III. Analysis
5.6.2 Interview
As another data collection method, interviews were essential at different stages of the research, namely before I began planning the ARCs and at the beginning of each ARC. Mostly teachers were interviewed in order to get their perceptions of students’ writing skills and what needed to be achieved, how they thought materials could be useful and how they implemented writing tasks. Interviews also played an important role when individual students could not find the time to participate in the focus groups being carried out at the end of each ARC. These were looked at as another source of feedback and proved essential in getting information about individual impressions on the module. This section will look at each of these stages and at the types of interviews carried out.
I. Methods and Aims
As referred to previously, interviews were part of the intervention at different stages. In this section, I shall define their importance at each one. My aims changed as did the types of questions asked. I worked with semi-structured interviews as these are more adjustable to my way of being and how I conduct research and develop relationships with the subjects. They allow room for further depth in certain subjects yet maintain a line of thought throughout the whole interview.
I adopted semi-structured interviews, which is seen by Gillham (2005) as having a structured but flexible nature. Gillham reveals that this interview’s flexibility is balanced by its structure. He claims that it should have a list of questions that is used for all interviewees involved. He also gives importance to the time that is allowed to each interviewee and claims that it should be equal for all the subjects. Nunan (1992) reinforces this idea when he delineates that the researcher has a general idea about where he wants this interview to go and his / her own interests, topics and issues dominate the interview. Most of my interviews took up to about half an hour each. Flexibilty is seen by Gillham to be balanced by structure. This is expected to be a part of the list of questions as these undergo a process of development that ensures their topic focus. Interviews that are semi-structured enable the researcher to plan an interview that is exploratory. On the one hand, one has a list of pre-prepared questions as well as follow-ups and probes, but on the other hand, this type of interview allows the setting of focus to be wide-angled due to the use of open-ended questions, as Gillham proposes. This stimulates the interviewee to speak more freely on the topic that he / she has been questioned on. It proved quite useful when talking to UMa lecturers, as lecturers generally enjoy talking about their areas of expertise and sharing their perspectives, so this allowed them space to speak freely.
Gillham believes the semi-structured interview anticipates analysis and facilitates the organisation of gathering data from interviews. All these points were considered whilst structuring the interview questions. He divides this interview into 5 main phases:
Preparation Phase: This happens before the interview and involves the definition
of the topic area to be explored, selecting the interviewees and the preparation of the list of questions for the interview. Preparing the questions presupposes a period of work whereby the researcher has to narrow down the focus after carefully analysing the questions’ logical sequence as well as follow-ups and probes that will enable the interview to have flexibility.
Initial Contact Phase: This phase takes place just before the actual interview
begins. The aim is to create a comfortable atmosphere so the interviewee feels at ease to talk to the researcher. This is viewed as primarily social interaction which may or may not have a link to the line of questioning which shall take place after a degree of empathy and ease has been created among the researcher and subject.
Orientation Phase: The interviewer has to be very conscious of this stage as it is
here that the interview is pointed in the direction the researcher wishes. This is done using follow-ups and probes when the researcher deems necessary.
Substantive Phase: Gillham describes this phase as the main empirical focus for
analysis, thus this would basically be most of the interview.
Closure Phase: As in the Initial phase, this is seen as partly social and cognitive,
for the researcher must round up the interview with a closing question, thank the interviewee for his / her time, leave the idea that a future interview may be useful for further research and make sure that the subject is willing to partake in other interviews (p.31-32).
Once the researcher has determined the topic that he / she wishes to cover and begins to think about how to conduct an interview that will provide his / her research with the desired data, scaffolding of the questions begins. Rubin (2005) refers to the importance of ‘scaffolding’ the interview in order to first come up with a skeleton for it. This step took me through a few drafts until getting to the list of questions I actually used for the interviews.
Rubin divides the question types into Main Questions, Follow-up Questions and Probes. I have chosen to use this terminology due to its simplicity. The main question types that Rubin explains and that were adopted in the interviews that were carried out included ‘Opening the Locks’. These are designed to encourage the subject to talk about his / her experience at length, so the data gathers depth on a particular topic and it helped to gather data about different perspectives related to students’ writing. ‘River and Channel’ makes the researcher feel the need to follow a certain train of thought and it is probable that some of the main questions may cease to be of importance when tagging along with the interviewee. It is however important to have the main questions planned out so the general aim is set and both the researcher and subject are aware of the research questions.
