Tag Archives: convergent interviewing

Adjusting my data analysis strategy

Posted on by
Reply

I have been reviewing and rethinking my data analysis approach. I have been coding in NVivo directly to the audio track. This has been revealing in terms of coding and memo writing but is extremely time consuming. After thinking through the process from my original strategy of using Convergent Interviewing and keeping short notes rather than transcribing the whole interview, I realised that I need to revisit this and code against my typed notes rather than the whole audio track.

According to Dick the original Convergent Interviewing process relied on note taking without transcription where the notes are linked to the under pinning theoretical framework.

I am now going to:

  1. code in NVivo against the notes I took during and after the interviews
  2. review the contract summary forms to determine themes
  3. query the codes and memos from the text coding and memo writing
  4. combine with coding from the interview recordings

This will return me to my original plan and speed up the data analysis process.

 

Data analysis – is hard

Posted on by
Reply

I’ve had to stop writing my thesis for the time being, I’ve come to the end of the first draft of the literature review but I have been told by my supervisor to concentrate on the data analysis otherwise I’m going to run out of time. I was going to start with In Vivo coding but this felt too loose and unstructured. I prefer to use more structured methods. So, I’ve gone back to basics and reviewed some of the literature on qualitative data analysis.

I’ve been reading The Coding Manual for Qualitative Researchers by Jonny Saldana and Qualitative Data Analysis by Miles and Huberman and have decided to avoid In Vivo coding and start again by creating a conceptual framework based on research questions, hypothesis, problem areas, and / or key variables that the researcher brings to the study

I’m going to create a set of codes based on:

Bogden and Biklen’s (1992) coding accounting scheme.

  1. Settings/Context: general information on surroundings that allows you to put the study in a larger context.
  2. Definition of the situation: how people understand, define, or perceive the setting or the topics on which the study bears.
  3. Perspectives: ways of thinking about their setting shared by informants (“how things are done here”).
  4. Ways of thinking about people and objects: understandings of each other, of outsiders, of objects in their world (more detailed than above).
  5. Process: regularly occurring kinds of behaviour.
  6. Activities: regularly occurring kinds of behaviour.
  7. Events: specific activities, especially ones occurring infrequently.
  8. Strategies: ways of accomplishing things; people’s tactics, methods, techniques for meeting their needs.
  9. Relationships and social structure: unofficially defined patterns such as cliques, coalitions, romances, friendships and betrayals.
  10. Methods: problems, joys, dilemmas of the research process – often in relation to comments by observers.

Libraries and customer service

Posted on by
Reply

It has become clear after extensively reviewing the literature on customer services in the Higher Education sector that student services have been slow to the party when it comes to changing processes and services to reflect changing student expectations.

I have found that the university library sector has been ahead of the game for years and there is a significant amount of published research, case studies and articles outlining what has been achieved over the past twenty to twenty five years.

Libraries have really followed on from IT Help Desk systems where staff in organizations became connected to networked Help systems. These allowed staff to place a ‘ticket’ that specified an issue or system failure. The response would likely be someone visiting the staff or a fix via network tools.

Over time the IT Help Desk was reconfigured into a Service Desk where service staff were able to provide more customer focused services not just system fixes. Service Desk staff were eventually able to provide training and other support.

These Service Desk and Help Desk systems have now been introduced into university student service desks and allow students to access 24 hour self service and receive streamlined services.

The issue for where I am carrying out my research project is the SID system (Student Information Desk) combined with a new physical student HUB has significantly changed the working practices of the Faculty support teams. The main impact has been a reduction in student foot-fall and fewer students coming in person.

This has had considerable impact on the Faculty teams in terms of how they now perceive their roles, the SID system has in fact imposed an alternative identity on teams and individuals. The system has demanded that the staff undertake additional training and changes to work process flows. Students are now more likely to go to the student HUB and issues are posted on the SID system where Faculty interventions are needed.

The system has changed the relationship between the staff and students and what were front line staff have now become back-office staff. These issues have generated a mass of data for my research project.

Reflections on the interviews

Posted on by
Reply

Convergent interviewing as a process is very effective for getting participants to provide a lot of detailed information in a reasonably short period of time. In my experience the participants found the process interesting because the majority of them thought that they would be asked a series of structured questions and when I told them that I only had one main question they found this slightly perplexing.

For the majority of cases the participants did not find it difficult staying on  topic or talking for long periods of time without prompting. There was no issue with any participant about the recording of the interviews. Everyone was happy to sign off on the confidentiality form.

