Training Special Education Teachers to Implement Evidence-Based, Technology-Supported Spelling Instruction for Students with Dysorthographia

Abstract

Special education teachers often lack training to implement research-evaluated writing programs with fidelity, which contributes to insufficient instruction for students with disabilities. This study addresses a research gap: the limited documentation of implementation fidelity in French spelling interventions that integrate assistive technologies (ATs) for learners aged 9–12 with dysorthographia. Grounded in a theoretical foundation that coordinates alphabetic, orthographic, and morphographic processes within an explicit instruction sequence (explanation, modeling, guided practice, and independent application), the program aligned text-to-speech and word prediction with targeted spelling goals. Using a mixed-methods design, six elementary students participated in a single-case protocol with a transformative sequential design over 20 weeks. Four teachers received targeted training (theoretical + practical) and delivered explicit, individualized instruction during a 10-week intervention. Content analysis of teacher and researcher logs showed high, yet context-responsive, fidelity with variations by student profile, school context, and teacher. Converging quantitative and qualitative patterns suggest improvements in word-level accuracy/fluency and highlight training/coaching as a driver of fidelity. The discussion provides actionable implications for professional learning, school scheduling and dosage protection, and future research that multimodalizes fidelity evidence and instruments AT orchestration across the writing cycle.

1. Introduction

The introduction first allows us to situate the research context. Then, it presents the theoretical framework, analyzing the main concepts of this research.

1.1. Context

A growing body of research underscores that training special education teachers to enact evidence-based practices (EBPs) is pivotal for improving outcomes for students with disabilities; programs grounded in research, supported by explicit teacher learning structures and coaching, are most likely to yield measurable gains at the student level (Mohr et al., 2024; Flanagan et al., 2025). In French, lexical spelling is semi-opaque: phoneme–grapheme relations are many-to-many and interact with morphemic units within a multi-systemic orthography (phonographic, morphographic, and logographic), adding complexity to writing instruction (Catach, 2008). Similar challenges in other opaque orthographies, such as English, suggest that these issues are not language-specific (Department for Education, 2025; Education Endowment Foundation, 2024). Recent integrative reviews confirm that the multisystemic nature of French orthography continues to pose significant challenges for learners with dyslexia–dysorthographia, particularly in upper elementary grades where morphological and orthographic demands intensify (Chaix, 2025; Hall et al., 2023).

Providing effective writing remediation for 9–12-year-old students with dyslexia–dysorthographia, therefore, requires the close alignment of four elements: (1) priority lexical-spelling word production process; (2) individualized adjustments for learners who have not responded to handwriting-centric interventions; a coherent, research-based program architecture; and (4) the explicit integration of assistive technologies—namely text-to-speech and word prediction—to support these instructional aims (Chaix, 2025; Hall et al., 2023; Troia, 2014; Svensson et al., 2019). This need for coordinated, multi-component intervention is echoed in recent clinical and intervention syntheses, which emphasize that phonological, orthographic, and morphological processes must be explicitly integrated within instruction to produce durable gains (Chaix, 2025; Fontaine et al., 2025). Aligning these components increases the likelihood that a research-based program will contribute to improved written-word production in authentic school settings.

Despite this, achieving such alignment remains difficult. Few specialized, research-based, school-embedded interventions exist for this population; many existing programs were developed for handwriting and therefore lack integration of assistive technologies now common in students’ writing; effects reported in the literature vary widely, with dosage emerging as a consistent moderator; and there is limited empirical clarity on mechanisms of success/failure and on how to train and coach teachers to adjust implementation to learner profiles and contextual constraints (Fontaine, 2019; Fontaine et al., 2025; Hall et al., 2023; Galuschka et al., 2019; Rouse, 2023). This gap is compounded by persistent challenges in preparing teachers to deliver multi-component literacy interventions with fidelity, a difficulty repeatedly highlighted in recent implementation and professional-learning research (Mohr et al., 2024; Rouse, 2023). Recent reviews also note that, despite the widespread availability of assistive technologies in classrooms, very few intervention studies document how these tools are pedagogically orchestrated within explicit instruction, particularly in French (Fernández-Batanero et al., 2022; Harrar Eskinazi, 2023).

Within this landscape, implementation fidelity functions as a bridge between an innovative remedial program and its measurable effects: it encompasses adherence, dosage, and quality of delivery, while documenting necessary adaptations (C. Carroll et al., 2007). Programs that combine structured teacher training, coaching/behavior-skills training, and fidelity monitoring are thus both more interpretable and more likely to produce documented contributions to students’ written-word development (Fixsen et al., 2005). Meanwhile, assistive technologies such as text-to-speech and word prediction show potential for supporting word-level performance, yet under-utilization is common when teacher preparation is insufficient and program structure is weak (Fernández-Batanero et al., 2022). Their contributions are therefore conditional on teacher training and fidelity-attentive implementation.

Despite converging insights, research-based remedial programs for French written-word production that integrate AT and are implemented with documented fidelity by trained specialized teachers in authentic school settings remain scarce; the literature thus provides limited evidence about what such a program contributes under conditions of structured training, coaching, and fidelity monitoring—particularly for learners aged 9–12 with dyslexia–dysorthographia (Fontaine, 2019; Fontaine et al., 2021). Addressing this gap requires examining a specific, research-based remedial program—implemented by trained specialized teachers—that integrates assistive technologies, namely speech synthesis and word prediction, and assessing its effects on French written-word production while evaluating implementation fidelity to determine the degree of delivery and to document in situ adaptations (Svensson et al., 2019; C. Carroll et al., 2007; Fixsen et al., 2005; Fontaine, 2019; Galuschka et al., 2019; Rouse, 2023). As several authors have recently argued, documenting both the effects of interventions and the fidelity conditions under which they are delivered is essential for advancing evidence-based practice in writing instruction (Hall et al., 2023; U.S. Department of Education, Institute of Education Sciences, What Works Clearinghouse, 2022). This issue raises the general research question: What are the contributions of a research-based remedial education program that incorporates technological aids and is faithfully implemented by trained specialized teachers to the development of written French word production in students aged 9 to 12 with dyslexia and dysorthography?

1.2. Theoretical Framework

This manuscript addresses the qualitative analysis of two core elements of implementation fidelity: (a) the extent to which the program was implemented as intended, including any adjustments made during delivery; and (b) the challenges related to the training and support of specialized teachers. The subsections below synthesize the theoretical and empirical bases that ground these focal points and lead to the study’s sub-questions.

1.2.1. French Written-Word Production: Process and Orthographic Depth

French lexical spelling is opaque: phoneme–grapheme relations are many-to-many and interact with morphemic units within a multi-systemic orthography (phonographic, morphographic, and logographic), which increases demands on instruction and assessment (Catach, 2008). These properties require coordinated alphabetic plausibility checks, orthographic pattern selection (including context-sensitive choices), and morphographic processing to achieve accurate and fluent written-word production (Galuschka et al., 2019).

Several theoretical models explain this development. Seymour’s developmental model (2008), validated in French, is considered the most relevant for explaining this process, particularly among the students targeted. It emphasizes the main processes that interact in text writing, unlike other models that focus on expert procedures (e.g., Barry, 1994), detailed procedures (e.g., Daigle et al., 2018), or stage-based procedures (Frith, 1985). Seymour (2007) identifies four fundamental processes that contribute to expanding spelling vocabulary and producing written words with accuracy (correctness) and fluency (speed and accuracy) (see

Table 1. Processes that contribute to spelling, according to Seymour (2007).

Type of Processing	Description	Examples
Logographic	Writing very frequent words	Maman (mom)
Alphabetical	Writing regular words by making the phoneme-grapheme association and checking the phonological plausibility of the written words	Chaton (kitten)
Orthographic	Writing words with irregular phoneme-grapheme associations, considering the position of graphemes	Choosing the grapheme “em” from several possibilities (e.g., en, em, am) to write embrasser (kiss)
Morphographic	Writing words that are derived from or contain silent letters	“in + visible” = invisible dent (tooth)

).

For the rehabilitation of these processes, the schematic representation of writing processes (Laplante, 2011) offers a high level of precision on the interactive dynamics that connect them in text production. This author identifies three key processes, based on influential theoretical models of writing acquisition (e.g., Berninger et al., 2008a; Hayes & Flower, 1980) (see

Table 2. Key processes of writing acquisition, according to Laplante (2011).

Type of Process	Description
Metacognitive	Planning, monitoring, and revising the writing task.
Non-specific	Abilities to: - Write well-structured sentences through syntactic analysis; - Write relevant and sufficient information through propositional analysis; - Link information via microprocesses; - Summarize and structure text through macroprocesses.
Specific	Spelling through logographic, alphabetic, orthographic, and morphographic processing.

).

This multi-component view is consistent with Kim’s Interactive Dynamic Literacy Model, which conceptualizes reading and writing as dynamically interacting systems in which phonological, orthographic, morphological, syntactic, and higher-order processes operate jointly and reciprocally during written-word production (Kim, 2020). The model reinforces the idea that spelling accuracy emerges from the coordinated activation of these linguistic and cognitive processes, a perspective that aligns closely with Laplante’s framework and with the specific treatments targeted in the present study.

In this article, ‘French written-word production’ refers to the specific processes that allow learners to spell words in accordance with French lexical orthography, consistent with Laplante’s schematic representation of principal writing processes and the developmental perspective of Seymour; these processes can be assessed in de-/contextualized, standardized or non-standardized tasks, with or without technology (Catach, 2008; Daigle et al., 2018; Kim, 2020; Laplante, 2011; Seymour, 2007; Federici & Scherer, 2017; Nelson, 2016).

