Abstract
In this thesis we describe the organization of speech recognition based writing processes. Writing can be seen as a visual representation of spoken language: a combination that speech recognition takes full advantage of. In the field of writing research, speech recognition is a new writing instrument that may cause a
... read more
shift in writing process research because the underlying processes are changing. In addition to this, we take advantage of on of the weak points of speech recognition, namely the misrecognitions in the text. As it is, speech recognition visualizes how writers deal with errors in the text produced so far. In addition, speech recognition may cause a different view on writing processes because the processes are easier to reveal. This thesis addresses three main research questions: What is the influence of speech recognition on cognitive writing processes and the organization of writing? How do errors in the text produced so far influence text production (when writing with speech recognition)? What (combination of) research methods are most adequate to observe and - quantitatively - analyze the writing processes in different writing modes without interfering in the writing process? The thesis consists of 7 articles. These 7 articles are related to the four main research projects as described above (see Figure 1). The first project was on the influence of speech recognition on the adaptation processes and writing processes of novice speech recognition users (chapters 2, 3 and 4). In this first section we approach the collected data from three different perspectives: adaptation strategy, learning style and previous (classical) dictating experience. Chapter 2 focuses mainly on the adaptation strategies and learning styles of writers with the same writing background (lawyers that already had previous dictating experience). Chapter 3 extensively takes this difference in previous writing experience into account and focuses on the adaptation processes and writing processes. Next to this, the chapter describes the classification model in greater detail. Chapter 4 focuses on the error correction strategies of novice speech recognition users. In section 2 we describe the research project on error correction strategies in isolated sentences (chapters 5 and 6). Chapter 5 describes the theoretical framework and the research methodology in great detail. The analyses performed in this chapter are based on aggregated data. In chapter 6 we describe a more detailed re-analysis of the dataset via multilevel analyses. Writing researchers can also read this chapter as an illustration of conducting this kind of analysis. The third section describes a research project on error correction strategies of professional speech recognition users (chapter 7). Finally in section 4, we wind up this thesis with an article on the logging program Inputlog (chapter 8). This chapter provides state-of-the-art information about Inputlog. The latest version of Inputlog can be found on the Inputlog website (www.inputlog.net). The website also contains extra documentation on Inputlog.
show less