Research Methods for Information Research
7. Beyond research methods
7.10 Using concept maps in research
I have been working with two colleagues, at King’s College, London to explore some possibilities in working though old interview transcripts and reconstructing them as concept maps (drawing on the approach developed by Novak43). By isolating concepts and specifying how they are linked we hoped to be able to translate sets of interview responses into diagrammatic form and, in doing so, to test the potential for this approach to research evidence collection. We chose to try this out on a set of interviews conducted to find out about the information-related behaviour of a group of managers and advisers at work.
Clearly there are marked differences between asking research respondents to construct a concept map (and, usually, helping them to do so) and trying to ‘second-guess’ the respondents’ conceptualisation of the matter under discussion and turn the ideas into map form. However, it proved to be relatively easy to refashion most sets of interview replies into individual concept maps. As expected, the more informative interviews yielded the more complex maps and vice versa. What the retrospective concept map construction revealed and what conclusions emerged will be reported later in a refereed journal article, but doing this work did raise some interesting possibilities in the area of data-gathering (using concept map creation as an alternative or adjunct to the traditional research interview). The work also highlighted some interesting issues about data analysis and data presentation.
Taking the possibility of adding value to the interview itself first, I hope it is reasonable to assume that most information research interviews are conducted to get a better understanding of how the respondents perceive a problem or issue, how they go about undertaking various types of tasks or how they deal with specific situations, and what they feel about all of these. Concept maps are designed to help people to convey how they conceptualise given issues, but they can also be adapted into flow charts (this time with specified links between processes or actions) to explore activity.
Using concept maps in or after interviews
This visual approach to data-gathering could be used as an activity within the interview itself, in which the respondent is encouraged to make a concept map or flow map. For some purposes it might be more productive to use the concept map as the main focus of the interview, in which any questioning would be confined to clarification of what was appearing in the map. Alternatively the information gleaned during the interview could be reworked after the interview and the resulting maps sent or shown to the respondent to check for accuracy and to encourage further elaboration and clarification. The same approach could be adopted at macro-level, by constructing a super-map showing the findings across all the interviews and then going back to each respondent with this map to further amplify and refine the picture.
One rule-of-thumb for researchers conducting interviews, especially when exploring a new issue or situation, is that they should continue to set up and conduct interviews with relevant informants until these stop providing new information. This stage is sometimes referred to as data saturation and deciding when that point is reached is highly subjective. Rather than hoping that researchers are sufficiently experienced or attuned to make this call, an overarching concept map could be constructed from all the interview replies during a research project, or during a specific phase of interviewing, to provide a consolidated view of what is being learnt. This would offer a rather more objective view of saturation since the time for interviews to stop would be when nothing was being added to or altered in the model.
Working the data
One traditional view of qualitative research data analysis is that exploration has ended and it is now time to make sense of the data by organising it into categories and, if necessary, changing headings and categories until the data make sense. This approach is sometimes characterised as ‘listening to the data’ but might be more accurately described as ‘massaging the data into the shoe until they fit’. (Increasingly, much of this process is performed electronically, often with levels of help and support that ease the researcher away from careful consideration of the implications of arbitrary category assignment or of exact meanings in words or phrases.)
Some researchers identify two processes here, those of data coding and display. Once the data have been arranged into manageable categories, the compressed information can be presented in one of a variety of forms. Concept maps have already been used in this form of display, but are not a prominent feature in LIS research.
This is not the only view of how data should be used. As noted in a previous column, advocates of Grounded Theory offer a rigorous approach based on three types of data categorisation. One of the progenitors of this movement, Anselm Strauss44 has gone further in taking a more radical view of qualitative data analysis. In this view, the emphasis is no longer on data compression and summarising, but on expanding the conceptual framework by asking further questions about the data. Concept mapping fits very readily into this type of approach: in this view of coding, concept maps can do more than merely present summaries of data. By translating data into diagrammatic form and emphasising links and relationships, concept maps can play an important part in raising questions about the data.
43. Novak, J.D. (1998) Learning, creating and using knowledge: concept maps as facilitative tools in schools and corporations Hillsdale, NJ: Lawrence Erlbaum. ↩
44. Strauss, A. L. (1987) Qualitative analysis for social scientists Cambridge: Cambridge U P ISBN 13: 9780521338066 ↩