You are on page 1of 19

Journal of Research in Reading, ISSN 0141-0423 DOI:10.1111/1467-9817.

12359
Volume 00, Issue 00, 2021, pp 1–19

Reading competence and its impact on


writing: an approach towards mental
representation in literacy tasks

Antonio P. Gutierrez de Blume


Department of Curriculum, Foundations, and Reading, Georgia Southern Univer-
sity, Statesboro, GA, USA

Christian Soto
Department of Spanish, Psycholinguistics Area, Universidad de Concepción,
Concepción, Chile

Camila Ramírez Carmona


Universidad de Concepción, Concepción, Chile

Fernanda Rodriguez
Department of Spanish, Psycholinguistics Area, Universidad de Concepción,
Concepción, Chile

Patricio Pino Castillo


Facultad de Educación, Universidad Santo Tomás, Santiago, Chile

Background: Reading and writing are both fundamental activities for successful
learning. However, little is known about the effect of reading comprehension perfor-
mance on writing, as well as the pedagogical guidelines that can be drawn from this
influence.
Method: Thus, the purpose of the present investigation was to examine the influence
of performance in reading comprehension, distinguishing between poor and proficient
readers (N = 105), who were enrolled in four eighth-grade classes between the ages of
12 and 14, on the writing of narrative and expository texts.
Results: Results revealed that proficient readers outperformed poor readers on objec-
tive measures of text production and informative/expository texts. Additionally, re-
gression models demonstrated that proficient readers relied more on deeper aspects
of reading and writing such as inferential skills, whereas poor readers tended to focus
on superficial aspects of texts, or what Kintsch referred to as text-base, and appeared
to perform better in reading and writing tasks related to narratives compared to
information-based, expository texts.
Conclusion: These results support the theoretical perspectives of Kintsch’s construc-
tion–integration model and Otero’s regulation model regarding the relation between
reading, writing and mental representations.

Keywords: mental representation, performance, reading comprehension, writing

© 2021 UKLA. Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ,
UK and 350 Main Street, Malden, MA 02148, USA
GUTIERREZ DE BLUME, SOTO, RAMÍREZ CARMONA, RODRIGUEZ & PINO CASTILLO

Highlights

What is already known about this topic

• Little is known about the relation between aspects of reading and writing.
• Effective readers and writers need to slow their processing to develop a clear
mental model of texts.
• Clear mental models of texts allow readers to invoke prior knowledge to fill
comprehension gaps in their reading and writing process.

What this paper adds


• Reading and writing are related processes to the extent that effective readers
also tend to be effective writers.
• Poor and proficient readers exhibit distinct writing profiles when producing
narrative or expository texts.

Implications for theory, policy or practice

• Proficient readers and writers rely on deeper, higher order processes during
reading and text production than poor readers, who rely on superficial aspects.
• Educational interventions should focus on both reading comprehension and
writing production to be most effective.

Reading comprehension is an effortful, reflective process that requires individuals to mar-


shal a series of cognitive (e.g. strategy selection) and metacognitive (e.g. planning, infor-
mation management, and comprehension monitoring) resources to comprehend texts
more fully (Bereiter & Scardamalia, 1987). Because effective reading comprehension of
texts involves such activities as negotiating the meaning intended by the author, the
intra-individual characteristics of the reader (e.g. prior knowledge of the topic[s] covered
in the text stored in long-term memory), the characteristics of the text itself (e.g. vocabu-
lary and sentence complexity; abstractness vs. concreteness; cohesion and coherence;
and lexical diversity; among others) and the nature of the text (e.g. expository in learning
of science text vs. narrative texts; e.g. Arya et al., 2011; Bereiter & Scardamalia, 1987;
Dunlosky & Lipko, 2007; Hiebert & Pearson, 2014), the process of reading should be
approached with deliberation and care. Readers, for instance, need to determine when they
have acquired sufficient information from a text. If they recognise that they have not
reached a sufficiently deep level of understanding, presumably, they should engage in
deeper processing to enrich understanding. Processes such as metacognitive monitoring
and control as well as self-explanations are necessary for successful reading comprehen-
sion (McNamara, 2017; McNamara & Magliano, 2009).
Such a dynamic and complex process as reading comprehension has the potential to lead
to complications in text comprehension as a function of reading proficiency. Recent re-
search on this topic has demonstrated that individuals exhibit differences in critical learning
outcomes, which are moderated by reading comprehension proficiency (e.g. Soto

© 2021 UKLA
READING, WRITING, AND MENTAL REPRESENTATION

et al., 2020). More specifically, reading proficiency appears to influence inference-making


(Kintsch & Rawson, 2007; Soto et al., 2019), metacomprehension (Soto et al., 2019;
McNamara & Magliano, 2009) and even the types of cues readers use during reading,
including metacognitive ones (Soto et al., 2020). A series of studies tentatively show that
proficient readers tend to use more sophisticated strategy-selection processes, engage in
inference-making by linking information about which they are reading with prior
knowledge (i.e. information that is not explicitly stated in the text) and more frequently
employ metacognitive skills such as information management, evaluation, and monitoring.
On the other hand, poor readers tend to employ more superficial strategies and tend to
focus on the text-base level of understanding, which leads to a less enriched understanding
of texts (Language and Reading Research Consortium & Logan, 2016; Otero, 2002; Soto
et al., 2019; Soto et al., 2019; Soto et al., 2020). Nevertheless, three gaps in these studies
are that they: (1) did not differentiate between narrative texts and expository texts; (2)
employed only one measure of reading comprehension; (3) analysed data based on
composite full-scale scores rather than individual sub-dimensions, which led to a loss of
information; and (4) did not examine the dynamic relation between reading and writing.
Thus, the purposes of the present study were: (1) to investigate the effect of reading
comprehension level (proficient, poor) on students’ general writing performance on spe-
cific narrative and informative writing tasks and (2) to explore whether microlinguistics
metrics of informative and narrative writing predicted writing performance in poor and
proficient readers.

Reading and writing


Reading and writing are clearly interwoven. Research shows that comprehension and pro-
duction are: (1) separated, meaning non-existence of influences in their representations and
processes; (2) separable, meaning these representations and processes are shared and they
commune under certain circumstances; and (3) inseparable, meaning these processes and
representations are shared and indivisible (Meyer et al., 2016). Although there is no evi-
dence thus far on the nature of reading and writing as cognitive activities with separate
and individual processes and representations, there is ample evidence on confluences be-
tween comprehension and production (Buz et al., 2016; Fricke et al., 2016; Guzzardo
Tamargo et al., 2016; Hsiao & MacDonald, 2016; Kittredge & Dell, 2016; Zamuner
et al., 2016). Similarly, Berninger et al. (2002) measured children’s handwriting, spelling,
word recognition, narrative and expository written composition and reading comprehen-
sion skills. They found that: (1) word recognition predicts transcription (handwriting and
spelling) and handwriting predicts word recognition only in Grade 2; (2) reading compre-
hension and spelling have reciprocal, bidirectional relationships in all grades; and (3) read-
ing comprehension predicts composition quality, but composition quality predicts
comprehension only in Grades 4–6. Likewise, a longitudinal study on reading and writing
and their component skills such as word recognition, subword letter writing, and spelling
in Grades 1–7 demonstrated that component skills relate more strongly to themselves than
their same-level counterparts in adjacent years (Abbott et al., 2010); that is, word recogni-
tion to word recognition, as opposed to word recognition to subword letter writing. This
shows that, at least at the lexical level, reading and writing are related, but not identical.
These findings are echoed by a more recent study in Grades 3–6 that attempted to explain
development patterns in reading (word reading and reading comprehension) and writing
© 2021 UKLA
GUTIERREZ DE BLUME, SOTO, RAMÍREZ CARMONA, RODRIGUEZ & PINO CASTILLO

