Social Sciences In Public Health

Question 1

The incompatibility thesis

Answer 1

Arguments against mixing methods often maintain that quantitative and qualitative methods represent fundamentally different approaches, and should best be understood as occupying separate and competing ‘paradigms’ (See, for example, Guba and Lincoln, 1994). In considering these arguments, we will briefly re-visit how key distinctions between quantitative and qualitative approaches are popularly characterised, including in relation to differences between ‘positivism’ and ‘interpretivism’, ‘deduction’ and ‘induction’, and ‘realism’ and ‘critical realism’ (see, for example, Lecture 1). Advocates of mixed-method approaches may also reproduce the idea that qualitative and quantitative designs are ‘paradigmatically’ distinct.

Question 2

Pragmatism

Answer 2

Arguments for mixed methods approaches often emphasise pragmatism, proffering a ‘horses for courses’ approach, wherein methods are envisaged as a kind of ‘tool box’, with methods selected according to what works, their feasibility and specific purpose. Methods are chosen on the basis of their suitability to address specific research needs. The pragmatic stance suggests that the epistemological differences between quantitative and qualitative approaches have been exaggerated.

Question 3

Different types of triangulation

Answer 3

 method triangulation – comparing data produced by different methods  data triangulation – comparing data in different settings, at different points in time or place  investigator triangulation – comparing multiple researcher perspectives within a team  theory triangulation – approaching data using different explanatory models or hypotheses

Question 4

The idea of triangulation as a way of maximising validity is open to a number of criticisms.

Answer 4

It can be argued that there is no such thing as a single valid account. Qualitative research is accepting of relativist multiple views of equal validity; and different methods and data come in different forms which sometimes defy comparison. For instance, it is sometimes not easy to compare data from treatment records or surveys with that of interviews or observations as the data that these produce can be quite different and may be measuring different things (see also, the paper by Wagner et (2012) in the Short Loan collection). Data produced by different methods provide ‘parallel datasets’, each giving a partial view. Similar findings provide corroboration or reassurance, but the absence of similar findings does not provide grounds for refutation. Taken together, triangulation does not guarantee validity or check for accuracy. Rather it deepens understanding through multiple perspectives.

Question 5

two distinct models for doing social science within larger health research studies (linked to mixed methods)

Answer 5

Sequential

Concurrent

Question 6

Sequential

Answer 6

Sequential – Here, mixing methods is a staged process, usually when qualitative data collection (stage 1) informs quantitative data collection (stage 2). This is probably the most common way in which qualitative methods are integrated into larger studies in applied health research. The priority here is quantitative data collection, with qualitative methods a junior partner, whereby qualitative data collection/analysis leads to the development of instruments for quantitative research. Qualitative-quantitative data integration occurs at the analysis stage (in theory: this perhaps happens less often that hoped for).

Question 7

Concurrent

Answer 7

Concurrent – Here, quantitative and qualitative data are collected at the same time, with equal weight to each. Qualitative-quantitative data are integrated during data collection or analysis to best understand a research problem. Investigators may present quantitative and qualitative data and their analyses separately or together (though often the former), exploring how findings corroborate or contradict each other. Some refer to the concurrent use of mixed methods as a ‘triangulation model’.

Question 8

Benefits mixed methods

Answer 8

Mixed-method approaches capitalise on the strengths of the different methods used to compensate for the weaknesses in any single method.

The benefits of mixing methods include: corroborating findings; minimising alternative explanations when interpreting findings; describing and explaining divergence or negative cases; and generating richer and deeper understandings. Examples of rationales for mixed-methods research include:  Using qualitative data to inform quantitative measures or instruments  Using qualitative data to question and/or aid the interpretation of quantitative findings  Using qualitative data in to explore process in evaluations or interventions  Using quantitative data to assess the generalisability of qualitative findings  Using multiple qualitative methods in combination