1.2.2. Dyslexia–Dysorthographia and Its Implications for Writing in 9–12-Year-Olds

A conceptual analysis of definitional frameworks (American Psychiatric Association, 2013; Lyon et al., 2004; Tunmer & Greaney, 2009), together with recent consensus statements and neurocognitive reviews (J. M. Carroll et al., 2025; Chaix, 2025; Hall et al., 2023), converges on the same core markers: persistent word-level accuracy and fluency difficulties in reading and spelling, unexpected given age and instruction, and defined by exclusionary criteria. These difficulties typically reflect underlying phonological, orthographic, and processing-speed weaknesses that continue to affect written-word production in late elementary grades.

The J. M. Carroll et al. (2025) consensus definition synthesizes these elements and is retained here; it implies that rehabilitation must target word-level processes with sufficient intensity, while anticipating attention/executive constraints that can affect writing:

Dyslexia is a set of processing difficulties that affect the acquisition of reading and spelling. In dyslexia, some or all aspects of literacy attainment are weak in relation to age, standard teaching and instruction, and level of other attainments. Across languages and age groups, difficulties in reading fluency and spelling are a key marker of dyslexia. Dyslexic difficulties exist on a continuum and can be experienced to various degrees of severity. The nature and developmental trajectory of dyslexia depend on multiple genetic and environmental influences. Dyslexia can affect the acquisition of other skills, such as mathematics, reading comprehension, or learning another language. The most observed cognitive impairment in dyslexia is a difficulty in phonological processing (i.e., in phonological awareness, phonological processing speed, or phonological memory). However, phonological difficulties do not fully explain the variability that is observed. Working memory, processing speed, and orthographic skills can contribute to the impact of dyslexia. Dyslexia frequently co-occurs with one or more other developmental difficulties, including developmental language disorder, dyscalculia, ADHD, and developmental coordination disorder (J. M. Carroll et al., 2025, p. 1072).

The analysis of implications of dyslexia and dysorthographia in French 9–12, shows that the typical spelling profiles include different alphabetic confusions/substitutions (e.g., /f/–/v/, /t/–/d/), orthographic mis-selections among graphemes (e.g., c/k/qu, en/em/an), and morphographic omissions (e.g., silent letters), with slower text production and frequent pauses that constrain output and coherence. These patterns are often remitted more slowly than other literacy components (Alegria & Mousty, 1994, 1996; Quémart & Casalis, 2017; Sumner et al., 2013). Because dyslexia–dysorthographia centers on word-level accuracy/fluency with language-specific alphabetic/orthographic/morphographic challenges in French, remedial efforts in this study explicitly target these processes (treatments) and measure their change in both de-/contextualized tasks and authentic text production (J. M. Carroll et al., 2025; Alegria & Mousty, 1994, 1996; Quémart & Casalis, 2017). Recent clinical reviews confirm that these profiles remain stable across contemporary cohorts and continue to require explicit, high-intensity intervention targeting phonological, orthographic, and morphographic processes (Chaix, 2025).

1.2.3. Research-Based Remedial Programs and Integration of Assistive Technology: Evidence-Based Practices and Teacher Training

Research syntheses consistently show that spelling interventions grounded in evidence-based practices (EBPs)—notably programs that are explicit, targeted, individualized, and intensive—lead to improvements in the accuracy of written-word production among students with dyslexia–dysorthographia. Recent meta-analysis confirms that spelling interventions for learners with dyslexia produce small-to-moderate effects overall, with substantial heterogeneity attributable to program structure, dosage, and fidelity of implementation (Galuschka et al., 2019; Hall et al., 2023). These effects underscore the need for clear program architecture, sufficient instructional dosage, and robust teacher preparation. Recent clinical syntheses also emphasize that dyslexia–dysorthographia requires structured, intensive, and multi-component interventions combining phonological, orthographic, and morphological treatments, with explicit routines and sustained feedback (Chaix, 2025). These recommendations converge with the need for programs that articulate clear instructional sequences and provide teachers with sufficient preparation to deliver them. This aligns with disciplinary literacy perspectives, which emphasize the need for explicit teaching of the linguistic and orthographic processes specific to written French (Shanahan & Shanahan, 2008).

Taken together, these findings reinforce the central role of explicit, structured, and process-aligned instruction in supporting written-word development among learners with dyslexia–dysorthographia. Building on these converging insights, recent intervention research consistently highlights the importance of explicit instructional routines. Across the literature, effective programs generally adopt a four-step instructional sequence—explanation, modeling, guided practice, and independent application—with feedback tied to alphabetic, orthographic, and morphographic processes (Troia, 2014). An integrative review of spelling intervention programs for dysorthographic students aged 9–12 highlights similar patterns: explicit, targeted, and intensive programs yield the most consistent gains, but implementation quality and teacher preparation remain major determinants of effectiveness (Fontaine et al., 2021).

However, despite the effectiveness of explicit instruction, much of the existing evidence, particularly in English, remains grounded in handwriting-centric contexts, which limits its applicability to technology-rich writing environments. This gap becomes particularly salient when considering the integration of assistive technologies in writing instruction. Although improvements are often observed in the spelling of trained items—whether isolated words (e.g., Arnback & Elbro, 2000; Berninger et al., 2013; Chapleau & Beaupré-Boivin, 2019) or short texts (Berninger et al., 2008b)—students rarely reach age-appropriate expectations, and difficulties persist despite sustained intervention. Progress in spelling does benefit text production and word identification (Berninger, 2009), but response to intervention remains slower for spelling than for other literacy components in handwriting-centric contexts (Abbott et al., 1997; Swanson et al., 2013). A 40-year systematic review of reading interventions for students with or at risk for dyslexia similarly shows that treatment effects depend heavily on instructional quality, teacher expertise, and the consistency with which evidence-based routines are enacted (Hall et al., 2023). These persistent challenges underscore the need to examine not only program design but also the quality and consistency of implementation. Indeed, implementation fidelity has emerged as a critical explanatory factor in understanding the variability of treatment effects.

A significant limitation in this body of research is the lack of documentation of implementation fidelity, a well-known issue in intervention studies involving students with learning difficulties (Gresham et al., 2000; John et al., 2025; Joly et al., 2005; Swanson et al., 2013). Without detailed information on adherence to program components, dosage, quality of instruction, or student engagement, it becomes difficult to determine whether observed effects—or lack thereof—reflect the program itself or its enactment in authentic school settings. To address persistent spelling difficulties and considering the central role of feedback and strategy instruction, many scholars recommend evaluating remedial programs that explicitly integrate assistive technologies, such as speech synthesis and word prediction, while systematically documenting their fidelity of implementation (Basham et al., 2010; Kennedy & Deshler, 2010; John et al., 2025; MacArthur, 2013; Smith & Okolo, 2010). Recent empirical and doctoral work further supports this position by demonstrating that multimodal interventions and strategy-focused instruction are most effective when teachers receive structured training and ongoing coaching (Harrar Eskinazi, 2023; Toste et al., 2019).

By linking assistive technologies to research-based rehabilitation for 9- to 12-year-old students with dysorthographia, two tools emerge as particularly relevant for the targeted treatments/processes. Text-to-speech supports alphabetic processing by enabling rapid phonological plausibility checks during composing and review (syllabic listening; verification of phoneme–grapheme correspondences), while partially offloading working memory during successive rereads and promoting self-detection of substitutions/omissions/inversions. Word prediction primarily supports orthographic and morphographic processing by constraining the choice set within orthographic neighborhoods (e.g., c/k/qu; en/em/an; gu/e/i) and by facilitating access to derivational morphemes and silent final letters, thus strengthening morpho-orthographic encoding and accelerating retrieval of correct lexical forms. Embedded within a four-step explicit teaching sequence (explanation, modeling, guided practice, and independent application), Text-to-speech and word prediction operate at distinct moments of the planning–composing–revising cycle, multiplying opportunities for strategy use with immediate feedback on the targeted treatments. However, realizing these benefits requires documenting implementation fidelity, particularly adherence to treatment-aligned objectives/activities and exposure/dosage and ensuring teacher preparation (tool calibration and level-appropriate lexicons/word banks aligned to the graphemes targeted in instruction) and high-quality interventions (clear instructions, modeling, guided support, and informative feedback), given the sensitivity of this population to the quality and quantity of remediation.

Translating evidence-based principles into high-quality classroom delivery requires professional development reinforced by coaching or behavior-skills training—modeling, rehearsal, and performance feedback—to strengthen teachers’ mastery of explicit instruction routines and the task-aligned use of tools (Kretlow & Bartholomew, 2010; Mohr et al., 2024). Implementation of fidelity thus becomes the bridge between program design and measurable effects. Key dimensions include adherence to objectives and activities, exposure/dosage, teacher competence (clarity of instructions, modeling, support, feedback), differentiation, student engagement, and documented adaptations (C. Carroll et al., 2007; Rouse, 2023). Implementation science further highlights the role of enabling drivers such as training, coaching, performance assessment, and leadership (Fixsen et al., 2005). Taken together, these updated findings reinforce the need for specialized teachers to master explicit instructional routines and to differentiate support according to students’ phonological, orthographic, and morphographic profiles—an area repeatedly identified as challenging in recent intervention research (Toste et al., 2019).

In the present study, the program’s instructional core aligns with the processes described in Section 1.2.1 and the axes outlined in the Context: alphabetic plausibility supported by speech synthesis, orthographic and morphographic patterning supported by word prediction, and explicit strategy instruction delivered through a structured four-step sequence. The program architecture, combined with dedicated teacher learning supports (PD + coaching/BST) and the fidelity constructs above, provides the framework for examining both the degree of implementation and the training challenges that shape outcomes in real school settings.

