Developmental Psychology © 2013 American Psychological Association

2014, Vol. 50, No. 4, 1026 –1036 0012-1649/14/$12.00 DOI: 10.1037/a0035187

Orthographic and Phonological Contributions to Reading Development:

Tracking Developmental Trajectories Using Masked Priming

Johannes C. Ziegler and Daisy Bertrand Bernard Lété

Aix-Marseille Université and Centre National de la Recherche Université de Lyon
Scientifique, Marseille, France

Jonathan Grainger
Aix-Marseille Université and Centre National de la Recherche Scientifique, Marseille, France
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

The present study used a variant of masked priming to track the development of 2 marker effects of
orthographic and phonological processing from Grade 1 through Grade 5 in a cross-sectional study.
Pseudohomophone (PsH) priming served as a marker for phonological processing, whereas transposed-
letter (TL) priming was a marker for coarse-grained orthographic processing. The results revealed a clear
developmental picture. First, the PsH priming effect was significant and remained stable across devel-
opment, suggesting that phonology not only plays an important role in early reading development but
continues to exert a robust influence throughout reading development. This finding challenges the view
that more advanced readers should rely less on phonological information than younger readers. Second,
the TL priming effect increased monotonically with grade level and reading age, which suggests greater
reliance on coarse-grained orthographic coding as children become better readers. Thus, TL priming
effects seem to be a good marker effect for children’s ability to use coarse-grained orthographic coding
to speed up direct lexical access in alphabetic languages. The results were predicted by the dual-route
model of orthographic processing, which suggests that direct orthographic access is achieved through
coarse-grained orthographic coding that tolerates some degree of flexibility in letter order.

Keywords: pseudohomophone, transposed letter, lexical decision, dual route, reading development

Theories of reading development highlight two core processes:
phonological decoding and orthographic development (Frith,
1985; Grainger & Ziegler, 2011; Share, 1995; Ziegler & Goswami,
2005; Ziegler, Perry, & Zorzi, in press). Although much is known
about phonological decoding and its fundamental role in reading
development (Goswami, Ziegler, Dalton, & Schneider, 2001;
2003; Treiman, Goswami, & Bruck, 1990), our knowledge about
This article was published Online First December 2, 2013.
Johannes C. Ziegler and Daisy Bertrand, Laboratoire de Psychologie
Cognitive, Aix-Marseille Université, Marseille, France, and Centre Na-
tional de la Recherche Scientifique, Marseille, France; Bernard Lété,
Laboratoire d’Etude des Mécanismes Cognitifs, Université de Lyon, Lyon,
France; Jonathan Grainger, Laboratoire de Psychologie Cognitive, Aix-
Marseille Université and Centre National de la Recherche Scientifique.
Daisy Bertrand is now at the Centre d’Etudes et de Recherche en Gestion
d’Aix-Marseille, Faculté d’Economie et de Gestion, Aix-Marseille Univer-
sité, Aix-en-Provence, France.
This research was supported by a grant from the European Research
Council (ERC Grant 230313) to Jonathan Grainger. It was carried out
within the Labex BLRI (ANR-11-LABX-0036) and has benefited from
support from the French government, managed by the French National
Agency for Research, Investments of the Future AⴱMIDEX (ANR-11-
IDEX-0001-02).
Correspondence concerning this article should be addressed to Johannes
C. Ziegler, Laboratoire de Psychologie Cognitive (LPC, UMR7290), Aix-
Marseille Université, 3 place Victor Hugo, 13331 Marseille, France.
E-mail: Johannes.Ziegler@univ-amu.fr
orthographic development remains rather sketchy (Castles & Na-
tion, 2006). The key idea of all major theories on reading devel-
opment is that once children know letters and their names, they
apply this knowledge to learn about letter–sound correspondences
(Treiman, Sotak, & Bowman, 2001). The initially serial applica-
tion of these letter–sound correspondences (serial decoding strat-
egy) allows the child to access whole word phonology and,
through this, meaning. Every successfully decoded word provides
the child with an opportunity to set up direct connections between
a given letter string (orthography) and the spoken word (phonol-
ogy), which results in the development of an orthographic lexicon.
This learning loop has been recently implemented in a computa-
tional model of reading development (Ziegler et al., in press).
The prediction that follows from this general framework is that
the development of an orthographic lexicon should progressively
reduce the impact of phonological decoding on reading as children
become skilled readers, whereas reliance on orthographic access
processes should increase. Although this prediction is straightfor-
ward, the available data seem to suggest a more complex picture.
First of all, robust effects of phonological manipulations—such as
when people incorrectly classify a pseudohomophone (roze) as a
member of a semantic category (flower)—are still found in skilled
adult readers (Booth, Perfetti, & MacWhinney, 1999; Braun, Hut-
zler, Ziegler, Dambacher, & Jacobs, 2009; Diependaele, Ziegler, &
Grainger, 2010; Ferrand & Grainger, 1992; Frost, 1998; Lukatela
& Turvey, 1994a, 1994b; Rastle & Brysbaert, 2006; Van Orden,
1987; Ziegler, Van Orden, & Jacobs, 1997). Thus, phonological

