Monday, 18 May 2009

Data Analysis

In my previous post I talked about my initial thoughts on the data I am collecting from my interviews. That was an exercise to warm my brain up to start thinking on qualitative data analysis, categorisation and coding of data. In this post I would like to briefly explain the methodology I am using to analyse that data. Be careful... this post has a bit of theory on methodology, but I'll try to keep it simple.

To start I have to say that I have been contacting administrative and academic staff from around the University. (You can see a classification of interviewees in my previous post.) I have done this by using the contact details other people I have previously met gave me. So for example if I interviewed Dr. X and he suggested I could contact Professor M, I will then contact Professor M via email, and say Dr X gave me your name and suggested I could talk to you... This has helped me a bit with getting a bit of trust and credibility from potential interviewees. It has also helped me with making sure I am meeting with the right people.

I have had interviews as short as 20 minutes and as long as 1:45hr. I have recorded all of them except one telephone conversation I had with a divisional research administrator. I was on the phone with him for 1hr! Interviews have been mostly semi-structured/unstructured. I took a flexible approach to account for Oxford's heterogeneity. I would always start with the same questions (I would ask them to tell me about their jobs) and then I would choose questions depending on their answers. However I always tried to keep my questions mainly in these three areas:
  1. Questions related to the creation and management of Research Management/Activity data.
  2. Questions related to the use of Research Management/Activity data (perhaps from other sources.)
  3. Questions related to issues and future uses of Research Management/Activity data.
Some of my respondents were able to cover these three areas some only one or two. This depended of course on their roles and experience.

Next step was to transcribe those interviews into Word documents. I've been doing that on the train. Surprisingly this is the perfect place for me to do such a boring and tedious task. So 1hr each way and I am able to transcribe possibly 30 to 45 minutes. (If you are transcribing audiofiles and dreading it, try doing it on the train.) I haven't done an exhaustive verbatim transcription but tried to capture all the ideas covered in every interview. Now I have enough material to start doing the analysis.

As I have carried out interviews, my data are qualitative, i.e., texts containing my interviewees’ ideas. The aim of qualitative data analysis is to abstract those ideas into one cohesive set of statements which could stand for similar pieces of data i. This is not a statistical generalization but an interpretive one ii. The way this works is by organising segments of text according to categories of data, or data codes. I then will go through an iterative process of rephrasing and writing summaries of all of the ideas contained in each category. At doing this I am abstracting the ideas from their original contexts (e.g. the interviews or the interviewees’ jobs) and assigning them new contexts, the one of their categories. Selecting categories is not a science but a kind of art. They depend on the way the researcher interprets the data, and they need constant reading and re-reading of the texts, to make sure categories and analysis reflect the phenomenon under study. The end result of the analysis will help me to draw specific implications, like for example the relationship between BRII stakeholders and the Research Information Infrastructure, characteristics of data needed, uses of research activity data, ways of accessing and viewing information which are most useful for different roles in Oxford, etc.

Anyway, having explained the (sort of) theory behind the analysis process, I will finish this post by explaining the first set of categories that have emerged from my data so far:
  • Perspectives on Research Management data/Research Activity data, what people think about its importance, benefits, relevance to their work, accessibility, visibility, and its management. This is also about the kinds of activities that they perform that involve this kind of data.
  • Research activities, what are the actual processes connected to research activities, types of activities, types of groups, etc, how are they reflected in data?
  • Content of Data/Sources of data – what "objects" are these data describing? (this category will also describe data contributors.) Other issues such as management, quality of content, difficulties at gathering data, difficulties at putting together a website, what is sensitive data, etc
  • Types of Stakeholders, descriptions of departments, functions, people’s roles and their activities
  • Notes for development - anything relevant for the design of the infrastructure or the web services, including my own thoughts.
To give you an idea of the kind of data I got, here you have an extract from an interview with someone from the Medical Sciences division. I have initially classified this text under Research Activities.

"Themes have no money (they are different from institutes and centres) Themes are purely a way of helping to sell their research in a way, showing where their strengths are in this university. A theme is a way to classify people. Themes are also a way of quantifying what they do."

i Tesch, R., (1990), Qualitative Research: analysis types and software tools, New York, The Farmer Press.
ii Walsham, G., (1995), 'Interpretive case studies in IS research: nature and method', European Journal of Information Systems, 4, no.2

Print this post

No comments:

Post a Comment