1.2.4. Specific Question, Objectives and Sub-Questions

The present study was initially designed to address a specific research question focused on the effects of a research-based remedial program integrating assistive technologies on the written-word production of students aged 9 to 12 with dysorthographia. However, during the implementation phase, the quantitative monitoring—based on more than 25 repeated measures per participant using multiple instruments aligned with the dependent variable—revealed rich, complex, and individualized patterns of progression. All students demonstrated a clear upward trend, yet each trajectory reflected unique response profiles shaped by contextual and instructional factors. These unexpected but robust findings required a complementary analytical lens to understand why the program produced these effects and under which conditions. This shift in analytical focus was therefore necessary to fully interpret the observed effects.

Consequently, the study was expanded to include a second analytical component focusing on implementation fidelity, which functions as the interpretive bridge between program design and observed outcomes. This addition was necessary to determine (a) the degree to which the program was implemented as intended and (b) the training and coaching dynamics that shaped teachers’ capacity to deliver the program with fidelity in authentic school settings.

The specific research question, therefore, comprises two interdependent components.

Specific research question:

What are the contributions of a research-based remedial program that integrates assistive technologies and is implemented with documented fidelity by trained specialized teachers to the development of written-word production in French among students aged 9 to 12 with dysorthographia?

To operationalize this expanded research focus, the study pursued two complementary objectives.

• Primary objective:
  To evaluate the individual effects of the program on students’ written-word production, based on repeated quantitative measures collected throughout the single-case protocol.
• Secondary objective:
  To evaluate the fidelity of implementation of the program and the conditions that shaped its delivery by specialized teachers.

Given the focus of the present article, the analyses reported here address the secondary objective while summarizing the quantitative tendencies that contextualize the fidelity findings.

To meet this objective, two sub-questions guide the qualitative inquiry:

(SQ1) To what extent was the program implemented as intended, and what purposeful adjustments were made during delivery with respect to adherence (objectives/activities), exposure (frequency/duration), teacher competence, and student engagement?

(SQ2) What training and coaching challenges influenced teachers’ ability to implement these components with fidelity in authentic school settings?

Although conceptually distinct, these sub-questions are closely connected in practice. It operates at distinct but complementary analytical levels.

• SQ1 is grounded in established conceptual dimensions of implementation fidelity (C. Carroll et al., 2007; Fixsen et al., 2005), which are reviewed in the theoretical framework.
• SQ2, in contrast, concerns the emergent challenges encountered during teacher training and coaching. These challenges are not predetermined by theory but arise inductively from the data; they are therefore not part of the literature review but are analyzed in the Results and Discussion.

These sub-questions structure the qualitative component of the study and guide the methodological choices detailed in the next section.

2. Materials and Methods

The methodology of this mixed experimental research favors a quantitative method to achieve the primary objective and a qualitative method to achieve the secondary objective. Given the purpose of the article, it is described and justified here, focusing primarily on the qualitative method. Consistent with the remit of this article (Section 1.2.4), the present paper emphasizes the qualitative examination of fidelity to answer two sub-questions: (SQ1) the degree to which the program was delivered as intended and the purposeful adaptations that occurred in situ (adherence to objectives/activities; exposure in frequency/duration; student engagement; teacher competence/quality), and (SQ2) the training and coaching challenges encountered by specialized teachers when planning and enacting these components. Thus, the research protocol and design, as well as the research plan, participants, measures, intervention program, procedure, and data analysis plan, are explained and justified.

2.1. The Research Protocol and Design

This research opts for a single-case protocol that closely aligns with remedial teaching practices by focusing on the individual development of targeted students in a personalized support context (Horner et al., 2005; Maggin et al., 2013; Neuman, 2011; Salkind, 2010). According to these authors, like group protocols, this protocol allows for the manipulation of an independent variable—in this case, an intervention program—and the evaluation of its impact on the dependent variable. It is particularly well-suited to the small number of students with dysorthographia and the diversity of their profiles, especially in Quebec, where dysorthographia is rarely identified. This protocol is based on continuous and repeated quantitative measurements, with each student serving as a control for themselves. Implemented in real-world conditions by typical teachers, the protocol provides valuable data for planning and regulating practices, especially when combined with a sequential transformative design to evaluate the fidelity of program implementation. According to Creswell (2009), this design promotes understanding of the phenomenon as it evolves, concurrent with the implementation of a program. In this study, it is updated during the intervention phase through continuous quantitative measurements followed by qualitative measurements, as well as standardized quantitative measurements before and after the program implementation phase. All these measures are directly related to the concepts involved, which favor the validity of the construction.

The sequential design also allows the qualitative data collected during and after implementation to directly document the dimensions of fidelity targeted in SQ1 (adherence, exposure, teacher competence, and student engagement) as they unfold in authentic school settings. In parallel, the design captures the training and coaching dynamics experienced by teachers, which are central to SQ2 and emerge from their logs, coaching notes, and reflective accounts.

2.2. The Research Plan

A single-case protocol with multiple subjects of the A-B-BC-A type is carried out over 20 weeks. Therefore, the three phases are as follows: (1) the baseline phase lasts 5 weeks to establish the students’ initial level and prepare them for the implementation of the program (A); (2) the implementation phase lasts 10 weeks. It is divided into two sub-phases: five rehabilitation sessions with speech synthesis (B) and rehabilitation sessions with speech synthesis and word prediction. Each phase consists of six intervention sessions (BC); (3) the learning observation phase lasts 5 weeks to assess stability (A). Continuous measurements are taken once a week during the three phases of the study to evaluate the effect of the program (primary objective), which is obligatory in a single-subject study. These are followed by qualitative measurements to evaluate the fidelity of its implementation (secondary objective) at key moments during the intervention phase.

These qualitative measurements (teacher logs, researcher logs, and observation grids) directly inform SQ1 by documenting how the program was enacted session by session, including adherence to planned objectives, dosage protection, and the nature of in situ adaptations. They also provide the empirical basis for SQ2 by capturing the challenges teachers encountered during training, planning, and delivery, as well as the coaching supports required to address them.

2.3. Participants and Intervention Settings

The participants are four special education teachers working with six students with dysorthographia, aged 9 to 12, in four different schools. Institutional ethics approval was obtained; parental consent, student assent, and teacher consent were collected. Three of the teachers are novices, and one is an expert. The novice teachers are fourth-year preservice teachers in a bachelor’s program in special education and social adaptation at a Canadian university in Quebec. The expert teacher, who has been teaching for 20 years, also contributed to feasibility validation, pre-testing, and coaching of preservice teachers, and participated in peer feedback during coaching meetings. These four special education teachers work with six students with a specific profile. More specifically, interventions were delivered by three pre-service teachers to participants with dysorthographia P1 and P2, P5, and P6, and by an expert teacher to participants with dysorthographia P3 and P4. The contrast between novice and expert teachers is particularly relevant for SQ2, as it provides natural variation in training needs, coaching responses, and implementation challenges.

The number of six students, similar to the average number of subjects in a single-case protocol (mean = 5.12, median = 3), is advantageous for external validity, which is, however, limited to individuals with similar profiles (Kazdin, 2011). These students are selected in collaboration with their school staff, using inclusion criteria and selection procedures that are justified as follows. The diversity of student profiles and school contexts also supports SQ1 by allowing examination of how fidelity varies across settings and learner needs.

As indicated in the definitions of dyslexia–dysorthographia, their characteristics involve specific difficulties in identifying and producing written words, due to a phonological deficit. They were confirmed using school assessments and questionnaires for remedial teachers, teachers, and parents, as well as standardized tests (school records, questionnaires, Fontaine, 2019; ODÉDYS 2 battery version 2008; Cogni-Sciences, 2008, Grenoble, France) was used for the assessment, to guide individualized objectives and the graded selection of activities (Section 2.5). Classroom instruction continued as usual; no competing specialized program in lexical spelling was initiated during the ten-week intervention (current manuscript). Despite remedial interventions in handwriting and word processing, the difficulties of the students persist and affect their alphabetical and orthographic processing. All have been introduced to the use of the keyboard (Type to Learn 3 test, 2012, Sunburst Digital, 2004, Pleasantville, NY, USA) and word processing (interview/observation grid, Evmenova et al., 2010), which facilitates faster writing and longer texts (Morphy & Graham, 2012). These students have no sensory, motor, intellectual, or behavioral impairments, and French is their native language (school records; Development Questionnaire, Douaire, 2006).

2.4. Measures and Instruments

In addition to qualitative measures of implementation fidelity, a structured set of quantitative measures was used to assess students’ written-word production throughout the study. These measures were aligned with the dependent variable and included (a) standardized pre- and post-intervention assessments and (b) repeated continuous measures administered weekly during the 20-week protocol.

Standardized assessments included spelling dictation tasks (e.g., Ortho 3 from the Batterie d’Évaluation du Langage Écrit et de ses troubles—Mousty et al., 1994 and word and non-word dictation tasks from the Batterie Analytique du Langage Écrit—Cogni-Sciences, 2010) and structured text production tasks (descriptive and narrative writing tasks; e.g., the “Paragraph Writing” subtest of the Wechsler (2005) Individual Achievement Test, and the “Abandoned House” task from the Writing Competency Assessment, Jalpert, 1995), as described in Fontaine (2019).

In addition, continuous measures were collected using a weekly experimental task in which students produced written texts under controlled conditions. Fluency was operationalized as the number of correctly written, grade-appropriate words per minute, while accuracy was assessed through error rates and correct grapheme selection.

The combined use of standardized and continuous measures provided both a normative perspective (comparison to age expectations) and an idiographic perspective (individual progress over time), ensuring a comprehensive assessment of program effects.