(spelling and written production) (Kim et al., 2018). Results showed a linear increase in
lexical-level skills, with non-linear trajectories in reading comprehension and writing
production.
A meta-analysis of 151 studies of blood flow by Indefrey (2018) has provided
evidence on the existence of a shared system of syntactic comprehension and production
with two cortical regions, which are activated differently in comprehension and
production and also in the processing of simple and complex sentences. This conceptu-
alisation has given rise to a proliferation of studies attempting to measure: (1) the impact
of writing on reading; (2) the impact of reading on writing; and (3) the impact of
combined reading–writing interventions on them. In turn, this propelled three important
meta-analyses on the same areas. A meta-analysis of 95 true or quasi-experimental
studies in Grades 1–12 that included at least one reading measure to assess the
impact of writing about material read, instruction on reading, or interventions aimed to
increase time spent in writing (Graham & Hebert, 2011). Graham and Hebert (2011)
showed that writing about material read increases narrative and expository text compre-
hension in different subject areas. They also demonstrated that multicomponent writing
instruction improved reading comprehension, particularly reading fluency and word
reading skills.
Concerning writing differences, it is well documented that there are stages in its
development, which emerge due to the growing automaticity of lower level processes,
allowing for the allocation of cognitive resources to metacognitive activities such as
planning, revision, editing and, overall, the replacement of a linear process for a multidi-
mensional one in which there is interaction between content problem solving and rhetorical
problem solving (Scardamalia & Bereiter, 1992). In this line of research, advanced writing
(knowledge-transforming) is a problem-solving process that transforms knowledge and
involves subobjectives and subprocesses in two areas, content problems and rhetorical
problems. Conversely, superficial writing (knowledge-telling) is characterised by a linear
text that is produced immediately when given a prompt, with no planning and no rhetorical
considerations.
In both reading and writing, metacognitive skills seem to play an essential role in
describing performance between skilled and poor readers and writers. In writing, research
suggests that, as compared with poor writers, skilled writers are more metacognitively
involved in the writing process. They spend more time planning and revising their text
(Harris & Graham, 1992; Hayes & Flower, 1980) and actively monitor their writing
(Beal et al., 1990). Also, skilled writers base their self-regulation on a deeper knowledge
of the task that poor writers seem to lack (Ferrari et al., 1998).

Narrative and expository text processing


Different text types require different cognitive mechanisms in their production and compre-
hension. Concerning comprehension of narrative and informative texts, León and
Peñalba (2002) devised a theoretical framework explaining cognitive functioning in the
construction of causal links in text comprehension, a chronological construction in narra-
tive texts and analytical construction in informative texts. These constructions are activated
in the readers’ prior knowledge. They also posited the prevalence of the ‘natural’ order of
antecedent–consequent and the temporal sequence as factors that facilitate comprehension.
This is especially important in science texts, in which readers may not have prior
© 2021 UKLA
READING, WRITING, AND MENTAL REPRESENTATION

knowledge as readily available as for narrative texts. An alternative perspective is offered


by a more recent study on think-aloud protocols of 107 sixth- and seventh-grade adoles-
cents from 16 different schools (Karlsson et al., 2018). Karlsson et al. (2018) showed the
existence of three reading profiles: literal, paraphrasing, and elaborative. Interestingly, chil-
dren were generally consistent with their profiles in both narrative and expository texts,
showing more incorrect inferences in expository texts and better overall performance from
elaborative readers than literal or paraphrasing readers. The extent to which this translates
into prevailing options in the writing process is as yet unknown.

Mental representation and regulation

Writers construct meaning when they compose texts, and readers construct meaning when
they understand and interpret texts (Spivey, 1990). The source of this meaning is the mental
representation, conceived as ‘configuration of concepts and relations’ (de Beaugrande &
Dressler, 1981). In reading, the mental representation consists of information from the text,
reader’s own information that is related to the text, and the inferences that are generated in
the process (Kintsch, 1988; Kintsch, 1998; McNamara & Magliano, 2009). On the other
hand, in writing, there is a movement from the ‘language of thought’ to the language of
formal prose. According to the multiple representation hypothesis of Flower and
Hayes (1984), as writers compose, they create multiple internal and external representa-
tions of knowledge, including images and text type schemas. Much of the writing work in-
volves creating and translating these mental representations of meaning into a prose. In
fact, in the conceptions of knowledge-telling and knowledge-transforming that appear in
the investigation of Scardamalia and Bereiter (1992), the main difference between mature
and immature composing is in how knowledge is brought into the writing process and what
happens to knowledge in that process.
The concept of regulation refers to the processes used to achieve cognitive consistency in
the knowledge elements of a text (Otero, 2002). This search for consistency can lead to a
modification of the mental representation generated. Otero (2002) proposes an interesting
approach to understand the regulatory procedure, which is activated after a constraint sat-
isfaction process. There is first a mechanism that evaluates the level of coherence in the
mental representation generated (the goodness of the representation, called ‘G’), and this
is subsequently contrasted with the reader’s own degree of tolerance for inconsistencies.
The latter is called MAL (minimum acceptable level). Thus, if the value of G is above
MAL, the reader will not detect any comprehension problem and will continue reading.
But if G falls below MAL, the reader detects a difficulty in understanding and activates reg-
ulatory processes. In this case, they invoke the generation of inferences that support the
comprehension of the text, modifying the initial mental representation. As it is possible
to observe, Otero’s model refers only to reading. However, something similar may happen
in writing. In this case, the mental representation under evaluation is the knowledge that is
being transformed into prose, and the regulatory processes could consist in more planning
or revision (Allal, 2000). In any case, more research is needed to confirm this hypothesis.
An integrative perspective is offered by the direct and indirect effects model of reading
(DIER; Kim, 2020), which posits that several lower and higher order variables act and in-
teract in the reading comprehension process, such as: language; cognition; knowledge;
component skills; reading affect; higher order cognition and regulation; vocabulary; gram-
matical, phonological, and morphological knowledge; spelling; and general cognition. This
© 2021 UKLA
GUTIERREZ DE BLUME, SOTO, RAMÍREZ CARMONA, RODRIGUEZ & PINO CASTILLO

