Creative Research Methods 2e. Kara, Helen
involved patients with dementia and their family carers, gathered qualitative and quantitative data in nine different ways and worked to integrate the findings from their analysis. The research team included expertise from the fields of arts and humanities, science and neuroscience, psychology and neuropsychology, nursing, social work, counselling and education. ‘These fields span both qualitative and quantitative research traditions – we found this essential for informed decision making and functioning in all stages of our mixed methods research’ (Robinson et al 2011: 335).
The point of combining qualitative and quantitative methods is that they offer us different ways to understand the world. Quantitative methods show us how much, what, when and where, based on a theory of normality and difference: is this within the curve or outside it? Qualitative methods show us why and how, based on a theory of interactions, events and processes. But multi-modal research is not inherently ‘better’ than single-method research. As always, it depends on the research question and its context. For some questions, in some contexts, only a single method will be necessary to find an adequate answer. Other questions, in other contexts, can be fully addressed only through a multi-modal approach.
One technique for combining qualitative methods is known as ‘bricolage’, from the French word meaning to make something using whatever materials are to hand. In research terms, this means drawing on theory from any discipline or disciplines, using a combination of data-gathering methods and analytic techniques and taking a similarly eclectic approach to the presentation and dissemination of research (Kincheloe 2005: 323–4). The researcher as ‘bricoleur’ can focus on the methods or techniques they prefer, or those that they feel are best suited to their research (Broussine 2008: 79). While some find this too haphazard, advocates of bricolage suggest that it provides more opportunities for sense making than other methods (Warne and McAndrew 2009: 857), perhaps because the researcher is not fettered by a particular method or approach. It may also be because the technique of bricolage is closer to the approach that an artist might use than to that which a scientist might use, offering more scope for creativity, as well as the chance to ‘make for making sense’. Indeed, scholars writing of bricolage often use arts-based metaphors like weaving, collage or patchwork (Wibberley 2012: 6) (Box 2.9).
Box 2.9: Remix in research by Markham
US researcher Annette Markham built on the concept of bricolage to develop the technique of ‘remix’. She uses this in cross-disciplinary workshops with scholars who work online and are new to qualitative research, to help them explore creative approaches to research. There are five elements to Markham’s conceptualisation of remix.
•Generate – expanding your perception of data beyond what you purposefully gather to include field notes, early drafts, doodles, photos, uncoded transcripts, coded transcripts and so on, any of which may trigger useful ‘connections among ideas’ in a ‘wonderful chaos of inquiry’ (Markham 2013a: 74).
•Play – either guided/rule driven or free form and open, but always using your curiosity and imagination to drive exploration and experimentation.
•Borrow – ideas, approaches, perspectives, techniques and so on – from other researchers, disciplines and professions.
•Move – whether forward or backward, leading or following, move and allow yourself to be moved, maintaining awareness that research is ‘always situated, but never motionless’ (Markham 2013a: 77).
•Interrogate – constantly questioning data, literature, context, power, your own motivation and so on, and all aspects of the research project and the subject under investigation.
Markham writes, ‘The concept of remix highlights activities that are not often discussed as a part of method and may not be noticed, such as using serendipity, playing with different perspectives, generating partial renderings, moving through multiple variations, borrowing from disparate and perhaps disjunctive concepts, and so forth’ (Markham 2013a: 65). Remix also implies creative reassembly of these disparate parts, although that may or may not lead to a cohesive final output; it may simply create a new connection between two hitherto unconnected elements. The process focuses on meaning, rather than method as such, so its marker of quality is the extent to which its results have resonance with their audiences.
In combining quantitative and qualitative methods, some researchers embed qualitative methods within a quantitative framework, and some do the opposite (Plano Clark et al 2013: 220). Even some of the most reified conventional methods, such as randomised controlled trials, are now being redesigned in some contexts to incorporate, or be incorporated into, multi-modal designs (Hesse-Biber 2012: 876) (Box 2.10).
Box 2.10: Multi-modal research by Robinson and Mendelson
Sue Robinson and Andrew Mendelson, at the University of Wisconsin–Madison in the US, studied the way photographs and text interacted for readers of a non-fiction magazine article. Their research embedded qualitative methods within a quantitative framework. For data gathering they used pre- and post-test surveys including open-ended questions and a two-stage test involving a randomised experimental stage, and then a focus-group stage. In the experimental stage participants read one of three versions of the article: text only, photos only or text and photos. Each focus group contained participants who had all read the same version of the article. For data analysis, Robinson and Mendelson used inferential statistics, frequencies and textual analysis for the surveys; discourse analysis and inferential statistics for the experimental conditions; and narrative, discourse, textual or content analysis for the focus groups. This enabled them to elicit rich information about the meanings constructed by participants from their readings of the article, and also to compare the ways in which those meanings changed between the different types of media (Robinson and Mendelson 2012: 341).
Quantitative methods can be embedded within a qualitative framework using the technique of quantitisation, where aspects of qualitative data are converted into numbers for analysis. There are a variety of methods for this, including:
•counting, for example, how many participants said X and how many said Y;
•dichotomising, that is, identifying whether a participant did or did not say anything within a particular theme;
•frequencies, for example, which code was used most frequently and which least often;
•statistical analysis, which can ‘highlight patterns and relationships between groups of participants, thus helping researchers identify meaningful comparisons between contrasting cases (for example, participants, social contexts, and events)’ (Collingridge 2013: 82).
It is possible to combine other aspects of research, such as theories or disciplinary perspectives. Some researchers are working towards ‘integrated methods’, where many different aspects and viewpoints are brought together (Box 2.11).
Box 2.11: Integrated methods by Franz et al
Psychologists Anke Franz and Marcia Worrell from the UK, and Claus Vögele from Luxembourg, studied adolescent sexual behaviour in Germany and England. These multiple investigators drew on multiple theories to underpin their use of mixed methods of data gathering, which led to multiple datasets. They used Q methodology, questionnaires and measurement scales to investigate teenage sexual health, discourses about gender roles, sexual assertiveness and sexual self-efficacy. Data was at first analysed separately for each method of gathering data, then the findings were integrated ‘to provide a holistic explanation of cultural and individual influences on adolescent behaviour’ (Franz et al 2013: 383). This enabled more robust conclusions than the initial separate analyses because ‘The quantitative part could not explain the influence of discourses on young people’s behaviour, whereas the discourse research could not make inferences about the relationship of the discourses to individual characteristics’ (Franz et al 2013: 383–4). Essentially, integrating their methods allowed Franz and her colleagues to gain a more complete picture of a very complex situation.