1026
 ORTHOGRAPHIC AND PHONOLOGICAL MASKED PRIMING
information continues to play a role in skilled reading even if the
nature of the phonological activation changes in the course of
reading development. For example, the slow and serial activation
of phonology that characterizes early decoding strategies must be
gradually replaced by a more rapid and parallel process, as sug-
gested by the dramatic reduction of the word length effects in the
course of reading development (Acha & Perea, 2008; Aghababian
& Nazir, 2000; Ziegler, Perry, Ma-Wyatt, Ladner, & Schulte-
Korne, 2003).
Second, there are problems with the prediction that reliance on
orthographic access processes should increase during reading de-
velopment. If we take the existence of the transposed letter (TL)
effect (caniso primes CASINO) as the hallmark of orthographic
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
processing (e.g., Acha & Perea, 2008; Castles, Davis, Cavalot, &
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Forster, 2007; Grainger, 2008; Grainger, Kiyonaga, & Holcomb,
2006; Lété & Fayol, 2013; Perea & Lupker, 2004; Ziegler et al.,
2013), the resulting picture is not straightforward because Castles
et al. (2007) found TL effects to diminish with reading age,
whereas Lété and Fayol (2013) found them to increase with
reading age. Cross-language differences can hardly explain this
discrepancy because Acha and Perea (2008) found the same pat-
tern as Castles et al. (2007) although their study was conducted in
Spanish, which is much more regular than English. Apart from
these discrepancies, none of the previously mentioned studies
investigated the joint effects of orthographic and phonological
development. Given that models of reading development predict a
shift from phonological decoding to parallel orthographic process-
ing (e.g., Grainger & Ziegler, 2011; Share, 1995), investigating the
development of these two processes within the same paradigm is
an important step ahead.
One recent study did directly compared an orthographic with a
phonological marker effect in the course of reading development
(Grainger, Lété, Bertrand, Dufau, & Ziegler, 2012). As all studies
discussed previously, Grainger et al. (2012) used the TL effect as
a proxy for orthographic processing. However, they did not use TL
priming (caniso primes CASINO) but investigated to what extent
participants could correctly reject TL nonwords, such as caniso, in
a lexical decision task. As a proxy for phonological processing,
they used the well-known pseudohomophone (PsH) effect, that is,
participants’ ability to correctly reject nonwords that sound like
real words, such as brane (e.g., Goswami et al., 2001; Ziegler,
Jacobs, & Klueppel, 2001). Using such PsH and TL nonwords in
the nonword trials of a lexical decision task, Grainger, Lété, et al.
(2012) indeed reported that the PsH interference effect diminished
with reading age, whereas the TL interference effect first increased
and then diminished.
One problem with Grainger, Lété, et al.’s (2012) study is that
PsH and TL effects in the lexical decision task cannot unequivo-
cally be attributed to either phonological or orthographic process-
ing. For example, the PsH effect was found to decrease with
reading age (i.e., fewer misclassifications of brane as a function of
age). However, can we attribute this pattern to a reduction in the
use of phonology, or is it not precisely the development of high-
quality orthographic representations that allows participants to
correctly reject a PsH in the lexical decision task? Similarly,
should stronger reliance on orthographic processing result in larger
TL interference effects (i.e., greater reliance on coarse grained
orthographic access) or should we expect smaller TL interference
effects as a consequence of more precise orthographic representa-
1027
tions? In fact, due to the complex decision operations that are
inherent to the lexical decision task (Dufau, Grainger, & Ziegler,
2012), using PsH and TL nonwords as stimuli for the “no” re-
sponses makes it rather difficult to interpret the size of these
effects as a pure reflection of either orthographic or phonological
processes that underlie task performance.
One solution to this dilemma is to use masked priming meth-
odology, as has been done by Castles et al. (2007); Lété and Fayol
(2013), and Acha and Perea (2008). The advantage of masked
priming is that PsH and TL nonwords are no longer used as stimuli
that participants have to reject but are used as subliminally pre-
sented primes, which require no response or decision. Priming
effects are measured as decrease (facilitation) in response speed
when the target is preceded by a related TL or PsH prime com-
pared with an unrelated but orthographically matched control
prime. That is, in both critical priming conditions, participants
respond “yes” to a target word that is preceded either by a PsH
prime, a TL prime, or an orthographically matched control prime.
Thus, facilitatory TL priming (caniso primes CASINO more than
its yoked control prime) can be unequivocally attributed to reliance
on coarse-grained or flexible orthographic processing, that is,
orthographic activation that tolerates some degree of mismatch in
letter positions and/or letter orders (Davis, 2010; Gomez, Ratcliff,
& Perea, 2008; Whitney, 2001). In contrast, facilitatory PsH ef-
fects (brane primes BRAIN more than its yoked control) can
unequivocally be attributed to the sublexical computation of pho-
nology that activates the whole word phonology of the target word
(Lukatela & Turvey, 1994b).
The present study goes beyond previous studies in several
important ways. First, it goes beyond the Grainger, Lété, et al.
(2012) study because it allows us to unambiguously interpret the
size of the marker effects in terms of either orthographic or
phonological influences. Second, it goes beyond previous devel-
opmental priming studies (Acha & Perea, 2008; Castles et al.,
2007; Lété & Fayol, 2013) because it contrasts orthographic with
phonological development within the same children. Third, it uses
a much broader range of reading levels (Grades 1–5) than previous
studies, which allows these effects to be studied in a parametric
fashion. Fourth, it builds on the advantages of masked priming
while increasing measurement sensitivity via the so-called sand-
wich priming variant of this technique (Lupker & Davis, 2009).
Finally, it is of theoretical interest because the dual-route theory
(Grainger & Ziegler, 2011) and the lexical-quality hypothesis
(Castles et al., 2007) make opposite predictions as to whether TL
priming effects should increase or decrease with reading skills.
The dual-route theory predicts an increase in the size of the TL
priming effect, whereas the lexical-quality hypothesis predicts a
decrease in the size of the TL priming effect.
Method
Participants
The initial sample consisted of 284 children. They were re-
cruited from a single public elementary school in a middle-class
environment in the center of Lyon, the third largest city in France,
with a population of about 1.3 million people. The sample was
ethnically diverse although only children with French as their
native language were selected to participate in the study. Children
 1028 ZIEGLER, BERTRAND, LÉTÉ, AND GRAINGER
were tested at the end of their school year in June. Therefore, even
first graders were able to decode novel words and read all of the
target words (which were selected according to grade-appropriate
norms; discussed later). For each grade level, two out of three
classes from the same school were selected. Reading instruction in
France is phonics based and follows a tight national curriculum.
Thus, we presumed teacher effects on masked PsH and TL priming
to be negligible.
To ensure that all readers were reading at the expected grade
level and in order to exclude children with potential reading
impairments, we assessed the reading level of each participant with
a standardized reading test (Alouette–R; Lefavrais, 2005). Chil-
dren with a formal diagnosis of dyslexia or those whose standard-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
ized reading score in the Alouette reading test was 18 months
This document is copyrighted by the American Psychological Association or one of its allied publishers.
below the expected level (i.e., formal criterion used by dyslexia
assessment centers in France) were excluded from the study.
Finally, children whose teachers or parents reported spoken lan-
guage impairments were also excluded. These criteria led us to
retain a total of 267 children. Further information on gender, age
and reading age per grade can be found in Table 1.
Stimuli
The experiment consisted of two critical conditions, the pseu-
dohomophone (PsH) and the transposed-letter (TL) condition.
Each condition contained word and nonword targets that were
primed by either PsH primes and their yoked controls or TL primes
and their yoked controls. All primes were nonwords (see later
sections). Target words were selected using MANULEX (Lété,
Sprenger-Charolles, & Colé, 2004), a computerized child lexical
database that provides word frequencies for each grade compiled
from over 1.9 million words taken from the major primary school
reading books. To ensure that the children were familiar with the
target words, we chose only words that appeared in each grade. For
the purpose of the lexical decision task, we selected a set of
nonwords that matched the target words in length and consonant–
vowel structure. All nonwords were orthographically legal and
pronounceable letter strings in French. They were preceded by the
same kind of primes as the word trials.
Pseudohomophone condition. Thirty-three French words
four and five letters in length served as targets (see Appendix A).
The word targets had an average frequency of 178 (SD ⫽ 174)
occurrences per million according to MANULEX (all grades com-
bined). Average orthographic neighborhood size (OLD20) was
1.43 (Yarkoni, Balota, & Yap, 2008). Each target word was
associated with a PsH prime and a control prime. PsH primes were
created with PHOM, a database of 14,000 French PsHs (Farioli,
Table 1
Characteristics of the Sample
Gender Age Reading age
Grade N (boys) (in months) (in months)
1 39 18 81.4 (3.4) 85.0 (4.9)
2 55 29 94.0 (4.4) 100.0 (14.6)
3 51 28 107.4 (4.2) 103.7 (12.7)
4 57 32 118.8 (4.9) 118.2 (19.4)
5 65 39 130.8 (4.5) 125.9 (18.3)
