Assessing the efficacy of a Tier 2 early intervention for transcription skills in Spanish elementary school students

The main objective of this study was to evaluate the efficacy of a Tier 2 intervention within the context of the Response to Intervention (RtI) model implemented by Spanish first- to second-grade primary school teachers to improve at-risk students’ transcription skills. Teachers were instructed in the administration of a writing curriculum-based measure composed of 5 isolates measures (allographs, dictated words with arbitrary spelling or non-ruled, dictated words with rule-based spelling, dictated nonsense words, and dictated sentences) to identify at-risk students and to monitor their progress; and in the implementation of a systematic and explicit instructional program to improve transcription skills in at-risk students. A total of 164 at-risk students from 12 schools from the Canary Islands participated in the study. Teachers were provided with a support system that included a web-based training program for writing. Implementation fidelity was analyzed using direct observations and self-reports. All students were assessed three times during the academic year. In a hierarchical linear growth modeling procedure, differences in growth rates of transcription skills were compared between students who received the intervention, those who did not, and their typically achieving peers. Additionally, the effects of the intervention in the experimental at-risk group were also analyzed, differentiating between those who left the risk status and those who remained at-risk. Children at-risk in the intervention condition appeared to benefit more in handwriting skills than at-risk children in the control condition, but to a lesser extent in spelling accuracy. Nevertheless, improvements in both transcription skills had only been made in the experimental risk group that had left the risk situation.


Introduction
Over the last decade, researchers' attention has been directed towards the study of early identification and intervention for writing (Berninger et al., 2006a(Berninger et al., , 2006b. In the early grades, writing development primarily takes the form of learning to write letters, spell words, and compose short texts; by fourth grade, pupils begin to apply more extended writing skills as a means of learning about a topic (Berninger et al., 1995(Berninger et al., , 2006a(Berninger et al., , 2006b. Therefore, beginning writers have to gain automaticity in the low-level skills of putting language to paper, which include retrieving phoneme-grapheme correspondences, controlling motor movements to make legible letter forms, and retrieving orthographic word representations (Graham et al., 1997;Kim et al., 2011). Difficulties with handwriting, grammar, punctuation, and spelling may further impact written expression (Saddler & Asaro-Saddler, 2013).
It has been suggested that a Response to Intervention (RtI) approach is probably the best opportunity to improve education for all students, in general, and in particular for those students at risk of experiencing learning disabilities (LD) (Tilly, 2006). In the United States, the Individuals with Disabilities Education Improvement Act (IDEA, 2004) emerged as an alternative to the discrepancy criterion, prioritizing the implementation of early intervention models for LD over the "wait-to-fail" model. As a consequence, the RtI model emerged with the goal of assessing the student's response to the intervention and adjusting pedagogical strategies through a multilevel prevention system organized in three levels or tiers (i.e., Tier 1, 2, and 3). The quality and intensity of the intervention levels increase over 1 to 3 tiers. Apart from this multi-level prevention system, the RtI model includes other three components: universal screening, progress monitoring, and data-based decision making ("National Center on Response to Intervention Essential Components of RTI-A Closer Look at Response to Intervention," 2010). Two RtI approaches for prevention and intervention of academic and social problems have emerged: standard-treatment protocols and problem-solving models (Lerner & Johns, 2012). In the present study, we followed a standard-treatment approach that requires using the same empirically validated treatment for all children with similar problems in a given domain (Fuchs et al., 2003).
In Spain, the detection and early intervention of LD have been included as a main priority in the Spanish educational legislation (Ley Orgánica 8/2013, de 9 de Diciembre, para la Mejora de la Calidad Educativa [Organic Law 8/2013, of December 9, for the Improvement of Educational Quality], 2013). This recognition in the Spanish legislation opens up the possibility that in Spain the criteria for identifying LD could be addressed on the basis of models based on RtI model. In fact, recent studies in Spain have shown the effectiveness of Tier-2 or secondary intervention in early reading skills (Jiménez et al., 2021), and early math skills (de León et al., 2021). In both studies, the at-risk children in the intervention condition showed a higher growth compared to at-risk children in the control 1 3 Assessing the efficacy of a Tier 2 early intervention for… condition. In early writing skills, a study provides evidence that parent-supported Tier 2 intervention was effective in bringing struggling students' performance back into line with their peers (Arrimada et al., 2018(Arrimada et al., , 2022. In the present study, we investigated whether Spanish first to second graders at risk of writing failure benefited from a Tier 2 intervention based on transcriptions skills, implemented by in-service teachers. Although the effectiveness of the RtI model in improving students' performance has been widely demonstrated (Burns et al., 2005), the presence of varied results in student performance after the intervention suggests that teacher training is a critical element in the effectiveness of the model (O'Connor & Freeman, 2012). The presence of the four essential components (i.e., multi-level prevention system; universal screening; progress monitoring; data-based decision making) might not be a guarantee of success per se. One aspect to highlight in the research in Spain is that the implementation has been accompanied by the training of teachers (Jiménez et al., 2021a). According to Castillo et al. (2016), teachers should be provided with professional development to achieve the following specific skills: (a) participate in decision-making based on data; (b) identify and implement practices based on empirical evidence across the different educational levels; (c) collect data to evaluate the fidelity and effectiveness of evidence-based practices; and (d) use the technology to analyze and extract data. In short, training teachers at the different levels of the RtI model is an essential requirement for its success (Berkeley et al., 2009;Fletcher & Vaughn, 2009;Mastropieri & Scruggs, 2005;Philippakos & Fitzpatrick, 2018;Reddy & Dudek, 2014).