hierarchical, interactive, and dynamic model implies that different variables contribute dif-
ferently to the comprehension process depending on the text, activity, and developmental
stage. Assuming that reading comprehension occurs as a result of different-level overlap-
ping representations, each level requires different components. For instance, surface-level
representation requires general cognition (working memory and attention) to process,
whereas text-base representation necessitates foundational oral skills such as vocabulary
and general grammatical knowledge. Finally, situation-model representation needs higher
order cognition and regulation, including comprehension monitoring, inferences, reasoning
and perspective-taking to establish local and global coherence. Even though this promising
model has been empirically tested, it delves neither into language production nor the inter-
actions between reading and writing.
In summary, a coherent mental representation is required to read comprehensively and to
write a quality text. The coherence standard applied varies between individuals and varies
according to the task (van den Broek et al., 2011). This standard influences the dynamic
pattern of automatic and strategic cognitive processes that take place during the literacy
task, as the regulatory processes that are activated depend on it (Otero, 2002; van den
Broek et al., 2011).
Studies that have analysed the performance in skilled and poor readers/writers confirm
the relevance of regulation. In writing, skilled writers spend more time recursively planning
and revising their text (Harris & Graham, 1992). They show the use of different writing
strategies, according to whether they were composing a narrative or an argumentative text
(Beauvais et al., 2011), and they also base their self-regulation on a deeper knowledge of
the task (Ferrari et al., 1998). Consistently throughout these studies, proficient readers ob-
tain better scores in detecting and repairing inconsistencies in a text (Cain et al., 2004;
Long & Chong, 2001; Soto et al., 2020).

The present study

Research questions and hypotheses. We proposed the following research questions to


guide the conduct of this study.

1 What is the effect of reading comprehension proficiency (proficient, poor) on


eighth-grade students’ performance in the production of narrative writing, informative
writing and general writing?
2 What aspects and to what degree do microlinguistics components of writing predict
eighth-grade students’ writing performance in narrative and informative writing as a
function of reading skill (proficient, poor)?

Hypothesis 1. We predicted that students with high reading proficiency would perform
significantly better than students with low reading skill on narrative writing, informative
writing, and general writing.

Hypothesis 2. We expected differences in predictive patterns of writing between those with


high and low reading skill. More specifically, we hypothesized that the writing predictors
for those with high reading proficiency would differ from those with poor reading

© 2021 UKLA
READING, WRITING, AND MENTAL REPRESENTATION

proficiency on reading measures because the literature shows that readers with low profi-
ciency tend to focus on superficial aspects of texts, whereas high-proficiency readers tend
to rely on deeper aspects of text during reading (Soto et al., 2019; Soto et al., 2020).

Method

Participants, sampling, and research design


Participants in this study consisted of 105 eighth-grade students (53 males and 52 females)
whose age ranged from 12 to 14 years (M = 12.89; SD = 0.96) selected via a convenience
sampling approach. We employed a quantitative non-experimental research design with
descriptive and inferential components. No student in the sample was diagnosed with a
reading or writing learning disability as of the time data were collected and their achieve-
ment across the various academic domains, gender distribution, and age range was typical
of middle school students in this region of Chile. Regarding the demographic characteris-
tics of the two groups of readers, proficient and poor (see below in Data Analysis for how
the groups were divided), the proficient readers group consisted of 57 individuals (29 males
and 28 females) whose age ranged from 12 to 14 years (M = 13.01; SD = 0.74), whereas
the poor readers group consisted of 48 individuals (26 males and 24 females) whose age
ranged from 12 to 14 years (M = 13.08; SD = 0.66). Thus, the two groups were similarly
distributed regarding demographic variables and academic achievement, except in reading
comprehension, which was used to divide the groups.

Materials and instruments


All the measures we administered as part of this study were objective in nature (i.e. we did
not employ self-report surveys of attitudes, opinions or perceptions, but rather measures of
achievement, some of which were standardised and norm referenced).

Reading and writing skills. Reading and writing proficiency were measured using Prueba
de Comprensión Lectora y Producción de Textos (CL-PT). This is a standardised reading
comprehension and text production test that evaluates different cognitive features of the
reading and writing processes (Medina & Gajardo, 2010). The reading component dissects
reading proficiency into seven dimensions, with their corresponding conceptual and oper-
ational definitions, as follows: (1) comprehension of textual structures, conceptualised as
the recognition of text structure and its function and operationalised as recognition of nar-
rative, informative, poetic, interactive, continuous, and discontinuous texts, from cues in
them and understanding of their function and purpose; (2) literal understanding, defined
as the retrieval of explicit information found in the text and operationalised as the retrieval
of details, main ideas, sequences, and relations, among others; (3) inferential understand-
ing, conceptually defined as the integration of textual information and the reader’s previous
knowledge and operationalised as the inference of details, main ideas, sequences, compar-
isons, cause–effect relationships, character details, the prediction of results, and contents,
among others; (4) critical understanding, defined as the construction of personal interpre-
tations, judgement, and recognition of the author’s intentions and operationalised as the
recognition of reality or fantasy appraisal, recognition of facts and opinions, appropriacy
© 2021 UKLA
GUTIERREZ DE BLUME, SOTO, RAMÍREZ CARMONA, RODRIGUEZ & PINO CASTILLO

judgements, value, convenience, and recognising the author’s purpose, among others; (5)
metacognitive understanding, defined as the awareness and regulation of the reader’s
own comprehension processes and operationalised as identifying and regulating one’s
own comprehension or absence thereof and using text structure to facilitate comprehen-
sion, among others; (6) reorganisation of information, understood as processing of infor-
mation according to the reader’s own comprehension schema with the purpose of
devising a comprehensive synthesis and operationalised as paraphrasing, building orga-
nisers, summarising, creating drawings, and other representations; and (7) code control,
which was conceptually defined as the recognition and control of words, conventional
signs, sentence structures, coordinating and subordinating connectors and deictic referents
and operationalised as recognising subject, predicate, and complements, recognising and
using connecting words, recognising and using words, and conventional symbols.
The writing component of the CL-PT test measures seven different indicators in writing
proficiency: (1) adaptation to the communicative situation, conceptualised as the perti-
nence of the text as it pertains to the parameters of language appropriateness, intended au-
dience, layout, support, and presentation, which contribute to the communicative intention,
and operationalised as communicative intention, relation with intended audience, and pre-
sentation and intelligibility; (2) ideas production, understood as the core of the message
and content of the text together with the details that enrich and deepen it and
operationalised as precision, variety, and topic comprehensibility; (3) voice or personal
seal, defined as manifestation of the sense given by authors, their creativity, perspective,
aesthetic appreciation, conviction, and commitment, and operationalised in writers’ crea-
tive capacity, reflexivity, and commitment to the topic; (4) word choice, conceptualised
as the use of language, richness, variety and accuracy of vocabulary, which moves the
reader and fosters reflection, and operationalised as the writer’s accuracy and range in vo-
cabulary use; (5) fluency and cohesion, conceptualised as the flow of sentences and rhythm
and the sound of linguistic patterns and operationalised in the use of cohesive devices and
logical connectors to enhance the fluidity of the text; (6) structure and organisation,
conceptualised as the use of logical and structural patterns conducive to the central mean-
ing of the text and operationalised as the accuracy in text structure; and (7) grammatical
conventions, conceptualised as the norms related to ordering of words in a sentence, gram-
matical correspondence and orthography and operationalised as the accurate use of syntac-
tic, morphological, and graphemic conventions.
The test presents a series of reading and writing tasks that depend on a common text. For
example, the first text presented in the test is a narrative text entitled ‘The businessman’.
The test includes a series of multiple-choice questions that assess comprehension of the
text, and then the student is asked to write a story assuming that the reader has the role
of the ‘businessman’ within the narrative. We decided to use the results of this test as a gen-
eral measure of performance in reading comprehension and writing because it is a
standardised and validated test in the Chilean school population.
Medina and Gajardo (2010) provided validity evidence for the instrument. An inter-rater
reliability of 0.80 was reported on all items of the test except for three, which were
redefined and retested for reliability. Validity of the instrument was assessed using expert
judges for its construction, and its criterion-related validity was confirmed using a previ-
ously validated test, Pruebas de Comprensión Lectora de Complejidad Lingüística
Progresiva (Alliende et al., 1987).
Each correct item was scored with a point, and the overall score of the test is the simple
summation of the correct answers. Scores were linearly transformed to percentiles (0–100)
© 2021 UKLA
READING, WRITING, AND MENTAL REPRESENTATION