Qualitative measures were used in the context of direct observation of interventions to assess the fidelity of program implementation, namely a logbook written by the teachers and an observation grid completed by the researcher From a comprehensive perspective, these instruments are combined to gather the perceptions of the various stakeholders regarding the components of program implementation fidelity (triangulation of tools and sources), which is essential to ensure the credibility of the approach (Noell et al., 2005; Wickstrom et al., 1998).

A coherent set of qualitative instruments was used to address both sub-questions. These instruments include teacher logs, researcher observation grids, fidelity checklists, coaching notes, and reflective journals. Detailed descriptions and blank versions of all instruments are provided in Appendix A and Appendix B.

For SQ1, these instruments document the degree of implementation fidelity by capturing adherence to planned objectives and activities, exposure (frequency and duration), teacher competence during explicit instruction routines, student engagement, and the nature of purposeful in situ adaptations. The fidelity dimensions and coding anchors are aligned with established frameworks (C. Carroll et al., 2007; Fixsen et al., 2005). For SQ2, the same instruments provide insight into teachers’ training and coaching experiences. Coaching notes, reflective journals, and planning logs allow the identification of emergent challenges related to instructional routines, tool calibration, planning constraints, and contextual barriers. Using a single, integrated set of instruments ensures methodological coherence while allowing the analysis to distinguish between deductive fidelity dimensions (SQ1) and inductive training/coaching themes (SQ2).

The logbook, designed by the researcher, is completed by the teachers at each rehabilitation session (see Appendix A). After each session, adherence is assessed based on the achievement of specific program objectives and the student’s individual goals using a color code (green = achieved, yellow = partially achieved, and red = not achieved), comments on their progress, and the material completed by the student. If additional activities are added, these are also noted. For the exhibition, the log specifies the number and duration of sessions and the weeks of implementation. Finally, comments on the student’s engagement are recorded.

The observation grid is completed by the researcher during two specific sessions: one session with speech synthesis (session 2) and one with speech synthesis and word prediction (session 4) (see Appendix B). Its purpose is to validate the information in the logbook through observation of activities and discussions with the teachers. It also serves to guide the future implementation of the program by validating quality interventions and suggesting avenues for intervention. In the grid, the researcher records the duration of the session (exposure), summarizes the achievement of individual objectives (adherence), and notes the student’s engagement. She also assesses the teachers’ ability to apply the recommended methods, particularly their ability to provide explicit, targeted, individualized, flexible, and intensive instruction. The quality of the instructions (clear, vague, or absent), modeling, support, and feedback (adequate, inadequate, or absent) is evaluated and commented on. These instruments, therefore, provide a coherent and comprehensive dataset from which both deductive (SQ1) and inductive (SQ2) analyses can be conducted.

2.5. Program Integrating Assistive Technology

This program lasts 30 h. It is implemented 3 × 1 h per week, individually or in small subgroups of two students. It is structured as a flexible sequence of micro-graded activities, meaning that out of 130 h of activities to choose from, 30 h are selected based on each student’s initial profile and progress assessment. In addition, individual objectives to support the development of writing processes in text production are planned and adjusted for all these activities. One objective relates to specific processes (accuracy or speed) and another to non-specific processes. These choices enhance social validity and transferability of the research, knowing that the students have different needs.

The activities begin with an introductory session that provides explicit instruction on the use of speech synthesis, specifying the following: What is it? When? Why? And how to use it? Next, an introductory session on speech synthesis, two rehabilitation sessions with speech synthesis (three one-hour sessions per session), an introductory session on word prediction, and then three rehabilitation sessions with speech synthesis and word prediction are offered (three one-hour sessions per session). Students used the WordQ software (version current at the time of the study, Quillsoft Ltd., 2018, Toronto, ON, Canada).

Sessions with speech synthesis focus on alphabetical processing. They therefore promote the production of regular words and non-words, chosen according to the student’s grade level and common difficulties. For each student, two sessions are planned to work on groups of graphemes that represent similar phonemes (possible specific objectives: f/v, ch/j, p/b, t/d, ou/on) (Horowitz-Kraus & Breznitz, 2011; Martinet & Valdois, 1999; Plisson et al., 2010). The remedial sessions with speech synthesis and word prediction focus on orthographic and morphographic processing. They therefore encourage the production of irregular words. For each student, two groups of graphemes are targeted to help them choose the correct grapheme associated with the phoneme (possible specific objectives: gu/e/i, s/c/t, c/k/qu, en/em/an/em, in/im/ein/ain, and graphemes derivable by morphology).

Each remedial session offers a sequence of 16 activities divided into three phases: preparation (four activities), implementation (eleven activities), and integration (one activity) (see Appendix B). In the preparation phase, students make connections with previous sessions and classroom activities, learn about the specific objective of the session, set two individual goals for the production of authentic texts, and begin to reflect on useful strategies for achieving them. They then benefit from explicit instruction tailored to their profile, immediate and informative feedback, and support in achieving their goals, using WordQ software. This instruction begins with decontextualized tasks (analysis and production of isolated words, fill-in-the-blank dictations), then moves on to the production of authentic texts to facilitate learning (description of images, writing and revision of humorous stories and autobiographical accounts). The integration phase helps students take stock of what they have learned, note key words in their reference guide, and apply what they have learned in a handwriting context.

The structure of the intervention program (explicit instruction sequence; alignment of activities with alphabetic, orthographic, and morphographic processes; integration of text-to-speech and word prediction) provides the analytical framework for SQ1, as fidelity is assessed relative to these planned components. The program also embeds teacher learning supports (PD sessions, modeling, rehearsal, performance feedback), which are central to SQ2 because they shape teachers’ capacity to implement the program with fidelity and reveal the challenges encountered during training.

2.6. The Research Process

The process, including initial validation, recruitment, and program implementation, took place over two years. These steps are described with reference to the researcher’s journal, in which she recorded her actions at each stage. This ongoing involvement promotes the achievement of the criteria of confirmation and reliability of the research.

The initial validation of the program took place in two stages: first, a pilot project in a university clinic involving two students was used to develop the first version of the program and select quantitative measures. Next, a 15-week pre-experiment in two schools with three students was used to adjust and validate the feasibility of the program in a real-world context. The A-B-BC research plan was then used to measure progress and refine the activities, logbook, and observation grid in preparation for wider implementation.

The recruitment of facilitators, schools, and participants took place in early fall, in parallel with the pre-experiment. Three fourth-year students from the university’s Bachelor of Education in Special Education and Social Adaptation program were approached (preservice teachers). They expressed interest in voluntarily participating in the project as part of their final internship. An experienced remedial teacher who had participated in the pre-experiment also joined the team to work with two additional students at her school.

In December, the program was launched with an initial meeting in the community, allowing the preservice teacher and, where applicable, the remedial teacher hosting the preservice teacher for an internship, to evaluate the child to plan the implementation of the program. The results of the quantitative measurements made it possible to target the rehabilitation sessions to the individual objectives.

Subsequently, in January, the initial training for teachers focused on program implementation, mastery of assistive technology software, theoretical foundations, and continuous measurement. This six-hour training, delivered at the university by the researcher, incorporated a variety of teaching strategies (e.g., explicit teaching, case studies, and interactive presentations). It was enhanced by filmed excerpts from the pre-experiment and a program implementation guide in accordance with the recommendations of experts in the field (Kretlow & Bartholomew, 2010; Noell & Gansle, 2009; Power et al., 2005; Sanetti & Kratochwill, 2009). Collaboration with the experienced practitioner enriched the training.

This initial training was followed by ongoing support, as recommended by these experts and researchers who have conducted similar studies (Chapleau, 2017; Ise & Schulte-Körne, 2010; Jung et al., 2018). The support took place from January to April, combining three group meetings and four individual meetings, held every two weeks. The group meetings, planned with the experienced practitioner after field observations, were intended to review the sessions, provide feedback, and adjust interventions according to the students’ progress. The individual meetings began with a joint analysis of follow-up measures, progress charts, logbooks, and student work to guide the process. Each meeting included feedback with the student on their learning, followed by a one-hour intervention allowing for both observation and explicit teaching. The researcher gradually transferred responsibilities to the practitioner, providing feedback and modeling as needed. Finally, a written report was given to each school team, accompanied by recommendations.

During each intervention session, the teacher and researcher logs documented adherence to planned objectives, the duration and frequency of activities, the quality of instructional routines, and student engagement. These procedural traces directly inform SQ1 by providing fine-grained evidence of how the program unfolded in practice. The procedure also included structured coaching meetings before and during the intervention. These meetings generated data relevant to SQ2, as they captured teachers’ questions, difficulties, and the supports required to maintain or improve fidelity.

2.7. The Analysis Plan

The analysis plan allowed for parallel analysis of quantitative and qualitative results, as in most mixed research (Teddlie & Tashakkori, 2009). It consists of quantitative analyses of intra-individual effects and is supplemented by qualitative analysis strategies to make connections between the effects of the program and the interventions offered to them (content analysis).

Quantitative analyses were conducted from two perspectives: a normative perspective and an idiographic perspective. To determine whether students produced French written words with greater accuracy in both isolated-word and text production, pre- and post-intervention standardized measures were compared and interpreted against available norms. Each hypothesis was confirmed or refuted depending on whether the direction of change matched the a priori verification criteria. When a student’s initial performance fell below age-expected standards and moved closer to expectations after the program, this was interpreted as a meaningful contribution to the targeted learning outcome (Fontaine, 2019). On an exploratory basis, performance with versus without assistive technology was compared to examine transfer of learning and the potential for compensation through these tools (current manuscript). To test whether students produced French written words with greater fluency in text production (third hypothesis), we analyzed weekly continuous measurements from the experimental task. The outcome was the number of accurately written, grade-appropriate words per minute across the A–B–BC–A phases. Interrupted-time-series procedures (mean, control chart, and C statistic) were used to complement visual inspection and trend analysis, following single-case design conventions (Kratochwill et al., 2010). These idiographic findings were interpreted together with the normative standardized-measure results and with the qualitative evidence.