Note. Standard deviations are in parentheses.
Grainger, & Ferrand, 2011). They were constructed in such a way
that their pronunciation, but not their spelling, was identical to the
target (e.g., target: vase [vase], pronounced /vaz/; PsH prime: vaze
pronounced /vaz/). Orthographic changes only involved one pho-
neme of the target. The control primes had the same level of
orthographic overlap with the target as the corresponding PsH
prime but were phonologically different from the target (e.g.,
target: vase (vase); PsH prime: vaze; control prime: vare).
Transposed-letter condition. Thirty-six French words five
and six letters in length served as targets1 (see Appendix B). The
word targets had an average frequency of 155 (SD ⫽ 134) occur-
rences per million according to MANULEX (all grades combined),
and an average orthographic neighborhood size of 1.44 (range
1.0 –1.9) measured with the OLD20 metric. Each target was asso-
ciated with two prime types: (a) TL prime constructed by swap-
ping two adjacent letters (only consonants were involved in the
transposition) in the target, either the third and the fourth letter (for
17 items) or the fourth and the fifth letter (19 items); and (b)
control primes constructed by substituting the transposed conso-
nants by two other consonants (e.g., target: table; TL prime: talbe;
control prime: tarfe).
Procedure and Apparatus
Participants were seated in front of a 17-in. color monitor
connected to a Pentium III laptop computer (Intel Corp., Santa
Clara, CA) running E-Prime software Version 2 at a viewing
distance of approximately 60 cm. The stimuli were displayed in
lowercase in 24-point Courier font with a 640 ⫻ 480 resolution.
To amplify priming effects, we implemented the sandwich prim-
ing procedure (Lupker & Davis, 2009). That is, before the presen-
tation of the prime, the target word appeared subliminally for 27
ms before being replaced by the prime. Each trial thus consisted of
the following sequence of events. Participants were presented with
a fixation point (ⴱ) for 1,000 ms, which was replaced by a target
stimulus (in lowercase letters) centered on the fixation point for 27
ms. The target was then replaced at the same localization by a
prime stimulus presented in lowercase. After 70 ms, the prime was
replaced by the target presented in uppercase. The target remained
on the screen until participants responded. They gave their re-
sponse by pressing one of two response keys on the keyboard.
Participants were instructed to respond as quickly and as accu-
rately as possible.
Each participant saw each target only once (no target repetition).
Given that we had 33 targets in the PsH condition and 36 in the TL
condition, there were 69 critical related trials. Because the task was
lexical decision, we also had to add 69 nonword trials for the “no”
decisions, which led to a total of 138 trials per participant. These
138 trials were divided into six blocks of 23 items (6ⴱ23 ⫽ 138)
with an approximately even number of trials per condition. The
order of blocks was permutated using a Latin square design. There
was a short 1-min break between each block. Two “warm-up”
1
Note that target words in the two conditions were not matched for
length (four or five letters in the PsH condition and five or six letters in the
TL condition). This is not a problem, however, because target words were
not compared directly across conditions. Only the priming effects were
compared across conditions, and priming effects were controlled for length
because related and unrelated primes within a given condition were
matched for length.
 ORTHOGRAPHIC AND PHONOLOGICAL MASKED PRIMING
trials (excluded from subsequent analyses) were added at the
beginning of each block. Because each target word was preceded
by either a related or an unrelated prime, we had to create two lists
such that a given target word was preceded by a related prime in
List 1 and an unrelated prime in List 2. List assignment was
counterbalanced across participants.
Children were tested in groups of four by two experimenters in
the same room in a single 25-min session. Children were given
child-friendly instructions, and they were given a short training
block to familiarize them with the lexical decision task. During the
practice trials, experimenters were careful about the position of
children in front of the screen, and their correct use of the keyboard
to give their response.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Results
Items and participants producing more than 40% errors were
excluded from the analysis.2 This criterion led to the exclusion of
nine items in Grade 1, two in Grade 2, one in Grade 3, and none
in the other grades (excluded items are marked in Appendices A
and B). Three participants were excluded in Grade 1, and two each
were excluded in Grades 2, 3, and 4. For the remaining data,
response times (RTs) that were three standard deviations (SDs)
beyond the mean of the participant were excluded as outliers (⬍
1% of the data). RTs were inverse transformed (RT ⫽ ⬎ 1/RT)
before the analysis in order to normalize the RT distributions
(Ratcliff, 1993). This means that we transformed the classic la-
tency measure (seconds per word) into a speed measure (words per
second). Nontransformed mean RTs and error rates for both prim-
ing conditions across the five grades can be found in Tables 2 and
3 for the critical word trials.
Global Analysis
Error rates and inverse RTs for word targets were submitted to
a three-factorial analysis of variance (ANOVA) with Grade (1–5),
Prime (prime vs. control), and Condition (PsH vs. TL) as factors.
Grade was a between-subjects factor, whereas Prime and Condi-
tion were within-subject factors.
The accuracy data showed a significant main effect of Grade,
F(4, 234) ⫽ 61.09, p ⬍ .0001, partial ␩2 ⫽ .510; Condition, F(1,
234) ⫽ 19.42, p ⬍ .0001, partial ␩2 ⫽ .077; and Prime, F(1,
236) ⫽ 7.41, p ⬍ .01, partial ␩2 ⫽ .031. These main effects reflect
the finding that error rate dropped from Grade 1 to Grade 5 (main
effect of Grade), that word targets in the PsH condition yielded
more errors than word targets in the TL condition (main effect of
Condition, 10.3% vs. 8.3%, respectively), and that priming showed
the expected facilitatory effect of reducing error rates for primed
targets (main effect of Prime). None of the interactions reached
significance, suggesting that both type of primes (TL and PsH)
reduced error rates in a similar fashion across grades (all Fs ⬍ 1).
Given that the triple interaction failed to reach significance in the
global analysis, the accuracy data were not further broken down by
condition.
The analysis of the inverse RT data (words/sec) revealed sig-
nificant main effects of Grade, F(4, 234) ⫽ 81.62, p ⬍ .0001,
partial ␩2 ⫽ .58, and Prime, F(4, 234) ⫽ 131.75, p ⬍ .0001, partial
␩2 ⫽ .36). The main effect of Condition was not significant, F ⬍
1. The two-way interaction between the effects of Condition and
1029
Prime was significant, F(1, 234) ⫽ 39.26, p ⬍ .0001, partial ␩2 ⫽
.14, reflecting the fact that priming effects were stronger for TL
than for PsH primes. None of the other two-way interactions was
significant, all Fs ⬍ 1.5, all ps ⬎ 20. It is important to note that the
triple interaction of the effects of Grade, Condition, and Prime was
significant, F(4, 234) ⫽ 3.42, p ⬍ .01, partial ␩2 ⫽ .06, suggesting
differences in the size of the PsH and TL priming effects across
grades. This triple interaction is illustrated in Figure 1. Indeed,
while the size of the PsH priming effect remained relatively
constant across grades, the size of the TL priming effect increased
monotonically across grades. The two conditions (PsH, TL) are
analyzed separately in the following sections.
PsH priming. The inverse RT data were submitted to a two-
factorial ANOVA with Prime (PsH vs. controls) and Grade (1 to 5)
as within-subject factors. The analysis showed significant main
effects of Grade, F(4, 253) ⫽ 80.71, p ⬍ .0001, partial ␩2 ⫽ .56,
and Prime, F(1, 253) ⫽ 20.04, p ⬍ .0001, partial ␩2 ⫽ .07, but no
significant interaction between the effects of Prime and Grade,
F(4, 253) ⫽ 0.33, p ⬎ .80, partial ␩2 ⫽ .005.
TL priming. The same ANOVA as described for PsH priming
was conducted for the TL priming effects. The analysis showed
significant main effects of Grade, F(4, 236) ⫽ 77.05, p ⬍ .0001,
partial ␩2 ⫽ .57, and Prime, F(1, 236) ⫽ 153.90, p ⬍ .0001, partial
␩2 ⫽ .40. The critical interaction between the effects of Prime and
Grade was significant, F(4, 236) ⫽ 4.34, p ⬍ .002, partial ␩2 ⫽
.07.
Effect Sizes
To analyze how the priming effects in the two conditions unfold
across grades, we calculated the effect sizes for the PsH and the TL
priming effect in two ways. First, we computed Cohen’s d (M1 ⫺
M2/spooled) using the inverse RT data for each grade level. These
results are plotted in Figure 2, Panel A. Second, because absolute
speed differences across different groups can lead to spurious
overadditive interactions (Faust, Balota, Spieler, & Ferraro, 1999),
we standardized the raw RT data of each participant in each
condition using z scores, such that each participant had a mean of
zero in each priming condition. We then plotted the difference
between the related and the unrelated prime condition expressed in
z-score differences (see Figure 2, Panel B). Despite the fact that the
first analysis was based on inverse RTs, whereas the second was
based on standardized raw RTs, both effect size analyses con-
verged on the same result, namely, that the TL priming effect
increased monotonically with grade, whereas the PsH priming
effect remained constant. This is an important finding because the
net priming effects using the untransformed RTs (see Table 2)
would have suggested that the PsH priming effect decreases with
grade while the TL priming effect remains constant. However, this
pattern was clearly due to differences in absolute response speed
across grades, which result in larger standard deviations in earlier
than in later grades. When taking these differences into account
through the computation of effect size measures, we clearly see
that the TL effect increases with age, whereas the PsH effect
remains constant.
2
The 40% cutoff was chosen because RT estimates for a given condi-
tion, which are based on correct trials only, tend to be unreliable if more
than 40% of the trials are excluded.
 1030 ZIEGLER, BERTRAND, LÉTÉ, AND GRAINGER