to ease interpretation and more readily compare results with other measures. Internal con-
sistency reliability coefficient, KR-20, for the sample on this measure was 0.713.

Narrative and informative writing test. We created a computer-based writing test based on
the tasks and evaluation criteria devised by the Chilean Education Quality Agency
(Agencia de Calidad de Educación) in the creation of a high-stakes writing test called Ed-
ucation Quality Measurement System (Sistema de Medición de la Calidad de la Educación
[SIMCE]). This writing test comprised two writing tasks.
The first is a narrative writing test in which instructions given to students were to write a
story for a literary context in their school with the title ‘And it/he/she was never the same
again’. The other was an informative writing test in which instructions for students were to
write a piece of news related to a means of transport (and the text was accompanied by pic-
tures of these).
A rubric was employed (based on the SIMCE writing evaluation guideline used by the
Ministry of Education) to obtain the scores in the observation guideline based on SIMCE
writing. This allows 2 maximum points in the communicative purpose criterion, 3 points
for the textual organisation criterion and 4 points to the criteria of development of ideas
and coherence. In addition, the spelling criterion was added, which evaluated spellings that
were correct in terms of the use of the correct letters, punctuation marks and accents of stu-
dents’ written productions. For this, a maximum score of 3 points was awarded to each
one. In each of the criteria, the minimum score was 1 point, which signified poor perfor-
mance. For a more objective analysis of the results, three experts separately evaluated each
of the responses of the participants, jointly analysing the cases in which there was a signif-
icant difference between the evaluations of writing.

Trunajod. Our team has already developed and validated a prototype (Trunajod) with the
capacity to analyse texts automatically and obtain various indices of textual complexity
(e.g. sentence length, clause length, lexical diversity, lexical density, lexical frequency,
overlap between word and sentences, different type of connectors, etc.). In the new devel-
oped version of Trunajod, we enhanced the capabilities of the tool using artificial intelli-
gence, moving beyond word and sentence, analysing information about linguistic
features underlying the local and global cohesion of the texts as well as their rhetorical
structure. Those different microlinguistics features can be captured in both ways to analyse
the complexity of the texts and to determine the characteristics of the written composition
of readers (see Soto et al., 2019; Soto et al., 2020, for a more comprehensive explanation of
Trunajod).
In this study, both the expert evaluation with a rubric and the traits found by Trunajod
contribute to the main analyses, as Trunajod automatically evaluates aspects of writing that
are difficult to capture manually.

Procedures
The CL-PT test and the text production observation guideline were administered in two
sessions, one for each instrument, and the approximate time of each session was
90 minutes. CL-PT, as a paper-and-pencil test, was applied in the classroom of each group
of students, in which the participants were able to answer questions related to both reading
comprehension and text production (CL-PT reading, CL-PT writing). The application of
© 2021 UKLA
GUTIERREZ DE BLUME, SOTO, RAMÍREZ CARMONA, RODRIGUEZ & PINO CASTILLO

the text production observation guideline, based on SIMCE, was administered in online
format through Google Forms in the computer room of each school. Students developed
two text productions, one of an informative type and the other of a narrative type.
Each of the instruments was evaluated according to its respective rubric, in the case of
CL-PT, with the rubric provided by the application manual, and in the case of the text pro-
duction observation guideline, based on SIMCE, with the standardised rubric correspond-
ing to the SIMCE test. Both evaluations were rigorously scored together by three of the
members of the research team to avoid biases that could exist with the interpretation of
the instruments.
Finally, the texts produced by the students in the SIMCE-based text production observa-
tion guideline were also analysed by Trunajod to identify the differences within each of the
productions and, likewise, the linguistic indices of each of them.

Data analysis
Data were evaluated for univariate normality using skewness and kurtosis values and his-
tograms with normal curve overlay and multivariate normality via a normalised estimate
of the linear combination of dependent variables (Tabachnick & Fidell, 2013). All vari-
ables approximated univariate and multivariate normality across both proficient and poor
readers. No cases were classified as outliers through box-and-whisker plots by group
(univariate) and via review of Cook’s D and Mahalanobis distance (multivariate), and thus,
all 105 cases were retained for analysis. There were no missing data, as all participants
completed all data points. Other assumptions such as homoscedasticity, homogeneity of
variance (univariate) and homogeneity of variance–covariance matrices (multivariate),
and lack of multicollinearity were also met. Therefore, we proceeded with the planned
analyses without making any adjustments to the data. The Bonferroni adjustment to statis-
tical significance was employed to control familywise Type I error rate inflation. All data
were analysed via IBM’s Statistical Package for the Social Sciences (SPSS) 23.
Descriptive statistics and bivariate correlation coefficients were requested for all vari-
ables. For the purposes of data analyses, we conducted a median split procedure to deter-
mine proficient readers (reading comprehension score ≥ 46) and poor readers (reading
comprehension score ≤ 45), as reading comprehension performance was expressed as
whole number percentiles (Min. = 2; Max. = 79; Med. = 46). Even though the median split
approach remains controversial among researchers, we employed it because we wanted to
facilitate meaningful interpretation of findings and because our intent was to continue to
understand the distinct profiles of more successful versus less successful readers. This ra-
tionale is supported by previous work on the use of transforming continuous variables into
meaningful categories (DeCoster et al., 2011) and the use of blocking variables and covar-
iates in research (Tabachnick & Fidell, 2013).
We answered the first research question by conducting a one-way multivariate analysis
of variance (MANOVA), in which reading proficiency (proficient, poor) served as the
between-subjects factor and narrative texts score, SIMCE informative test score, and
CL-PT writing score served as the dependent variables. The second research question
was answered by conducting a series of ordinary least squares (OLS) linear regressions.
Given the large number of Trunajod micro-indices, we opted to employ stepwise regres-
sion in our modelling approach, which we believe is a better choice to backward elimina-
tion and forward selection. In these stepwise regression models, Trunajod indices served as
© 2021 UKLA
READING, WRITING, AND MENTAL REPRESENTATION

predictors, and narrative and informative writing test performance served as the criterion
separately. We conducted a separate stepwise regression for proficient and poor readers
to ascertain whether distinct predictive patterns existed among the two groups with respect
to the reading comprehension skills we evaluated in the present study.
Effect sizes for the MANOVA were reported as η2, and those for the regression analyses
were reported as the R2. Cohen (1988) provided the following interpretive guidelines for
η2: 0.010–0.059 as small; 0.060–0.139 as moderate; and ≥0.140 as strong. For R2, these
values were: 0.010–0.249 as small; 0.250–0.499 as moderate; and ≥0.500 as strong.

