Developing revision skills: Studying real-time writing processes to inform instruction practices

This contribution aims to give an overview of how revision during writing is used by writers across the school curriculum until adulthood. In doing so it wants to identify revision patterns typical for specific age groups, and explore some qualitative differences in revision practices.

In the seminal cognitive model of writing by Hayes and Flower (1980), revision is one of the three main processes, alongside planning and translation/transcription. Revision research during the past decades include examples from Faigley & Witte, 1981; an overview on revision development in Chanquoy (2009), an updated writing model in Chenoweth and Hayes (2003), and recent approaches to analyze revisions in real-time in Lindgren, Westum, Outakoski & Sullivan (2019).

Writers need to develop an understanding of on the one hand the possibilities and on the other hand the limitations of written language. For the child learning to write the new activity will by necessity be contrasted to the typical situation of spoken interaction: two or more interlocutors who share the ‘here and now’, and who can provide immediate reactions to each other’s utterances, and remedy misunderstandings. Compared to this, a limitation of writing is that it takes place in a different here and now than the reading of the same text. For the written text to function in another context, the writer is required to anticipate the reader’s needs, and adapt the language, content and structure to those needs. This is a highly complex task, but in solving this task, the writer is helped by revision, i.e. the possibility to write things down, evaluate the written text, and revise it if it does not meet the intended goal. Thus, through revision – the ability to resolve “mistakes” in the already written text before it reaches the future reader is a kind of superpower. The learning writer needs to understand how to use this superpower, and to make use of the benefits of written language to compensate for its limitations. Young writers who are linear in their writing, i.e., writing their texts from the beginning to end following the so-called knowledge translating strategy (Bereiter & Scardamalia, 1987) also adapt a superficial revision pattern (see Chanquoy, 2010), similar to the linearity of spoken language (see Johansson, 2009). They have not started to use the full possibilities of written language. Mid-teenage writers will often adapt other strategies, with more extensive revising of their texts, following the knowledge transforming strategy, with more deep revisions, moving from a more spoken-like process of writing to a more mature behavior, making use of the delayed reception of the written message. Studying the development of revision processes is further connected to aspects of general linguistic and cognitive development. Equally it is related to practice: The writer’s ability to take the reader into consideration concurrent with text production, and concurrently adapt the content and rhetoric goals of the text to the future reader has been discussed among others by Kellogg (2008), who proposes that it takes years of training for a writer to successfully and simultaneously integrate text content, linguistic form and rhetoric adaptation to the reader – in fact, the ability to do so is what characterizes an expert writer. We can thus expect revision skills to develop long into adulthood.

This study explores how students from grade 3 to upper secondary school, as well as adults, revise their text, how much time they revise and what revision strategies they use. We ask qualitative and quantitative research questions: Does the amount of revision (in proportion of time on task, and in number of revised characters) differ across the school age? In which contexts do participants revise, and does this differ between the ages? Are there any qualitative differences between the revision strategies of children, adolescents and adults?

The cross-sectional comparisons are made through empirical data consisting of experimentally collected texts, using keystroke logging, which allows for the in-depth study of what has been deleted and when, and for analyzing revisions and planning patterns (through pauses) at the same time. In total, 170 narrative and expository texts by writers age 9–18, and adults are analyzed.

The results show no linear connection between increase in age and amount of deleted text. The analyses of revision strategies reveal that students until 15 years of age are linear in both text generation and revision. In the later part of adolescence more students rearrange the texts more, both through text deletion, reworking what is written, and adding information. The adult controls adapt a more global revision perspective, and often rework a greater part, if not all of their texts.

This extended description of how revision strategies develop across school age into adulthood acts as input for writing instruction at all school levels. It demonstrates great variation in quality ¬– more than quantity – across the age span, and it further underlines that it takes a long time to develop global revision strategies. The result can guide writing instructors on how to reinforce age-typical revision strategies and in doing so preparing students for the next level of revision development.