Research Approaches & Qualitative Data Analysis
So tonight we held the first (maybe not the last) #TwitEdRes Twitter chat session. It was all inspired by @thosethatcan really (Emma Kell – soon to be doctor… on the road to being, anyway ;-))
The point of the Twitter chat was to do with Emma being involved with a conference at Middlesex University later this summer, and having been invited to speak on the topic of the advantages and disadvantages of using Twitter in Ed research. Tonight’s chat included those topics (advantages and disadvantages) as well as other emergent chat and room for people to spew out ideas.
Ideas I put forth were to do with research methods you could think about using if you were using Twitter as a data collection tool (not just a forum for attracting research participants, but also actually using the data from people’s timelines and the chat feed.
There are research types of merit that you could consider and adapt for use in conjunction with that type of research if you wanted. A summary of these are as follows:
- Ethnography (‘structural’, Gubrium): ‘Classifies and highlights the social organisation and distribution of subjective meanings as native and diverse field realities’, being ‘concerned with … cataloguing their forms and relationships in time and space’ (Gubrium, 1988:26),
- Ethnography of communication (‘microethnography’): Focus ‘on the patterns of social interaction among members of a cultural group or among members of different cultural groups’ in order to ‘specify the processes of interaction and understand how these ‘micro’ processes are related to larger ‘macro’ issues of culture and social organisation’ (Jacob, 1987:18).
- Ethnomethodology (‘articulative ethnography’, Gubrium): ‘Study how members of society, in the course of ongoing social interaction, make sense of ‘indexical’ expressions. Indexicals are terms whose meaning is not universal, but is dependent on the context’ (Bailey, 1978:249), ‘how members of situations assemble reasonable understandings of the things and events of concern to them and, thereby, realise them as objects of everyday life’ (Gubrium, 1988:27), ‘how people in society organise their activities in such a way that they make mutual sense, how people do things in such ways that others can recognise them for what they are’ (Sharrock and Anderson, 1986:56).
- Ethnoscience (cognitive anthropology): ‘To understand participants cultural categories and to identify the organising principles that underlie these categories… through the study of semantic systems’ (Jacob, 1987:22), ‘to define systematically the meaning of words, or labels – in short the names of things in the context of their use’, in order to ‘construct lexical-semantic fields of linked propositions’ (Werner and Schoepfle, 1987:29, 38).
- Event structure analysis: ‘To examine and represent series of events as logical structures, i.e., as elements and their connections (including the assumptions that govern these connections) that can serve as explanatory models for interpreting actual or folkloristic sequences of events (Heise and Lewis, 1988).
- Symbolic interactionism: ‘To see how the process of designation and interpretation [participants are defining and interpreting each other’s acts] is sustaining, undercutting, redirecting and transforming the ways in which the participants are fitting together their lines of action’ (Blumer, 1969:53), ‘understanding how individuals are able to take one another’s perspective and learn meanings and symbols in concrete instances of interaction’ (Jacob, 1987:29).
3.3 – RESEARCH TYPES AND DATAANALYSIS METHODS ANDCHOICE
After you have done your data collection (no mean feat defining your unit of analysis if you are using Twitter by the way!), you will need to decide on an approach for data analysis. Here are a few ideas:
‘Classic content analysis’ (Krippendorf, 1980) would enable you to make contextual inferences from data, while at the same time, ‘ethnographic content analysis’ (Altheide, 1987) would allow reflexive document analysis, the documenting and understanding of the meaning of communication, as well as the verification of any possible theoretical relationships.
‘Discourse analysis’ (Stubbs, 1983:1), the ‘linguistic analysis of naturally occurring connected spoken or written discourse’, could also provide insight into communication and interaction (van Dijk, 1985:4).
‘Document study’ (Bailey, 1978), as an unstructured and non-quantitative approach using a Twitter timeline for example, may result in typologies ‘through which to examine and analyse the subjective experience of individuals and their construction of the social world’ (Jones, 1983:147).
In order to make sense of the collected data according to research type, you will need to determine a suitable data analysis ‘tool set’.
Relevant data analysis techniques you could choose from include variations on ‘coding’, ‘categorising’, and the contextualisation of words and phrases, via indices, word listing and structural analysis. These techniques and their variances are described briefly below (adapted from Tesch, 1991:26-28):
Qualitative Analysis Methods Overview
Locating individual words and phrases: For language orientated research, as a first exploratory step, to classify and look for correlations of same word usage and synonyms, in proximity of one text or several related ones.
Creating alphabetic word lists, counting the frequency of the occurrence of words: To get an overview of the vocabulary used and to enlarge the word/phrase location ‘picture’ by judging where emphasis of a text lies by word/phrase occurrence frequency.
Creating indices (attaching source information to each occurrence) and ‘key word in context’ concordances: To compare the vocabulary of one text with others, by looking at a list of the location of each word or phrase, in order to take context of word/phrase into consideration, by creating a KWIC concordance (key word in context).
Attaching key words to segments of texts: To examine a text for topics where precise words may not appear to sum these up, and then attach an appropriate keyword. Segments dealing with same topic can be assembled by key words and interpreted.
Attaching codes (categorisation symbols) to segments of text: Codes are abbreviations of category names, where categories can be pictured as the conceptual equivalent of file folders, each labelled with the name of one aspect of the research project or one topic found in the data, serving to organise data pieces. These categories may emerge during the analysis or be developed beforehand, or partially either way. Organising data this way is common in ethnographic content analysis, classical ethnography, life history studies, oral history, document case studies and grounded theory, and is a prerequisite for creating well-ordered narratives about the nature of the phenomenon investigated.
Connecting codes (categories): Goes beyond classification and explores whether or not the phenomenon possesses a discernible structure, or whether or not linkages exist between/among particular categories. Most notably this is done in grounded theory, but also in event structure analysis, ethnographic content analysis, discourse analysis, ethnography of communication, ethnoscience and structural ethnography, symbolic interactionism and ethnomethodology. The purpose is to develop propositional statements or to make assertions regarding the structure of the linkages, or to relate concepts in order to discover underlying principles (also referred to as ‘hyposthesis generation’). Researchers may also aim to find support for assertions or to verify them by negative case analysis, etc.
Ok… got that? Put your thinking cap on and play then ;-))