Results

Descriptive statistics for the relevant variables for the sample and each reading proficiency
group are displayed in Table 1. Bivariate zero-order correlations for each group and the
sample are displayed in Tables 2 and 3, respectively.
Descriptive and correlational data show the clear distinction between proficient readers
and poor readers on the variables of interest to this study. More specifically, the means
of the proficient reading group are higher than the poor reading group, and the correlation
coefficients are also higher for the proficient reading group than the poor reading group.
RQ1: Differences in writing metrics as a function of reading proficiency

Results of the one-way MANOVA revealed statistically significant findings regarding


the linear combination of dependent variables, multivariate F12,196 = 4.85, P = 0.01,
η2 = 0.101. Review of the univariate findings, however, demonstrated that statistically sig-
nificant differences existed only in CL-PT writing, F1,103 = 5.14, P = 0.007, η2 = 0.092,
and informative writing test, F1,103 = 4.57, P = 0.01, η2 = 0.081, but not for narrative writ-
ing test, P = 0.15. Across the two significant dependent variables, general writing and in-
formative writing performance, the proficient reading group significantly outperformed the
poor reading group (see Table 1 for descriptive statistics).

RQ2: Differences in predictive patterns by reading proficiency regarding narrative and in-
formative writing

Results of the models with narrative writing performance as the criterion were statisti-
cally significant for proficient readers, F9,47 = 13.96, P < 0.001, R2 = 0.728, and poor
readers, F4,43 = 7.50, P < 0.001, R2 = 0.411. Evidently, there were more significant

Table 1. Descriptive statistics for the sample and for proficient and poor readers

Sample Proficient (n = 57) Poor (n = 48)


Variable M SD M SD M SD
CL-PT reading 45.13 14.61 55.95 7.09 39.24 10.10
Narrative writing test 76.43 14.19 77.88 15.12 74.69 12.94
Informative writing Test 52.99 13.60 55.86 13.46 49.58 13.09
CL-PT writing 28.49 23.96 31.30 25.48 25.15 21.83
N = 105.

© 2021 UKLA
GUTIERREZ DE BLUME, SOTO, RAMÍREZ CARMONA, RODRIGUEZ & PINO CASTILLO

Table 2. Zero-order correlation matrix of measures for the sample

Variable 1 2 3 4
1. CL-PT reading - 0.20* 0.23** 0.29**
2. CL-PT writing - 0.40** 0.32**
3. Narrative writing test - 0.49**
4. Informative writing test -
N = 105.
*
P < 0.05.
**
P < 0.01.

Table 3. Zero-order correlation matrix of measures for proficient and poor readers

Variable 1 2 3 4
1. CL-PT reading - 0.43** 0.18 0.14
2. CL-PT writing 0.53** - 0.33* 0.26*
3. Narrative writing test 0.51** 0.43** - 0.52**
4. Informative writing test 0.51** 0.33** 0.44** -
Correlations above the diagonal are for poor readers (n = 48), and those below the diagonal are for proficient
readers (n = 57).
N = 105.
*
P < 0.05.
**
P < 0.01.

predictors accounting for greater explained variance for the proficient reading group than
for the poor reading group. For the models with informative writing performance as the cri-
terion, results were significant for proficient readers, F5,51 = 8.12, P < 0.001, R2 = 0.443,
and poor readers, F4,43 = 8.03, P < 0.001, R2 = 0.428. As with narrative writing perfor-
mance, the greatest explained variance by the predictors was found in the model for the
proficient reading group, albeit the model for poor readers remained consistent and the
model for the proficient reading group declined regarding the explained variance of the
model. Table 4 presents the models for narrative and informative writing performance
for each reading proficiency group, respectively.

Discussion

The purpose of the present study was to investigate the effect of reading comprehension
level (proficient, poor) on students’ general writing performance on specific narrative
and informative writing tasks. Moreover, we investigated whether microlinguistics metrics
(measured by Trunajod) of informative and narrative writing predicted writing perfor-
mance in poor and proficient readers.
Our results provide partial support for our first research hypothesis because reading com-
prehension skill level only significantly impacted performance on general writing and in
informative writing, but not in narrative writing performance. This finding is congruent
© 2021 UKLA
READING, WRITING, AND MENTAL REPRESENTATION

Table 4. Stepwise regression results of Trunajod linguistic indices for narrative and informative writing tests
by reading proficiency

Predictor B+ (CI95%) β t P
Narrative writing test
Proficient readers
Cause_dm 5.31 ( 6.95, 3.67) 0.60 6.50 0.001*
Noun density 201.67 ( 295.00, 108.34) 0.39 4.35 0.001*
Prom. long. syllable words 59.05 (23.90, 94.19) 0.29 3.38 0.001*
Frustration 87.76 ( 127.25, 48.27) 0.37 4.47 0.001*
Polysemic_dm 9.25 ( 13.38, 5.13) 0.38 4.51 0.001*
Function_ttr 62.04 (14.04, 110.04) 0.23 2.60 0.01*
Adj. density 262.01 ( 401.48, 122.55) 0.37 3.78 0.001*
Nom. sentence density 21.48 (5.04, 37.92) 0.25 2.63 0.01*
Argument_ttr 40.37 (7.14, 73.60) 0.27 2.44 0.02
Poor readers
Verb_ttr 55.69 (25.69, 85.68) 0.49 3.74 0.001*
Fear 142.46 ( 253.03, 31.89) 0.31 2.60 0.01*
Density 1,2 people 87.90 (2.16, 173.64) 0.25 2.07 0.04
Prom. long. letter words 0.26 ( 0.52, 0.01) 0.26 2.04 0.04
Informative writing test
Proficient readers
Fear 54.91 (22.25, 87.56) 0.37 3.38 0.001*
Lexical overlap 2.49 (.61, 4.37) 0.29 2.66 0.01*
Equality_dm 15.48 (6.43, 24.53) 0.37 3.44 0.001*
Imageability 8.67 ( 15.52, 1.82) 0.27 2.54 0.01*
Density ADV/ADP 55.35 ( 102.28, 8.42) 0.25 2.37 0.02
Poor readers
Content_ttr 39.06 (14.12, 63.69) 0.38 3.16 0.003*
Disim. PoS per sentence 20.10 (8.53, 31.66) 0.43 3.50 0.001*
Nom. sentence density 27.70 (9.39, 46.02) 0.36 3.05 0.004*
Noun density 63.82 (10.86, 116.77) 0.29 2.43 0.01*
Note: The bold entries represents statistically significant predictors.
N = 105.
+
B = unstandardized regression coefficients and their 95% confidence interval (CI95%); β = standardised regres-
sion coefficients.
*
P ≤ 0.01.

