Theoretical Arguments For Mixed Methods Research
The principal objection to mixed methods research – as we shall see later – is that it violates the belief that specific methods are inextricably linked to certain ontological and epistemological ideas, and that the data derived from each must therefore be interpreted in different ways and are not compatible. Arguments for mixed methods therefore have to find ways to overcome this objection. Strangely perhaps, some of these arguments are based on differing and at times even conflicting views about ontology and epistemology. This section looks at the following arguments in turn:
- Complementarity of approaches (put simply - differences between paradigms exist but it’s productive to work with different theories and methods if they are seen as parallel but separate)
- Dialectical position (differences between paradigms exist and should be worked with and faced head on)
- Paradigm differences do not determine difference in method (differences between paradigms exist but particular methods are not tied to particular paradigms)
- Differences between paradigms are over-stated
- Pragmatism (arguments about differences between paradigms have reached an impasse: it’s more important to get on with research)
- Transformative-emancipatory approach (development of a new ‘paradigm’ which claims to overcome previous debates or make them less relevant).
Complementarity Of Approaches
The quantitative approach linked to the positivist paradigm and the qualitative approach linked to the interpretivist paradigms can be employed within one study to explore differing research questions or different aspects of the problem. This approach is consistent with the view that no one truth exists: mixing methods can be seen as a response to the multi-faceted nature of reality. According to this argument, differences emerging in the findings from different kinds of data tell us just as much as any similarities. However, these differences are not seen as problematic in that one can help throw light on the other and, once explained, they are not in conflict with each other.
For example, a good theoretical case can be made for combining qualitative and quantitative methods in order to conduct different levels of enquiry. What is happening at micro level can be better understood by examining what’s going on at macro level and vice versa. Have a look at Sarah Irwin’s paper (2006) Combining Data, Enhancing Explanation (especially pages 4-7). In this paper, she reflects on the perennial theoretical problem of reconciling the competing notions of structure and agency, arguing that analysis of different data sets can help bridge the gap between subjective orientation and social circumstances. For instance, people ‘objectively’ located on a particular point in the social class structure very often do not see themselves as such. Irwin argues that what researchers say about this duality does not come from data alone, but from theory. Therefore we need to bring together different methods associated with different theories to understand what she calls ‘puzzles in sociological explanation.’
The Dialectical Position
Arguably, the dialectical position is intellectually the most challenging but potentially the most exciting approach to mixed methods. It is similar to the argument for complementarity of approaches, in that it examines distinct yet related questions, recognising different ontological and epistemological traditions. Rather than seeing differences, both at paradigm level and in relation to specific research findings, as resolvable through what Jennifer Mason denigrates as ‘parallel logic’, the dialectical position tries to grapple with these very contradictions. From this perspective, such discrepancies should spur us on to look more closely at the data, think harder about the implications and, ideally, go on to generate new theory. Researchers make synergistic use of methods, working with the creative tensions thus engendered (Tashakkori and Teddlie, 2003b, Mason 2006). Mason calls for an interplay between different ways of perceiving the social world rather than trying to unhappily accommodate them or even ‘squash’ them into one dominant approach.
Have a look at pages 9 and 10 of Mason’s (2006) paper which sets out this position (which she calls ‘multi-dimensional logic’). She concludes:
The opportunities for harnessing creative tensions and building on rather than ironing out the distinctive strengths of different approaches are substantial. Such an approach, like no other, can facilitate the researcher in asking new kinds of questions, ‘thinking outside the box’, developing multi-dimensional ways of understanding and deploying a creative range of methods in the process (p10).
Paradigm Differences Do Not Determine Differences In Method
An alternative view is that while there are significant theoretical differences between paradigms, it does not follow that there are differences in method because particular methods are not tied to particular paradigms. Peter Halfpenny (1997) is a proponent of this position: you can read his six point argument in his paper The Relation between Quantitative and Qualitative Social Research.
He acknowledges that social science is conducted within incompatible theoretical frameworks, which differ fundamentally in how they perceive both the social world and human action and thus, in how they go about understanding them. However, Halfpenny goes on to argue that ‘paradigms or approaches are logically independent from procedures’ and that qualitative and quantitative data are not fundamentally different, in the sense that each involves ‘stripping away content and context from the richness of lived experience’ (p.12). He concludes that researchers should be more imaginative in the range of procedures they select, rather than restricted to particular methods they mistakenly associate with their ‘favourite paradigm’.
