In an era increasingly characterised by generative AI, text-to-text software such as ChatGPT, Gemini or Claude can produce rich written content with minimal human input. Such AI advances are prompting a re-examination of writing pedagogy: why teach writing when a machine can generate it? The function of writing instruction, as the ability of teachers to foster critical and creative thinking and authentic expression, urgently needs to be examined in order not to lose sight of why written instruction is still crucial.

The evaluation of in-service teacher training on writing and its impact on teaching remains limited in its systematic scope, especially in the Italian context. This study is a response to the call for evidence on how to be better prepared to teach writing in an effective and cognitively active way. Therefore, the impact and the implications of a training intervention is assessed.  

Drawing on process models of written composition (Hayes & Flower, 1980; Bereiter & Scardamalia, 1987; Kim & Graham, 2022) and socioconstructivist perspectives (Dyson, 1993), the aim is to investigate how lower secondary teachers working in a northern Italian metropolitan area can implement an explicit, process-oriented approach in their teaching (it has been shown, for example, that the revision process is rarely taught in italian classrooms; Boscolo & Zuin, 2015). This study focuses on a specific training program for three in-service teachers (impacting 91 students) adapted from the Writing Workshop model (Calkins, 1986). Drawing on previous evidence-based educational approaches (Darling-Hammond et al., 2017), this study addresses how best to integrate practical and theoretical content so that teachers examine not only “what” they do in the classroom from a different perspective, but also “why” they do it.

At the same time, the research emphasises the importance of local and contextual factors (Boylan et al., 2018). Although national regulations mandate teacher professional development (D.L. 107/2015; D.L. 36/2022), many teachers still see limited access to robust, research-based training, particularly for writing instruction. The motivational aspects of why teachers seek (or do not seek) professional development are also critical (Appova & Arbaugh, 2018). In the face of generative AI, educators must also grapple with ethical and practical questions: how to design tasks that can promote true authorship? How to cultivate cognitive skills when an external tool can seemingly perform the writing process instead of the students?

Anchored in Kirkpatrick’s four-stage evaluation model (Kirkpatrick & Kirkpatrick, 2016), the research is guided by four main questions:

1. Teacher Satisfaction: do teachers find the writing workshop-based training model engaging and relevant?

2. Teacher learning: does participation lead to measurable gains in knowledge, skills, and sense of efficacy?

3. Behavior change: do teachers adapt their writing instruction (e.g., planning, facilitating revision, integrating peer collaboration) according to the training content?

4. Impact: does the newly introduced instructional approach lead to significant improvements in students’ written production (Ventista & Brown, 2023) and cognitive development??

By exploring these questions, the study aims to fill gaps in both the national literature on in-service training and the emerging global discourse on AI in writing education. The study contributes to a broader discussion about how schools can balance technological changes with foundational literacies, ensuring that writing remains a human-centred ability for expressing oneself, forming thoughts, and developing creative knowledge.

A mixed-methods approach (Trinchero & Robasto, 2019) was chosen to capture both the quantitative and qualitative dimensions of training impact. The above-mentioned three volunteer teachers received 16 hours of training in a workshop format focusing on the writing workshop method. The focus was on the explicit teaching of prewriting, drafting and revising, as well as digital integration and a creative use of AI for learning to write. Data collection was based on Kirkpatrick’s model (levels 1–4): Stage 1 (satisfaction): Questionnaires at the end ofthe training - likert-type plus open-ended responses - to determine participants’ views on relevance and delivery. Stage 2 (Learning): A written post-test on knowledge of writing processes and a self-assessment questionnaire on teaching competences, triangulated with a test of orientation toward writing instruction (Poch et al., 2020). Stage 3 (behavior change): Classroom observations documented using field notes to see if teachers implemented the new strategies. Stage 4 (Impact): Student writing samples were collected before and after the intervention. A rubric adapted from Calonghi and Boncori (2006) was used to measure changes in cognitive processes involved in learning, coherence, text organization, and linguistic accuracy. A comparable control group, whose teachers had not participated in the training, enabled quasi-experimental comparisons. To deepen the findings, the researcher also kept an observational log of limiting and facilitating factors. By converging multiple data sources, the aim was to examine whether the changes in teacher practice actually led to improved teacher knowledge, abilities, self-efficacy, and student outcomes, while responding to the new challenges of generative AI.