Qualitative content analysis was used to identify salient themes from textual data collected across stakeholders, a strategy widely recommended for analyzing multi-source educational data (Fortin & Gagnon, 2016; Noell & Gansle, 2009). The qualitative data were analyzed using two complementary strategies aligned with the study’s sub-questions. For SQ1, a directed content analysis was conducted based on established fidelity dimensions (C. Carroll et al., 2007; Fixsen et al., 2005). Deductive codes were applied to all qualitative sources using a structured coding grid (Appendix C), allowing systematic examination of adherence, exposure, teacher competence, student engagement, and documented adaptations. For SQ2, an inductive thematic analysis was conducted to identify emergent challenges related to teacher training and coaching. This analytic strategy was selected because these challenges are not predetermined by theory but arise from teachers’ lived experiences during program delivery. The thematic development process and coding steps are documented in Appendix C. Together, these analytic strategies ensure alignment between the methodological choices and the dual structure of the research question.

The units of analysis (applied to both instruments for SQ1 and SQ2) are:

-

Primary unit: The intervention session (teacher × student[s], 60 min) as recorded in the teacher session logs, used to appraise fidelity dimensions and adjustments over time.

-

Secondary unit: The instructional episode (activities 1–19) within the two observed sessions per student captured in the researcher’s observation checklist (one text-to-speech session; one text-to-speech plus word-prediction session), used to examine the quality of explicit teaching routines and moment-by-moment engagement.

The first author coded the full corpus (both instruments and student artifacts). A second coder independently analyzed a stratified subset; discrepancies were discussed and resolved by consensus, with codebook updates as needed. Triangulation (logs, checklist, and student work), peer debriefing (including the expert remedial teacher), and analytic memos formed an audit trail, supporting credibility and confirmability (naturalistic inquiry).

A three-step synthesis applied to both instruments was conducted:

• Step 1 (within-source coding and grouping): For each instrument, data were coded against the fidelity framework (with emergent themes noted), then grouped by fidelity component and examined vertically (by participant) and horizontally (across participants).
• Step 2 (cross-instrument integration): Findings for adherence and engagement from both instruments were compared vertically and horizontally to align teacher-reported and researcher-observed evidence.
• Step 3 (planning-to-delivery linkage): Results were re-examined, considering planning and implementation activities to determine the degree of implementation achieved for each fidelity component and to specify where and why adjustments or training supports were necessary.

Finally, idiographic quantitative evidence (continuous measures and visual inspection), normative quantitative evidence (pre/post standardized measures), and the qualitative fidelity analysis from both instruments were interpreted concurrently and complementarily to address SQ1 (degree/quality of implementation; nature of adjustments) and SQ2 (specific training/coaching challenges and the levers that addressed them) (mixed-methods guidance).

For SQ1, qualitative data were analyzed using a directed content analysis approach anchored in established fidelity dimensions (C. Carroll et al., 2007; Fixsen et al., 2005). Codes were deductively derived from adherence, exposure, teacher competence, and student engagement, allowing systematic examination of the degree of implementation and documented adaptations. For SQ2, an inductive thematic analysis was conducted to identify emergent challenges related to teacher training and coaching. This analytic strategy was selected because these challenges are not predetermined by theory but arise from teachers’ lived experiences during program delivery. The combination of deductive (SQ1) and inductive (SQ2) analyses ensures alignment between the methodological choices and the dual structure of the research question. The following section presents the results of these analyses, organized to reflect the integrated nature of SQ1 and SQ2.

3. Results

3.1. Overview of Quantitative Tendencies (See Fontaine, 2019, for Full Statistical Details)

Across the 20-week protocol, the repeated quantitative measures revealed clear upward trends in written-word accuracy and fluency for all six students. To synthesize the quantitative findings and clarify the criteria used to confirm each hypothesis,

Table 3. Summary of Hypothesis Testing and Quantitative Results.

Hypothesis	Outcome Variable	Measure	Confirmation Criteria	Result
H1. Accuracy (isolated words)	Correct spelling of isolated words	Standardized dictation tasks (Fontaine, 2019)	Improvement from pre- to post-test and movement toward age-expected norms	Confirmed for all students
H1 (with Assitive Technology)	Correct spelling with assistive technology	Standardized dictation tasks with text-to-speech and word prediction (Fontaine, 2019)	Increased accuracy and effective use of assistive technology	Confirmed for all students
H2. Accuracy (texts)	Error rate in written texts	Standardized descriptive and narrative writing tasks (Fontaine, 2019)	Reduction in errors and progression toward expected performance levels	Confirmed; all students showed improvement
H3. Fluency (texts)	Writing fluency	Weekly continuous measures (experimental task)	Upward trend across phases (A–B–BC–A), supported by visual and statistical analyses	Confirmed for all students

presents a summary of the results across outcome variables and measurement types.

As shown in Table 3, all hypotheses were confirmed, with consistent improvements observed across participants in both accuracy and fluency measures. These quantitative trends provide essential context for interpreting the qualitative findings related to implementation fidelity. Although the magnitude and trajectory of progress varied across individuals, each student demonstrated measurable improvement aligned with the targeted alphabetic, orthographic, and morphographic processes supported by assistive technology.

• Hypothesis 1 (isolated words; accuracy): Across all students, standardized dictations showed pre-to-post gains in accuracy for French written words in isolated-word tasks. Before the program, performances were typically below grade-level expectations and revealed heterogeneous weaknesses across alphabetic, orthographic, and morphographic processing. After explicit, individualized instruction aligned to these processes, gains were observed precisely on the graphemes and items that were targeted (e.g., acontextual graphemes addressed with text-to-speech; inconsistent contextual and morphographically derivable graphemes addressed with text-to-speech plus word prediction), moving all students closer to age-expected norms, albeit with residual needs on the most complex items (Fontaine, 2019). Hypothesis 1 (isolated words with assistive technology; accuracy): With text-to-speech and word prediction, post-program performance improved for all students; those previously unfamiliar with the tools shifted from limited/no compensation to effective compensation (often outperforming handwriting), while students already familiar with the tools reached very high or perfect accuracy on trained content (Fontaine, 2019).
• Hypothesis 2 (texts; accuracy): In text-writing tasks, error rates decreased after the program in handwriting and with assistive technology, with several students improving by two to three quartiles in the descriptive task and one to three grade levels in the narrative task administered with tools, bringing all students within age expectations (Fontaine, 2019).
• Hypothesis 3 (texts; fluency): Weekly continuous measures documented higher rates of correctly written, grade-appropriate words per minute across the A–B–BC–A phases for every student. Visual inspection, trend analysis, control-chart deviations, and the C statistic jointly supported statistically meaningful fluency gains, consistent with single-case design conventions for interrupted time-series interpretation (Kratochwill et al., 2010; Fontaine, 2019).

Although the quantitative results are not the focus of the present article, they provide essential context for interpreting the qualitative findings. Significant gains in accuracy and fluency were observed for all students across all measures, with assistive technologies amplifying instruction-aligned progress, particularly among initially unfamiliar users. The variability of these trajectories across profiles and school contexts underscores the need to examine how the program was implemented in authentic settings and under which conditions teachers were able to deliver it with fidelity (Kratochwill et al., 2010; Fontaine, 2019).

3.2. Qualitative Results

This qualitative content analysis is presented according to the components of fidelity, namely adherence, exposure, participant engagement, and the competence of teachers in using the recommended methods (Perepletchikova & Kazdin, 2005; Power et al., 2005). Although SQ1 and SQ2 are analytically distinct—SQ1 examining fidelity dimensions through a deductive lens and SQ2 examining training and coaching challenges through an inductive lens—the two sub-questions are empirically intertwined. In practice, adherence, exposure, and teacher competence (SQ1) were directly shaped by the training and coaching processes documented in SQ2. For this reason, the Results section presents the qualitative findings in an integrated manner, while clearly indicating which elements relate primarily to SQ1 or SQ2 in the summary of each component.

3.3. Adherence

In summary, analysis of the adherence indicators reveals a high degree of fidelity to the initial plan, with most of the specific objectives having been achieved (SQ1). The implementation process revealed several challenges in training and supporting specialized teachers, particularly with respect to formulating individualized objectives that aligned with each student’s profile, mastering the assistive technologies integrated into the program, and continuously adjusting instructional sequences in response to students’ evolving needs (SQ2).

Initial planning linked diagnostic assessment to session design and was co-constructed with teachers and school staff. Support primarily addressed two challenges: translating individual needs into specific (accuracy or fluency) and non-specific objectives (e.g., syntactic or propositional processes), and sequencing activities accordingly. In practice, objectives were regularly adjusted through short feedback cycles based on weekly observations. Thus, in order to meet the needs of two contrasting cases of students in alphabetical processing, support was offered to the pre-service teacher working with participants with dysorthographia P1 and P2 (PREP1P2) to target sessions with speech synthesis, allowing them to work on both consistent, context-free graphemes that were difficult for P1and complex syllabic structures for P2. Thus, to begin the sessions with speech synthesis, the expert teacher working with participants P3 and P4 (EXP3P4), the pre-service teacher working with participant with dysorthographia P5 (PREP5), and the pre-service teacher working with participant with dysorthographia P6 (PREP5) set goals related to accuracy, while in non-specific processes, they focused on syntactic analysis. Indeed, the student can check the phonological plausibility of written words and the syntactic acceptability of sentences by listening to the text with this tool. In addition, in specific processes, PREP1 P2, PRE5, and PREP6 set goals related to speed, while for non-specific processes, they focused on propositional analysis, given that word processing can help to quickly correct errors and offers the possibility of moving or adding information. The implementation of the plan was formalized and discussed with the student, school staff, and parents prior to implementation. Support continued through regular meetings aimed at regulating interventions according to the student’s progress. The observations recorded made it possible to trace this complex process (see

Table 4. Excerpts from the preservice teachers’ logbooks and the researcher’s observation grid relating to adherence for P2.