Differences Between Paradigms Are Over-Stated
This argument has a different starting point but reaches a similar conclusion – that mixed methods is not theoretically problematic and can be a great strength in research. Rather than focusing on theoretical differences, which are seen as exaggerated or even false, this approach identifies similarities in certain aspects of the two main paradigms.
Martyn Hammersley has been writing about research methodology for many years, challenging the traditional division between qualitative and quantitative approaches, ‘one representing the true way, the other the work of the devil’ (Hammersley 1992: 39). Like Halfpenny, Hammersley rejects the view that qualitative methods are based on idealism and quantitative methods on realism, arguing that not all quantitative researchers are realists, and that some qualitative researchers are realists.
He differs significantly from Halfpenny, however, in characterising the differences in ontological and epistemological assumptions as ‘no more than working assumptions’, which researchers should try to overcome. For instance, Hammersley refers to arguments between proponents of different paradigms about the relative merits of natural and artificial settings for conducting fieldwork. He counters that social institutions, the site of much qualitative research, are not ‘natural’ while reactivity, or the effect of the researcher’s presence on data collection, may apply in either setting. To get a fuller flavour of this position, it’s worth looking at Hammersley’s response (2005) to Halfpenny’s paper above, entitled Continuing the Dialogue: Peter Halfpenny on paradigms and methods.
A relatively recent position promotes pragmatism as a ‘leading contender for the philosophical champion’ of mixed methods research (Greene 2008:8). Put simply, pragmatism suggests that the likely empirical and practical consequences of any idea should be taken into account when assessing its value or usefulness. Pragmatism developed as a response to philosophy which tried to solve problems by reason alone: its pioneers, the American philosophers Peirce, James and Dewey, wanted to solve problems in the real world. The key tenets of pragmatism, and the benefits of using it as a theoretical underpinning to social research, are set out in accessible format by Johnson and Onwuegbozie (2004) in a paper entitled ‘Mixed Methods Research: A Research Paradigm whose Time has Come’. You can find this in the journal Educational Researcher, 33, 7, 14-26.
The pragmatic rule or maxim or method states that the current meaning or instrumental or provisional truth value…of an expression (eg: "all reality has a material base” or “qualitative research is superior for uncovering humanistic research findings”) is to be determined by the experiences or practical consequences of belief in or use of the expression in the world (p. 16).
These authors emphasise that pragmatism represents a full philosophical system so you may be puzzled to see that one of its characteristics, set out in Table 1 of their paper, is ‘prefers action to philosophizing (pragmatism is, in a sense, anti-philosophy)’! Nevertheless, they contend that pragmatism presents a way forward when arguments between paradigmatic opposites have become stuck. Brannen (2005:10) summarises helpfully:
What distinguishes the pragmatic researcher from the paradigm-oriented one? In the paradigmatic version of the world, the former is more interested in ideas and their origin, in the ideas which drive the research and the ideas upon which the research should be founded. The concern of the pragmatist is to open up the world to social enquiry and hence to be less purist in terms of methods and preconceptions (about theory and method)… Pragmatic rationality will more readily embrace a mix of methods if the research questions and practicalities of the research context allow it.
The Transformative-Emancipatory Approach
The transformative-emancipatory approach shares pragmatism’s desire to solve problems in the real world. There has been a burgeoning of emancipatory research over recent years, covering many fields including disability, race, gender, post-colonialism, critical and queer theory. The key aim of this work is to challenge traditional power relations both in research and the wider social world. Mertens (2003) argues that the transformative-emancipatory approach is a paradigm in its own right (raising interesting questions about what constitutes a paradigm). Mertens also suggests that pragmatism does not provide a satisfactory value base for social research in that practicality does not necessarily involve challenge of the status quo. She argues that the transformative-emancipatory perspective forms a more appropriate paradigm for this field of research since, she says, it is now generally accepted that science is both empirically based and value-influenced:
Transformative scholars assume that knowledge is not neutral but is influenced by human interest, that all knowledge reflects the power and social relationships within society, and that a important purpose of knowledge is to help people improve society. (2003: 139)
At an ontological level, reality is located within historical, political and cultural settings. At epistemological level, value is placed on developing trusting and equitable relationships between researcher and research participants. At methodological level, mixed methods are seen as an appropriate (but not obligatory) way to address research problems relating to diverse groups.