CHAPTER THREE: METHODOLOGY
I- statement coding
10. Feelings and affect
‘I hate writing in French, it’s so difficult, so many accents’ (Writing experience, Angeles)
‘and I really feel like quite happy with that story’ (Writing experience, Derek) ‘I just drew a boundary in a way, like hatred’ (Language experience, Anna)
After all the I-statement coding was finished, for each participant, the number of I-statements falling under each category was counted and the percentage each category takes out of the total number of I-statements was also calculated. Then, under each of the above ten categories, the number of I-statements falling under each Content Area (i.e. writing experience, language experience, educational experience, professional experience, and reading experience) was further counted and the percentage each content area takes out of the total number of I-statements was also
calculated.
In the following, I shall explain my coding scheme for the We- and You- statements uttered by the fifteen participants in the in-depth interview.
3.2.4.1.1.2 We- and You-statement analysis
I have developed We- and You-statement analysis in the present research for the purpose of gaining insight into the L2 creative writers’ sense of membership of particular communities as revealed in their interview comments. We- or/and You- statement analyses have been scarcely used in identity research. My conceptualization of this quantitative discourse analysis tool in my attempt to unravel the L2 creative writers’ sense of belongingness to particular CoPs was inspired by two investigative endeavours. The first one is Fairclough’s (2003) discussion on how the representations and classifications of social actors are achieved by speakers through them naturally shifting among pronouns like ‘I’, ‘You’, and ‘We’ when talking about their sociocultural experiences. Fairclough illustrated how ‘we-community’ and ‘you-community’ are depicted in texts to signify social relations and to ‘represent and construct groups and communities’ (ibid, p. 149), for example, exactly which social group is represented by a particular ‘We’ and how this ‘we-community’ is positioned to other social groups (p. 148-p. 150); in short, the social significance of inclusion and exclusion. Secondly, my development of the We- and You-statement analysis was also inspired by Wei’s presentation at a conference at Warwick University of her PhD work, a practitioner research investigating her EFL students’ learner Autonomy in an university in Taiwan (Wei, 2011). In her research, through analysing the ‘We-statements’ shown in her EFL students’ self-
reflective diaries on their individual English language learning experiences, Wei teased out her students’ identifications and alignments with specific communities. In the present research, We- and You-statement analysis was used to gain insight into the nature and variety of the communities of practice (CoPs) that the speaker perceived him/herself to be a member of. Thus, the purpose of conducting We- and You-statement analysis in the present research is different from its purpose behind the I-statement analysis. As shown previously, I-statement analysis was conducted to reveal the participants’ self-portrayal and self-fashioning in five content areas, regarding how the individuals represented who they were and how they interacted with the world; differently, the We- and You-statement analysis concentrates on teasing out the types of social groups with which the L2 creative writers identified themselves.
It is not difficult to see that when people speak in the first person plural ‘We’, they instantly identity themselves as belonging to a certain community and signal this sense of belongingness to the listener(s). On the other hand, the type of You- statements focused on in the present study is when people speak in the generic ‘You’. For example, in the in-depth interview conducted with one creative writer participant Maggie, she commented on the discomfort she felt in reading out her work in front of other people and tried to explain such feelings by saying that:
‘if you chose to write a short story of course it’s not always very good or very nice,
and you always feel if you are presenting like a story to someone, you always feel like you have this strange claim of a strange idea it needs to be very nice’.
As demonstrated in Maggie’s above comment and as stated by Fairclough (2003), the generic ‘You’ is employed to reference ordinary experience (p. 150), which, I
would argue, further conveys the speaker’s belief in the well-accepted nature of the practice concerned or the pervasiveness of the particular phenomenon under discussion. In Maggie’s case, she employed the generic ‘You’ to talk about a particular creative writing experience of her own. That is to say, she perceived her personal experience of feeling ill at ease when reading out her work in public as ordinary and shared by all the other creative writers. By aligning all the other creative writers with her, Maggie thus immediately identified herself as one of them. Therefore, when people speak in the generic ‘You’, they are simultaneously expressing a strong identification with particular values, beliefs, practices and experiences by claiming that such values, beliefs, practices or experiences are of common sense. This subsequently shows the speakers’ self-perception of belonging to the CoP which defines such values, beliefs, practices or experiences. The following procedures were adopted to carry out We- and You-statement analysis in the present research.