Table 2
Response Latencies (in Milliseconds) for Target Words Preceded by Either Pseudohomophone Primes and Their Respective Controls
or Transposed-Letter Primes and Their Respective Controls

Pseudohomophone priming Transposed-letter priming

Grade PsH Control Priming TL Control Priming

1 1783 (516) 1860 (551) ⫺77 1875 (606) 1947 (533) ⫺72
2 1241 (335) 1279 (326) ⫺38 1300 (497) 1401 (459) ⫺101
3 962 (203) 990 (212) ⫺28 933 (212) 1033 (219) ⫺100
4 858 (180) 887 (182) ⫺29 824 (202) 886 (172) ⫺62
5 755 (130) 771 (147) ⫺16 716 (157) 807 (162) ⫺91
Note. Standard deviations are in parentheses. PsH ⫽ pseudohomophone; TL ⫽ transposed-letter.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Correlations With Reading Age
All previous analyses were based on grade. However, reading
competence is a continuous variable, which was measured for each
individual with a standardized reading test (Alouette–R; Lefavrais,
2005). We therefore performed linear regression analyses to find
out whether we could replicate the pattern we have described using
reading age (RA) of each child rather than grade level. The
analyses were performed using the standardized effect size mea-
sures for each priming effect and each child (z-score differences
between the related and the unrelated condition) as the dependent
variable (both for RTs and inverse RTs) and RA as the predictor
variable. On the standardized raw RTs, the results showed a
significant correlation between RA and TL priming (r ⫽ .304, p ⬍
.0001), but no correlation between RA and PsH priming (r ⫽ .052,
p ⬎ .40). The same pattern was found on standardized inverse RTs
with a significant correlation between RA and TL priming (r ⫽
.355, p ⬍ .0001) but no such correlation between RA and PsH
priming (r ⫽ .057, p ⬎ .30). Thus, the correlation analyses join the
effect size analyses (Figure 2) to suggest that the TL effect in-
creases with RA, whereas the PsH effect remains constant.
Nonword Trials
Because half of the trials contained nonword targets that were
either preceded by a PsH prime (or its yoked control) or a TL
prime (or its yoked control), the data from the nonword trials could
be analyzed in the same way as the word trials. Thus, error rates
and inverse RTs for nonword targets were submitted to a three-
factorial ANOVA with Grade (1–5), Prime (prime vs. control), and
Condition (PsH vs. TL) as factors. Grade was a between-subjects
factor, whereas Prime and Condition were within-subject factors.
The results appear in Tables 4 and 5.
The accuracy data showed a significant main effect of Grade,
F(4, 233) ⫽ 10.11, p ⬍ .0001, partial ␩2 ⫽ .15, and Condition,
F(1, 233) ⫽ 35.96, p ⬍ .0001, partial ␩2 ⫽ .13, but no effects of
Prime and no interactions with this factor (all Fs ⬍ 1). As concerns
the inverse RT data, there were main effects of Grade, F(4, 234) ⫽
82.43, p ⬍ .0001, partial ␩2 ⫽ .585; Condition, F(1, 234) ⫽ 5.99,
p ⬍ .05, partial ␩2 ⫽ .025; and Prime F(1, 234) ⫽ 5.82, p ⬍ .05,
partial ␩2 ⫽ .024, but no interactions with priming (all Fs ⬍ 1). In
sum, the nonword data did not show any interesting effects apart
from a significant priming effect on inverse RTs. Most important
is that priming did not interact with Grade or Condition. The fact
that TL priming did not increase with grade for nonword targets
(unlike for word targets) strengthens the claim that the increase in
the size of the TL priming effect for words was due to the
development of orthographic representations used to identify
known words.
Discussion
In the present study, we used the masked priming paradigm to
shed light on the role of orthographic and phonological processing
during reading development. Masked priming is an ideal paradigm
for this purpose as it allows one to compare net priming effects that
directly reflect the activation of orthographic and phonological
codes. It is also an ideal paradigm to use with children because the
primes are not visible (no response strategies), and responses in a
critical prime condition are made to the same target word that is
preceded by either a “related” or an “unrelated” prime (thus
avoiding problems of comparing different groups of items). In the

Table 3
Error Rates (in Percentages) for Target Words Preceded by Either Pseudohomophone Primes and Their Respective Controls or
Transposed-Letter Primes and Their Respective Controls

Pseudohomophone priming Transposed-letter priming

Grade PsH Control Priming TL Control Priming

1 18.8 (12.1) 17.7 (10.9) 1.2 17.6 (13.0) 17.8 (12.9) ⫺0.2
2 12.6 (10.3) 16.0 (11.6) ⫺3.4 10.0 (10.5) 13.0 (12.7) ⫺3.0
3 12.5 (7.8) 14.7 (11.8) ⫺2.2 8.5 (8.4) 8.7 (9.8) ⫺0.2
4 2.5 (2.2) 2.9 (2.0) ⫺0.4 1.2 (1.8) 2.9 (4.5) ⫺1.7
5 2.9 (2.2) 2.8 (2.3) 0.1 1.3 (1.6) 2.4 (2.2) ⫺1.1
Note. Standard deviations are in parentheses. PsH ⫽ pseudohomophone; TL ⫽ transposed-letter.
 ORTHOGRAPHIC AND PHONOLOGICAL MASKED PRIMING 1031