The importance of transcription skills in learning to write
According to the 'Simple View of Writing' (Juel, 1988;Juel et al., 1986), and the 'Not-So-Simple View of Writing' (Berninger & Winn, 2006), transcription skills (i.e., translating words, sentences, and higher levels of discourse into print, which requires handwriting and spelling) interact with high-order, executive and self-regulatory function to produce writing through its planning, reviewing, and revision phases. From a developmental perspective, transcription and text generation (i.e., selecting words and formulating sentences, paragraphs, and longer units of discourse) predominate early writing, as executive functions are not prominent until students attain self-regulation (Berninger et al., 2006a(Berninger et al., , 2006b. When students struggle with transcription skills (i.e., handwriting or spelling) they have limited resources to devote to other processes like word selection and sentence construction involved in text generation. Handwriting and spelling are two different abilities and it is not surprising that they rely on different cognitive processes. This theoretical foundation has been supported by factor analyses indicating that handwriting and spelling load on separate but correlated factors (Berninger, 2000).
Competence in handwriting is usually described in terms of legibility and speed (Graham & Weintraub, 1996;Graham et al., 1998). Graham et al. (1997) suggested that handwriting is critical to the generation of creative and well-structured written text and has an impact not only on fluency but also on the quality of composing.
A review of research undertaken over the last 10-15 years (Berninger., 1998;Berninger et al., 1994Berninger et al., , 2006aBerninger et al., , 2006b has investigated the role of handwriting in writing and it has been established that handwriting is far from a purely motor act, and orthographic and memory processes (i.e., the ability to recall letter shapes) contribute more to handwriting than do motor skills. Therefore, the role of orthographic motor integration and automaticity in handwriting is now seen as of key importance in composing.
Proficiency in spelling is the other critical component of writing in the developmental models of writing, and it is considered a lower level mechanical skill that is necessary to allow higher level composition processes such as generating content and planning during writing (Graham, 1990). Spelling component of transcription is the ability to use sound-symbol relations and orthographic rules to write words using conventional spelling. That is, it refers to the ability to correctly map graphemes to phonemes in dictation of single real words and pseudowords (Sun-Alperin & Wang, 2011).
In sum, both skills (i.e., handwriting and spelling) are correlated and automation of both is necessary for successful transcription (Gosse et al., 2021). The longitudinal study done by Gosse et al. (2021) provided a comprehensive picture of the development and the longitudinal relationship between spelling, handwriting speed, and handwriting quality at the word level. Based on correlational analyses, there is an association between spelling accuracy and handwriting speed in primary grades throughout the development process. Successful transcription skills will give students' the power to master written expression. On the contrary, impairments in transcription skills will not only limit the content and discourse aspects of writing but may also be the cause of a lack of motivation for the writing activity (Hidi & Boscolo, 2006).

Evidenced-based approaches to intervention in writing
A number of evidence-based interventions have been proposed to address written expression through its different developmental stages. Some of these interventions have been devoted to high-level skills for writing composition such as self-regulation strategy (Burke, 2017;Harris et al., 2012Harris et al., , 2015Johnson et al., 2013) while others have been focused on low-level skills such as transcription skills (Santangelo & Graham, 2016). In this sense, it is important to ensure students develop legible and fluent handwriting. Santangelo and Graham (2016) conducted a meta-analysis that examined true-and quasi-experimental intervention studies conducted with K-12 students to determine if teaching handwriting enhanced legibility and fluency and resulted in better writing performance. When compared to no instruction or nonhandwriting instructional conditions, teaching handwriting resulted in statistically greater legibility (ES = 0.59) and fluency (ES = 0.63). Additionally, Graham and Santangelo (2014) conducted a comprehensive meta-analysis of experimental and quasiexperimental studies to answer the controversial issue raised by theorists who question the effectiveness of teaching spelling directly and systematically. They provided support for the contention that formal spelling instruction is efficient, at least when compared to methods common to the spelling is caught approach. For another study, Graham et al. (2018) analyzed the impact of supplemental handwriting and spelling instruction on learning to write in an experimental study with first grade students who were not acquiring these skills as rapidly as their classmates. Thirty students (16 boys, 14 girls) were randomly assigned to a handwriting and spelling instructional condition or a phonological awareness instructional control condition. Students in each condition received 16 h of one-on-one instruction. The treatment condition was designed to improve children's handwriting fluency and spelling accuracy. In comparison to students in the phonological awareness control condition, students who received supplemental handwriting and spelling instruction made greater gains in handwriting fluency, handwriting legibility, and spelling accuracy.
However, since most of the current empirical evidence on training transcription skills cited above comes from English, a language with deep orthography, it is very relevant to investigate whether similar results are found in a language with a shallow and fine-grained orthography such as Spanish. In fact, comparative studies show differences in the speed at which basic spelling mechanisms are acquired, demonstrating that learning is faster in consistent orthographic systems than in inconsistent systems (Alegría & Carrillo, 2014;Carrillo et al., 2013). To date, several studies have examined the efficacy of interventions to support the development of writing skills in Spanish struggling writers who received twice-weekly, parent-delivered training in transcription skills (handwriting and spelling) (Arrimada et al., 2018). These authors found some evidence that Tier 2 intervention resulted in improved performance on the phase-end handwriting speed task, but Tier 2 intervention did not appear to result in improved performance, relative to peers, on the spelling task. Arrimada et al. (2022) included a post-Tier 2 follow-up in their second study. The only measure in the Tier 2 sample that had not improved in Phase 2, relative to the comparison group, was the accuracy of the spelling in the students' texts. In the study presented here, a web-based training program for writing was offered to inservice teachers, and assessment and instruction were provided by the in-service teachers within the school context rather than relying on parents for Tier 2 implementation (Arrimada et al., 2018).