Participants	Sessions	Logbooks of the Practitioner (ISP1P2)			Researcher’s Observation Grid
		O	A	Observations/Comments	Observations/Comments
P2	Speech synthetis	Short text		Session 1: ④ p/b (Student’s name) wrote several words in their text in a short amount of time.	Session 2: ③ ou/on Essentially, this student’s challenges are to continue: - increase writing speed. - Organize (and select) your ideas. Before the session: I gave the student feedback on their goals myself (…) to set an example and give the facilitator free rein afterwards. The student has a clear understanding of their goal.
		Text developed		Session 2: ③ ou/on In his texts, he has several ideas, but he tends not to make connections between them or develop them. It is therefore important that he learns to organize his ideas and clarify them (…), e.g., When I open the door, I go in. There was a basement with a grandmother sitting on the sofa, and she was tapping on the living room floor. He had an ugly pink bedroom, and it was very worn out (…)
	Speech synthesis + word prediction	Short text		Session 4: ② s/c/t His ideas are connected. (…) However, he has difficulty grasping sentence structure and punctuation (…) He made several corrections and moved sentences around to make the meaning clearer.	Session 4: ② s/c/t Student who produces more words in a more organized manner. His goals: to validate himself, particularly with regard to: - sentence structures. - punctuation. I am helping the student review part 2: sentence structures that are difficult to grasp. I suggested encouraging the student to practice syntax (also done in class).
		Text developed		Session 6: ⑥ graphemes that can be derived from morphology. His ideas are linked together. By rereading his work (with speech synthesis), he added several punctuation marks. However, this still needs work. The structure of his sentences is good. He added a few details to them (…), e.g., In this beautiful weather, I had a great life because I didn’t go to school. I went to daycare and I was very annoying (…)

O: Objectives. A: Assessments. Achieved. N/A: Not observable.

).

The sessions with speech synthesis enabled the teachers to confirm that the objectives had been achieved. The support provided helped to adapt the intervention to the student’s progress. In specific processes, EXPP3P4 and PREP6 observed that the student developed accuracy by teaching him to use speech synthesis to check the phonological plausibility of written words, while in terms of non-specific processes, he developed syntactic analysis by teaching him to listen to sentences and the entire text with this tool. Similarly, in specific processes, PRE and PREP5 assessed that the student gained speed by teaching him to use word processing to validate the lexical spelling of frequent words with the spell checker. Thus, depending on their profile, students were encouraged to record their ideas during the planning stage, to transpose them into writing, leaving traces of their spelling doubts during the writing process, and then to check them with this tool during the revision and correction stage. When writing short texts, they managed the time available with a timer. In addition, in non-specific processes, they found that students developed macro-processes and propositional analysis by being encouraged to use this tool to move or add information. Thus, at the end of these sessions, the support provided made it possible to validate that these objectives had been achieved or partially achieved. It also made it possible to observe more marked progress among participants who were new to assistive technology (EXPP3P4 and PRE5) compared to those who were familiar with it (PREP1P2 and PREP6).

For sessions with speech synthesis and word prediction, the support provided led the teachers to maintain certain objectives and formulate others. By adding this tool, which quickly provides word suggestions based on the graphemes typed by the student to help them choose a contextual grapheme or one that can be derived from morphology, they formulated an individual objective relating to specific processes concerning either accuracy (PREP3P4, PREP5, and PREP6) or speed (PREP1P2). They formulated a second objective related to non-specific processes, namely propositional analysis (PREP1P2, PREP6), syntactic analysis (EXP3P4), or macroprocesses (PREP1P2; PREP5). As a result, they felt confident about pursuing these goals in subsequent sessions, especially since they anticipated progress in their students; they also had the opportunity to use word prediction in class for text production. The objectives were indeed consolidated or achieved. Thus, with regard to specific processes, the support led PREP1P2, EXPP3P4, and PREP6 to help the student consolidate his individual objectives, and PREP1P2, EXPP3P4, and PREP5 to develop speed. Speed was then developed by maintaining the strategies put in place during the rehabilitation sessions with speech synthesis, while explaining how to refer to word prediction. Similarly, in terms of non-specific processes, the support helped PREP1P2, EXPP3P4, and PREP6 assess whether the student was achieving their individual goals, as they had developed syntactic and propositional analysis and vocabulary. In addition, this support supported the decision to add activities to support the transfer of learning for students newly introduced to word prediction. These activities consisted mainly of reminders of the strategies taught during classroom writing activities, as agreed with their teacher.

3.4. Exposure

The analysis shows that the initial plan was generally followed, with the identified support challenges being addressed. Thus, the number and duration of remedial sessions were generally respected (SQ1). The analysis shows that the initial plan was generally followed, with the identified support challenges being addressed. Thus, the number and duration of remedial sessions were generally respected (SQ2).

In the initial planning phase, the challenge for the support team was to validate the choices made by the trainee teachers for the initial planning of the sessions, in collaboration with the school’s remedial teacher. This led them to find solutions for implementing the number and duration of sessions (3 × 1 h per week, for 10 weeks), taking into account the organization of remedial teaching and classroom services. As for the number of sessions, the support helped PREP1P2 and PREP5 to plan them into the school’s next block of intensive interventions. It also allowed PREP6 to offer an additional session per week, as EXPP3P4 did. For session duration, this support helped IEP1P2 and ISP6 maintain the 60 min sessions already offered at the school. In addition, since planning the sessions meant that the student would be absent from class, the support helped the teachers to plan the remedial sessions closely with the student’s teacher, taking into account the planned activities.

In practice, the challenge was to help the trainee teachers find solutions to comply, as far as possible, with the number and duration of rehabilitation sessions, given the constraints encountered in the school context. After two weeks of implementing the sessions, they had all had to make adjustments due to classroom activities while respecting the planned schedule for their internship (e.g., canceled, rescheduled, or shortened sessions). Therefore, from the first group meeting onwards, the support provided helped to reinforce the adjustments that had been made and those planned for the future, by helping them to continue analyzing the student’s school context. It also encouraged them to reschedule rather than cancel sessions, while maintaining the agreed duration in order to comply with approximately 90% of the planned sessions. Subsequently, during individual meetings at the schools, the support helped to validate the adjustments made to the sessions and to plan future adjustments using the same guidelines.

3.5. Participant Engagement

Overall, the assessment shows that initial planning was implemented during sessions using speech synthesis and effectively refined through the integration of assistive technology. Student engagement was consistently observed, with adjustments particularly in stimulating writing tasks, while teachers adapted their intervention strategies to students’ profiles, school contexts, and the instructional possibilities afforded by the program (SQ1). The implementation process revealed several challenges in training and supporting preservice teachers, particularly in guiding them to select and sustain engagement-enhancing strategies, to provide continuous and strategy-oriented feedback (including discussing students’ approaches, reviewing progress graphs with them, and connecting remediation to classroom learning) and to adjust these practices responsively as writing tasks became more complex (SQ2).

In planning, the support challenge was to help the trainee teachers analyze the students’ profiles, identifying the students’ abilities and needs, as well as a strategy likely to promote their engagement (see

Table 5. Excerpts from the initial plan regarding participant engagement.

Teacher	Individual Profile	Preferred Strategies
ISP1P2	P2: cheerful, not very attentive to details	- Pay attention to the other person’s strategies

) Thus, like EXPP3P4, PREP1P2 planned a strategy to encourage discussions between the two students about their writing processes, while PREP5 and PREP6 focused primarily on discussions with their students on this issue. This plan was also discussed with the student, his school staff, and his parents with a view to its implementation. The planning took shape during implementation, when the teachers adapted and adjusted their strategies to encourage the students’ engagement (see

Table 6. Excerpts from the teachers’ logbooks and the researcher’s observation grid relating to engagement.

Participants	Sessions	Logbooks	Observation Grid
		Observations-Comments	Observations-Comments
P2	SV	(Student’s name) works hard and listens to my instructions.	(The student is) very attentive to his story (production of a humorous story).
	SV + PM	(Student’s name) works very well (…) However, they need to take the time to reread their work and make sure that the meaning of their text is clear.	(The student shows) great interest in the task (elaborate text production).

Commitment, Non-commitment, Pursuit of objectives, SVPM: Speech synthesis Word prediction.

).

During rehabilitation sessions with speech synthesis, the challenge was to provide direct support by guiding teachers to offer continuous feedback, promoting active dialogue with the student about his writing process. From the outset, it was noted that the student was more engaged in tasks he found stimulating, particularly when he received immediate feedback from the speech synthesis. To maintain this engagement, group meetings were used to reaffirm the importance of continuous feedback integrated into discussions about the students’ processes. Individual sessions then provided an opportunity to model ways of offering personalized feedback, based on the student’s concrete results (progress charts and completed work) and by discussing his writing process. This atmosphere of sharing encouraged the student to invest himself in the task and to collaborate.

