System for categorisation and colours codes. Relations between analytical and descriptive categories, including assignment components. 

Contexts in source publication

Context 1

... make it possible to answer the first research question, how do students make meaning and represent knowledge about civil rights in the English speaking world in their written essays and audiovisual presentations, the merged transcriptions (of essay and audiovisual presentation) were marked with colour codes ( figure 6) after the descriptive categories information (subdivided into explanations and exemplifying), relating to the reader (subdivided into arguments and opinions) and textual devices (subdivided into linking and paratext) (refer to figure 9). Each set of descriptive category (information, relating to the reader, textual devices) was linked to the analytical categories the metafunctions, ideational meaning, interpersonal meaning and textual meaning. In this way the metafunctions were made visible in the transcriptions, which could then be read after colour codes in analyses. Blue was used to code idea units where the ideational metafunction was dominant. Pink marked idea units where the interpersonal metafunction was dominant, and turquoise shades were used to mark idea units where the textual metafunction of communication was identified as dominant ( figure 6). Grey was used to mark an absence of information in one channel or mode of communication. Yellow was used to mark episodes, and black to visually delimit essay from audiovisual presentation, or the end of a written or audiovisual text. Figure 6. Example of merged transcription. This is an example from the merged transcriptions of Emma's ...

Context 2

... et al.'s (2016) study was used as inspiration and a starting point for creating descriptive categories, across written and audiovisual representations of knowledge, which could be used for comparative analyses. Out of the categories used by Smith et al. (2016), it was 'descriptions' and 'building relations to the reader' (p. 142) that turned out to be useful and applicable to the empirical material in this study. After several adjustments, and rethinking both scope of the thesis and categorisations, descriptive categories for this thesis were finally divided into information (including explanations and exemplifying), relating to the reader (including arguments and opinions) and textual devices (including linking and paratext) ( figure 9). Categorisation was applied to subunits of text labelled idea units. The concept idea unit was borrowed from Smith et al. (2016) and Mayer (1981Mayer ( , 1985. The subunits of text used by Smith et al. (2016), going back to Mayer (1981,1985), were adapted and adjusted further, also using ideas from , for dividing the students' texts (table 2). Kress writes about subtextual entities as "units which serve to make up texts and other semiotic entities" (2010, p. 147). Here the term idea unit is used for such subtextual units (Kress uses other terminology but the purpose is the same). Kress defines text as the largest semiotic entity, which is made up of subtextual units where these are "derived from their functions and uses within the text" (Kress, 2010, p. 148). Mayer (1981Mayer ( , 1985 together with Smith et al. (2016) was used for the delimitation of idea units. Dividing the written and audiovisual texts into episodes and idea units helped in the fined grained comparative analyses. The use of idea units made it possible to compare content, how meaning was made and transduced across written and audiovisual texts at a microlevel. Episodes made it possible to compare sequences of text that were one logical unit (such as paragraphs in written text) that supplied a comparative framework. The use of episodes and idea units made it possible to analyse sequentiality and organisation, including cohesion and coherence in the written and audiovisual texts, and then how these were transduced. Multiple transcriptions were constructed and adjusted in a continuous and ongoing process. The different stages of processing the essays and audiovisual presentations, including different types of selections and delimitations, were all part of a process of analysis. It was not possible to separate the process of transcription from the process of analysis, and data should in this study be seen as produced rather than collected. Table 3 illustrates the data produced and processed in two stages, where the second stage of produced data are the transcriptions created and used for analysis of transduction processes. ...

Context 3

... students' essays and audiovisual presentations are with this theoretical and methodological framework seen as material instantiations of the ideational, interpersonal and textual metafunctions (Jewitt, 2014c); the students (re)construct a social reality, interpreting and shaping signs (Selander & Kress, 2010) in a process of transferring information and shaping representations of knowledge, in a process of transduction (Kress, , 2017. The metafunctions were in this study used as analytical tools. For analysing the ideational metafunction there were three areas of focus used; what has been included, what meaning was being made and, what processes are in action -classifying, analytical or symbolical Kress & van Leeuwen 2006). The reader is reminded that the classificational, analytical and symbolical processes were originally used to reflect conceptual representation in images ( van Leeuwen, 1996/2006). The students' written essays were analysed for similar processes, where classifying (retelling facts, stating information) was used as corresponding to the classificational process in images, and when students analysed in writing that was used as corresponding to the analytical process in images, lastly symbolical writing was used for the written mode where a lexical items or a phrase had symbolical meaning rather than a straight forward literal denotation. Hence, in this study the use of these processes in action is expanded to include conceptual ideas in idea units where the modes can be visual and/or verbal, including both speech and writing. For the interpersonal metafunction, subjective positioning, style, language, closeness-distance, placement and gaze were considered. For the textual metafunction, principles of sequencing, coherence, cohesion, composition, linking information, given/new, centre/margin and framing were analysed (see table 4). Three types and levels of coding were used, starting with the assignment components, background, comparison, reflection and paratext. The analytical categories the ideational, relational and textual metafunctions, and their related descriptive categories, information (including explanations and exemplifying), relating to the reader (including arguments and opinions) and textual devices (including linking and paratext) are shown in figure 9. Together these three 'dimensions' of coding make out the basis for analysis and interpretation. Colour codes were used for visually marking categories in the merged transcription matrices, and were used both for descriptive and analytical categories, where blues were used for the ideational metafunction and related categories, pinks for interpersonal metafunction and related categories, and shades of turquoise for textual metafunction and related categories. Analyses were carried out in several steps using a comparative visual method adapted after inspiration from Smith et al. (2016). The students' essays and audiovisual presentation were analysed separately and then together, starting with an iterative process . Then transcription and analysis was a process of going back and forth between the empirical material, matrices and coding categories, making it difficult to describe the process as clear cut and separate. There was a general order of working, which has been described here, but the reader should keep in mind that for every step along the way there has been sidesteps and retakes to make adjustments, and to note down observations and temporary conclusions. As a result both transcripts and analyses have been reworked multiple times, and in parallel to a (top-down) analysis of content and meaning making, analyses were carried out on detailed ...