & Magnan, A. (2013). Computer-assisted learning in young poor
readers: the effect of grapho-syllabic training on word reading and reading
comprehension. Computers in Human Behavior, 29(4), 1368e1376. http://
dx.doi.org/10.1016/j.chb.2013.01.041.
Ecalle, J., & Magnan, A. (2002). The development of epiphonological and metaphonological processing at the start of reading: a longitudinal study. European
Journal of Psychology of Education, 17, 47e62.
Ecalle, J., & Magnan, A. (2007). Development of phonological skills and learning to
read in French. European Journal of Psychology of Education, 22, 153e167.
Ecalle, J., Magnan, A., & Biot-Chevrier, C. (2008). Alphabet knowledge and early
literacy skills in French beginning readers. European Journal of Developmental
Psychology, 5, 303e325. http://dx.doi.org/10.1080/17405620600901714.
Edmonds, M. S., Vaughn, S., Wexler, J., Reutebuch, C., Cable, A., Tackett, K. K., et al.
(2009). A synthesis of reading interventions and effects on reading comprehension outcomes on older struggling readers. Review of Educational Research,
79, 262e287. http://dx.doi.org/10.3102/0034654308325998.
Ehri, L., Nunes, R. S., Willows, D., Schuster, B. V., Yaghoub-Zadeh, Z., & Shanahan, T.
(2001). Phonemic awareness instruction helps children learn to read: evidence
from the national reading panel's meta-analysis. Reading Research Quarterly, 36,
250e287. http://dx.doi.org/10.1598/RRQ.36.3.2.
r, S., Keenan, J. M., Olson, R. K., Byrne, B., & Samuelsson, S. (2013). Longitudinal
Elwe
stability and predictors of poor oral comprehenders and poor decoders. Journal
of Experimental Child Psychology, 115, 497e516. http://dx.doi.org/10.1016/
j.jecp.2012.12.001.
, G., & Levorato, M. C. (2009). Listening comprehension in
Florit, E., Roch, M., Altoe
preschoolers: the role of memory. British Journal of Developmental Psychology,
27, 935e951. http://dx.doi.org/10.1348/026151008X397189.
Fluss, J., Ziegler, J., Ecalle, J., Magnan, A., Warszawski, J., Ducot, C., et al. (2008).
valence des troubles d'apprentissages du langage e
crit en de
but de scolarite
:
Pre
conomique au travers de trois zones d'e
ducations
l'impact du milieu socioe
distinctes [Prevalence of reading disabilities in early elementary school: impact
of socioeconomic environment on reading development in 3 different educational zones]. Archives de P
ediatrie, 15(6), 1049e1057. http://dx.doi.org/10.1016/
j.arcped.2008.02.012.
Foulin, J.-N. (2005). Why is letter-name knowledge such a good predictor of
learning to read. Reading and Writing: An Interdisciplinary Journal, 18, 129e155.
http://dx.doi.org/10.1007/s11145-004-5892-2.
Fricke, S., Bowyer-Crane, C., Haley, A. J., Hulme, C., & Snowling, M. J. (2013). Efficacy
of language intervention in the early years. Journal of Child Psychology and
Psychiatry, 54, 280e290. http://dx.doi.org/10.1111/jcpp.12010.
Fuchs, D., Fuchs, L. S., Mathes, P. G., & Simmons, D. C. (1997). Peer-assisted
learning strategies: making classrooms more responsive to diversity. American Educational Research Journal, 34, 174e206. http://dx.doi.org/10.3102/
00028312034001174.
, P., Theurel, A., Gurgant, M., Huron, C., et al.
Gentaz, E., Sprenger-Charolles, L., Cole
chelle
(2013). Evaluation quantitative d’un entrainement a la lecture a grande e
s en Re
seaux d’Education Prioritaire : apports et
pour des enfants de CP scolarise
limites [A large-scale study of intervention in reading in first graders of priority
education networks: contribution and limitations]. Approche Neuropsychologique des Apprentissages chez l'Enfant (ANAE), 123, 172e181.
Gough, P. B., & Tunmer, W. E. (1986). Decoding, reading, and reading disability.
Remedial and Special Education, 7, 6e10. http://dx.doi.org/10.1177/
074193258600700104.
de Graaff, S., Bosman, A. M. T., Hasselman, F., & Verhoeven, L. (2009). Benefits of
systematic phonics instruction. Scientific Studies of Reading, 13, 318e333. http://
dx.doi.org/10.1080/10888430903001308.
Greenwood, C. R., Tapia, Y., Abbott, M., & Walton, C. (2003). A building-based case
study of evidence-based literacy practices. The Journal of Special Education, 37,
95e110. http://dx.doi.org/10.1177/00224669030370020401.
Griffiths, Y., & Stuart, M. (2013). Reviewing evidence-based practice for pupils with
dyslexia and literacy difficulties. Journal of Research in Reading, 36, 96e116.
http://dx.doi.org/10.1111/j.1467-9817.2011.01495.x.