with research on metacomprehension in reading that concludes that students who are pro-
ficient at reading employ more sophisticated learning strategies that focus on depth of un-
derstanding compared with poor readers, who tend to use strategies focused on a
superficial grasp of the information in texts (Soto et al., 2019; Soto et al., 2020). Expository
texts communicate objective information through a less predictable structure, the
© 2021 UKLA
GUTIERREZ DE BLUME, SOTO, RAMÍREZ CARMONA, RODRIGUEZ & PINO CASTILLO

conceptual density is greater, and, often, the information they contain is far from the
reader’s prior knowledge (Saenz & Fuchs, 2002). All these components deliver a higher
degree of complexity. In contrast, narrative texts follow a consistent hierarchy (story gram-
mar), are situated in a temporal logic, and have a frequently used vocabulary (Dymock &
Nicholson, 2010; Hall, 2004). We postulate that the use of more complex metacognitive
and learning strategies expected in the group of proficient readers favours performance
in the writing of expository texts. On the other hand, the nature of narrative thinking helps
the poor reader to generate an adequate level of comprehension comparable to the profi-
cient reader’s, explaining the lack of significant impact in narrative writing.
Regarding our second research question, we found support for our hypothesis here as
well. The predictive patterns differed between proficient and poor readers. Across both nar-
rative and informative writing, the models for the proficient readers captured more
explained variance than those for the poor reading group, albeit the explained variance
for the proficient readers was higher for narrative than for the informative writing. Regard-
ing narrative writing, only two Trunajod metrics were significant predictors for poor
readers. For proficient readers, there were eight significant predictors of narrative writing.
This suggests that proficiency in reading exerts a powerful influence on the linguistic fea-
tures that predict performance on narrative writing. Interestingly, the diversity of verb
lemmas (verb_ttr) and fear of the task itself (fear) were the only significant predictors of
narrative writing performance for poor readers. Conversely, for proficient readers, apart
from the mean length of words in terms of number of syllables (prom. long. syllable), di-
versity of functional words used (function_ttr) and density of noun phrases (nom. sentence
density), which were positive predictors, the remaining five predictors negatively predicted
performance on narrative writing. Proficient readers use different types of cues in their nar-
rative writing that are a combination of lexical, such as noun and adjective density in the
text (noun density, adj. density), connections such as connecting words and use of polyse-
mous words (cause_dm, polysemic_dm) and emotional factors (frustration words). On the
other hand, the reading comprehension performance of poor readers is linked just to the
presence of the diversity of the action words on the sentences (verb_ttr) and words related
to fear (fear). The nature of these differences explains the better prediction from writing to
reading for proficient readers using more sophisticated cues to interpret the text and link
ideas.
With respect to informative writing performance, the regression model for poor readers
and proficient readers yielded four significant predictors. Nevertheless, the pattern of pre-
dictors is completely different for each group, with no overlap, which is a monumental
finding that supports the novel theoretical claims proposed in the present study. Whereas
diversity of content words (content_ttr), variety in parts of speech used in different
sentences (disim. PoS per sentence), noun phrase density (nom. sentence density) and com-
mon noun density (noun density) were significant predictors for poor readers, emotional
factors (fear), discourse markers signalling equality of information (equality_dm), referen-
tial cohesion markers such as argument overlap (lexical overlap) and use of words that
readily lend themselves to be visualised (imageability) were significant predictors for pro-
ficient readers.
In fact, it is particularly interesting, in comparing proficient readers and poor readers,
that there is no overlap in the writing cues they employ in their reading performance.
Whereas for proficient readers the relations of connections between the words were impor-
tant, for poor readers the diversity and density of the word and sentence was more predic-
tive of their reading performance.
© 2021 UKLA
READING, WRITING, AND MENTAL REPRESENTATION

Contribution to understanding the relation between reading and writing


Our results support the hypothesis of the relation between reading and writing and that it
becomes stronger as individuals have a better mental representation of the text. This be-
comes evident for the relationships between reading and writing about different discursive
genres. As the reader becomes more proficient, that relation becomes more intense and sta-
ble. The existence of different predictive patterns in the writing cues they employ according
to their reading performance elucidates some interesting ways regarding what dimensions
of writing are linked to reading in each different type of reader.
The proficient readers base their writing performance on linguistic markers about the
connection between words and sentences while activating emotional elements (e.g. fear
and frustration). This shows that proficient readers also generate a deeper mental represen-
tation during writing tasks, which results in flexible and integrated writing patterns. This
representation is based on the internal relations of the text (inferences) and on the activation
of the individual’s own elements (emotional factor), as suggested by the C–I model. For
their part, poor readers base their writing performance on the nature of the words/sentences
(e.g. density and diversity), which allude to a more textual dimension (text-base), without
connection between the components of their composition. A similar trend has been found
in previous research investigating only the reading process (Language and Reading
Research Consortium & Logan, 2016; McNamara, 2004; Soto et al., 2019; Soto
et al., 2020).
Based on our results, we propose that the two different performance patterns, generated
from the interaction between performance in reading comprehension and the linguistic in-
dices of written production, depend directly on the quality of the mental representation
elaborated in both types of operations, reading and writing. Proficient readers who, while
reading, establish a high-quality mental representation through the generation of inferences
and the appropriate application of metacognitive strategies will base their writing on the
combination of connecting cues between concepts and the use of words with emotional rel-
evance. In turn, these elements account for the elaboration of a rich mental representation
prior to and during writing. Presumably, as the mental representation is enriched, the rela-
tions between comprehension and production become stronger and more stable.

Avenues for future research


Although the present study provided tentative evidence for a new theoretical explanation
and direction regarding the influence between reading, writing, and the role of mental rep-
resentations, future research should replicate this study with larger sample sizes to ensure
the stability and precision of our findings in multiple samples of this population. In addi-
tion, future research should employ latent variable modelling approaches to examine the
explanatory power of theoretical principles we propose. Researchers should also consider
employing multiple reading and writing tasks in future studies to examine the complex dy-
namic relation more closely between reading and writing and how they influence mental
representations during learning episodes in different domains. It would also be worthwhile
for researchers to employ reading and writing tasks on the same topic to examine whether
the theoretical claims we make regarding mental representation bear out. Finally, the find-
ings of the present investigation have bearing on applied research as well. Future research
should develop more specific, finer grained, and comprehensive educational interventions
that are tailored to improve both reading and writing processes.
© 2021 UKLA
GUTIERREZ DE BLUME, SOTO, RAMÍREZ CARMONA, RODRIGUEZ & PINO CASTILLO

Methodological reflections and limitations


First, our research design was non-experimental in nature, thereby limiting the inferences
and conclusions we can draw from our data, as we cannot infer any causal effects from
our findings. Second, our sample size was relatively small. Finally, our OLS regression ap-
proach (i.e. stepwise) is a mathematical one. We selected this specific approach for our re-
gression models because Trunajod measures over 50 unique indices of readability. As one
can imagine, it would be unwieldy and unfeasible to include that many predictors, some of
which may be collinear, in an OLS regression using a standard/simultaneous or hierarchi-
cal approach. Thus, to avoid model misspecification, and given the relatively small sample
size, we selected the stepwise regression approach as the best, most feasible option. It may
be that this modelling approach neglected to capture Trunajod metrics that were significant
predictors, however marginally. Finally, the use of a single grade level also limits the gen-
eralizability of our findings.
Despite these limitations, we wish to highlight some of the strengths of our study. First,
we employed objective measures in our study rather than subjective ones such as
self-report surveys. This strengthens the claims and inferences we draw from the findings.
Second, the effect sizes produced by the analyses are robust, including moderate to large
effects, thereby minimising the likelihood of the commission of Type II errors in inferential
decision-making. Finally, no study to date, to our knowledge, examines reading compre-
hension the way we have in this study. Therefore, we believe our study is a worthwhile
contribution to the extant research literature on reading comprehension as a function of
reading proficiency.