Subsequently, the combined introduction of speech synthesis and word prediction enriched the strategies for promoting student engagement, leading the teachers to conclude that the student needed increased support through continuous feedback, especially in more complex writing contexts in the classroom. Thus, after discussion with the teachers, PREP1P2 continued to invite her students to pay attention to their peers’ strategies with these tools, while encouraging them to identify which ones to use in class, as she observed that P1 was less concerned with classroom assessments when proceeding in this way and P2 was more attentive to details. Similarly, EXPP3P4 and PREP5 had the opportunity to suggest that students work on their classroom writing with these tools, which helped P3 to meet the requirements, P4 to be willing to do the proposed task, and P5 to use her strategies. In addition, PREP6 suggested that P6 pay attention to the number of words she was producing, which encouraged her to improve.

3.6. The Competence of the Teachers in Updating the Program

At the end of the two sessions observed, the teachers’ ability to implement the recommended program seems very satisfactory, as the accompanying researcher’s comments on this subject in the observation grid are generally positive. These observations also reveal differences between the preservice teacher and the experienced facilitator in modeling, support, and feedback. These differences led to the suggestion of adjustments to them from the first session onwards, and then to help them develop this skill further during the second session (SQ1). The implementation highlighted several challenges in training and supporting preservice teachers, particularly in helping them appropriate and mobilize the instructional skill set required for explicit teaching—clarifying instructions, modeling strategies, providing targeted support, and delivering immediate, informative feedback—while simultaneously mastering the integration of speech synthesis and word prediction in increasingly complex writing situations (SQ2).

In planning, the challenge was to help the preservice teachers prepare to use this skill to implement the program in a remedial education context, knowing that this context was new to them, as was the program as a whole and the integration of assistive technology. During the training, the support helped the teachers familiarize themselves with this skill by learning the theoretical foundations of the program and analyzing its application and effects. This training also gave them the opportunity to familiarize themselves with the guide for implementing the program and its implementation, supported by video excerpts from the pre-experimentation phase. After using speech synthesis in a writing task, the support helped them analyze how instructions, modeling, support, and feedback had been provided in the pre-experiment. This allowed them to assess whether the instructions were clear and stated according to the guide, whether sufficient time was allowed for modeling to help the student observe specific strategies, and whether the support was focused on these strategies and provided continuous feedback. By proposing the same exercise with word prediction, the facilitator led them to make the same observations. She also helped them realize that this skill is more complex to mobilize during sessions with speech synthesis and word prediction, since students must coordinate their strategies using these two tools in increasingly complex writing situations.

In practice, during the first intervention session, i.e., the introductory session on speech synthesis, the support provided allowed the teachers to model this skill in the presence of the student and then answer their questions, helping them to master the technical aspects of speech synthesis and word prediction. Subsequently, after observing a rehabilitation session with speech synthesis and then a session with speech synthesis and word prediction, the challenge was to support the development of this skill by evaluating it and offering suggestions. This analysis also highlights that these steps were taken based on observations made with the experienced practitioner, as she was able to easily apply the targeted skill (see

Table 7. Excerpts from the researcher’s observation grid relating to the teachers’ competence.

Participants	Sessions	Activity	Observations and Comments
P2	SV	10	Support: Reactivate knowledge. Modeling: Verbalize strategies to fluently produce words suggested for the story (using tools). Support: Help them produce words even if they don’t always know exactly how to spell them, checking with other tools (including word prediction) if necessary.
		13.	Instructions/support: (Give a) reminder of the instructions to students to encourage them to think about improvements and the rest of the text. Support: (Give a) reminder of the text structure.
	SV + PM	13.	Feedback: (Give) feedback to students individually, based on their goals (using the progress chart and their work). Modeling: (Give) a reminder of how to plan writing (…) give examples and counterexamples. Support: Encourage them to practice syntax.

Validation, Suggestion, SV: Speech synthesis, PM: Word prediction.

).

First, the support provided made it possible to verify that the instructions were clear and in line with the guide, while emphasizing the importance of ensuring that students understood them, particularly during their initial learning phase. During the second session, the instructions were further adapted to the student’s specific needs. Second, modeling was deemed adequate, although it was sometimes absent among some facilitators during the first session. The researcher encouraged the adaptation of modeling according to the student’s objectives and the importance of devoting sufficient time to this step before moving on to guided and independent practice. Positive feedback was provided during the second session after observing appropriate modeling. Third, during the first session observed, based on observations of EXPP3P4, the coach encouraged the trainee teachers to focus on the student’s objectives, reminding them of the strategies to be used at each stage of the task with speech synthesis and providing feedback on these strategies once the stage was completed. As these suggestions seemed to be taken on board, the supervisor allowed time for reflective feedback on their support interventions during the second session, observed with speech synthesis and word prediction. She also encouraged them to expand on this based on the new goals agreed upon for the student, so that he could coordinate his strategies more effectively using these tools. Finally, the feedback seemed adequate and positive during the two sessions observed. However, it seemed to be more informative and immediate in the second session, encouraging them to comment on their results and approaches, as in EXPP3P4.

3.7. The Synthesis of the Results

In short, the results show that the degree to which the program was implemented as planned is very satisfactory, considering the level of adherence, exposure, participant engagement, and staff competence. It specifies that each component was initially planned and implemented, requiring adjustments with school stakeholders. It also shows that challenges in supporting trainee teachers were identified at each stage, based on strategies developed by the experienced practitioner. The analysis indicates that these support challenges were met by helping the facilitators to make connections between the students’ progress in writing and the interventions offered to them in this program, and that explicit, targeted, individualized, flexible, and intensive instruction was provided over an extended period of time, as planned.

More specifically, adherence to planned objectives was consistently observed, supported by systematic planning and ongoing regulation of instruction based on students’ individual progress (SQ1). Nevertheless, the adherence dimension surfaced training challenges for preservice teachers, particularly in translating diagnostic profiles into coherent, individualized, specific and non-specific objectives, calibrating objectives to the targeted alphabetic/orthographic/morphographic processes, and sequencing activities consistently across sessions (SQ2). Most objectives related to accuracy, fluency, and higher-level writing processes were achieved or partially achieved, demonstrating close alignment between program design and implementation. Exposure to the intervention was also largely maintained, with the majority of planned sessions delivered at the intended frequency and duration, following adjustments negotiated with school staff to preserve program integrity (SQ1). Yet maintaining this exposure required recurring support to help preservice teachers plan and sustain the intended dosage within authentic school constraints—protecting 60 min blocks, prioritizing rescheduling over cancellation, and coordinating with classroom timelines (SQ2). High levels of student engagement were documented throughout the intervention, particularly during authentic writing tasks supported by speech synthesis and word prediction. Engagement was actively fostered through explicit instruction, continuous feedback, and links between rehabilitation and classroom writing activities (SQ1). Concurrently, training challenges emerged in guiding preservice teachers to select and sustain engagement-enhancing strategies and to provide continuous, strategy-oriented feedback—discussing students’ approaches, reviewing progress graphs with them, and explicitly linking remediation to classroom learning—while adapting these practices as task complexity increased (SQ2). Finally, teachers, including those with limited prior experience, demonstrated progressive competence in modeling strategies, providing targeted support, and delivering informative feedback (SQ1). However, uneven competence in explicit-teaching routines highlighted continuing support needs, notably in clarifying instructions, modeling strategies, delivering targeted support and immediate, informative feedback, and in coordinating the simultaneous use of speech synthesis and word prediction in real-time instructional contexts (SQ2).

4. Discussion

The discussion allows us to interpret the results presented above and to cross-reference them with the theories and research findings outlined in the introduction. It also allows us to understand the scope of these results and to interpret them. Thus, after interpreting the results, this discussion outlines the limitations and implications of this research, then presents avenues for further research.

4.1. Interpretation of the Results

These results are difficult to compare with those of similar previous studies, as implementation fidelity is rarely assessed in experimental research involving students with learning difficulties (Capin et al., 2018; Gresham et al., 2000; Lenker et al., 2010; McIntyre et al., 2007; Swanson et al., 2013). In this sense, the present findings are unprecedented. Given that remedial programs are among the most challenging to implement in schools and often fail to produce sufficient progress in writing, the results obtained here are particularly significant (Fuchs et al., 2014). Evaluating fidelity for a complex, multicomponent program therefore provides a unique opportunity to interpret the results across four dimensions—adherence, exposure, participant engagement, and teacher competence—while distinguishing the contributions of SQ1 (degree of implementation) and SQ2 (training/coaching challenges).

4.1.1. Adherence

Consistent with prior research on response-to-intervention frameworks, the findings show that effective implementation depended on systematic analysis of multiple data sources to plan and regulate instruction (Coolong-Chaffin & McComas, 2015; Jenkins et al., 2017; Jung et al., 2018; Stecker et al., 2005; Reddy et al., 2016; Lembke et al., 2024; Baco et al., 2023) (SQ1). What is particularly noteworthy is that novice teachers were able to do so because of the structured support provided—an element rarely documented in the literature (SQ2).

The individualized nature of the program also distinguishes it from existing interventions, which are typically standardized for groups of students. The results suggest that a training model grounded in instructional planning and regulation contributed directly to student progress. They also show how teaching, learning, and assessment were aligned to support written-word development—an essential requirement given the complexity of French orthography and the scarcity of programs that integrate alphabetic, orthographic, and morphographic processes within a structured sequence supported by assistive technologies (Galuschka et al., 2019; Hall et al., 2023; Dahl-Leonard et al., 2023; Lembke et al., 2024; Keelor et al., 2023).

The alignment observed echoes the theoretical framework: effective remediation requires coordinated activation of multiple subprocesses (Seymour, 2007; Laplante, 2011), consistent with integrative models such as Kim’s Interactive Dynamic Literacy Model (Kim, 2020). Teachers were able to target objectives supported by the same assistive technology, combining accuracy and fluency. This synergy reflects core evidence-based principles—structured modeling, guided practice, and immediate feedback—systematically reinforced by text-to-speech and word prediction.

