Developmental and acquired writing difficulties – synthesizing and deepening 20 years of research.
The aim of this sabbatical is to synthesize and deepen research from 20 years studies, by our group about writing processes and writing difficulties. With the digitalization of western societies during the last decades, writing skills have become increasingly important. Questions we have asked in our research are: What distinguishes written production from oral, how is it affected by genre and writing technology, why are writing difficulties even more challenging than reading dificulties and how can we support writing in persons with developmental and acquired language difficulties? Several of our projects have included methological development. Together can our datasets contribute to answering more complex questions than each and one of them. During this sabbatical I plan to write four different articles with the aim of increasing our understanding of the writing processes by persons who are challenged by writing.
Final report
Background
In 2020, I was awarded RJ sabbatical funding for the project "Developmental and Acquired Writing Difficulties – Synthesis and Deepening of 20 Years of Research." This project was carried out during 2022-2023 and included two month-long stays with the research group led by Professor Vincent Connelly at Oxford Brookes University in the United Kingdom - one of the few research groups worldwide dedicated to studying writing processes in individuals with language difficulties. Additionally, the project provided me with the opportunity to strengthen collaboration with a research group at the University of Antwerp, Belgium, led by Professor Luuk van Waes, which, like our own, focuses on the development of keystroke logging software for studies of writing processes and writing difficulties.
The project resulted in numerous scientific publications and conference contributions, including a special issue in Frontiers of Psychology titled “Analysing Writing Processes of People with Language, Mental, Cognitive or Physical Disorders” (Van Waes, Wengelin & Henriksson, forthcoming). It also initiated a series of workshops focusing on methods to investigate writing processes and writing difficulties. Furthermore, it led to a thematic symposium on writing difficulties at the SIG Writing 2022 conference, two individual conference presentations, and a planned thematic symposium on writing technologies at SIG Writing 2024. Research findings were disseminated through journals and events run by the Swedish dyslexia associations.
Results
The purpose of this project was to shed light on how spelling difficulties and strategies in individuals with language difficulties manifest in their writing processes, how temporal patterns in the writing process differ between individuals with and without language difficulties, such as dyslexia or aphasia after a stroke, and how modern and accessible writing technologies can facilitate writing for individuals facing significant writing challenges.
My point of departure was that previous research on writing difficulties, including our own and that of the Oxford Brookes group, has suggested that both children and adults with language difficulties invest significant cognitive effort at the word level, especially in spelling and word choice to avoid spelling errors. In the writing process, this has been evident through frequent word-internal pausing and extensive spelling revision (e.g. Behrns et al., 2008; O'Rourke, 2020; Sumner et al., 2013; 2014; 2015; Wengelin, 2002; 2007; Wengelin et al., 2014). This has been suggested to explain why these writers often produce texts with lower lexical diversity and/or lower text quality (Kraft et al., 2018; Sumner, 2013). These findings are in line with theoretical models of the writing process (e.g., Hayes & Berninger, 2014), where the hypothesis of cognitive capacity (McCutchen, 1996) plays a central role. In brief, it implies that if low-level processes like spelling have not been sufficiently automated and thus demand too much of working memory, resources for higher-level processes, such as creating structure and coherence, become inadequate. This would also suggest that if the use of a reliable spell check or a good speech recognition program reduces the resources allocated to spelling, it can contribute to higher text quality.
However, results from previous research are somewhat conflicting. For example, Connelly et al. (2006) did not find that writers with dyslexia produced texts with lower lexical diversity. Furthermore, Torrance et al. (2016) observed that participants with reading and writing difficulties made fewer spelling revisions than a similar-aged control group. There are many possible explanations for these conflicting results, such as differences in the orthographic depth of languages, participants' ages, and the specific writing tasks used.