In order to gather the most information out of well designed main questions, Rubin recommends ‘Comparison and Contrast questions’, whereby one questions what is thought to be the best and worst characteristics of certain situations. ‘Confirmatory main questions’ also help to certify that certain information that one already has is in fact correct. This helps solidify the data of the research and was used often throughout the interviews conducted.
McGill and Beatty (2001) add a few more question types which enabled me to better my interviewing techniques:
Affective questions: ‘How do you feel about…?’
Probing questions: ‘What aspects of … do you think might be relevant?’
Checking questions: ‘What you plan to do is… Is that right?’
Reflective questions: ‘In what way was… confusing?’
Follow-up questions come after the main questions and are also prepared before-hand. When wording follow-ups, Rubin believes that these should make the main questions either more specific or general in order to pinpoint information or to search for a general picture.
Probes are techniques that keep a conversation flowing whilst ensuring clarification. Rubin refers to various types of probes:
Continuation Probes: these are used to encourage people to talk and to
illicit more information. (hand gesture to continue)
Elaboration Probes: these suggest the need for elaboration in detail of a concept. (‘Please can you specify ...’)
Attention Probes: these are meant to let on that the researcher is paying
close attention to what is being said. (Leaning forward; ‘I understand.’ ‘Can I quote you on that?’)
Clarification Probes: these indicate that something should be explained. (‘Sorry?’ ‘Could you explain what you mean by…?’)
Steering Probes: these are used when the conversation gets too far off track. (You mentioned…, can we go back to that?’)
Sequence Probes: these help place things into the correct order. (‘So, first you…. And then?’)
Evidence Probes: These probes aim at certifying whose position is the
most reliable or which argument weighs more. (‘How do you know that?’)
Slant Probes: these provide clues as to what side your interviewee has
taken and help interpret answers. (‘So you agree with…?’) (p.175). So, on the whole interviewers need to understand the diverse functions of main questions, follow-ups and probes as well as how each function in certain types of interviews. Probes are, for example, very useful in unstructured interviews as a means to stimulate the subject’s ideas, whereas follow-ups help to obtain the desired depth in semi-structured interviews. Despite the confusion that arises from distinguishing probes and follow-ups, I understand the difference to lie in the fact that follow-ups should be prepared and foreseen prior to the interview. Probes are instinctive reactions to keep the conversation going and no prior preparation is needed.
II. Context, Content and Participants
Each set of the above-mentioned interviews took place in different contexts with different participants and this section shall demonstrate how each one was carried out. After looking into many types of interviews as well as their objectives, I decided upon the semi-structured interview to begin collecting my data. I carried out a pilot study and then moved onto interviews with lecturers and undergraduates who are part of the English and German Studies Department (DEAG – Departamento de Estudos Anglísticos e Germanísticos) at the University of Madeira (UMa). The semi-structured interview seems to be the one that is most appropriate for my research. Despite the preparation that is necessary, this type of interview leaves room for introducing questions the interviewer thinks appropriate providing the interview with flexibility to probe into an issue that seems interesting. This is of utmost importance to me as I was
trying to understand how writing was viewed in DEAG. Having room to explore areas that may be touched upon by the interviewees and that I may not have had the sensitivity or awareness about is of utmost importance to enable a more in-depth approach to the research.
Interviews were also carried out at high schools with Engineering and English teachers before the ARCs began and learning about what the students wrote at the previous educational level helped define the direction of the writing as well as the teaching methodologies.
In order for data retrieved from interviews to be reliable and valid, there are a few issues that have to be taken into account. These include the type of questions, their design; the diverse steps followed from the trial interview through to the final interview itself; the structure, organisation and ordering of the questions; the way in which the questions are pruned so they prove to be relevant to the research; how the interviewees narrative is extended by putting them at ease and linking the questions so they see the logic behind the questioning and how the data is presented.