Conclusion

Results of the present study revealed that proficient readers outperformed poor readers on
objective measures of general writing and informative/narrative writing. Additionally, pre-
dictive models demonstrated that proficient readers relied more on deeper aspects of read-
ing and writing such as inferential skills, whereas poor readers tended to focus on
superficial aspects of texts, or what Kintsch (1988, 1998) referred to as text base, and ap-
peared to perform better in reading and writing tasks related to narrative compared to infor-
mative writing. These results form the foundation for a novel theoretical perspective
regarding the relation between reading, writing, and mental representations that has never
been proposed in the literature on these topics. More specifically, the cues that proficient
readers invoke during reading and writing appear to be more sophisticated and better suited
to the complex nature of informative writing and that poor readers tend to employ more su-
perficial text-base cues that are better suited to narrative discourse. By borrowing tenets of
Kintsch’s (1988) construction–integration model and Otero’s (2002) regulation model, the
present investigation tentatively shows how comprehension improves regulation during
reading and writing, and thus, enhances learning.

Source of Funding

Award Title ‘Desarrollo de una herramienta computacional para la evaluación automática


de textos en el sistema escolar chileno’ (Development of a computational tool for the au-
tomatic evaluation of texts in the Chilean school system; Award No. Fondef IT17I0051).
© 2021 UKLA
READING, WRITING, AND MENTAL REPRESENTATION

Conflicts of Interest

None of the authors have any conflicts of interest to report.

Data Availability Statement


Data associated with this research are available via the Open Science Framework (OSF) by
clicking here.

References

Abbott, R.D., Berninger, V.W. & Fayol, M. (2010). Longitudinal relationships of levels of language in writing and
between writing and reading in grades 1 to 7. Journal of Educational Psychology, 102(2), 281–298.
Allal, L. (2000). Metacognitive regulation of writing in the classroom. In G. Rijlaarsdam & E. Espéret (Series
Eds.) & A. Camps & M. Milian (Vol. Eds.), Studies in writing: Metalinguistic activity in learning to write
(pp. 145–167). Amsterdam University Press.
Alliende, F., Condemarín, M. & Milicic, N. (1987). Pruebas de comprensión lectora de complejidad lingüística
progresiva (CLP): Formas paralelas. Ediciones Universidad Católica.
Arya, D.J., Hiebert, E.H. & Pearson, P.D. (2011). The effects of syntactic and lexical complexity on the compre-
hension of elementary science texts. International Electronic Journal of Elementary Education, 4(1), 107–125.
Beal, C.R., Garrod, A.C. & Bonitatibus, G.J. (1990). Fostering children’s revision skills through training in com-
prehension monitoring. Journal of Educational Psychology, 82(2), 275–280.
Beauvais, C., Olive, T. & Passerault, J.-M. (2011). Why are some texts good and others not? Relationship between
text quality and management of the writing processes. Journal of Educational Psychology, 103(2), 415–428.
Bereiter, C. & Scardamalia, M. (1987). The psychology of written composition. Lawrence Erlbaum Associates Inc.
Berninger, V.W., Abbott, R.D., Abbott, S.P., Graham, S. & Richards, T. (2002). Writing and reading: Connections
between language by hand and language by eye. Journal of Learning Disabilities, 35(1), 39–56.
Buz, E., Tanenhaus, M.K. & Jaeger, T.F. (2016). Dynamically adapted context-specific hyper-articulation: Feed-
back from interlocutors affects speakers’ subsequent pronunciations. Journal of Memory and Language, 89,
68–86.
Cain, K., Oakhill, J. & Lemmon, K. (2004). Individual differences in the inference of word meanings from con-
text: The influence of reading comprehension, vocabulary knowledge, and memory capacity. Journal of Educa-
tional Psychology, 96(4), 671–681.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Lawrence Erlbaum & Associates.
de Beaugrande, R. & Dressler, W. (1981). Introduction to text linguistics. Longman.
DeCoster, J., Gallucci, M. & Iselin, A.M. (2011). Best practices for using median splits, artificial categorization,
and their continuous alternatives. Journal of Experimental Psychopathology, 2(2), 197–209.
Dunlosky, J. & Lipko, A.R. (2007). Metacomprehension: A brief history and how to improve its accuracy. Current
Directions in Psychological Science, 16(4), 228–232.
Dymock, S. & Nicholson, T. (2010). "High 5!" Strategies to enhance comprehension of expository text. The Read-
ing Teacher, 64(3), 166–178.
Ferrari, M., Bouffard, T. & Rainville, L. (1998). What makes a good writer? Differences in good and poor writers’
self-regulation of writing. Instructional Science, 26(6), 473–488.
Flower, L. & Hayes, J. (1984). Images, plans and prose: The representation of meaning in writing. Written Com-
munication, 1(1), 120–160.
Fricke, M., Kroll, J.F. & Dussias, P.E. (2016). Phonetic variation in bilingual speech: A lens for studying the
production-comprehension link. Journal of Memory and Language, 89, 110–137.
Graham, S. & Hebert, M. (2011). Writing to read: A meta-analysis of the impact of writing and writing instruction
on reading. Harvard Educational Review, 81(4), 710–744.
Guzzardo Tamargo, R.E., Valdés Kroff, J.R. & Dussias, P.E. (2016). Examining the relationship between compre-
hension and production processes in code-switched language. Journal of Memory and Language, 89, 138–161.
Hall, L.A. (2004). Comprehending expository text: Promising strategies for struggling readers and students with
reading disabilities? Reading Research and Instruction, 44(2), 75–95.