The present study
This study aims to explore the effectiveness of response-to-intervention approach for providing support for students who struggle to learn to write in their first and second year of formal education. In the present intervention study, we had two goals. On the one hand, as research started by providing professional development to in-service teachers, it was necessary to analyze the effects of professional development for transcription instruction followed by teacher implementation at Tier 2, with followup observation and online support. On the other hand, the purpose of this study was to determine whether first and second graders would benefit from the intervention in terms of improving their transcription skills (i.e., handwriting and spelling). Specifically, this study addressed the following research questions: Research Question 1: Will in-service teachers gain subject matter knowledge about how and what to teach early writing skills as part of RtI?
Research Question 2: Are there significant differences in the growth of handwriting skills (i.e., measure of fluency in allograph selection) and spelling skills (i.e., word and sentence dictation measures) of at-risk first and second graders who have received a Tier 2 RtI intervention, and no intervention?

Design
A quasi-experimental design with school-based assignment was used to conduct a Tier 2 intervention to improve the transcription skills of at-risk students.

Participants
A total of 12 schools (three subsidized schools and nine public schools) were invited to participate in this study through a collaboration agreement between La Universidad de La Laguna and the Department of Education and Universities of the Canary Islands Government (Boletín Oficial de Canarias, 2017). A sample of 901 first to second graders participated in the study. Students' risk status was established using the 25th percentile of the composite measure of the Indicators of Basic Early Writing Skills (IPAE, Jiménez & Gil, 2019), calculated by averaging unweighted standardized subtest scores (Gil et al., 2020). A sample of 164 at-risk students (females = 77; males = 87; M age = 6.86, SD age = 0.65), from 1st (subsidized schools n = 41; public schools n = 32) to 2nd (subsidized schools n = 43; public schools n = 48) grade, was considered eligible for the Tier 2 intervention. The students were randomly assigned into two groups, an experimental at-risk group (ER-G; n = 71), and a control at-risk group (CR-G; n = 93). In addition, the typically achieving peers (TAP-G) group was composed of 45 students who were also randomly chosen (females = 26; males = 19; M age = 6.70, SD age = 0.47), from 1st (subsidized schools n = 10; public schools n = 17) to 2nd (subsidized schools n = 9; public schools n = 9) grade. There were no significant differences between the three groups in age and gender, F (2, 206) = 3.08, p = 0.56; χ 2 (2) = 2.20, p = 0.33, respectively. Additionally, the effects of the intervention in the experimental at-risk group were also analyzed, differentiating between those who left the risk status (ER-G (OR); n = 30) and those who remained at-risk (ER-G (RR); n = 41) once the intervention ended. Ten volunteer interventionist teachers were in charge of implementing the Tier 2 intervention with students identified as students at risk for writing failure.

Fidelity of implementation
Interventionist teachers' fidelity of implementation was conducted for the screening and intervention. These fidelity observations were conducted using a checklist in its two modalities: (1) teachers' self-report and (2) research assistants' direct observations. The fidelity of using the Indicators of Basic Early Writing Skills (IPAE, Jiménez & Gil, 2019) was assessed using the Fidelity Assessment Scale (FAS), and the fidelity of using the Instructional Activities for Early Writing Improvement (IAEWI,  was assessed using the Fidelity of the Intervention Scale (FIS). Both will be described in more detail in the measures section. Seven research assistants were trained in three sessions of four hours each in the administration of the IPAE and IAEWI and the use of the FAS and FIS scales. Each research assistant was in charge of evaluating specific teachers through direct observation and giving them feedback to improve the quality of their implementation. Every teacher was observed three times for the administration of the IPAE, and five times (monthly) for the IAEWI usage.
An average score was calculated for each teacher for the self-report and direct observation modalities of the fidelity of the assessment and intervention scales (i.e., FAS and FIS). The overall mean implementation fidelity rating across time and conditions was high: (a) self-report (

Screening and progress monitoring measures
Students were administered five writing screening and progress monitoring measures included in the IPAE for 1st and 2nd grade. The IPAE includes three alternate forms (i.e., fall, winter, and spring) of each measure: Allographs (ALLO), dictated words with arbitrary spelling (DWAS) (i.e., non-ruled), dictated words with rulebased spelling (DWRBS), dictated nonsense words (DNW), and dictated sentences (DS). These measures have shown adequate indexes of reliability, validity, and diagnosis accuracy (Gil et al., 2020). ALLO is a 1 min timed measure that includes the 27 letters of the alphabet. Students had to write all the letters in lower case as accurately and as quickly as possible. All allographs written properly and corresponding to the upper-case letter presented were considered as correctly copied letters. The raw score was the total number of letters correctly written. DWAS spelling includes 20 familiar words with arbitrary spelling. A word is considered correctly written if its spelling is correct. DWRBS includes a total of 20 words. One point is given to the student when they write the spelling rule correctly. DNW includes 20 nonsense words created taking into account their length and syllable structure. A score is obtained from the total number of nonsense words with the correct graphic representation of the sounds. Finally, for the DS task, sentences are dictated to the student, containing words used both in arbitrary spelling and rule-based spelling. A word with no spelling mistakes was considered a correct word. For all the dictated tasks, the raw score is the total number of words correctly written.