1.6 1.6
1.5 1.5

Response Speed (worrds/sec)

Response Speed (worrds/sec)

1.4 1.4
1.3 1.3
1.2 1.2
1.1 1.1
1.0 1.0
0.9 PSH 0.9 TL
Control Control
0.8 0.8
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

07
0.7 07
0.7
0.6 0.6
1 2 3 4 5 1 2 3 4 5
Grade Grade
Figure 1. Response speed (words/sec) for target words preceded by either pseudohomophone primes and their
respective controls (PsH, left) or transposed-letter primes and their respective controls (TL, right) from Grades
1 through 5. Error bars are standard errors.

adult literature, masked priming has been extensively used to study
the time course of various processes involved in reading (e.g.,
Ferrand & Grainger, 1992; Forster & Davis, 1984; Rastle &
Brysbaert, 2006). By and large, the masked priming paradigm has
made it possible to uncover the time course of effects that have
been perfectly replicated in more natural reading tasks using online
measures, such as eye movements (e.g., Rayner, Pollatsek, &
Binder, 1998) or event-related brain potentials (Grainger et al.,
2006).
The results of the present study are straightforward. While the
size of the TL priming effect (i.e., the orthographic marker effect)
increased monotonically with grade or RA, the size of the PsH
priming effect (i.e., the phonological marker effect) remained
stable across grades and independent of RA. Interestingly, had we
only looked at the net priming effect in the untransformed data
(Table 2), we would have come to the wrong conclusion that
reliance on phonological activation decreases with RA. However,
as nicely demonstrated by Faust and collaborators (1999), in the
presence of large differences in global response speed between
groups—RTs in our study dropped from 1,866 ms (SD ⫽ 551) in
Grade 1 to 762 (SD ⫽ 149) in Grade 5—it is inappropriate to draw
conclusions about effect sizes based on absolute RT differences.
When the data were z-transformed, as advocated by Faust et al.
(1999) as one solution to this problem, it became clear that PsH
effects in the present task do not decrease with RA. Normalizing
the data by using an inverse RT transformation had the same effect
(see Figure 2). The inverse RT transformation (i.e., 1/RT) normal-
izes RT data (i.e., it reduces the rightward skew of the RT distri-

A B
0.8 0.8
ohen’s d)

0.7 0.7
Efffect Size (z-sscores)

0.6 0.6
05
0.5 05
0.5
Efffect Size (Co

0.4 TL 0.4
0.3 PsH 0.3 TL
0.2 0.2 PsH
0.1 0.1
0 0
1 2 3 4 5 1 2 3 4 5

Grade Grade

Figure 2. Effect sizes for the pseudohomophone (PsH) and transposed-letter (TL) priming effects. Panel A
displays effect sizes expressed in terms of Cohen’s d (M1 ⫺ M2/spooled) using the inverse reaction time (RT) data
for each grade level. Panel B displays effect sizes in terms of z-score differences between the related and the
unrelated conditions, which were based on standardizing the raw RT data of each participant in each condition.
 1032 ZIEGLER, BERTRAND, LÉTÉ, AND GRAINGER

Table 4
Response Latencies (in Milliseconds) for Nonwords Preceded by Either Pseudohomophones and Their Respective Controls or
Transposed-Letter Primes and Their Respective Controls

Pseudohomophone priming Transposed-letter priming

Grade PsH Control Priming TL Control Priming

1 2368.7 (803.6) 2475.4 (870.9) ⫺106.7 2651.6 (925.5) 2622.3 (833.6) ⫺29.4
2 1628.5 (605.1) 1622.3 (599.9) 6.2 1650.9 (577.3) 1696.5 (659.6) ⫺45.6
3 1175.6 (336.8) 1226.1 (342.7) ⫺50.5 1227.4 (381.2) 1219.6 (350.8) 7.8
4 1018.3 (244.6) 1032.6 (281.8) ⫺14.3 1013.6 (308.9) 1027.7 (307.0) ⫺14.1
5 873.3 (184.0) 907.6 (236.0) ⫺34.3 881.2 (197.4) 895.6 (191.9) ⫺14.4
Note. Standard deviations are in parentheses. PsH ⫽ pseudohomophone; TL ⫽ transposed-letter.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

bution). This suggests that unequal variances across grades in the
raw data are mainly due to the much longer tails of the RT
distributions in younger than older readers. On the basis of these
results, we advocate the use of inverse RTs in developmental
studies. Conceptually, inverse RT is a speed measure (items per
second), whereas standard RT is a latency measure (seconds per
item). Conclusions can therefore be drawn in terms of effects of
factors on untransformed reading speed.
One important finding of the present study is a significant PsH
priming effect that remains stable across development. This result
challenges the view that more advanced readers should rely less on
phonological activation than do younger readers (e.g., Doctor &
Coltheart, 1980; Frith, 1985). The idea that there is an age-related
reduction in reliance on phonology mainly comes from studies
which showed a reduction of the effects of word length on reading
(Acha & Perea, 2008; Aghababian & Nazir, 2000; Bijeljac-Babic,
Millogo, Farioli, & Grainger, 2004; Ziegler et al., 2003) or a
reduction in error rate when children have to reject pseudohomo-
phones in lexical decision or homophones in semantic categoriza-
tion tasks (Doctor & Coltheart, 1980; Grainger, Lété, et al., 2012;
Sprenger-Charolles, Siegel, Bechennec, & Serniclaes, 2003).
However, the disappearance of the length effect in reading aloud
does not necessarily mean that children rely less on phonology as
they become better readers but rather that the phonological com-
putation has become more efficient and parallel, possibly relying
on the computation of larger chunks than individual letters and/or
graphemes. Furthermore, the age-related reduction in error rate
when participants have to reject pseudohomophones or homo-
phones might well result from better specified orthographic rep-
resentations rather than weaker phonological activation. This pos-
sibility is not only consistent with a number of developmental
priming studies, which showed fast, automatic, and general pho-
nological activation that did not decrease with age (Alario, De
Cara, & Ziegler, 2007; Booth et al., 1999) but also with the great
bulk of studies on skilled reading, which have shown robust
phonological influences in adults (e.g., Braun et al., 2009; Ferrand
& Grainger, 1992; Lukatela & Turvey, 1994a, 1994b; Rastle &
Brysbaert, 2006; Ziegler et al., 1997). Still, it is worth noting that
the phonological priming effect was overall smaller than the or-
thographic priming effect (Figure 2), which suggests that more
attention needs to be given to orthographic processes in developing
readers, which are far less studied than phonological processes
(Castles & Nation, 2006; Castles et al., 2007; Grainger & Ziegler,
2011).
The second important finding is the monotonic increase in the
size of the TL effect across grades and reading age. Lété and Fayol
(2013) showed a similar increase in TL priming between Grades 3
and 5. Moreover, in a recent study, Kohnen and Castles (2013)
showed that children made many more errors on “migratable”
words (e.g., parties ¡ pirates; three ¡ there) than on nonmigrat-
able words. As in our study, the proportion of migration errors
(i.e., the TL effect) tended to increase with increasing grade level.
It is notable that the proportion of migration errors was correlated
with lexical reading skills, but not with other subcomponents of
reading or guessing strategies. This finding is consistent with
Ziegler et al.’s (2013) observation that the size of TL effects in a
word-learning experiment with nonhuman primates (Grainger, Du-
fau, Montant, Ziegler, & Fagot, 2012) was strongly correlated with
their overall word accuracy. Together, these findings converge to
suggest that direct orthographic access is achieved through coarse-
grained orthographic coding that tolerates some degree of flexi-
bility in letter order. Indeed, all recent theories of letter-position