© 2021 UKLA
GUTIERREZ DE BLUME, SOTO, RAMÍREZ CARMONA, RODRIGUEZ & PINO CASTILLO

Harris, K.R. & Graham, S. (1992). Self-regulated strategy development: A part of the writing process. In M.
Pressley, K.R. Harris & J.T. Guthrie (Eds.), Promoting academic competence and literacy in school,
(pp. 277–309). Academic Press.
Hayes, J.R. & Flower, L.S. (1980). Identifying the organization of writing processes. In L.W. Gregg & E.R.
Steinberg (Eds.), Cognitive processes in writing, (pp. 3–30). Erlbaum.
Hiebert, E.H. & Pearson, P.D. (2014). Understanding text complexity. The Elementary School Journal, 115(2),
153–160.
Hsiao, Y. & MacDonald, M.C. (2016). Production predicts comprehension: Animacy effects in mandarin relative
clause processing. Journal of Memory and Language, 89, 87–109.
Indefrey, P. (2018). The relationship between syntactic production and comprehension. In S.-A. Rueschemeyer &
G. Gaskell (Eds.), The Oxford handbook of psycholinguistics, (pp. 483–505). Oxford University Press.
Karlsson, J., van den Broek, P., Helder, A., Hickendorff, M., Koornneef, A. & van Leijenhorst, L. (2018). Profiles
of young readers: Evidence from thinking aloud while reading narrative and expository texts. Learning and In-
dividual Differences, 67, 105–116.
Kim, Y.S.G. (2020). Toward integrative reading science: The direct and indirect effects model of reading. Journal
of Learning Disabilities, 53(6), 469–491.
Kim, Y.S.G., Petscher, Y., Wanzek, J. & Al Otaiba, S. (2018). Relations between reading and writing: A longi-
tudinal examination from grades 3 to 6. Reading and Writing, 31(7), 1591–1618.
Kintsch, W. (1988). The use of knowledge in discourse processing: A construction-integration model. Psycholog-
ical Review, 95, 163–182.
Kintsch, W. (1998). Comprehension: A paradigm for cognition. Cambridge University Press.
Kintsch, W. & Rawson, K.A. (2007). Comprehension. In M.J. Snowling & C. Hulme (Eds.), The science of read-
ing, (pp. 209–226). Blackwell.
Kittredge, A.K. & Dell, G.S. (2016). Learning to speak by listening: Transfer of phonotactics from perception to
production. Journal of Memory and Language, 89, 8–22.
Language and Reading Research Consortium & Logan, J. (2016). Pressure points in reading comprehension: A
quantile multiple regression analysis. Journal of Educational Psychology, 109(4), 451–464.
León, J.A. & Peñalba, G.E. (2002). Understanding causality and temporal sequence in scientific discourse. In J.
Otero, J.A. León & A.C. Graesser (Eds.), The psychology of science text comprehension, (pp. 155–178).
Lawrence Erlbaum & Associates.
Long, D.L. & Chong, J.L. (2001). Comprehension skill and global coherence: A paradoxical picture of poor
comprehenders’ abilities. Journal of Experimental Psychology: Learning, Memory, and Cognition, 27(6),
1424–1429.
McNamara, D.S. (2004). SERT: Self-explanation reading training. Discourse Processes, 38(1), 1–30.
McNamara, D.S. (2017). Self-explanation and reading strategy training (SERT) improve low-knowledge students’
science course performance. Discourse Processes, 54(7), 479–492.
McNamara, D.S. & Magliano, J.P. (2009). Towards a comprehensive model of comprehension. In B. Ross (Ed.),
The psychology of learning and motivation, (pp. 297–284). Elsevier Science.
Medina, A. & Gajardo, A. (2010). Pruebas de comprensión lectora y producción de textos (CL-PT). Ediciones
Universidad Católica de Chile.
Meyer, A.S., Huettig, F. & Levelt, W.J.M. (2016). Same, different, or closely related: What is the relationship be-
tween language production and comprehension? Journal of Memory and Language, 89, 1–7.
Otero, J. (2002). Noticing and fixing difficulties in understanding science texts. In J. Otero, J.A. León & A. Graesser
(Eds.), The psychology of science text comprehension, (pp. 281–307). Lawrence Erlbaum & Associates.
Saenz, L.M. & Fuchs, L.S. (2002). Examining the reading difficulty of secondary students with learning disabil-
ities: Expository versus narrative text. Remedial and Special Education, 23(1), 31–41.
Scardamalia, M. & Bereiter, Y.C. (1992). Dos modelos explicativos de composición escrita. Infancia y
Aprendizaje, 58, 43–64.
Soto, C., Gutierrez de Blume, A.P., Carrasco Bernal, M.A. & Contreras Castro, M.A. (2020). The role of
metacognitive cues on the comprehension of proficient and poor readers. Journal of Research in Reading, 43
(3), 272–289.
Soto, C., Gutierrez de Blume, A.P., Jacovina, M., McNamara, D., Benson, N. & Riffo, B. (2019). Reading com-
prehension and metacognition: The importance of inferential skills. Cogent Education, 6, 1–20.
Soto, C., Gutierrez de Blume, A.P., Rodriguez, M.F., Asún, R., Figueroa, M. & Serrano, M. (2019). Impact of
bridging strategy and feeling of knowing judgments on reading comprehension using COMPRENDE: An ed-
ucational technology. TechTrends, 63(5), 570–582.

© 2021 UKLA
READING, WRITING, AND MENTAL REPRESENTATION

Spivey, N.N. (1990). Transforming texts: Constructive processes in reading and writing. Written Communication,
7(2), 256–287.
Tabachnick, B.G. & Fidell, L.S. (2013). Using multivariate statistics. (6th edn). Pearson.
van den Broek, P., Bohn-Gettler, C., Kendeou, P., Carlson, S. & White, M.J. (2011). When a reader meets a text:
The role of standards of coherence in reading comprehension. In M.T. McCrudden, J. Magliano & G. Schraw
(Eds.), Relevance instructions and goal-focusing in text learning, (pp. 123–140). Information Age Publishing.
Zamuner, T.S., Morin-Lessard, E., Strahm, S. & Page, M.P.A. (2016). Spoken word recognition of novel words,
either produced or only heard during learning. Journal of Memory and Language, 89, 55–67.

Antonio P. Gutierrez de Blume, Ph.D., is currently an associate professor of research at Georgia


Southern University, where he teaches quantitative research methods and statistics. He is interested
in researching metacognition under the theory of self-regulated learning. More specifically, he is in-
terested in how learners monitor their comprehension during learning episodes. His programme of
research includes examining the effects of dispositional characteristics (e.g. various aspects of moti-
vation) and learning strategy training on learners’ calibration (accuracy and bias), confidence in per-
formance judgements and performance as well as investigating the latent dimensions of calibration to
improve its measurement.

Christian Soto, Ph.D., is currently investigating metacomprehension in reading and how this is re-
lated to other metrics of metacognitive monitoring, most notably learners’ ability to accurately report
what they know or do not know about a topic. We are also examining whether intelligent tutoring
systems based on artificial intelligence can effectively and efficiently train reading comprehension
skills in children and adolescents. We believe these lines of inquiry are essential to inform not only
educational practice of teachers in classrooms but educational policy as well such as funding deci-
sions for school systems and educational research.

Camila Ramírez Carmona, B.S., is currently a practicing teacher in a middle school in Chile, where
she teaches Spanish.

Fernanda Rodriguez, M.S., is a professional in Dr Christian Soto’s lab. Her expertise is in reading
comprehension, especially among individuals with learning and intellectual disabilities.

Patricio Pino Castillo, B.S., is currently an English instructor at the Universidad Santo Tomás in
Chile. He is knowledgeable in pyscholinguistics.

Received 4 October 2020; revised version received 25 February 2021.

Address for correspondence: Antonio Gutierrez de Blume, Department of Curriculum,


Foundations, and Reading, Georgia Southern University, Statesboro, GA, USA
Email: agutierrez@georgiasouthern.edu

© 2021 UKLA

You might also like