The project was based on keystroke-logged texts that capture the entire writing process through time-stamped keyboard and computer mouse activities, which can reveal pausing and editing. I began by examining to what extent temporal patterns in 4th-6th-grade students with reading and writing difficulties aligned with the patterns we had previously identified in older students and adults with similar difficulties. Somewhat surprisingly, the effects of word-internal pauses and spelling edits were very weak for this group, both in terms of distinguishing students with reading and writing difficulties from those without difficulties and in explaining the linguistic characteristics of their texts. Currently, we cannot identify clear temporal patterns that robustly differentiate between these groups. This prompted an article summarizing the knowledge on writing processes in groups with various language-related difficulties and identifying the need for further research (Wengelin, Van Waes & Henriksson, under review), as well as a meta-narrative review article (Wengelin, Van Waes, Leijten, & Johansson, forthcoming).
The former has been submitted to Frontiers of Psychology and will, if accepted, be part of a special issue on writing difficulties that I am co-editing, "Analyzing Writing Processes of People with Language, Mental, Cognitive, or Physical Disorders" (Van Waes, Wengelin, & Henriksson, forthcoming). The latter sheds light on how research on temporal patterns related to writing processes and writing difficulties has developed theoretically and methodologically in our and others' research since a discussion on the concept of pauses that I initiated in Wengelin (2006). It is currently in manuscript form and will be submitted shortly.
What did, however, align with previous research was that the 4-6-grade-student students with reading and writing difficulties left behind a significantly higher number of misspelled words in their texts than their peers, both in relation to the total word count and the number of errors made. This could either be due to them not detecting their errors, supported by the fact that the weakest readers left behind the most errors, or simply that they are less bothered by spelling concerns while writing than our older participants have been and wait to review the text until it is completed. These results are presented in Wengelin, Frid, Johansson, and Johansson (under revision) In this work, I also create a taxonomy of the various ways in which spelling difficulties manifest in the writing processes of individuals with different writing difficulties.
To shed light on how we can facilitate writing for individuals with significant writing difficulties, we explore the use of writing technologies. Such research also contributes to understanding the writing processes of individuals with language difficulties by allowing us to observe how different subprocesses are affected. During my sabbatical, I text quality and linguistic characteristics of texts dictated with the assistance of speech recognition. Dictating one's text can be advantageous for writers with severe spelling difficulties, as the risk of spelling errors in the final text is significantly reduced. Furthermore, according to the hypothesis of cognitive capacity, dictation could potentially reduce the risk of spelling difficulties overloading working memory and negatively impacting the writing process. However, using speech recognition presents new challenges, one of which is that instead of misspelling, speech recognition systems may misinterpret the dictated word and write a different one. Thus, it places significant demands on the writer's ability to proofread and edit their text.
An interesting result of this sub-study was that the writers who were the weakest readers were the ones who benefited the most from using speech recognition. Differences in text length and text quality that we observed in texts written by writers with and without reading and writing difficulties were reduced when they used speech recognition, despite the fact that the speech recognition system's output required some correction. We report these results in Wengelin, Kraft, Thurfjell & Rack (forthcoming).
The somewhat contradictory results from the above-mentioned studies, where weak readers produced better texts using a speech recognition system that was by no means perfect and required a fair amount of correction (Kraft, 2023), despite leaving many spelling errors when typing on a keyboard, underlines the need for further research on writing difficulties in general and indicate particularly that we need more investigation into the relationship between reading and writing to understand writers' challenges and their strategies to address them. In Wengelin, Johansson, Frid & Johansson (2023), we propose a method to investigate how and to what extent writers continue to write while visually attending to different aspects of the already-written text. The results also suggest that analyses of writing processes can be valuable in writing pedagogical contexts, a topic we discuss in Wengelin & Johansson (2023). It is clear that a significant research gap exists in this area.