Preliminary results indicate a high level of satisfaction among teachers (Level 1), with participants praising the collaborative nature of the workshop and its immediate applicability in the classroom. Questionnaire data (Level 2) indicate a significant increase in teachers’ conceptual and procedural knowledge of writing instruction - particularly in relation to planning explicit lessons for drafting and feedback - although uncertainties remain in relation to comprehensive assessment practices. Observational data (Level 3) indicate notable changes: two out of three teachers consistently implemented the newly designed writing activities, provided more class time for revision, and engaged students in discussions to raise awareness towards the use of AI text-to-image for learning how to write. While the final results (level 4) are still being evaluated, preliminary analyses of the student texts show slight improvements in structural cohesion and creativity. A closer qualitative look shows increased motivation among learners, who reported that they enjoyed the autonomy and creativity of the workshop activities. Ongoing analyses are comparing these results with a control group to determine whether the observed gains are statistically significant and sustainable. In summary, the project provides promising evidence that well-structured, context-sensitive teacher training, even on a small scale, can positively affect teachers’ approach to writing instruction and, potentially, students' writing skills and cognitive abilities in a statistically significant manner (through the infusion model approach; Trinchero, 2022). The analysis is also analysed critically, considering that professional development can have challenges in implementation and effectiveness due to a huge number of reasons (McChesney & Aldridge, 2019).

Asterhan, C. S., & Lefstein, A. (2023). The search for evidence-based features of effective teacher professional development: A critical analysis of the literature. Professional Development in Education, 50(1), 11–23. Bereiter, C., & Scardamalia, M. (1987). The psychology of written composition. Lawrence Erlbaum. Boscolo, P. & Zuin, E. (2015), Come scrivono gli adolescenti, il Mulino: Bologna. Boylan, M., Coldwell, M., Maxwell, B., & Jordan, J. (2018). Rethinking models of professional learning as tools: a conceptual analysis to inform research and practice. Professional Development in Education, 44(1), 120–139. Calkins L. (1986). The art of teaching writing. NH: Heinemann. Calonghi, L., & Boncori, L. (2006). Guida per la correzione dei temi. LAS. Darling-Hammond, L., Hyler, M. E., & Gardner, M. (2017). Effective teacher professional development. Learning Policy Institute. Dyson, A. H. (1993). Social worlds of children: Learning to write in an urban primary school. Teachers College Press. Hayes, J. R., & Flower, L. (1980). Writing as problem solving. Visible Language, 14(4), 388–399. Kim, Y. S. G., & Graham, S. (2022). Expanding the Direct and Indirect Effects Model of Writing (DIEW). Journal of Educational Psychology, 114(2), 215–231. Kirkpatrick, J. D., & Kirkpatrick, W. K. (2016). Kirkpatrick’s four levels of training evaluation. ATD Press. McChesney, K., & Aldridge, J. M. (2019). What gets in the way? A new conceptual model for the trajectory from teacher professional development to impact. Professional Development in Education, 47(5), 834–852. https://doi.org/10.1080/19415257.2019.1667412 Poch, A. L., Hamby, M., & Chen, X. (2020). Secondary teachers’ beliefs about teaching writing. Reading & Writing Quarterly, 36(6), 497–520. Ventista, O. M., & Brown, C. (2023). Teachers’ professional learning and its impact on students’ learning outcomes: Findings from a systematic review. Social Sciences & Humanities Open, 8(1), 100565 Trinchero, R. (2022). Metodo, atteggiamento, consapevolezza. Per una didattica orientata allo sviluppo dell’intelligenza. In Percorsi di ricerca didattica e docimologica: studi in onore di Cristina Coggi (pp. 17-41). FrancoAngeli Trinchero, R., & Robasto, D. (2019). I Mixed Methods nella ricerca educativa. Mondadori.