Huang, F. L., Tortorelli, L. S., & Invernizzi, M. A. (2014). An investigation of factors
associated with letter-sound knowledge at kindergarten entry. Early Childhood
Research Quarterly, 29, 182e192. http://dx.doi.org/10.1016/j.ecresq.2014.02.001.
Hulme, C., Bowyer-Crane, C., Carroll, J. M., Duff, F. J., & Snowling, M. J. (2012). The
causal role of phoneme awareness and letter-sound knowledge in learning to
read: combining intervention studies with mediation analyses. Psychological
Science, 23, 572e577. http://dx.doi.org/10.1177/0956797611435921.
Jonson-Reid, M. (2011). Disentangling system contact and services: a key pathway
to evidence-children's policy? Children and Youth Services Review, 33, 598e604.
http://dx.doi.org/10.1016/j.childyouth.2010.10.010.
Justice, L. M., McGinty, A., Piasta, S. B., Kaderavek, J., & Fan, X. (2010). Print-focused
read-aloud in preschool classrooms: intervention effectiveness and moderators
of child outcomes. Language, Speech, and Hearing Services in Schools, 41,
504e520. http://dx.doi.org/10.1044/0161-1461(2010/09-0056).
Kearns, G., & Biemiller, A. (2010). Two-questions vocabulary assessment: developing a new method for group testing in kindergarten through second grade.
Journal of Education, 190, 31e41.
Kendeou, P., Bohn-Gettler, C., White, M. J., & Van Den Broek, P. W. (2008). Children's
inference generation across different media. Journal of Research in Reading, 31,
259e272. http://dx.doi.org/10.1111/j.1467-9817.2008.00370.x.
Kendeou, P., Savage, R., & Van den Broek, P. (2009). Revisiting the simple view of
reading. British Journal of Educational Psychology, 79(2), 353e370. http://
dx.doi.org/10.1348/978185408X369020.
Kendeou, P., Van Den Broek, P. W., White, M. J., & Lynch, J. S. (2009). Predicting
reading comprehension in early elementary school: the independent contributions of oral language and decoding skills. Journal of Educational Psychology,
101, 765e778. http://dx.doi.org/10.1037/a0015956.
Labat, H., Ecalle, J., Baldy, R., & Magnan, A. (2014). How can low-skilled 5-year-old
children benefit from multisensory training on the acquisition of the alphabetic
principle. Learning and Individual Differences, 29, 106e113. http://dx.doi.org/
10.1016/j.lindif.2013.09.016.
Levin, I., Shatil-Carmon, S., & Asif-Rave, O. (2006). Learning of letter names and
sounds and their contribution to word recognition. Journal of Experimental Child
Psychology, 93, 139e165. http://dx.doi.org/10.1016/j.jecp.2005.08.002.
Loftus, S. M., & Coyne, M. D. (2013). Vocabulary instruction within a multi-tier
approach. Reading & Writing Quarterly, 29, 4e19. http://dx.doi.org/10.1080/
10573569.2013.741942.
Lonigan, C. J., Burgess, S. R., & Anthony, J. L. (2000). Development of emergent literacy and early reading skills in preschool children: evidence from a latent
variable longitudinal study. Developmental Psychology, 36, 596e613. http://
dx.doi.org/10.1037/0012-1649.36.5.596.
J. Ecalle et al. / Teaching and Teacher Education 50 (2015) 102e113
Lonigan, C. J., Purpura, D. J., Wilson, S. B., Walker, P. M., & Clancy-Menchetti, J.
(2013). Evaluating the components of an emergent literacy intervention for
preschool children at risk for reading difficulties. Journal of Experimental Child
Psychology, 114, 111e130. http://dx.doi.org/10.1016/j.jecp.2012.08.010.
Magnan, A., & Ecalle, J. (2006). Audio-visual training in children with reading disabilities. Computers & Education, 46, 407e425. http://dx.doi.org/10.1016/
j.compedu.2004.08.008.
Melby-Lervåg, M., Lyster, S. A. H., & Hulme, C. (2012). Phonological skills and their
role in learning to read: a meta-analytic review. Psychological Bulletin, 138,
322e352. http://dx.doi.org/10.1037/a0026744.
Mirgalet, F., & Zorman, M. (2011). Entraînement au code alphab
etique [Alphabetic
code training]. Grenoble, France: Editions de la Cigale.
Morais, J. (2003). Levels of phonological representation in skilled reading and in
learning to read. Reading and Writing, 16, 123e151. http://dx.doi.org/10.1023/A:
1021702307703.
Muter, V., Hulme, C., Snowling, M. J., & Stevenson, J. (2004). Phonemes, rimes, vocabulary and grammatical skills as foundations of early reading development:
evidence from a longitudinal study. Developmental Psychology, 40, 665e681.
http://dx.doi.org/10.1037/0012-1649.40.5.665.
Nation, K., & Snowling, M. J. (2000). Factors influencing syntactic awareness skills in
normal readers and poor comprehenders. Applied Psycholinguistics, 21,
229e241. http://dx.doi.org/10.1017/S0142716400002046.
New, B., Brysbaert, M., Veronis, J., & Pallier, C. (2007). The use of film subtitles to
estimate word frequencies. Applied Psycholinguistics, 28, 661e677. http://
dx.doi.org/10.1017/S014271640707035X.