An empirical study that I had originally planned was to empirically investigate the similarities and differences between acquired and developmental writing difficulties. For various technical reasons, it proved to be more challenging to carry out in a systematic and robust manner than I had anticipated, and thus, that study has only been initiated. We compiled a brief overview of the current status of the field in Wengelin, Van Waes & Henriksson (under review) and I presented an initial analysis at the Writing Research Across Borders 2023 conference (Wengelin, 2023). Another important area to explore further is how writing technologies can be utilized to support writing in individuals with reading and writing difficulties, as well as to contribute to untangle the mechanisms underlying writing difficulties. Work on a project proposal relating to writing technologies for people with writing difficulties has been initiated.
In 2020, I was awarded RJ sabbatical funding for the project "Developmental and Acquired Writing Difficulties – Synthesis and Deepening of 20 Years of Research." This project was carried out during 2022-2023 and included two month-long stays with the research group led by Professor Vincent Connelly at Oxford Brookes University in the United Kingdom - one of the few research groups worldwide dedicated to studying writing processes in individuals with language difficulties. Additionally, the project provided me with the opportunity to strengthen collaboration with a research group at the University of Antwerp, Belgium, led by Professor Luuk van Waes, which, like our own, focuses on the development of keystroke logging software for studies of writing processes and writing difficulties.
The project resulted in numerous scientific publications and conference contributions, including a special issue in Frontiers of Psychology titled “Analysing Writing Processes of People with Language, Mental, Cognitive or Physical Disorders” (Van Waes, Wengelin & Henriksson, forthcoming). It also initiated a series of workshops focusing on methods to investigate writing processes and writing difficulties. Furthermore, it led to a thematic symposium on writing difficulties at the SIG Writing 2022 conference, two individual conference presentations, and a planned thematic symposium on writing technologies at SIG Writing 2024. Research findings were disseminated through journals and events run by the Swedish dyslexia associations.
Results
The purpose of this project was to shed light on how spelling difficulties and strategies in individuals with language difficulties manifest in their writing processes, how temporal patterns in the writing process differ between individuals with and without language difficulties, such as dyslexia or aphasia after a stroke, and how modern and accessible writing technologies can facilitate writing for individuals facing significant writing challenges.
My point of departure was that previous research on writing difficulties, including our own and that of the Oxford Brookes group, has suggested that both children and adults with language difficulties invest significant cognitive effort at the word level, especially in spelling and word choice to avoid spelling errors. In the writing process, this has been evident through frequent word-internal pausing and extensive spelling revision (e.g. Behrns et al., 2008; O'Rourke, 2020; Sumner et al., 2013; 2014; 2015; Wengelin, 2002; 2007; Wengelin et al., 2014). This has been suggested to explain why these writers often produce texts with lower lexical diversity and/or lower text quality (Kraft et al., 2018; Sumner, 2013). These findings are in line with theoretical models of the writing process (e.g., Hayes & Berninger, 2014), where the hypothesis of cognitive capacity (McCutchen, 1996) plays a central role. In brief, it implies that if low-level processes like spelling have not been sufficiently automated and thus demand too much of working memory, resources for higher-level processes, such as creating structure and coherence, become inadequate. This would also suggest that if the use of a reliable spell check or a good speech recognition program reduces the resources allocated to spelling, it can contribute to higher text quality.
However, results from previous research are somewhat conflicting. For example, Connelly et al. (2006) did not find that writers with dyslexia produced texts with lower lexical diversity. Furthermore, Torrance et al. (2016) observed that participants with reading and writing difficulties made fewer spelling revisions than a similar-aged control group. There are many possible explanations for these conflicting results, such as differences in the orthographic depth of languages, participants' ages, and the specific writing tasks used.
The project was based on keystroke-logged texts that capture the entire writing process through time-stamped keyboard and computer mouse activities, which can reveal pausing and editing. I began by examining to what extent temporal patterns in 4th-6th-grade students with reading and writing difficulties aligned with the patterns we had previously identified in older students and adults with similar difficulties. Somewhat surprisingly, the effects of word-internal pauses and spelling edits were very weak for this group, both in terms of distinguishing students with reading and writing difficulties from those without difficulties and in explaining the linguistic characteristics of their texts. Currently, we cannot identify clear temporal patterns that robustly differentiate between these groups. This prompted an article summarizing the knowledge on writing processes in groups with various language-related difficulties and identifying the need for further research (Wengelin, Van Waes & Henriksson, under review), as well as a meta-narrative review article (Wengelin, Van Waes, Leijten, & Johansson, forthcoming).
