=Paper=
{{Paper
|id=Vol-2769/60
|storemode=property
|title=Does Finger-Tracking Point to Child Reading Strategies?
|pdfUrl=https://ceur-ws.org/Vol-2769/paper_60.pdf
|volume=Vol-2769
|authors=Claudia Marzi,Anna Rodella,Andrea Nadalini,Loukia Taxitari,Vito Pirrelli
|dblpUrl=https://dblp.org/rec/conf/clic-it/MarziRNTP20
}}
==Does Finger-Tracking Point to Child Reading Strategies?==
https://ceur-ws.org/Vol-2769/paper_60.pdf
Does finger-tracking point to child reading strategies?
Claudia Marzi Anna Rodella
ILC - CNR Pisa Università di Pisa
claudia.marzi@ilc.cnr.it a.rodella@studenti.unipi.it
Andrea Nadalini Loukia Taxitari Vito Pirrelli
ILC - CNR Pisa ILC - CNR Pisa ILC - CNR Pisa
andrea.nadalini@ilc.cnr.it loukia.taxitari@ilc.cnr.it vito.pirrelli@ilc.cnr.it
Abstract children are learning to read. Despite the undoubt-
edly different dynamics of the two types of text
The movement of a child’s index finger exploration, finger-pointing to text helps children
that points to a printed text while (s)he learn to look at print, and supports critical early
is reading may provide a proxy for the reading behaviours: directional movement, atten-
child’s eye movements and attention fo- tion focus, and voice-print match (Mesmer and
cus. We validated this correlation by Lake, 2010; Uhry, 2002).
showing a quantitative analysis of pat- ReadLet (Ferro et al., 2018a; Ferro et al.,
terns of “finger-tracking” of Italian early 2018b) is a web application with a tablet front-end,
graders engaged in reading a text dis- designed to support online monitoring of silent
played on a tablet. A web application in- and oral reading abilities through finger-tracking.
terfaced with the tablet monitors the read- Finger-tracking consists of recording the time se-
ing behaviour by modelling the way the ries of touch events on the tablet screen where
child points to the text while reading. The a child is reading a short story, while the child
analysis found significant developmental is pointing to the text with the index finger of
trends in reading strategies, marking an in- her dominant hand.2 Preliminary analyses of our
teresting contrast between typically devel- finger-tracking data (Pirrelli et al., 2020) high-
oping and atypically developing readers. lighted a diminishing influence of word frequency
and word length on reading time as readers get
older and more proficient (from 3rd to 6th grades).
1 Introduction With increasing exposure to written words, differ-
ences in tracking time between high and low fre-
Recent experimental evidence in visual perception
quency words gradually tend to decrease, suggest-
analysis (Lio et al., 2019) shows that eye move-
ing a ceiling effect in the entrenchment of both
ments and finger movements strongly correlate
high- and low-frequency lexical representations in
during scene exploration, at both individual and
long-term memory (Zoccolotti et al., 2009). Simi-
group levels. In Lio et al.’s (2019) experiment,
larly, word length was found to significantly inter-
subjects are invited to explore a blurred image dis-
act across grades. Younger readers show increas-
played on a touchscreen by moving their fingers
ing difficulty with longer words, with a steeper
on the display. Picture areas that are located im-
time increment for word length > 6, while older
mediately above the touch point of the subject’s
readers are slowed down when words are longer
finger on the screen are automatically shown in
than 8 letters. This integrates previous evidence
high resolution, thereby simulating the subject’s
(De Luca et al., 2008), confirming that not even
central (foveal) vision. The experiment proves that
the most experienced readers can avoid the slow-
the subjects’ image-exploring patterns in the two
ing down effect of word length.