Web-based training (WBT) program
Teachers were enrolled in the TRAZO WBT program . TRAZO aims to provide teachers with knowledge and strategies to carry out the RtI model effectively in their classrooms when teaching writing skills. TRAZO is a 120 h WBT program organized in a gap of 16 weeks (i.e., from September to December) that includes four modules: theoretical, intervention, assessment, and experiences. The theoretical module includes twelve tutorials around six themes (i.e., Learning theories and beliefs in the teaching of writing; What is writing? Handwriting/typing, spelling, and writing sentences and texts; Prevention and RtI Model). Likewise, basic principles of explicit instruction to deal with students atrisk of LD in writing are provided. The intervention module provides information about how to teach within the RtI Tier 2 and use the instructional activities for early writing improvement (IAEWI). The assessment module is based on the usage of the CBM IPAE. Finally, in the fourth module, experiences and video recordings are presented on implementing good teaching practices when teaching children about vocabulary, alphabetical knowledge, phonological awareness, handwriting, spelling, and text production. In our study, teachers were able to ask questions using the forums included in each module, along with all the intervention program. In each of the theoretical training tutorials, teachers had to repeat the following steps: 1) answer the previous knowledge questionnaires on that section (i.e., pretest); 2) see at least three times each theoretical tutorial; 3) answer the evaluation questionnaires of each section (i.e., posttest).

RtI website
Teachers used the RtI website (http:// webrti. ull. es/) to identify students' risk status and monitor their progress. They introduced the raw score obtained for each student in each IPAE measure. The RtI website identifies students' risk status using the unweighted mean of the standardized measures and a specific cut-off (Gil et al., 2020). For ease of interpretation, the RtI website identifies with a red dot those students who are at-risk, and with a yellow dot those who are at low-achievement.

Instructional activities for early writing improvement (IAEWI)
The IAEWI ) is a Tier 2 writing intervention designed based on the principles of systematic and explicit instruction and the use of modeling, scaffolding, and feedback (Gersten et al., 2009). It aims to help students at-risk achieve the skills needed for early success in writing. The materials have been designed both for students and teachers to guide teachers' instruction and guarantee that all teachers are conducting the same instructional methodology.

Materials
Teachers are supported with different materials (i.e., teachers' manuals and students' workbooks) to use in their classrooms. The teachers'manual includes several sheets with all the content and activities they can use in their classrooms. They will choose, from the proposed activities, those that better fit the students' needs. The teachers' manual is directly related to the students'workbooks. Therefore, each of the activities included and explained by the teachers'manual will have a worksheet associated with it in the students' workbook. They will be related by using the same code (e.g., activity 1.1). Teachers' sheets (included in the teachers' manual) explain how to implement each one of the students' sheets step by step (see an example in Appendix A and Appendix B).

Model of instruction
As was set out before, the IAEWI is based on the use of systematic and explicit instruction (Gersten et al., 2009). In order to guarantee that all teachers followed the same instruction, teachers' sheets were organized accordingly with the following instructional sequence to carry it out: 1. Activity presentation (give to the student information about the content of the lesson before it begins); 2. Modeling (provide the student examples of how to represent domain content); 3. Group guided practice (every child watches a teacher model how to solve the task. Students ask questions until they are able to understand. Then, the teacher turns over the task to the students); 4. Corrective feedback (students receive feedback aimed to improve their performance); 5. Individual guided practice (each student solves the task alone in their own sheet); 6. Corrective feedback (students receive feedback aimed to improve their performance); 7. Independent practice (each student solves the task alone in their worksheet); and 8. Corrective feedback (students will receive feedback aimed to improve their performance).
The IAEWI lessons were designed to be administered on a small group basis (3-5 students) in about 35-40 min, 3 to 4 days per week. Session length was modified for this study to meet schools timetables, lasting between 45-55 min per session. However, the integrity of the methodology proposed was preserved.