Table 5
Error Rates (in Percentages) for Nonwords Preceded by Either Pseudohomophones and Their Respective Controls or Transposed-
Letter Primes and Their Respective Controls

Pseudohomophone priming Transposed-letter priming

Grade PsH Control Priming PsH Control Priming

1 25.8 (14.5) 26.3 (19.0) ⫺0.5 17.8 (16.1) 21.7 (19.5) ⫺3.9
2 16.9 (16.9) 17.7 (17.7) ⫺0.8 15.2 (17.3) 15.0 (13.1) 0.2
3 15.5 (15.3) 14.5 (15.4) 1.0 13.2 (15.6) 13.7 (13.8) ⫺0.5
4 11.0 (12.9) 10.6 (10.0) 0.4 7.9 (12.1) 7.9 (10.6) 0.3
5 10.4 (11.1) 9.1 (9.0) 1.3 8.8 (10.2) 7.7 (10.2) 1.1
Note. Standard deviations are in parentheses. PsH ⫽ pseudohomophone; TL ⫽ transposed-letter.
 ORTHOGRAPHIC AND PHONOLOGICAL MASKED PRIMING
coding predict facilitatory TL effects due to some form of flexible
position coding (Davis, 2010; Gomez et al., 2008; Hannagan &
Grainger, 2012; Whitney, 2001). Thus, in line with the predictions
of the dual-route model of orthographic development (Grainger &
Ziegler, 2011), the size of the TL effects increased with reading
development, which suggests greater reliance on coarse-grained
orthographic coding in the course of development.
It remains to be discussed why some of the previous masked
priming studies did not show an increase in the size of the TL
effect across grades. There are at least three key differences
between the present study and prior research using masked prim-
ing, which might explain some of the differences. One is the use of
the sandwich priming variant of masked priming, known to pro-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
vide a more sensitive measure of orthographic priming than stan-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
dard masked priming (Ktori, Grainger, Dufau, & Holcomb, 2012;
Lupker & Davis, 2009). When comparing the TL priming effects
obtained with children of comparable age in the Acha and Perea
(2008) study, which was the only study to use the same double-
letter substitution control condition as used in the present work, it
becomes clear that priming effects in our study were twice as large as
priming effects in their study. Another key difference is the use of
transformed data in order to reduce contamination from absolute
RT differences across different groups of readers (for a similar use
of z scores, see Acha & Perea, 2008, and Lété & Fayol, 2013). As
suggested earlier, when absolute reading speed varies across dif-
ferent groups or reading levels, conclusions on the basis of abso-
lute RT differences are problematic (Faust et al., 1999). Finally,
our study is the only one in which children in all five grades of
primary education were tested. On this point it is interesting to note
that if we had only examined performance in Grade 2 and Grade
4 using untransformed RTs, then we would have been led to draw
the opposite conclusion. That is, the net TL priming effect using
untransformed RTs was 101 ms in Grade 2 and 62 ms in Grade 4
(see Table 2). Taken at face value, this reduction in net priming
could have been taken to suggest that TL priming effects diminish
with age, which is not the case when the results are analyzed
properly using normalized data. More research is needed to inves-
tigate to what extent the present findings can be replicated using
different sets of items in different languages. Note, however, there
is no reason to believe that the present findings are specific to
French because PsH and TL effects have been reported in many
alphabetic languages (e.g., English, Dutch, Spanish, and German;
for a review, see Frost, 2012).
In sum then, although the present study was not longitudinal, it
nevertheless draws a clear picture about the development of or-
thographic and phonological marker effects in silent reading. The
data show that phonological information not only plays an impor-
tant role in early reading development (Bosman & de Groot, 1996)
but continues to exert a stable effect throughout reading develop-
ment. At the same time, orthographic development takes place,
which manifests itself as an increase in the size of TL priming
effects, which, in turn, is in line with the hypothesized develop-
ment of coarse-grained orthographic coding as children become
more skilled readers (Grainger, Lété, et al., 2012; Grainger &
Ziegler, 2011). Finally, from a purely methodological perspective,
the present work highlights how important it is to consider issues
of measurement sensitivity, data normalization, and number of age
groups when evaluating developmental trajectories.
1033
References
Acha, J., & Perea, M. (2008). The effects of length and transposed-letter
similarity in lexical decision: Evidence with beginning, intermediate,
and adult readers. British Journal of Psychology, 99, 245–264. doi:
10.1348/000712607X224478
Aghababian, V., & Nazir, T. A. (2000). Developing normal reading skills:
Aspects of the visual processes underlying word recognition. Journal of
Experimental Child Psychology, 76, 123–150. doi:10.1006/jecp.1999
.2540
Alario, F. X., De Cara, B., & Ziegler, J. C. (2007). Automatic activation of
phonology in silent reading is parallel: Evidence from beginning and
skilled readers. Journal of Experimental Child Psychology, 97, 205–219.
doi:10.1016/j.jecp.2007.02.001
Bijeljac-Babic, R., Millogo, V., Farioli, F., & Grainger, J. (2004). A
developmental investigation of word length effects in reading using a
new on-line word identification paradigm. Reading and Writing, 17,
411– 431. doi:10.1023/B:READ.0000032664.20755.af
Booth, J. R., Perfetti, C. A., & MacWhinney, B. (1999). Quick, automatic,
and general activation of orthographic and phonological representations
in young readers. Developmental Psychology, 35, 3–19. doi:10.1037/
0012-1649.35.1.3
Bosman, A. M. T., & de Groot, A. M. B. (1996). Phonologic mediation is
fundamental to reading: Evidence from beginning readers. The Quar-
terly Journal of Experimental Psychology, Series A: Human Experimen-
tal Psychology, 49, 715–744. doi:10.1080/027249896392568
Braun, M., Hutzler, F., Ziegler, J. C., Dambacher, M., & Jacobs, A. M.
(2009). Pseudohomophone effects provide evidence of early lexico-
phonological processing in visual word recognition. Human Brain Map-
ping, 30, 1977–1989. doi:10.1002/hbm.20643
Castles, A., Davis, C., Cavalot, P., & Forster, K. (2007). Tracking the
acquisition of orthographic skills in developing readers: Masked priming
effects. Journal of Experimental Child Psychology, 97, 165–182. doi:
10.1016/j.jecp.2007.01.006
Castles, A., & Nation, K. (2006). How does orthographic learning happen?
In S. Andrews (Ed.), From inmarks to ideas: Current issues in lexical
processing (pp. 151–179). Hove, United Kingdom: Psychology Press.
Davis, C. J. (2010). The spatial coding model of visual word identification.
Psychological Review, 117, 713–758. doi:10.1037/a0019738
Diependaele, K., Ziegler, J. C., & Grainger, J. (2010). Fast phonology and
the bimodal interactive activation model. European Journal of Cognitive
Psychology, 22, 764 –778. doi:10.1080/09541440902834782
Doctor, E. A., & Coltheart, M. (1980). Children’s use of phonological
encoding when reading for meaning. Memory & Cognition, 8, 195–209.
doi:10.3758/BF03197607
Dufau, S., Grainger, J., & Ziegler, J. C. (2012). How to say “no” to a
nonword: A leaky competing accumulator model of lexical decision.
Journal of Experimental Psychology: Learning, Memory, and Cogni-
tion, 38, 1117–1128. doi:10.1037/a0026948
E-Prime Software (Version 2) [Computer software]. Sharpsburg, PA: Psy-
chology Software Tools.
Farioli, F., Grainger, J., & Ferrand, L. (2011). PHOM: Une base de
données de 14 000 pseudohomophones [PHOM: A database of 14,000
pseudohomophones]. L’Année Psychologique, 111, 725–751. doi:
10.4074/S0003503311004052
Faust, M. E., Balota, D. A., Spieler, D. H., & Ferraro, F. R. (1999).
Individual differences in information-processing rate and amount: Im-
plications for group differences in response latency. Psychological Bul-
letin, 125, 777–799. doi:10.1037/0033-2909.125.6.777
Ferrand, L., & Grainger, J. (1992). Phonology and orthography in visual
word recognition: Evidence from masked non-word priming. The Quar-