By planning the questions and then submitting them to a process of development whereby the focus on the topic narrowed, the interview is guided towards providing the adequate information on the topic in question. Due to the open-endings of the questions, it is easier to ensure that the interviewee talks about his/her own experience and puts across a personal opinion. If the open questions are insufficient to obtain the information required, the flexibility of this type of interview will allow probing until the information needed is disclosed. This flexibility also enables further probing into new or interesting topics that might be brought up and that were not initially contemplated. This type of flexibility also helps in ensuring that the data given is of a truthful nature. If one realises that an answer is dubious, this flexibility allows to cross-interrogate by
posing the same question in another manner or allows the interviewer to probe further into the matter so it may become clearer whether the interviewee’s stand point is in fact truthful and honest or not.
III. Analysis
In this section, analyses of the interviews are described. The software and hardware used and how they enabled analysis is discussed. The use of Mp3 players/ recorders and the ease of copying files onto a computer are part of this discussion. The importance of Soundscriber, as freeware that makes transcribing a less arduous task is also looked at. Coding of these interviews is an inevitable part in this section too.
Transcribing and translating some of the interviews involve various steps that take up quite a large amount of time. This is begun by writing up, as soon as possible, what was said during the interview and then conducting a content analysis of the transcription. As to the translation process, I do that simultaneously to transcribing. Some of the interviews carried out with high school teachers and some university lecturers needed to be carried out in Portuguese as they were unwilling or unable to do the interview in English.
Gillham (2000) provides a few tips on how to transcribe interviews. Transcriptions should be well identified with names, which are mostly codified, positions of interviewees, dates, place or codes because they might be needed at a later stage in the research. One must not place too many words on a page allowing generous margins for all the coding references and notes that will be later added in. 350 words with double spacing is considered adequate by this academic. He advises researchers to use different typeface for his/ her own questions and other meaningful parts of the interview, such as interjections. However, when using NVIVO8, which was probably unavailable when Gillham wrote this book, it does most of the above automatically so formatting is not
such a concern anymore. Soundscriber proved to be extremely useful to transcribe most of the interviews. I used NVIVO8 for transcribing excerpts but Soundscriber is much more effective when transcribing whole interviews. Soundscriber makes controlling the timing of each walk cycle and rewinding much simpler as they can be programmed to the speed that each transcriber needs to write down what is being heard.
Once the recorded discourse has been written up, the researcher goes through the transcripts trying to pinpoint the substantive statements. At this stage, repetitions and other irrelevant information are ignored. This will make it possible to mark up the categories. It is quite a difficult task and requires practice. Breaks should be taken when finding the categories so that the brain is not given time to start forming its own groups. These categories can then be grouped according to the questions asked and can thus be given a group heading. Comparing the substantive statements with the categories takes the researcher to the next stage. An analysis grid is filled in this phase. Gillham provides some interesting models of grids (p.65).
The analysis sheet is composed of the list of categories as well as the codes for each interview. Each substantive statement is written up against the referent category. If a count assignment needs to be done, this can be completed on the same sheet or two separate ones. By this I mean counting, for example, the number of instances something occurs within a category. This shall indicate how many people share a certain idea. By tabulating the statements, the interviews are attributed a range of responses and this provides material for qualitative analysis. Once again NVIVO simplifies this whole process at the touch of a few buttons.
Gillham goes on to elucidate the importance of tabulating statements as this brings ‘the summary category to life’ and helps ‘convey a range of responses’ (p.70) that will provide organised material for the analysis. He differentiates surface and meaning
analysis when doing content analysis. By meaningful analysis, it is meant that categories are created which can be used to bring together different ideas relating to a single topic. These are of high inference and normally include things like human behaviour or feelings. Descriptive categories, for example, of objects, fall into surface analysis and are of low inference. A range of responses helps analysis when a category can be further divided into sub-categories, thus providing more detail about a single category. These additional links help add depth to the major category. For example, when looking at Blended-learning, some subcategories might be ‘interacting, communicating and guiding. These are seen as high inference, whilst a surface analysis could include ‘hardware, Moodle and message’.