Contents
The intervention sequence employed for the IAEWI for first and second grade comprised two components: handwriting and spelling.
The handwriting component includes 41 lessons organized into 4 sections: postural control, phonemic awareness, alphabetic knowledge and writing with fluency (i.e., write legibly and quickly). The development of a good posture for writing included sitting down well, paper position and pencil grip. At the sub-word level, students were instructed in phonemic awareness and alphabetic knowledge using images and games (e.g., identify words with specific phonemes, say aloud words with a specific phoneme, naming the phonemes in a word, naming a letter and it´s phoneme…). Once they were able to associate a phoneme with its grapheme, they were instructed in the recognition (e.g., identify a letter with a mix of letters, identify a letter within a word…) and writing of the worked letter, both in upper-and lower-case format. Writing of letters was performed firstly isolated, repeating the same letter several times. Students worked with letter tracing sheets before to be able to write the letters on their own. After that, children were able to complete words using the worked letter. Letters were organized following a preestablished sequence. At the word level, students were asked to copy words, write words from an image, write dictated words or complete a sentence with a word, among other activities.
The spelling component includes 43 lessons organized into 4 sections: morphology, regular spelling, ruled based spelling, and non-ruled spelling. Morphology activities were focused on word families or word formation (e.g., find words from the same family, look for words from the same family as a given word, complete the words, complete the sentences…), prefixes and suffixes (e.g., complete words using prefixes and suffixes, classify words…), verb conjugation (e.g., complete sentences using the present, the past or the future of the given verbs, transform verbs from present to future, from past to present…), and number and gender (e.g., complete sentences, write words in singular, write words in plural…).
Regular spelling activities aimed to develop phonemic awareness (e.g., identify sounds, segment a word according to its sounds…), alphabetic knowledge (i.e., spell a word, name the first letter of a given word…), writing words (e.g., copying words, dictated words, writing words that begin with a specific letter…), and writing sentences.
Rule-based spelling activities were based on identifying and automatizing spelling rules through games, writing words, and writing sentences. Punctuation marks were also included through activities focused on the use of capital letters, full stops, question marks, exclamation marks and commas.
Finally, non-ruled spelling activities worked the use of homophones (e.g., /b/-/v/, /h/, and /ll/-/y/) (e.g., choose the correct homophone for each word, match each homophone with its meaning, write words based on an image, dictation of sentences…).
The words included in the program followed different syllable structures (i.e., CV, VC, CCV, and CVC) and length (i.e., from 3 to 9 letters per word).

Intervention sheet
Teachers were given a sheet to daily register students' attendance, and components or activities worked out during the sessions.

Scales for the fidelity of the implementation
The fidelity of the assessment scale (FAS) and fidelity of the intervention scale (FIS) are 12 items 5-points Likert-type scales (0 = none to 4 = all) designed to analyze the fidelity of the implementation. The FAS assesses the accuracy with which teachers follow the administration instruction of the IPAE; while the FIS assesses the accuracy with which teachers implemented the IAEWI. Both the FAS and FIS were designed in two modalities: self-report and direct observation. Note that both modalities included the same items but the self-report is present in 1st person, while the direct observation scale is presented in 3dr person (e.g., "I do" or "The teacher does"). Reliability indexes of those scales have been reported in previous studies (see de León et al., 2021;Jiménez et al., 2021b)

Procedure
Teachers' theoretical training was carried out from September to November, students assessments were performed in November (i.e., fall), February (i.e., winter), and May (i.e., spring), and the intervention was conducted from January to May.
All measures were group-administered in the classrooms by four research assistants and 10 teachers. Each class was assigned to a single person, either a research assistant or teacher, guaranteeing that the same person assessed the students throughout the school year. Research assistants and teachers were trained to administer and correct the IPAE in three sessions of four hours each, before each of the assessments (i.e., fall, winter, and spring). Each session included a theoretical presentation, roleplaying practice, and solving doubts and questions final section.
Interventionist teachers were in charge of the assessments in their schools, identifying students at risk and selecting the students joining the intervention. The requirements to enter the intervention group were the following: (1) the group must be composed of three to five students with similar writing difficulties, (2) students' selection must be made firstly, according to the screening results, and secondly, in agreement with the management team and the teaching staff, (3) students diagnosed with special needs could not be part of the intervention group.
Students' screening was performed using the IPAE in the fall (November), winter (February) and spring (May). Teachers choose those students in higher need in the fall to be included in the intervention group. The intervention was carried out from January to May, lasting approximately 17-weeks, with a recommended frequency of three to four sessions per week and 45 to 55 min per session (excluding holiday periods such as Christmas, carnivals, and Easter week). Winter and Spring assessments were used both to identify students' risk status, but also to assess students' progress monitoring throughout the school year both in the control and experimental groups. The winter and spring measures were also used to decide whether the students included in the experimental group were out of risk status and, therefore, could leave the intervention group, and to decide which new students at risk should be included in the intervention group. For those children who received the intervention outside the classroom, once the session ended, they returned to the ordinary classroom to receive the core program.
Progress monitoring was conducted for the intervention group in the months between the screening assessments (i.e., January, March, and April). This information was only used for the interventionist teachers to modify their intervention accordingly to students' improvements.
Teachers were trained with the TRAZO WBT program for 16-weeks (i.e., from September to December) to guarantee a correct assessment and implementation of Tier 2 of the RtI model. Moreover, teachers attended four face-to-face training sessions throughout the school year. Every session lasted four hours and was carried out every three months (i.e., beginning of the school year, and before each assessment time). In those sessions, teachers received training in the IPAE administration and the use of the IAEWI.
A General Linear Model (GLM) repeated-measures ANOVA was used to explore teachers' learning improvement (i.e., time: pretest-posttest) throughout the TRAZO WBT program theoretical module, which included six themes. The sphericity assumption was checked with Mauchly's sphericity test.
The efficacy of the Tier 2 intervention was tested by examining changes in the screening measures raw scores over time using Hierarchical Linear Modeling analysis (HLM, Raudenbush & Bryk, 2002). Several HLM analyses were conducted in this study to examine differences between the typically achieving peers group, the experimental at-risk group, and the control at-risk group, as well as to distinguish between those no longer at risk and those who are still at risk after the intervention. Three-level models were estimated with repeated measures of student outcome at the first level, differences among students at the second level, and schools at the third level. A stepwise method was followed (Peugh & Heck, 2017;Raudenbush & Bryk, 2002) to choose the most parsimonious model. The likelihood-ratio test and deviance statistics were used to choose the HLM model with the best fit (Raudenbush & Bryk, 2002).