modalities (optical and tactile) are highly congru-
The two-fold interaction of word frequency and
ent. The result is not surprising. A familiar context
word length with grade levels strongly suggests
which exploits this synergistic behaviour is when
that Italian children use a lexical route to decod-
Copyright c 2020 for this paper by its authors. Use per-
2
mitted under Creative Commons License Attribution 4.0 In- https://www.readlet.it/: see section publica-
ternational (CC BY 4.0). tions for materials related to this paper.
ing even at early stages of their reading develop- processing, where audio and finger-tracking time
ment, despite the transparency of Italian orthog- series are aligned with the text. Recorded and
raphy (Bates et al., 2001; Davies et al., 2013). post-processed data are exposed through a set of
It also suggests that young readers can make use web services offered by the cloud server.
of sublexical information, whenever they are con-
fronted with words that are not contained in their 3 The Data
orthographic lexicon. This developmental dy-
For our present goal, we focus on reading data of
namic would account for the stronger sensitivity of
237 children, sampled from entire classes rang-
less skilled readers to both lexical frequency and
ing from 3rd through to 6th school grades, in Ital-
word length. As lexical information increases with
ian and Italian-speaking Swiss schools.3 Partic-
age, with rarer and longer words finding their way
ipants included both typically developing readers
into the reader’s orthographic lexicon, the reader
(N=214) and children screened and reported in
makes an increasingly prominent use of lexical in-
schools as atypically developing readers (N=23),
formation and an increasingly sparser use of sub-
but who did not receive a clinical diagnosis. Eight
lexical information.
short stories were created for the pilot study, one
In this paper, we provide a finer-grained quan- for each of the four school grades, and for each
titative analysis of the finger-tracking profile of experimental condition (silent and aloud reading).
typically developing readers, offering further ev- Children were asked to read a story while finger-
idence that their reading strategy results from an tracking the text. After reading in the silent con-
optimal, interactive combination of both lexical dition only, children were asked a few multiple-
and sublexical information. The evidence is com- choice questions, to ascertain they actually carried
pared with the finger-tracking profile of difficult out the task.
readers. To provide a more realistic developmen- Texts were automatically annotated for part-of-
tal profile of these effects, we restrict our fo- speech, word token frequency, and word typical-
cus on nouns only, which are less likely to be ity (measured as either the size of the word’s lex-
skipped while finger-tracking, and present a nar- ical neighbourhood, or N-size, or the mean Lev-
rower range of variability in both length and fre- enshtein distance from its top 20 neighbouring
quency. words, or Old20 (Yarkoni et al., 2008)).
For each child, in both reading conditions, we
2 ReadLet calculated the token tracking time as the total time
ReadLet is a tablet-based application that com- spent in finger-tracking each word token while
bines an objective assessment of a child’s read- reading. To ensure reliability and precision in the
ing fluency and comprehension skills with care- alignment of finger-tracking data with the text be-
ful collection of large-scale behavioural data, and ing read, we selected reading trials with ≥75% of
quantitative modelling of the specific factors af- finger-tracked text pages. From the original set of
fecting reading development. It leverages an ICT tokens making up the 8 short stories, we selected
infrastructure with a cloud-based back-end expos- 97 lemmas for 109 noun tokens, by intersecting
ing a battery of web services acting as an inter- our data with age of acquisition and imageabil-
face between the central repository and the users. ity assessments by Italian speakers (Montefinese
The ReadLet front-end is an ordinary tablet, where et al., 2014; Montefinese et al., 2019). In the re-
short stories are displayed for children to read, ei- sulting data sample, word frequency4 is observed
ther silently or aloud. In both cases, the child 3
Data come from a pilot data collection carried out by
is asked to finger-point to the text while reading. Claudia Cappa and Sara Giulivi in schools of Southern Tus-
cany and Canton of Ticino in 2018 and 2019 for the AEREST
Texts are displayed on a 10” tablet screen in Lato project, funded by the Department of Teaching and Learn-
font 17pt in black against a white background. ing of the University of Applied Sciences and Arts of South-
During each reading session, the behaviour of the ern Switzerland. Data were encrypted and pseudonymized
locally, to then be delivered through a https protocol to a
child is captured through large streams of time- centralized server, where they are accessed with authentica-
aligned signals including voice recording, time- tion credentials and dedicated web services. A background
stamped finger-tracking patterns, reading time and history of children including information about any clinical
and/or psychological disorders was provided by the schools.