All of the interviews lasted a minimum of 50 minutes and it would have been possible to have continued with the interviews for longer in some cases.

The interviews needed a lot of pre-planning, listening to the last recording, designing the next question, preparing the location, contacting the next candidate to confirm that they are happy to participate. I arranged for each participant to select the location, the date and time.

The snowballing technique worked well. Every participant was able to provide the details of someone who would be useful for me to speak to next or at some point in future. I also put together a list of possible participants.

At the start of the process I intended sticking as closely as possible to the way that Dick had described but I found out quite quickly that I had to compromise the approach. I had intended putting together a reference group, as described by Dick and did recruit a small representative group who agreed to work with me. Very quickly I encountered a problem, the reference group could not agree when to meet as a group mainly due to time constraints and work responsibilities.

Eventually I realised that I needed to get on with the interviews as time was running out. I spoke to each of the people in the representative group and asked them to recommend a group of participants individually rather than meeting as a group. In the end the compromise worked well.

As the interview process has carried on I have not referred to the reference group. I have interviewed three members of the reference group because they line manage several of the subsequent interviewees.

The main thing that has worked is the open-ended unstructured interview and the note taking during the interviews. I have not compromised the process that much only in so far as it makes practical sense. So far the experience of convergent interviewing has confirmed to me the usefulness of semi-or unstructured questions for eliciting answers that are broad based but also contain enough specificity to get enough data to make the outputs useful.

Convergent interviewing thoughts on the process

Posted on by
Reply

After completing the assignment on using Convergent Interviewing (CI) and really reading thoroughly the papers and guidance in the literature (especially the one’s by Dick, 1998) I feel that CI is definitely a method that is worth further consideration for my thesis. What I like most about it are the issues of higher validity and reliability, the focus on taking small amounts of notes rather than making full transcripts and the fact that it can be implemented as a project in its own right.

The project aspect is quite interesting in terms of getting the Reference Group together which for me is similar to having a Project Board as part of a PRINCE2 or similar project process. Organizing the Reference Group will I think be an interesting – in terms of challenging – process. I am assuming that the Reference Group would be composed mainly of departmental managers or team leaders as they would be people who would have the ability or authority to make decisions and suggestions about who should be included in the interviewing rounds.

The other interesting or challenging issue would be deciding on whether to use (at least) one other interviewer. On large projects CI tends to make use of more than one interviewer and in some of the papers I’ve looked at up to  six. Although there is always the option of carrying out all the interviews myself the big advantage is using one other interviewer would enhance the reliability of the research by allowing cross referencing and reducing bias.

Having another interviewer would add an interesting dynamic to the process of the research project. How to recruit someone to work with is an interesting problem. One thought would be to get the help of a Masters student who might be working on a similar project and to use them as a research assistant. There are possible opportunities for this in the social sciences department. This  would add an interesting additional challenge to the ethical approval process.

With CI the use of two interviewers lends reliability and reduces bias through running interviews in parallel and then immediately after the interviews have finished the interviewers immediately compare notes and decide what the key issues or themes are. From these the next round of interviews are developed.

Sharing ideas and experiences I believe is a good way of building a creative dynamic throughout the interview process. The biggest issue that I found with doing just two interviews was typing up the transcripts from the voice recording. The first interview I did was 37 minutes long and took about 2 – 3 hours to transcribe. The second was only 15 minutes long but took about an hour to transcribe. The interesting thing about transcribing that I found was making sure that I did not add or remove any words. I think dealing with interviewing 20 or 30 people would require a serious amount of organization of the material and careful planning prior to starting the work. As already stated CI technically does not make use of long transcriptions of the interviews. In fact according to Dick voice recorders are not recommended. Dick recommends only brief note taking or using mnemonics to identify themes and issues during the interview and having only a page of notes at most.

To me this implies that there is a need for quite a disciplined approach to the process. The key issue for CI is ensuring that the initial interview question is further extended and more focused in order to gather information on more specific issues. For the assignment I only focused on one of the three theories that I am aiming to use in my full research project. These are, Sensemaking, Institutional Theory and Organizational Justice. For the assignment I only referred to Organizational Justice but after the interviews I carried out I believe that all three theories are relevant to the thesis at least as drivers into a direction for the project. CI is normally or often used in areas where there is little existing theory or where an area is under researched.

For the research project I believe that CI has the potential to lead to new theory development or an extension of an existing theory.