Oakhill, J., & Cain, K. (2012). The precursors of reading ability in young readers:
evidence from a four-year longitudinal study. Scientific Studies of Reading, 16(2),
91e121. http://dx.doi.org/10.1080/10888438.2010.529219.
Oakhill, J., Cain, K., & Bryant, P. E. (2003). The dissociation of word reading and text
comprehension: evidence from component skills. Language and Cognitive Processes, 18, 443e468. http://dx.doi.org/10.1080/01690960344000008.
Perfetti, C., & Stafura, J. (2014). Word knowledge in a theory of comprehension.
Scientific Studies of Reading, 18, 22e37. http://dx.doi.org/10.1080/
10888438.2013.827687.
Puolakanaho, A., Ahonen, T., Aro, M., Eklund, K., Lepp€
anen, P. H. T., Poikkeus, A.-M.,
et al. (2007). Very early phonological and language skills: estimating individual
risk of reading disability. Journal of Child Psychology and Psychiatry, 48, 923e931.
http://dx.doi.org/10.1111/j.1469-7610.2007.01763.x.
Puranik, C. S., Petscher, Y., & Lonigan, C. J. (2013). Dimensionality and reliability of
letter writing in 3- to 5-year-old preschool children. Learning and Individual
Differences, 28, 133e141. http://dx.doi.org/10.1016/j.lindif.2012.06.011.
R Core Team. (2013). R: A language and environment for statistical computing. Vienna,
Austria: R Foundation for Statistical Computing, ISBN 3-900051-07-0. http://
www.R-project.org/.
Roth, F. P., Speece, D. L., & Cooper, D. H. (2002). A longitudinal analysis of the
113
connection between oral language and early reading. Journal of Educational
Research, 95, 259e272. http://dx.doi.org/10.1080/00220670209596600.
Savage, R., Blair, R., & Rvachew, S. (2006). Rimes are not necessarily favoured by prereaders: evidence from meta- and epi-linguistic phonological tasks. Journal of
Experimental Child Psychology, 94, 183e205. http://dx.doi.org/10.1016/
j.jecp.2006.03.005.
Schatschneider, C., Fletcher, J. M., Francis, D. J., Carlson, C. D., & Foorman, B. R.
(2004). Kindergarten prediction of reading skills: a longitudinal comparative
analysis. Journal of Educational Psychology, 96, 265e282. http://dx.doi.org/
10.1037/0022-0663.96.2.265.
Snow, C., Burns, M., & Griffin, P. (1998). Preventing reading difficulties in young
children. Washington, DC: National Academy Press.
Sullivan, A. L., & Field, S. (2013). Do preschool special education services make a
difference in kindergarten reading and mathematics skills? A propensity score
weighting analysis. Journal of School Psychology, 51, 243e260. http://dx.doi.org/
10.1016/j.jsp.2012.12.004.
Torgesen, J. K., Wagner, R. K., Rashotte, C. A., Lindamood, P., Rose, E., Conway, T.,
et al. (1999). Preventing reading failure in young children with phonological
processing disabilities: group and individual responses to instruction. Journal
of Educational Psychology, 91, 579e593. http://dx.doi.org/10.1037/00220663.91.4.579.
Trabasso, T., & Bouchard, E. (2002). Teaching readers how to comprehend text
strategically. In C. C. Block, & M. Pressley (Eds.), Comprehension instruction:
Research-based best practices (pp. 176e200). New York: The Guilford Press.
Treiman, R. (2006). Knowledge about letters as a foundation for reading and
spelling. In R. M. Joshi, & P. G. Aaron (Eds.), Handbook of orthography and literacy
(pp. 581e599). Mahwah, NJ: Erlbaum.
Vellutino, F. R., Tunmer, W. E., Jaccard, J. J., & Chen, R. (2007). Components of
reading ability: multivariate evidence for a convergent skill model of reading
development. Scientific Studies of Reading, 11, 3e32. http://dx.doi.org/10.1080/
10888430709336632.
Verhoeven, L., van Leeuwe, J., & Vermeer, A. (2011). Vocabulary growth and reading
development across the elementary school years. Scientific Studies of Reading,
15, 8e25. http://dx.doi.org/10.1080/10888438.2011.536125.
Wen, X., Leow, C., Hahs-Vaughn, D. L., Korfmacher, J., & Marcus, S. M. (2012). Are
two years better than one year? A propensity score analysis of the impact of
head start program duration on children's school performance in kindergarten.
Early Childhood Research Quarterly, 27, 684e694. http://dx.doi.org/10.1016/
j.ecresq.2011.07.006.
Yuill, N., & Oakhill, J. (1988). Effects of inference awareness training on poor reading
comprehension. Applied Cognitive Psychology, 2, 33e45. http://dx.doi.org/
10.1002/acp.2350020105.
Zorman, M., & Jacquier-Roux, M. (2002). Phono GS-CP soutien [Support to phonological awareness in kindergarten and first grade]. Grenoble, France: Editions de
la Cigale.