terly Journal of Experimental Psychology, Series A: Human Experimen-
tal Psychology, 45, 353–372. doi:10.1080/02724989208250619
Forster, K. I., & Davis, C. (1984). Repetition priming and frequency
attenuation in lexical access. Journal of Experimental Psychology:
 1034 ZIEGLER, BERTRAND, LÉTÉ, AND GRAINGER
Learning, Memory, and Cognition, 10, 680 – 698. doi:10.1037/0278-
7393.10.4.680
Frith, U. (1985). Beneath the surface of developmental dyslexia. In K. E.
Patterson, J. C. Marshall, & M. Coltheart (Eds.), Surface dyslexia (pp.
301–330). Hillsdale, NJ: Erlbaum.
Frost, R. (1998). Toward a strong phonological theory of visual word
recognition: True issues and false trails. Psychological Bulletin, 123,
71–99. doi:10.1037/0033-2909.123.1.71
Frost, R. (2012). Towards a universal model of reading. Behavioral and
Brain Sciences, 35, 263–279. doi:10.1017/S0140525X11001841
Gomez, P., Ratcliff, R., & Perea, M. (2008). The overlap model: A model
of letter position coding. Psychological Review, 115, 577– 600. doi:
10.1037/a0012667
Goswami, U., Ziegler, J. C., Dalton, L., & Schneider, W. (2001). Pseudo-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
homophone effects and phonological recoding procedures in reading
This document is copyrighted by the American Psychological Association or one of its allied publishers.
development in English and German. Journal of Memory and Language,
45, 648 – 664. doi:10.1006/jmla.2001.2790
Goswami, U., Ziegler, J. C., Dalton, L., & Schneider, W. (2003). Nonword
reading across orthographies: How flexible is the choice of reading
units? Applied Psycholinguistics, 24, 235–247. doi:10.1017/
S0142716403000134
Grainger, J. (2008). Cracking the orthographic code: An introduction.
Language and Cognitive Processes, 23, 1–35. doi:10.1080/
01690960701578013
Grainger, J., Dufau, S., Montant, M., Ziegler, J. C., & Fagot, J. (2012, April
13). Orthographic processing in baboons (Papio papio). Science, 336,
245–248. doi:10.1126/science.1218152
Grainger, J., Kiyonaga, K., & Holcomb, P. J. (2006). The time course of
orthographic and phonological code activation. Psychological Science,
17, 1021–1026. doi:10.1111/j.1467-9280.2006.01821.x
Grainger, J., Lété, B., Bertrand, D., Dufau, S., & Ziegler, J. C. (2012).
Evidence for multiple routes in learning to read. Cognition, 123, 280 –
292. doi:10.1016/j.cognition.2012.01.003
Grainger, J., & Ziegler, J. C. (2011). A dual-route approach to orthographic
processing. Frontiers in Psychology, 2. doi:10.3389/fpsyg.2011.00054
Hannagan, T., & Grainger, J. (2012). Protein analysis meets visual word
recognition: A case for string kernels in the brain. Cognitive Science, 36,
575– 606. doi:10.1111/j.1551-6709.2012.01236.x
Kohnen, S., & Castles, A. (2013). Pirates at parties: Letter position pro-
cessing in developing readers. Journal of Experimental Child Psychol-
ogy, 115, 91–107. doi:10.1016/j.jecp.2012.08.011
Ktori, M., Grainger, J., Dufau, S., & Holcomb, P. J. (2012). The “electro-
physiological sandwich”: A method for amplifying ERP priming effects.
Psychophysiology, 49, 1114 –1124. doi:10.1111/j.1469-8986.2012
.01387.x
Lefavrais, J. (2005). Alouette–R: Test d’analyse de la lecture et de la
dyslexie [Alouette–R: Test for the analysis of reading and dyslexia].
Paris, France: Editions du Centre de Psychologie Appliquée.
Lété, B., & Fayol, M. (2013). Substituted-letter and transposed-letter
effects in a masked priming paradigm with French developing readers
and dyslexics. Journal of Experimental Child Psychology, 114, 47– 62.
doi:10.1016/j.jecp.2012.09.001
Lété, B., Sprenger-Charolles, L., & Colé, P. (2004). MANULEX: A
grade-level lexical database from French elementary school readers.
Behavior Research Methods Instruments, & Computers, 36, 156 –166.
doi:10.3758/BF03195560
Lukatela, G., & Turvey, M. T. (1994b). Visual lexical access is initially
phonological: 1. Evidence from associative priming by words, homo-
phones, and pseudohomophones. Journal of Experimental Psychology:
General, 123, 107–128. doi:10.1037/0096-3445.123.2.107
Lukatela, G., & Turvey, M. T. (1994a). Visual lexical access is initially
phonological: 2. Evidence from phonological priming by homophones
and pseudohomophones. Journal of Experimental Psychology: General,
123, 331–353. doi:10.1037/0096-3445.123.4.331
Lupker, S. J., & Davis, C. J. (2009). Sandwich priming: A method for
overcoming the limitations of masked priming by reducing lexical
competitor effects. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 35, 618 – 639. doi:10.1037/a0015278
Perea, M., & Lupker, S. J. (2004). Can CANISO activate CASINO?
Transposed-letter similarity effects with nonadjacent letter positions.
Journal of Memory and Language, 51, 231–246. doi:10.1016/j.jml.2004
.05.005
Rastle, K., & Brysbaert, M. (2006). Masked phonological priming effects
in English: Are they real? Do they matter? Cognitive Psychology, 53,
97–145. doi:10.1016/j.cogpsych.2006.01.002
Ratcliff, R. (1993). Methods for dealing with reaction time outliers. Psy-
chological Bulletin, 114, 510 –532. doi:10.1037/0033-2909.114.3.510
Rayner, K., Pollatsek, A., & Binder, K. S. (1998). Phonological codes and
eye movements in reading. Journal of Experimental Psychology: Learn-
ing, Memory, and Cognition, 24, 476 – 497. doi:10.1037/0278-7393.24
.2.476
Share, D. L. (1995). Phonological recoding and self-teaching: Sine qua non
of reading acquisition. Cognition, 55, 151–218. doi:10.1016/0010-
0277(94)00645-2
Sprenger-Charolles, L., Siegel, L. S., Bechennec, D., & Serniclaes, W.
(2003). Development of phonological and orthographic processing in
reading aloud, in silent reading, and in spelling: A four-year longitudinal
study. Journal of Experimental Child Psychology, 84, 194 –217. doi:
10.1016/S0022-0965(03)00024-9
Treiman, R., Goswami, U., & Bruck, M. (1990). Not all nonwords are
alike: Implications for reading development and theory. Memory &
Cognition, 18, 559 –567. doi:10.3758/BF03197098
Treiman, R., Sotak, L., & Bowman, M. (2001). The roles of letter names
and letter sounds in connecting print and speech. Memory & Cognition,
29, 860 – 873. doi:10.3758/BF03196415
Van Orden, G. C. (1987). A ROWS is a ROSE: Spelling, sound, and
reading. Memory & Cognition, 15, 181–198. doi:10.3758/BF03197716
Whitney, C. (2001). How the brain encodes the order of letters in a printed
word: The SERIOL model and selective literature review. Psychonomic
Bulletin & Review, 8, 221–243. doi:10.3758/BF03196158
Yarkoni, T., Balota, D., & Yap, M. (2008). Moving beyond Coltheart’s N:
A new measure of orthographic similarity. Psychonomic Bulletin &
Review, 15, 971–979. doi:10.3758/PBR.15.5.971
Ziegler, J. C., & Goswami, U. (2005). Reading acquisition, developmental
dyslexia, and skilled reading across languages: A psycholinguistic grain
size theory. Psychological Bulletin, 131, 3–29. doi:10.1037/0033-2909
.131.1.3
Ziegler, J. C., Hannagan, T., Dufau, S., Montant, M., Fagot, J., & Grainger,
J. (2013). Transposed-letter effects reveal orthographic processing in
baboons. Psychological Science, 24, 1609 –1611. doi:10.1177/
0956797612474322
Ziegler, J. C., Jacobs, A. M., & Klueppel, D. (2001). Pseudohomophone
effects in lexical decision: Still a challenge for current word recognition
models. Journal of Experimental Psychology: Human Perception and
Performance, 27, 547–559. doi:10.1037/0096-1523.27.3.547
Ziegler, J. C., Perry, C., Ma-Wyatt, A., Ladner, D., & Schulte-Korne, G.
(2003). Developmental dyslexia in different languages: Language-
specific or universal? Journal of Experimental Child Psychology, 86,
169 –193. doi:10.1016/S0022-0965(03)00139-5
Ziegler, J. C., Perry, C., & Zorzi, M. (in press). Modelling reading devel-
opment through phonological decoding and self-teaching: Implications
for dyslexia. Philosophical Transactions of the Royal Society B: Bio-
logical Sciences.
Ziegler, J. C., Van Orden, G. C., & Jacobs, A. M. (1997). Phonology can
help or hurt the perception of print. Journal of Experimental Psychology:
Human Perception and Performance, 23, 845– 860. doi:10.1037/0096-
1523.23.3.845
 ORTHOGRAPHIC AND PHONOLOGICAL MASKED PRIMING 1035