question-answering time. Data are automatically 4
Token frequency is given as the natural logarithm of oc-
captured and sent to a centralised server for post- currences in the Subtlex corpus (http://crr.ugent.be/subtlex-
to vary between min=5.61 and max=11.77, and
aloud silent
word length between 4 and 10 letters (median=5,
mean=5.62, sd=1.40). 2.0
4 Typical and atypical reading
tracking time
1.5
3
development 4
5
1.0
6
The main goal of the ReadLet project is to propose
and validate an ICT methodology for assessing the
0.5
typical reading development of children in Italian
schools. In this section we focus on the finger- 6 8 10 12 6 8 10 12
token frequency
tracking behaviour of typically developing chil-
dren engaged in reading a short text. The idea is to
Figure 1: Linear interaction of token frequency
provide evidence that finger-tracking patterns ex-
and grade level (from 3rd to 6th level), for typ-
hibit lexical effects that are well-established in the
ically developing readers in both aloud and silent
reading literature: namely word frequency, word
reading. Shaded areas indicate 95% confidence in-
length and word similarity (or N-size).5
tervals.
Figure 1 shows the effects of word frequency
across grade levels, in both aloud (left panel) and
silent (right panel) reading, for typically develop- typical atypical
ing readers. A linear mixed model fitting token 2.0
tracking time as a function of reading type, word
frequency and grade levels shows shorter tracking
tracking time
1.5
3
times in reading more frequent words. The model 4
also highlights a significant interaction between 1.0
5
6
years of schooling and word frequency, with fa-
cilitation effects getting smaller for older graders,
0.5
particularly in silent reading. The difference in fa-
cilitation rate between the two reading tasks is not 6 8 10 12 6 8 10 12
token frequency
statistically significant.
Figure 2 compares the developmental patterns
Figure 2: Linear interaction of token frequency
of token tracking time of typically (left panel) and
and grade level (from 3rd to 6th level), for typi-
atypically (right panel) developing readers, mod-
cally and atypically developing readers.
elled as a linear function of word token frequency
and grade level. The two patterns exhibit a clear
facilitatory effect of token frequency on reading readers is contrasted with the same effect in atyp-
speed, confirming that frequency makes reading ical readers. In both groups, children take more
consistently easier for both typical and atypical time to read longer words, but only typically de-
populations of young readers, who appear to en- veloping children exhibit a less strong sensitivity
tertain the same lexical reading strategy. How- to word length as grade level increases. The sta-
ever, only in typically developing children the ef- tistical significance of this interaction disappears
fect tends to diminish across grade levels, with in atypical readers, with the only exception of 3rd
slopes getting less steep as grade levels increase graders, compared with all remaining graders.
(Figure 2, left panel).6 Figure 4 shows how grade levels interact with
A similar overall pattern is shown in Figure N-size in affecting aloud (left panel) and silent
3, where the sensitivity to word length of typical (right panel) reading time. The dominant effect
is facilitatory, with a clear incremental advantage
it/) plus one. For our set of noun data the mean frequency is
9.45 (sd=1.61). in reading times for words with a high number
5
All figures in the section show regression plots of the of neighbours. Words are finger-tracked more
interaction of main effects, using the ggplot function. quickly when they belong to more dense neigh-
6
Regression slopes for 4th and 5th grades are not statis-
tically different from 3rd grade, but there is a significant dif- bourhoods, and this facilitatory effect is stronger
ference when comparing slopes for 3rd and 6th grades. for younger (3rd and 4th grade) than older (5th
typical atypical typical atypical
2.0 2.0
tracking time
tracking time
1.5 1.5
3
4 aloud
5 silent
1.0 1.0
6
0.5 0.5
4 6 8 10 4 6 8 10 0 5 10 15 20 25 0 5 10 15 20 25
word length number of neighbours
Figure 3: Linear interaction of word length and Figure 5: Linear interaction of N-size and read-
grade level (from 3rd to 6th ) for typically and atyp- ing task (aloud and silent) for typical and atypical
ically developing readers. readers.