In each transcript, among the participant’s own comments, the We-statements and You-statements (only the generic ‘You’) were segmented and counted at basic clause level. When ‘you know’ serve as fillers, they were not counted. The We- and You-statements were coded and counted together rather than separately. Each We- or You-statement was first coded based on its indication of the specific type of CoP, e.g. student writer group in a particular (disciplinary or sociocultural) context, member of an L1 or L2 student writer community, or member of a particular nationality or ethnicity. The coding categories of the community types are meant to be descriptive and explanatory. Again, similar to the approach I had adopted in examining and coding the I-statements, I employed inductive analysis of the We- and You- statements. The process of establishing the community types went simultaneously
with the process of data examination and coding; and both processes were recursive. Eventually, a total of 19 communities were identified (illustrated below). Each participant expressed his/her belongingness to some, not all (ranging from six to twelve communities), of the 19 communities. These 19 communities are displayed below in Table 3.2 (For an illustration of the 19 communities through the participants’ We- and You-statement examples, see Appendix E).
Table 3.2
The 19 communities identified in the entire We- and You-statement coding
1 2 3 4 5 6 7 8 9 Professional identity Educational community Community of L2 speakers
Student writer group in a particular context
Community of L1 student writers Community of L2 (or L3) student writers
Community of bilingual (or multilingual) student writers Community of creative writers Member of a particular nationality
10 11 12 13 14 15 16 17 18 19
Member of a particular ethnicity Member of a socializing community Community of readers
‘Individuals with a particular skill’ ‘Individuals with insight’
Member of Immigrants
Member of computer game players Member of experienced Internet users
Member of British society Gender identity
After my initial coding of the We- and You-statements, I coded the entire dataset again after an interval of one month, and revisions were made to the coding categories (the original 17 types of communities were developed into 19 types; the above 19 categories represent the final version). Then I randomly chose one participant (Eliza)’s We- and You-statement data (which had already been segmented with the We- and You-statements presented in context) and asked the same person, whom I mentioned previously, to independently code Eliza’s We- and You- statements. Again, before that, I explained to him my definition of the 19 types of
communities and showed him some examples of the other participants’ We- and You- statements. A 90% agreement was achieved between our coding results. Disagreements were all settled through discussion, but no revision was made to the coding categories. After the discussion, for a third time I coded the entire We- and You-statement dataset of the fifteen participants. In this process, revisions were made to the coding of some We- and You-statements but not to the coding categories. After all the We- and You-statement coding was finished for each participant the number of We- and You-statements falling under each of the 19 types of communities was counted and the percentage each community represents out of the total number of We- and You-statements was also calculated.
So far, I have explained the coding schemes for quantitatively analysing the transcripts of the in-depth interviews conducted with the fifteen participants, i.e. I- statement analysis, We- and You-statement analysis. In the following, I shall explain the data analysis method the present research has employed to quantitatively examine the think-aloud protocols generated by each participant in the two story- writing tasks.
3.2.4.1.2 Coding of the think-aloud protocols
The fifteen participants’ think-aloud verbalizations generated in the two differently-conditioned story writing tasks were transcribed verbatim. That is to say, in total, thirty think-aloud protocols were available for coding. Before I started examining and coding the think-aloud protocols, I applied a general conceptual framework to the think-aloud data, i.e. Flower and Hayes’s (1981) influential cognitive writing processes model (as demonstrated in Figure 3.4 below), which
consists of ‘Planning’, ‘Translating’, ‘Reviewing’, and ‘Monitor’ (p. 370). Their counterparts in the present research are: planning, composing (instead of ‘Translating), revising (instead of ‘Reviewing’, as will be explained later), and monitoring (instead of ‘Monitor’, as will be explained later). The Flower-Hayes cognitive process model of the composing processes has three major components— task environments, the writer’s long-term memory, and writing processes—as illustrated in Figure 3.4 below. Although the Flower-Hayes model has been criticized for not adequately addressing the emotional and social influences on how writers compose (see Atkinson, 2003; Matsuda, 2003), for the purpose of catching the moment-by-moment cognitive writing activities of L2 writers, I found the component ‘Writing Processes’ of the Flower-Hayes model particularly useful for analysing the think-aloud protocols.