Teacher knowledge
Descriptive statistics analyses for each one of the themes included in the TRAZO WBT Theoretical module are presented in Table 1. GLM repeated-measures

Intervention efficacy
Students were grouped in groups of 3 to 5 students, and received an average of 26.81 (SD = 17.64, minimum = 4, maximum = 67) intervention sessions, with a session length varying from 15 to 30 min (12.1%), 30 to 45 min (81%) and 45 to 60 min (6.9%). Regarding the components worked out through the intervention, 95.1% of students were instructed in handwriting skills, 48.7% in morphology, 47.5% in regular spelling, 17.1% in arbitrary spelling (non-ruled), and 21.7% in rule-based spelling. Descriptive statistics for each screening measure are presented in Table 2.
In order to analyze whether the words used in the assessment and intervention were equivalent in terms of word length, word frequency, and orthographic neighbors, these lexical properties of words were analyzed through EsPal, a web repository of Spanish words (Duchon et al., 2013). T-tests were performed for independent sets of words (i.e., assessment versus intervention) to compare frequency, frequency of substitution neighbors, and length. In addition, in these comparisons, t-tests were performed for each of the alternative forms of assessment (i.e., fall, winter, and spring) when compared to the intervention. Descriptive statistics are shown in Table 3.

Table 2
Descriptive statistical analysis for the screening measures All descriptive data represent accuracy but for the Allograph task, which measures fluency. ER-G = Experimental at-risk group; CR-G Control at-risk group; TAP-G Typically achieving peers group Measures HLM analyses were conducted to explore differences in the rate of improvement between the different groups. Table 4 shows the results obtained after performing the HLM models of each IPAE measure, following the step-by-step procedure (Peugh & Heck, 2017;Raudenbush & Bryk, 2002), throughout the school year. The residual analysis revealed that some models did not meet the assumption of normality; therefore, data transformation was required for the allograph task. Figure 1 shows a graph with the growth in the IPAE measures of the different groups. The unconditional baseline model (i.e., model 0) showed that the ICC value due to the school ranged from 0.54 to 0.68, highlighting the need to control for variability among schools.
Once the group variable was included in the model (i.e., model 1), the analysis revealed significant differences (p < 0.001) between the groups at the beginning of the school year (i.e., intercept). More specifically, it is observed how the typically achieving peers group (TAP-G) obtained a higher starting point in all IPAE measures, while the ER-G obtained a lower starting point than the CR-G on all tasks, except dictated nonsense words.
Regarding the growth slopes, the ER-G revealed greater growth coefficients than the TAP-G. The interaction of the ER-G and TAP-G slopes (i.e., Time x ER-G) was significant (p < 0.01) and positive for the allograph measure. However, the interaction of the CR-G and TAP-G slopes (i.e., Time x CR-G) was not significant. That means that the ER-G showed significantly greater growth than the TAP-G (i.e., 2.72 vs. 4.22 points per measurement time), while the CR-G showed a decrease (i.e., 2.1 points per measurement time), although not significant.    Table 5 shows the results obtained after performing the HLMs following the step-by-step procedure. In this case, the analysis was carried out to distinguish between those who are no longer at risk (ER-G(OR)) and those who are still at risk (ER-G(RR)) after the intervention. As a result of the residual analysis, some models failed to meet the assumption of normality; consequently, data transformation was required for the allograph task. Figure 2 shows a graph with the growth in the IPAE measures of the different groups. The unconditional baseline model (i.e., model 0) showed that the ICC value due to the school ranged from 0.54 to 0.68, highlighting the need to control for variability among schools.
Once the group variable was included in the model (i.e., model 1), the analysis revealed significant differences between the groups at the beginning of the school year (i.e., intercept). More specifically, it is observed how the TAP-G obtained a significant (p < 0.001) higher starting point in all IPAE measures than any of the other groups.
Regarding the growth slopes, the interaction of the ER-G (OR) and TAP-G (i.e., Time x ER-G (OR)) was significant (p < 0.05) and positive. That means that the ER-G (OR) showed significantly greater growth than the TAP-G in all IPAE tasks (i.e., allographs, 7.02 points per measurement time; dictated words Assessing the efficacy of a Tier 2 early intervention for… with arbitrary spelling, 3.18 points per measurement time; dictated words with rule-based spelling, 3.92 points per measurement time; dictated nonsense words, 3.18 per measurement time; and dictated sentence, 4.13 points per measurement time). The interactions of the ER-G (RR) and TAP-G slopes (i.e., Time x ER-G (RR)) were significant (p < 0.05) for some measures but negative. The ER-G (RR) showed a decrease in dictated words with arbitrary spelling and dictated nonsense words. Additionally, the interaction of the CR-G slope and the TAP-G slope (Time x CR-G) was significant only for the dictated sentences measure (p < 0.05), but negative, the CR-G showing a decrease in the dictated sentences measure.