The former has been submitted to Frontiers of Psychology and will, if accepted, be part of a special issue on writing difficulties that I am co-editing, "Analyzing Writing Processes of People with Language, Mental, Cognitive, or Physical Disorders" (Van Waes, Wengelin, & Henriksson, forthcoming). The latter sheds light on how research on temporal patterns related to writing processes and writing difficulties has developed theoretically and methodologically in our and others' research since a discussion on the concept of pauses that I initiated in Wengelin (2006). It is currently in manuscript form and will be submitted shortly.
What did, however, align with previous research was that the 4-6-grade-student students with reading and writing difficulties left behind a significantly higher number of misspelled words in their texts than their peers, both in relation to the total word count and the number of errors made. This could either be due to them not detecting their errors, supported by the fact that the weakest readers left behind the most errors, or simply that they are less bothered by spelling concerns while writing than our older participants have been and wait to review the text until it is completed. These results are presented in Wengelin, Frid, Johansson, and Johansson (under revision) In this work, I also create a taxonomy of the various ways in which spelling difficulties manifest in the writing processes of individuals with different writing difficulties.
To shed light on how we can facilitate writing for individuals with significant writing difficulties, we explore the use of writing technologies. Such research also contributes to understanding the writing processes of individuals with language difficulties by allowing us to observe how different subprocesses are affected. During my sabbatical, I text quality and linguistic characteristics of texts dictated with the assistance of speech recognition. Dictating one's text can be advantageous for writers with severe spelling difficulties, as the risk of spelling errors in the final text is significantly reduced. Furthermore, according to the hypothesis of cognitive capacity, dictation could potentially reduce the risk of spelling difficulties overloading working memory and negatively impacting the writing process. However, using speech recognition presents new challenges, one of which is that instead of misspelling, speech recognition systems may misinterpret the dictated word and write a different one. Thus, it places significant demands on the writer's ability to proofread and edit their text.
An interesting result of this sub-study was that the writers who were the weakest readers were the ones who benefited the most from using speech recognition. Differences in text length and text quality that we observed in texts written by writers with and without reading and writing difficulties were reduced when they used speech recognition, despite the fact that the speech recognition system's output required some correction. We report these results in Wengelin, Kraft, Thurfjell & Rack (forthcoming).
The somewhat contradictory results from the above-mentioned studies, where weak readers produced better texts using a speech recognition system that was by no means perfect and required a fair amount of correction (Kraft, 2023), despite leaving many spelling errors when typing on a keyboard, underlines the need for further research on writing difficulties in general and indicate particularly that we need more investigation into the relationship between reading and writing to understand writers' challenges and their strategies to address them. In Wengelin, Johansson, Frid & Johansson (2023), we propose a method to investigate how and to what extent writers continue to write while visually attending to different aspects of the already-written text. The results also suggest that analyses of writing processes can be valuable in writing pedagogical contexts, a topic we discuss in Wengelin & Johansson (2023). It is clear that a significant research gap exists in this area.
An empirical study that I had originally planned was to empirically investigate the similarities and differences between acquired and developmental writing difficulties. For various technical reasons, it proved to be more challenging to carry out in a systematic and robust manner than I had anticipated, and thus, that study has only been initiated. We compiled a brief overview of the current status of the field in Wengelin, Van Waes & Henriksson (under review) and I presented an initial analysis at the Writing Research Across Borders 2023 conference (Wengelin, 2023). Another important area to explore further is how writing technologies can be utilized to support writing in individuals with reading and writing difficulties, as well as to contribute to untangle the mechanisms underlying writing difficulties. Work on a project proposal relating to writing technologies for people with writing difficulties has been initiated.