Figure 3.4
The Flower-Hayes cognitive process model of the composing processes (Flower and Hayes, 1981, p. 370)
participants’ think-aloud verbalizations word by word and as accurately as I could manage, I had already gained a relatively thorough understanding of the think-aloud data. I then employed inductive analysis on the thirty sets of think-aloud protocols by focusing on Flower and Hayes’ four major categories of writing processes as mentioned above. The process of examining and coding the think-aloud protocols went simultaneously with the process of developing the coding scheme. Descriptive subcategories were then established under each of these four major categories. In each coded think-aloud protocol, the think-aloud verbalizations were segmented into think-aloud units. The segmentation was carried out based on two simultaneous conditions. That is, each think-aloud unit is distinguished by the mention of only one writing behaviour defined by one of the 42 subcategories of the coding scheme (shown below on p. 135), and is also distinguished by the mention of only one coherent and integral topic. Through recursive and on-going experiments of segmenting and coding the think-aloud protocols, and also through a series of reviews of the coded data, the final coding scheme was set up.
Before presenting the details of my think-aloud coding scheme, however, I would like to point out here that there are two crucial differences between Flower and Hayes’s conceptualization of the component ‘Writing Processes’ in their Cognitive Process Model (as shown above in Figure 3.4) and the theoretical framework adopted by the present research in order to examine the writers’ think-aloud protocols.
Firstly, as shown in Figure 3.4 in the box of ‘Writing Processes’, the writing activities of ‘Planning’, ‘Translating’, and ‘Reviewing’ are arranged in a linear order with no double-direction arrows between any two of them. On the other hand, between the ‘Monitor’ and all the other three writing activities, there are double-
direction interactions going on, and the ‘Monitor’ is seen as linking ‘Planning’, ‘Translating’, and ‘Reviewing’ into cohesively progressed writing processes. Flower and Hayes’s formulation of the ‘Writing Processes’ has been regarded by some researchers as failing to address the recursive nature of the cognitive writing process (e.g. Wang and Wen, 2002, Witte, 1987). In the present research, the recursive nature of writing is reflected in all fifteen participants’ coded think-aloud protocols, i.e. there is no linear order among the four major writing activities of planning, composing, monitoring and revising. In fact, any one of these four major writing activities (all of which occur recursively) could be followed by any of the other three.
In relation to this first difference in the conceptual stances taken by the present research and Flower and Hayes’s model, the second difference lies in what is perceived as the role of the ‘Monitor’. As pointed out by Flower and Hayes (1981), ‘[t]he monitor functions as a writing strategist which determines when the writer moves from one process to the next’ and thus ‘promote[s] [writer’s] switching between processes or encourage[s] the sustained generation of ideas’ (p. 374). That is to say, the ‘Monitor’ controls and ensures smooth and effective transitions between the different stages of writing, as exemplified by Flower and Hayes in saying that the ‘Monitor’ ‘determines how long a writer will continue generating ideas before attempting to write prose’ (ibid). However, the present research chose the action- oriented terminology of monitoring (i.e. the ING form of the verb ‘monitor’) instead of Flower and Hayes’s status-oriented naming ‘the monitor’ (i.e. the noun ‘monitor’) and thus treated monitoring as doing, rather than as an executive mechanism. Therefore, the writing activity of monitoring is conceived as on an equal footing with planning, composing, and revising, and mingled in with the other three writing
activities extensively and recursively, instead of being considered as a hierarchically higher ‘writing strategist’ allocated with executive power. Different from Flower and Hayes’s formulation of the ‘Monitor’ and the ‘Reviewing’ writing processes, in my coding scheme, writers’ ‘evaluative’ comments are put under monitoring, and thus the category revising is entirely constituted by writers’ specific and concrete revising behaviours and attempts at revising. Consequently, my definition of the monitoring category addresses the evaluative comments, metacomments or any kind of monitoring comments uttered by the L2 creative writers in the writing processes. In the following, I will explain in detail my coding of the think-aloud protocols and illustrate the coding scheme.