aloud silent typical atypical
2.0 2.0
tracking time
tracking time
1.5 1.5
3 3
4 4
5 5
1.0 1.0
6 6
0.5 0.5
0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25
number of neighbours number of neighbours
Figure 4: Linear interaction of N-size and grade Figure 6: Linear interaction of N-size and grade
level (from 3rd to 6th ) for aloud and silent reading level (from 3rd to 6th ) for typically and atypically
of typical readers. developing readers.
and 6th grade) readers. No significant difference ers does not replicate the trend of typical readers.
is found in the interaction between reading type First, the tracking time of atypical readers is more
and N-size in typical readers (Figure 5, left panel). strongly - and significantly - affected by token fre-
Atypical readers show equal slopes in both aloud quency, when compared with the typical tracking
and silent reading, but different intercepts, which time of their age-matched peers. This is espe-
capture the additional processing demands of con- cially true for the youngest readers in our sample
current articulation (Figure 5, right panel). This (3rd graders in the right plot of Figure 2). In ad-
evidence suggests a sublexical reading strategy dition, sensitivity to frequency appears to persist
that relies on orthographic similarity: words that with age, as there are no significant differences
are not read lexically (because they are too long or in the facilitatory effect of frequency across later
less frequent), are read by decoding and combin- grade levels. This suggests a delay in developing
ing the smaller parts they share with other neigh- and integrating lexical information. A nearly iden-
bouring words. Fitting a mixed model with N-size, tical developmental pattern is replicated with N-
frequency ranges, and grade levels, as variables size effects (Figure 6, right panel): younger chil-
predicting the token tracking time, and with sub- dren read words in denser neighbourhoods more
jects as random effect, we find that all predictors easily, taking advantage of the recurrent sublex-
and interactions are highly significant for typically ical parts shared by neighbouring words. Once
developing children (Figure 6). more, no significant developmental pattern is ob-
The behaviour of atypically developing read- served across grade levels, as atypical readers do
not appear to be able to increasingly rely on lexical This strategy remains in operation through read-
reading as they get more experienced (Figure 2). ing development, as shown by the decreasing
Finally, their reading time is persistently slowed tracking times of 6th graders as a function of lex-
down by longer words, suggesting a difficulty in ical frequency (Figure 2, left panel). Nonethe-
memorizing and making them accessible through less, the impact of word frequency is less strong in
the lexical route (Figure 3, right panel). older readers, whose orthographic lexicon makes
room for increasingly rarer (and longer) words.
5 General discussion Also atypical readers appear to use a similar
“chunking” strategy, but their developmental pat-
Facilitatory effects of lexical frequency on reading tern fails to show a clear interaction between grade
reaction time have been reported for Italian chil- level and frequency. In Figure 2 (right panel), 3rd
dren (Barca et al., 2006; Burani et al., 2002) as graders show a robust word frequency effect, but
well as adults (Barca et al., 2002; Burani et al., the diminishing role of frequency on tracking time
2007). The effects are argued to reflect the work- across grades turns out not to be significant. This
ing of the lexical route in dual-route models of suggests that atypical readers have problems with
reading (Coltheart et al., 2001): word items are ac- developing orthographic representations for rarer
cessed in the reader’s orthographic lexicon, to then (and longer) words, and they are not quite as suc-
be pronounced after their full phonological code cessful as typical readers in optimally integrating
is retrieved. The faster reading of high-frequency lexical and sublexical information.