Finally, the results show that teachers selected and adapted tasks to optimize transfer to classroom writing, consistent with RTI-based recommendations (Jung et al., 2018; Reddy et al., 2016). In doing so, the program operationalized a key challenge identified in the introduction: enabling specialized teachers not only to master program components but also to adapt them responsively to learner profiles and school contexts (Capin et al., 2018; Swanson et al., 2013; Dahl-Leonard et al., 2023; Lembke et al., 2024; Sawyer, 2025).

4.1.2. Exposure

The findings indicate that the support provided helped teachers balance fidelity with the flexibility required in authentic school settings—a major challenge in implementing intensive interventions over extended periods (Dahl-Leonard et al., 2023; Sawyer, 2025; Gresham et al., 2000; Harn et al., 2013; Lembke et al., 2024; McIntyre et al., 2007; (SQ1). Exposure is rarely analyzed in depth in studies with struggling writers, yet it is central to the feasibility of research-based practices (Fuchs et al., 2014; Reddy et al., 2016; Stoiber & Getting, 2016).

The results show that support optimized the articulation between remedial instruction and classroom teaching, improving the use of time and resources (SQ2). This aligns with recent evidence that coaching and behavioral skills training are key levers for sustaining planned dosage and instructional quality (Lembke et al., 2024; Sawyer, 2025). Moreover, integrating text-to-speech within explicit writing instruction is associated with improved comprehension and engagement for students with dyslexia (Keelor et al., 2023). This perspective moves beyond traditional indicators such as number and duration of sessions, highlighting instead how dosage interacts with instructional quality and technological scaffolding.

4.1.3. Participant Engagement

The findings reveal how participant engagement was actively supported—an aspect rarely documented in rehabilitation studies (SQ1 and SQ2). Collaboration with school stakeholders ensured alignment between student needs and prioritized interventions, consistent with prior research (Swanson et al., 2013). Teachers strengthened engagement by linking program learning to classroom writing through assistive technologies, supporting both motivation and transfer.

Variability in engagement across writing tasks suggests the value of offering task choices or providing in-class support, consistent with recent work on AT-supported writing (Almgren Bäck et al., 2024; Keelor et al., 2023; Fernández-Batanero et al., 2022). The results also highlight the determination of students facing persistent difficulties, supported by highly engaged adults.

4.1.4. Competence of Teachers

Although rarely examined in experimental research, teacher competence is a central dimension of implementation fidelity, and the present findings show how it can be supported and adjusted within a writing rehabilitation program for students with dysorthographia (Dane & Schneider, 1998; Gresham et al., 2000; McIntyre et al., 2007; Swanson et al., 2013; Reddy et al., 2016; Dahl-Leonard et al., 2023; Lembke et al., 2024; Sawyer, 2025) (SQ1, SQ2). This is particularly important given the theoretical framework, which highlights both the complexity of French written-word production—requiring coordinated alphabetic, orthographic, and morphographic processing (Seymour, 2007; Laplante, 2011; Catach, 2008)—and the need for high-quality explicit instruction to support these processes through modeling, guided practice, and immediate feedback. Because students with dyslexia–dysorthographia present persistent word-level difficulties and benefit from intensive, process-aligned feedback (Alegria & Mousty, 1994, 1996; Quémart & Casalis, 2017; J. M. Carroll et al., 2025; Kim, 2020), strengthening teachers’ ability to deliver explicit, targeted instruction is essential.

Working with teachers of varied profiles—most of them novices—led to the development of an explicit, flexible, individualized, and intensive training structure aligned with both teacher needs and intervention demands (Jung et al., 2018; Reddy et al., 2016). This structure directly reflected evidence-based principles, notably the four-step explicit instruction sequence validated in writing research (Troia, 2014; Berninger et al., 2008a). From the planning phase onward, training supported the analysis of student progress, while ongoing accompaniment enabled real-time adjustments (SQ1), reinforced by peer exchanges and continuous support from the researcher.

Because the program integrates assistive technologies whose effectiveness depends on strategic, process-aligned use (MacArthur, 2013; Svensson et al., 2019), teacher competence in modeling and coordinating these tools was also essential. The results show that novice teachers were able to adjust their modeling, support, and feedback—likely contributing to the program’s effects. These adjustments reflect a distinctive feature of the program: the integration of assistive technology guided by expert observation and transparent planning-regulation processes (Kretlow & Bartholomew, 2010).

While most studies describe training programs without documenting teachers’ implementation competence, the present findings reveal both a high level of competence and specific challenges in supporting preservice teachers (Berninger et al., 2013; Chapleau, 2017; Ise & Schulte-Körne, 2010; Dahl-Leonard et al., 2023; Torgesen, 2005) (SQ2). As Chapleau (2017) noted, adjusting support to student progress is particularly challenging; here, instructors were able to do so, in part because the program’s AT-supported structure made student progress more observable and actionable.

5. Conclusions

5.1. Limitations of the Research

First, the qualitative corpus was limited to textual traces (logs and checklists). The absence of multimodal data (audio/video, co-planning artifacts, and training interactions) restricts the depth of analysis of real-time regulation—an enhancement aligned with recent fidelity-measurement guidance (Dahl-Leonard et al., 2023).

Second, structuring every session around text-production tasks generated variable engagement. Some students would have benefited from task options or in-class scaffolds to strengthen transfer. This reinforces the importance of co-planning with classroom teachers and engineering engagement through explicit links to classroom writing, consistent with recent AT research (Fernández-Batanero et al., 2022; Smith & Okolo, 2010).

Third, external validity is limited: the single-case protocol in specific school contexts restricts generalization. Two axes follow: (a) longitudinal support for practitioners and (b) broader identification pathways for dysorthographia. These align with evidence that fidelity improves with structured coaching, BST, and iterative feedback cycles (Lembke et al., 2024; Sawyer, 2025; Baco et al., 2023).

5.2. Impact

On a social level, the research has enabled students, teachers, and school stakeholders to experience an innovative and engaging approach. It has led to the creation of a unique writing rehabilitation program, accompanied by an explicit, individualized, flexible, and intensive training system. These impacts reveal the potential of students with dysorthographia, which becomes apparent as we make connections between their marked progress and the interventions offered to them by the people involved. For training, findings asl inform the design of professional-learning systems: (a) resource-rich initial training (implementation guides, observation rubrics, and video exemplars), (b) in situ accompaniment (modeling, supervised practice, and BST), and (c) iterative observation–feedback cycles targeting explicit teaching routines and the TTS/word-prediction workflow (phonological plausibility checks and morpho-orthographic constraints). These features are associated with higher implementation fidelity and improved instructional quality (Fixsen et al., 2005; Kretlow & Bartholomew, 2010; MacArthur, 2013). In addition, school teams (special educator, classroom teacher, and school leadership) re-experienced collaboration by aligning teaching–learning–assessment in writing with AT use; such local governance proved decisive for preserving dosage and program integrity, in line with evidence that intensity moderates outcomes (Galuschka et al., 2019; Hall et al., 2023).

On a scientific level, this study stands out for the evaluation of the program’s impact and implementation, contributing to the advancement of knowledge in education and methodological innovation. Its highlight is the use of a mixed methodology that ensures a dynamic between teaching, learning, and evaluation, promoting optimal progress for students with difficulties and specialized teachers. This research, the result of an integrated approach to intervention, training, and pedagogical adaptation, aims to transform practices to enable students to succeed in diverse ways. The study helps operationalize implementation fidelity as a multidimensional construct in AT-supported writing, emphasizing the importance of dosage and of reporting beyond adherence to include instructional quality and participant responsiveness—elements advocated in the implementation and literacy studies (Gresham et al., 2000; Hall et al., 2023).

5.3. Future Research Directions

To move beyond the limited qualitative corpus and better explain real-time adjustments, a subsequent study should multimodalize evidence (audio/video of sessions and co-planning, interviews, and student artefacts) and cover all fidelity dimensions—beyond adherence and dosage, include instructional quality and participant responsiveness—in line with recent fidelity-reporting guidance in literacy and WWC standards. This extension would strengthen the traceability of practices and the interpretation of effects (Dahl-Leonard et al., 2023).

On the professional-learning side, a structured training-with-accompaniment design combining model–rehearsal–feedback (BST) and in situ coaching is a priority to raise and stabilize implementation quality beyond the study window. An iterative design-based (research-development) approach would set fidelity targets for each explicit instruction routine (clear instructions, modeling, and informative feedback) and monitor them over time (Sawyer, 2025; Lembke et al., 2024). A longitudinal follow-up (3–6–12 months) is recommended.

To reduce the observed variability in engagement, teams should co-plan with the classroom teacher a menu of tasks (genre options, graded scaffolds) and in-class supports explicitly linked to ongoing activities, then estimate their impact (quasi-experimental) on engagement and transfer. Evidence on assistive technologies (ATs) confirms that benefits depend strongly on context and teacher preparation (Fernández-Batanero et al., 2022; Almgren Bäck et al., 2024).

Two complementary strands follow. First, an earlier intervention integrating text-to-speech (TTS) for word identification should combine outcome measures (accuracy, fluency, and comprehension) with process measures tracing when and how TTS and word prediction are orchestrated across the explicit cycle (planning–drafting–revising). Second, a dyslexia/dysorthographia identification pathway co-designed with school teams would both broaden access and train/accompany staff for earlier screening. Both strands directly address externality limits and the need to instrument AT orchestration in authentic settings (Keelor et al., 2023; Lembke et al., 2024).

In summary, articulating teaching and training to support students with dyslexia–dysorthographia, this research represents a significant innovation. However, the single-case protocol and transformative sequential design limit generalization. Future research should extend assessment–intervention systems to support students and teachers over time and align practices more closely with current knowledge.