As previously mentioned, the audio-taped think-aloud verbalizations were transcribed verbatim in standard English, with the inclusion of pauses. In each coded think-aloud protocol, the underlined parts are writers’ verbalizing of their writing; the double underlined parts are the revisions actually made to the written text; the [parts within square brackets] are writers’ reading of what has been written down; a comma indicates pausing. Each think-aloud protocol was segmented into think-aloud units and each unit coded according to the 42 coding items and numbered in the order of its occurrences. For a sample of the segmented, coded and numbered think- aloud protocol, please refer to Appendix F where a portion of Maggie’s think-aloud protocol generated in the prompted writing task is provided. In the following, the 42 coding items under the four major categories are explained and illustrated with examples. Words within ‘< >’ alongside the think-aloud examples are my explanations. When necessary, in order to illustrate the context where the particular think-aloud example is embedded, the adjoining think-aloud unit is provided within ‘( )’. In the ‘{ }’ following each think-aloud example, I have specified, in this order:
the participant who made the think-aloud utterance, the specific writing task, and the sequence of the specific think-aloud unit in the think-aloud protocol.
Planning
(1) Idea-generating of local event
e.g. ‘emm, why why why would she why would she say that how can I how can I give
introduction, emm how can I give an introduction to to Noah to the story’ {Jingjing,
Autobiographical writing, 133}
(2) Idea-generating of global event
e.g. ‘emm also I want, is it a happy ending well we’ll see’ {Jingjing, Autobiographical writing, 49}
(3) Looking for vocabulary
e.g. ‘oh what do you say, relaxing no that’s not relaxing what do you say when you, it’s err healthy no it’s er well I would say, therapeutical thera there we go, I hope it exists, no it doesn’t so let’s see the options therapeutic oh, therapeutical therapeutic, oh there we go yes therapeutic just a matter of’ {Derek, Autobiographical writing, 100}
(4) Looking for ideas: its difference from ‘idea-generating of local/global event’ lying in that the idea was already out there, either in writer’s memory or knowledge or in the external resources, and the writer was trying to retrieve it
e.g. ‘and well it’s 1973 what happened in 73, 73 is after Vietnam War I think, 60’s ok and 73 after Vietnam War’ {Jingjing, Prompted writing, 89}
(5) Looking for grammar
(6) Looking for phrasing
e.g. ‘the two groups, what I want to write emm one making fun of the other being weird beatniks weird beatniks or, sh, emm, couldn’t be listen no,’ {Maggie, Autobiographical writing, 330}
(7) Goal-setting on local event
e.g. ‘and yeah we can have a little description of the wedding’ {Jingjing, Prompted writing, 21}
(8) Goal-setting on global event
e.g. ‘so I prefer to use numeric first point ok, body text is about emm what is game- based learning and what a game-based learning specialist do,’ {Ho, Autobiographical writing, 13}
(9) Goal-setting on literary technique, e.g. genre elements, or narrative structure e.g. ‘ok there should be some emm, problem faced, problem faced and then climax of the story, then final confrontation and then the ending’ {Ho, Autobiographical writing, 17}
(10) Goal-setting on text format: ranging from mechanical features such as punctuation, italicization, underlining, boldness and capitalization for particular effect, to syntactic features such as order of clauses or sentences, and finally to textual length and textual structure
e.g. ‘ok and then I can I normally play plan the number of words ok, require for example normally the body text will take err 80% and this is 10% 10% that means it’s 40 words 40 words and then this is 80% 320 words’ {Ho, Autobiographical writing, 5}