lexical items thus reflects the well-established sen-
sitivity of lexical access to word frequency. In Ital- This interpretation is supported by the analysis
ian, the systematic nature of letter-to-sound map- of two other lexical effects on child reading devel-
ping rules makes the operation of a sublexical opment: word length and neighbourhood size (N-
reading strategy a reliable alternative to the lexi- size). As expected, longer words elicit longer re-
cal route. However effective, sublexical reading is sponse latency and reading duration, but the effect
nonetheless less efficient, since it requires the on- is bigger for younger, typically developing read-
line, serial decoding of a word by its parts (e.g. ers compared to older ones (Figure 3, left panel),
n-grams or syllables). We conjecture that Italian and for atypical readers compared to their age-
children optimize reading efficiency at early stages matched peers (Figure 3, right panel). The use
of their reading practice, through dynamic integra- of sublexical information and serial n-gram decod-
tion of sublexical and lexical reading. Whenever ing appears to be more prominent in younger and
possible, they resort to word-sized orthographic atypical readers than in the older and more skilled
information in their lexicon (e.g. short and fre- group of readers. Once more, the effect can be
quent words), and make it up for missing ortho- argued to reflect the absence of fully specified or-
graphic items through sublexical information. thographic representations for longer words in the
lexicon of less skilled readers, and a related diffi-
Such an opportunistic strategy is in keeping
culty in building up complex orthographic chunks.
with the idea that early readers strive, through
reading practice, to “chunk” letter n-grams into Facilitatory effects of N-size on reading time
longer orthographic units. Chunked units are are reported for atypical Italian readers by
stored and made accessible in the readers’ lexicon, Marinelli et al. (2013), who, however, found
where they are associated with their fully speci- no significant facilitation in age-matched typical
fied phonological code. The length of stored items readers. They argue that atypical readers over-
is a function of their frequency, and the reader’s rely on co-activation of word neighbours during
processing efficiency, reading practice and age. reading to make it up for their poorly entrenched
Our data confirm that this strategy is consistently lexical representations. Conversely, access to in-
adopted by typically developing readers in both dividual lexical representations by typical readers
silent and aloud reading, suggesting that the in- is fast enough to make N-size effects hardly de-
fluence of lexical frequency is not confined to the tectable. Our data are consistent with Marinelli et
retrieval and planning stage of the word phonolog- al.’s evidence, but integrate it in two important re-
ical code, but appears to extend beyond response spects. First, the speeding-up influence of N-size
initiation, to affect full articulation of the code is detected in both aloud and (for the first time to
(Balota and Yap, 2006). our knowledge) silent reading of Italian, with no
significant difference between the two (Figure 4 groups of readers are sensitive to the same lexical
and Figure 5, left panel). This supports an inter- effects, but that atypical readers rely on an impov-
pretation of the N-size effect as having an impact erished lexicon. We take this evidence to show that
on both phonological planning and overt articula- although the two groups adopt the same strategy,
tion. Secondly, our data show that the effect is not they differ in their global ability in serial informa-
limited to the reading pace of younger and atypi- tion processing, which has a boosting influence on
cal readers, as observed by Marinelli et al., but it lexical development and reading speed.
also holds for typically developing readers (Figure Despite our promising results, one could le-
6), with an interesting modulation by grade level. gitimately wonder why we propose using finger-
This is mainly due to our focus on nouns, which tracking as a proxy of a more established tech-
include longer and less frequent words, for which nology such as eye-tracking. Portability and task
N-size effects are known to be stronger and easier ecology are our strongest arguments. ReadLet
to detect (Davies et al., 2013). Finally, the dimin- can be used in almost any environment with no
ishing impact of N-size for increasing grade lev- data-acquisition specialist or invasive, anxiety-
els confirms a sparser use of the sublexical route provoking equipment. This has practical con-
by more skilled readers, who are equipped with a sequences for research in education, computer
richer and more efficient orthographic lexicon. science, human cognition and medical sciences.