Discussion
This study examined the effectiveness of a Tier 2 early intervention for transcription skills with first and second grade Spanish students who were not acquiring these skills as rapidly as their classmates. The treatment condition was designed to improve children's handwriting fluency and spelling accuracy. As the research began by providing professional training, we will first discuss the results Fig. 2 Slopes of growth predicted through the Hierarchical Linear Model in the IPAE scores as a function of group. TAP-G = typically achieving peers group; ER-G (OR) = Experimental at-risk group (out of risk status); ER-G (RR) = Experimental at-risk group (remains at-risk); CR-G = control at-risk group regarding the effects of teacher training, and then we will do the same for the effects of Tier-2 intervention found in this study to improve at-risk students' transcription skills.

Will in-service teachers gain subject matter knowledge about how and what to teach early writing skills as part of RtI?
RTI is dependent on several factors, which include the professional development of teachers (Castillo et al., 2016). This is particularly important in the writing area since teachers report not receiving enough training on teaching writing (Phillipakos & Moore, 2017; Troia & Graham, 2016). With regard to the effects of web-based training on teachers' knowledge, the role of the teacher has been defined as an essential element in student learning (Heck, 2009;Weschke et al., 2011), as well as the importance of teacher training for effective teaching Myrberg et al., 2019). In this sense, the development of students' writing could be guaranteed, in part, with adequate teachers training. Results indicated that all teachers benefited from the WBT (i.e., TRAZO). Hence, all of them had the competencies to teach early writing skills (i.e., handwriting and spelling) to at-risk students, and to use the CBM IPAE, and the IAEWI materials adequately. These results are consistent with previous studies. For instance, McMaster et al. (2020) demonstrated that after professional development, teachers gained critical knowledge and skills in early writing (p < 0.000; g = 2.92). Furthermore, when children experience difficulties learning to write is essential to support them with explicit instruction (Berninger, 2009). As well as at the elementary level, in which the provision of explicit and systematic instruction is aforementioned. Thus, teachers will need to provide high-quality instruction aligned to meaningful research which can only be achieved by providing professional development to teachers.
Are there significant differences in the growth of handwriting skills (i.e., measure of fluency in allograph selection) and spelling skills (i.e., word and sentence dictation measures) of at-risk first and second graders who have received a Tier 2 RtI intervention, and no intervention?
In the present study, we followed a standard-treatment approach. The advantage of this approach is a clear lack of ambiguity, which streamlines decision-making for school staff. With scripted protocols, everyone knows what to do; thus, treatment fidelity can be maximized (Fuchs et al., 2010). To guarantee that in-service teachers implemented the Tier 2 of the RtI model with fidelity, they counted with a support system that included a formative program to ensure that they have the necessary knowledge and the skills, and forums and feedback to solve their doubts. Implementation fidelity is broadly defined as the extent to which an intervention is delivered as planned (Wolgemuth et al., 2014). In a review conducted by Capin et al. (2018), nearly 90% of studies that reported treatment fidelity data included an indicator of treatment adherence. This dimension was the only dimension of treatment fidelity identified as an "essential quality indicator" for group research in special education (Gersten et al., 2005, p. 152). Results found in this study revealed that both teachers' self-reports and research assistants' direct observations were coincident across time and conditions, indicating that teachers could implement the Tier 2 of the RtI model with adequate fidelity indexes. Therefore, the quality of both assessment and intervention administration was guaranteed.
Previous studies provided promising evidence that explicit systematic instruction in handwriting and spelling improves student performance on these specific skills. Students in the experimental condition, that received small group instruction by trained teachers using instructional procedures for teaching handwriting and spelling, were successful to support struggling writers in elementary grades (e.g., Graham et al., 2000;Jones & Christensen, 1999;Wanzek et al., 2017). Most of these studies have been conducted in languages with opaque orthography. Our expectation has been to find similar results in a language with a shallow and fine-grained orthography such as Spanish.
Many of the activities included in the IAEWI program have been shown to be effective to improve transcription skills in first and second grade in the above studies. So, for instance, Jones and Christensen's (1999) found that instruction focused on letter formation, corrective feedback, and letter writing fluency was effective. In a later study conducted by Graham et al. (2000), the handwriting instruction focused on correct, accurate writing of letters through letter name learning, letter formation through teacher modeling, letter tracing, and letter practice through copying. Additionally, Wanzek et al. (2017) gave specific instruction on handwriting fluency. When it comes to spelling instruction, Graham and Santangelo (2014), in their metaanalysis, found that activities such as teaching letter-sound relationships, spelling patterns, the use of these relationships and patterns to spell or build words, as well as instruction in spelling words, that influenced spelling performance.
Our findings seem somewhat consistent with the ones of the studies cited above. We hypothesized that a Tier 2 or secondary intervention would be able to cause differences in the growth rate of handwriting and spelling skills. This hypothesis was partially supported by the data. In comparison to students in the control condition, students who received supplemental handwriting and spelling instruction made greater gains in handwriting fluency, but to a lesser extent in spelling accuracy. This finding is coincident with previous studies conducted in Spain. For instance, Arrimada et al. (2018) also found some evidence that Tier 2 intervention resulted in improved performance on the handwriting speed task, but Tier 2 intervention did not appear to result in improved performance, relative to peers, on the spelling task. An explanation provided by these authors was that spelling instruction focused strongly on direct teaching of spelling rules, instead of teaching spelling in the context of a composition task.
In the present study, different explanations could be offered to explain the results. One reason for the lack of improvement in spelling skills could be that the words employed in the intervention did not share the same orthographic properties as the ones in the screening assessment. However, our analysis showed that the words used in the screening and training were matched in terms of word length, word frequency, and orthographic neighbours, since no statistically significant differences were found.
Another possible explanation for the significant differences over time found between the control and the experimental group for the allograph selection, but not for spelling accuracy, could be related to the components worked by teachers in their classrooms. As they reported in the intervention sheet, most of the teachers focused their intervention more on the instruction of handwriting skills, and to a lesser extent on the spelling words. Nonetheless, the present findings suggest that the improvement in the fluency of allograph selection should not be interpreted only in terms of improvement in motor processes, because research focused on this ability supports the idea that it shares central and peripheral components (i.e., orthographic access) (Peake et al., 2017). However, we suppose that spelling skills are more difficult to learn for struggling writers because word spelling requires specifying an entire sequence of letters and not just some of them. The theories of spelling development maintain that orthographic knowledge does not begin to affect a child's spelling until the child has accumulated a considerable number of words that are recognized by sight (Georgiou et al., 2020;Sharp et al., 2008). From a developmental perspective, different Spanish studies have shown that the most significant increase in orthographic skills takes place after Grade 2 (Defior et al., 2009;Sanchez Abchi et al., 2018). As opposed to allograph selection, which involves selecting the appropriate grapheme for a specific context, this ability is mastered between Grades 1 and 2 (Peake et al., 2017). That means that a much more intensive intervention would be necessary for struggling writers to improve their spelling skills. This would not apply to other struggling writers who advance faster and leave the risk situation earlier. A different scenario was encountered when we examined the effects of the intervention in the at-risk group, distinguishing between those who no longer were at risk and those who continued to be at risk after the intervention. One of the main findings was that struggling writers who are no longer at-risk have improved their transcription skills as well as brought those children closer to their peers who have a normal writing achievement. In intervention research, effective strategies have been identified for improving transcription (e.g., Berninger et al., 1997), but not all evidence-based writing interventions are likely to benefit all students. Even when students are identified as at-risk and receive a standard protocol intervention, they may still have differential needs (McMaster et al., 2016). In this sense, at-risk children may also present other types of difficulties (e.g., phonological recoding, implicit learning of orthographic structures, working memory difficulties). Transcription skills are also related to executive functions and to writingspecific processes, according to the NSVW model (Berninger & Winn, 2006). In fact, working memory, inhibition, and updating predict handwriting fluency and spelling, since difficulties with any of these processes can affect transcription (Drijbooms et al., 2015).
In sum, results indicated that teachers improved their subject matter knowledge about what to teach and how to teach the early writing skills within RtI. Furthermore, this study demonstrates that explicit and supplemental handwriting and spelling instruction in a transparent orthography contributes to improve handwriting fluency in young children who acquire text transcription skills more slowly than their peers. Nevertheless, the lack of improvement in spelling skills observed in those struggling Assessing the efficacy of a Tier 2 early intervention for… writers who did not leave the risk situation at the end of the intervention may have been mediated by domain-general skills that underlie low spelling performance but were not controlled or evaluated in this study.