Appendix A
Stimulus Material in the Pseudohomophone (PsH) Condition

Target word PsH Control Target nonword PsH Control

blanc blenc blunc drans drens druns

brosse broce brofe blisse blice blife
bruit bruie bruif gluit gluie gluif
casse cace cabe cusse cuce cufe
cave kave fave cune kune fune
chose choze chore frose froze frore
danse dance danve runse runce runve
farce farse farle funce funse funle
forme phorme blorme farbe pharbe blarbe
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

frite fritte frille frate fratte fralle

This document is copyrighted by the American Psychological Association or one of its allied publishers.

froid froit froil praud praut praul

genrea jenre venre giste jiste viste
largeb larje larte durge durje durte
lentb lend lenc bunt bund bunc
linge linje linte donge donje donte
loup lout louf maip mait maif
mere mert merf lère lert lerf
neige naije noide teige taije toide
nordb nore norf zird zire zirf
pain pein poun hain hein houn
place plasse plamme droce drosse dromme
plage plaje plame troge troje trome
prise prize prine cluse cluze clune
rose roze roge tase taze tage
rouge rouje roude reige teije teide
sage saje sare nige nije nire
seizec seise seibe roize roise roibe
singe sinje sinfe ringe rinje rinfe
sucre sukre subre ticle tikle tible
trace trasse tralle bloce blosse blolle
train trin trun scain scin scun
vase vaze vare dese deze dere
vrai vrei vroi glai glei gloi
a b c
Excluded in Grades 1 and 2. Excluded in Grade 1 only. Excluded in Grades 1, 2, and 3.

(Appendices continue)
 1036 ZIEGLER, BERTRAND, LÉTÉ, AND GRAINGER

Appendix B
Stimulus Material in the Transposed-Letter (TL) Condition

Target
Target word TL Control nonword TL Control

bouche bouhce bounle gaiche gaihce gainle

chance chacne charbe gronce grocne grorbe
comptea comtpe comfre vempte vemtpe vemfre
corde codre colne narde nadre nalne
corpsa coprs cotns marps maprs matns
couche couhce coulme veiche veihce veilme
course cousre coufpe tuirse tuisre tuifpe
ferme femre fepse durme dumre dupse
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

gauche gauhce gaulpe voiche voihce voilpe

This document is copyrighted by the American Psychological Association or one of its allied publishers.

herbe hebre hetme vorbe vobre votme

jambe jabme jaspe cumbe cubme cuspe
lampe lapme larbe rumpe rupme rurbe
langue lagnue lartue lungre lugnre lurtre
livre lirve litne nuvre nurve nutne
marche macrhe malphe borche bocrhe bolphe
monde modne molre jande jadne jalre
montre motnre molgre cantre catnre calgre
ombrea omrbe omtle usbre usrbe ustle
onclea onlce onrge iscle islce isrge
pauvre paurve paunte deivre deirve deinte
poche pohce povse liche lihce livse
pointe poitne poisme gounte goutne gousme
porte potre polde sirte sitre silde
poudre pourde pougse suidre suirde suigse
prince pricne prifme stonce stocne stofme
reste retse rebde muste mutse mubde
sable salbe sapne muble mulbe mupne
sieste sietse sierpe louste loutse lourpe
table talbe tarfe poble polbe porfe
tante tatne tamle sunte sutne sumle
tarte tatre tabse dorte dotre dobse
touche touhce toulbe caiche caihce cailbe
trompe tropme trobse stimpe stipme stibse
vache vahce vasme tuche tuhce tusme
ventre vetnre vepsre fintre fitnre fipsre
viande viadne viampe tounde toudne toumpe
a
Excluded in Grade 1 only.

Received February 16, 2013

Revision received October 9, 2013
Accepted October 23, 2013 䡲