To sum up, typical and atypical readers alike Our architecture supports highly parallel and dis-
strive to optimally integrate lexical and sublexi- tributed processes of data acquisition, which can
cal input patterns while reading, using the former be delivered in real time to research, clinical and
whenever possible for efficient decoding, and the education centers as terminals for data harvest-
latter as a fall-back strategy, whenever the lexi- ing and quantitative analysis. Large-scale studies
cal route fails. This dynamic, however straight- can be conducted, paving the way to more gen-
forward, has non-trivial consequences. In a de- eralizable results than ever in the past. In addi-
velopmental perspective, the orthographic lexicon tion, the possibility to take single-subject measure-
gets richer with practice, boosted by an age-driven ments on more occasions and in different environ-
improvement of children’s global ability in infor- ments makes finger-tracking evidence usable not
mation processing (Zoccolotti et al., 2009), which only in group studies but also for individual di-
makes longer and rarer words easier to store. As a agnostic purposes. Furthermore, the fine-grained,
result, the dynamic balance is shifted towards lex- multimodal evidence of different signal streams
ical reading. Conversely, atypical readers find it which are aligned with time and with linguistically
more difficult to develop and store detailed map- annotated texts provides invaluable training data
pings between orthographic and phonological se- for artificial neural networks and classification al-
quences, as confirmed by their greater sensitiv- gorithms designed to solve engineering problems
ity to frequency and length effects (Figures 2 and or simulate neurophysiological correlates of cog-
3) and by a prolonged, larger effect of N-size on nitive tasks. Last but not least, we know that read-
finger-tracking (Figure 5, right panel). ing probes are a commonly used for monitoring
progress in reading fluency and text comprehen-
6 Concluding remarks sion (Miura Wayman et al., 2007), but take huge
time and effort to collect. The use of a tablet for
We provided evidence that finger-tracking data of extended reading enables deriving this informa-
reading children can highlight congruent develop- tion unobtrusively and continuously, wherever the
mental patterns in the acquisition of literacy skills. child fancies reading, even at home.
We only replicated established benchmark effects
reported in the psycholinguistic literature on de- Acknowledgments
coding transparent orthographies. Nonetheless, to
our knowledge, this is the first time that finger- This work is supported by PRIN grant
tracking patterns are shown to significantly corre- 2017W8HFRX ReadLet: reading to under-
late with more established reading data. stand. An ICT driven, large-scale investigation
Unsurprisingly, typically developing readers of early grade children’s reading strategies
were shown to read at a faster rate than atypical (2020-22), from the Italian Ministry of University
readers. Our comparative analysis shows that both and Research.
References Guillaume Lio, Roberta Fadda, Giuseppe Doneddu,
Jean-René Duhamel, and Angela Sirigu. 2019.
David A Balota and Melvin J Yap. 2006. Attentional Digit-tracking as a new tactile interface for vi-
control and the flexible lexical processor: Explo- sual perception analysis. Nature Communications,
rations of the magic moment of word recognition. 10(5392):1–13.
From inkmarks to ideas: Current issues in lexical
processing, 229. Chiara Valeria Marinelli, Daniela Traficante, Pierluigi
Zoccolotti, and Cristina Burani. 2013. Ortho-
Laura Barca, Cristina Burani, and Lisa S Arduino. graphic neighborhood-size effects on the reading
2002. Word naming times and psycholinguistic aloud of italian children with and without dyslexia.
norms for italian nouns. Behavior Research Meth- Scientific Studies of Reading, 17(5):333–349.
ods, Instruments, & Computers, 34(3):424–434.
Heidi Anne E. Mesmer and Karen Lake. 2010. The
Laura Barca, Cristina Burani, Gloria Di Filippo, and role of syllable awareness and syllable-controlled
Pierluigi Zoccolotti. 2006. Italian developmental text in the development of finger-point reading.
dyslexic and proficient readers: Where are the dif- Reading Psychology, 31(2):176–201.
ferences? Brain and Language, 98(3):347–351.