Study limitations and perspectives
A number of limitations should be considered when interpreting the results of this study. Firstly, due to small sample sized, the results of the analysis are underpowered, so more research is needed to confirm the effects of Tier-2 intervention found in this study to improve at-risk students' transcription skills. Secondly, due to significant differences between the experimental and control groups at the intercept of the allograph measure, results must be carefully considered. Future research should ensure the assessment of the control group's intervention to guarantee the lack of differences in the intercept. Thirdly, no information could be collected about the type of instruction the control group at-risk was receiving as well as the lack of control over Tier 1 intervention in the experimental and control groups. Fourthly, future research on Tier-2 implementation should to consider incorporating a broader variety of domain-general skills that underlie low spelling performance. This would be essential for attaining a more holistic understanding of the unique influences that each of these skills may have on writing. Finally, further research is needed to analyze the effects of training transcription skills on written expression in a transparent orthography such as Spanish.

Conclusions
This study focused on evaluating the efficacy of a Tier 2 intervention within the context of the Response to Intervention (RtI) model implemented by Spanish first-to second-grade primary school teachers to improve at-risk students' transcription skills. Our study demonstrates that transcription skills in a transparent orthography such as Spanish can be improved through instructional activities that have also been effective in studies conducted in English. However, spelling skills are somewhat more resistant to intervention, especially in students who are reluctant to leave the risk situation. Since the Spanish language has a transparent orthography, students can learn how to write regular words without specific instruction about spelling. Teachers aimed to ensure that students learnt how to write, prioritizing legibility and speed (fluency) over spelling instruction. Therefore, only those students who improved enough in their handwriting skills were allowed to begin with the spelling activities. The findings imply that for struggling writers who show greater resistance it would be advisable for teachers to invest more time dedicated to improving spelling skills.