Miya Miura Wayman, Teri Wallace, Hilda Ives Wi-
Elizabeth Bates, Cristina Burani, Simona D’Amico, ley, Renáta Tichá, and Christine A Espin. 2007.
and Laura Barca. 2001. Word reading and picture Literature synthesis on curriculum-based measure-
naming in italian. Memory & Cognition, 29(7):986– ment in reading. The Journal of Special Education,
999. 41(2):85–120.
Cristina Burani, Stefania Marcolini, and Giacomo Maria Montefinese, Ettore Ambrosini, Beth Fairfield,
Stella. 2002. How early does morpholexical reading and Nicola Mammarella. 2014. The adaptation of
develop in readers of a shallow orthography? Brain the affective norms for english words (anew) for ital-
and language, 81(1-3):568–586. ian. Behavior research methods, 46(3):887–903.
Cristina Burani, Lisa S Arduino, and Laura Barca. Maria Montefinese, David Vinson, Gabriella
2007. Frequency, not age of acquisition, affects ital- Vigliocco, and Ettore Ambrosini. 2019. Ital-
ian word naming. European Journal of Cognitive ian age of acquisition norms for a large set of words
Psychology, 19(6):828–866. (itaoa). Frontiers in psychology, 10:278.
Max Coltheart, Kathleen Rastle, Conrad Perry, Robyn Vito Pirrelli, Claudia Cappa, Davide Crepaldi, Viola
Langdon, and Johannes Ziegler. 2001. Drc: a dual Del Pinto, Marcello Ferro, Sara Giulivi, Claudia
route cascaded model of visual word recognition and Marzi, Andrea Nadalini, and Loukia Taxitari. 2020.
reading aloud. Psychological review, 108(1):204. Tracking the pace of reading with finger movements.
Canada, 16-18 October, 2020. Words in the World
Robert Davies, Javier Rodríguez-Ferreiro, Paz Suárez, International Conference.
and Fernando Cuetos. 2013. Lexical and sub-lexical
effects on accuracy, reaction time and response du- Joanna K. Uhry. 2002. Finger-point reading in kinder-
ration: impaired and typical word and pseudoword garten: The role of phonemic awareness, one-to-one
reading in a transparent orthography. Reading and correspondence, and rapid serial naming. Scientific
Writing, 26(5):721–738. Studies of Reading, 6(4):319–342.
Tal Yarkoni, David Balota, and Melvin Yap. 2008.
Maria De Luca, Laura Barca, Cristina Burani, and Pier-
Moving beyond coltheart’s N: A new measure of or-
luigi Zoccolotti. 2008. The effect of word length
thographic similarity. Psychonomic bulletin & re-
and other sublexical, lexical, and semantic variables
view, 15(5):971–979.
on developmental reading deficits. Cognitive and
Behavioral Neurology, 21(4):227–235. Pierluigi Zoccolotti, Maria De Luca, Gloria Di Fil-
ippo, Anna Judica, and Marialuisa Martelli. 2009.
Marcello Ferro, Claudia Cappa, Sara Giulivi, Claudia Reading development in an orthographically regu-
Marzi, Franco Alberto Cardillo, and Vito Pirrelli. lar language: Effects of length, frequency, lexicality
2018a. ReadLet: an ICT platform for the assessment and global processing ability. Reading and Writing,
of reading efficiency in early graders. page 61, Ed- 22(9):965–992.
monton, Alberta (Canada), 25-29 September, 2018.
11th International Conference on the Mental Lexi-
con.
Marcello Ferro, Claudia Cappa, Sara Giulivi, Clau-
dia Marzi, Ouaphae Nahli, Franco Alberto Cardillo,
and Vito Pirrelli. 2018b. Readlet: Reading for
understanding. In 2018 IEEE 5th International
Congress on Information Science and Technology
(CiSt), pages 1–6.