Insight through integration: Mixing methods in psychological research
Insight through integration: Mixing methods in psychological research
By Carole Gardener
Anyone who has recently taken the cognitive or critical social psychology modules may be surprised to learn that a mixed methods approach to research design has been described as the new star in the research sky. This development comes from a growing group of researchers prepared to challenge the widely accepted idea that quantitative and qualitative research tools are philosophically incompatible, adopting instead a more practical approach to research design. For anyone thinking about taking on some form of independent research this approach opens up a number of new opportunities. However, as my own experience shows, it can also present a number of practical and philosophical challenges.
The mixed methods approach is, very simply, a methodology that allows researchers to use both quantitative and qualitative tools within the same study. Thus in-depth interviews or observations can sit alongside the use of semi-structured questionnaires or assessment scales. Most psychology students will be aware that there is already an established, but limited, tradition of combining these approaches, usually involving the use of qualitative tools to generate areas of interest in the planning stages of quantitative projects. However, the mixed methods approach takes this further, reflecting a new belief that there are some circumstances in which fully combining qualitative and quantitative tools in the research design produces more valuable findings than could be achieved using a single approach.
One example of how the mixed methods approach can enhance the research process is shown in a study by Sutton and Rafaeli (1988). These researchers wanted to test the hypothesis that there is a positive link between shop workers displaying positive emotions towards customers and higher retail sales. In order to achieve this they carried out structured observations in a large number of shops, but surprisingly found that displaying positive emotions was negatively correlated with higher sales. As the authors had combined this quantitative approach with both qualitative interviews and unstructured observations they were able to explore this relationship further, discovering in the process that increased sales put staff under more pressure and created an environment in which they had less opportunity to be friendly towards the customers. In his chapter on mixed methods, Bryman (2012) argues that the above study illustrates how mixed methods are not only useful in identifying relationships between variables, but can also be used to explore wider explanations concerning the processes underpinning the relationship. This, he notes, is particularly useful when the study produces somewhat puzzling results.
My own experience in health research has also highlighted further applications of the mixed method approach. Health research is an area where researchers are often required to evaluate complex interventions - patients may be at different stages of their illness, require different levels of intervention and have multiple health problems. In addition, funders not only want to know what works and why, but are increasingly interested in patient and carer experiences. Farquhar et al (2011) have described the value of using mixed methods in this context, drawing on their evaluation of a new service for people experiencing chronic breathlessness. At different points in the study they drew on a variety of research tools including validated scales, case studies, randomized control trials and in-depth interviews. Farquhar et al argue that using such a variety of research tools enabled the researchers to develop a comprehensive understanding of the service, addressing not only provider outcomes but also exploring patient perspectives. In addition, using differing approaches enabled the research team to cross-check the results (triangulation) and generate new areas of interest to be followed-up as the study progressed (initiation). Faced with the complexity of health needs a mixed methods approach has an intuitive appeal and is becoming increasingly popular in a health research context. However, the advantages apply equally to smaller scale studies in psychology - generating explanations, comprehensiveness, triangulation and initiating further areas of exploration are all models that can be successfully applied in this context.
But how can mixed methods researchers justify combining these approaches when the widely accepted views about psychological knowledge suggest this is simply not possible? In other words, how can studies, such as the above, employ a mixed design when both quantitative and qualitative methodologies are each seen as embedded in fundamentally incompatible philosophical viewpoints? Or when the recognized rule of thumb is that quantitative methods are the only acceptable tools within 'scientific research' and qualitative tools the preserve of critical psychology? Supporters of mixed methods have responded to this 'elephant in the room' with two useful arguments. Firstly, they point to the difficulty in demonstrating any concrete link between research methods and specific philosophies, arguing instead that research methods should be seen as independent entities. As an alternative, they have proposed the idea of a more pragmatic framework where research tools can be assessed in terms of their technical strengths and weakness and combined to overcome any limitations. For those of us contemplating integrating qualitative and quantitative tools, such counter-arguments provide further encouragement to move away from the restrictions imposed by the on-going paradigm wars and increasing support for the idea that combining methods is not only possible, but also at times extremely advantageous.
Of course, in addition to the above philosophical concerns, anyone using this approach is also likely to face a number of practical considerations. When I incorporated this approach into my Masters dissertation I immediately encountered a couple of obstacles. Firstly, although combining methods draws on the same techniques as applying each approach individually, it requires the researcher to be skilled in both areas. As someone who naturally leant more towards a qualitative research style, combining the two approaches, meant I had to take time out of my research to get up to speed with both quantitative data collection and analysis. In addition, using two different approaches is potentially more time consuming, increasing pressure on deadlines (as I found to my cost) and, where relevant, on research budgets.
Whilst the above obstacles can be overcome with good planning and preparation the real challenge for me came in attempting to ensure that the methods I used were truly related, rather than just two studies under one umbrella. Having collected quantitative data on Emotional Intelligence (EI) and coping, as well as qualitative data about coping strategies, I was keen to imitate the studies above and use the qualitative data to provide a more comprehensive account of the relationship between the quantitative data sets. But how do you go about this in practice? In this project I used the EI scores to categorize the participants and then compared the strategies emerging from the qualitative data across these categories. However, this raised the further problem of how to represent these findings. In a purely qualitative study you would not use quantification to describe results, or vice versa, but in a combined approach is there an argument for greater integration at this stage as well as within the broader research design? In the end I went down the tempting quasi-numerical route, even representing my findings graphically. Many would argue however, that this was mixing methods too far.
Despite the potential pitfalls, and the on-going debate about how appropriate it is to combine methods, there is no doubt that the popularity of this approach is on the increase. The number of mixed methods studies being published has been growing. There are now journals dedicated to this approach and the issue is increasingly being addressed in psychology textbooks. For those considering this approach it is an exciting time. Choosing to use mixed methods in a research project may cause some headaches but, as the above hopefully shows, it can also provide the opportunity to be part of a new way of thinking about research design.
Bryman, A. (2012). Social Research Methods, Oxford: Oxford University Press.
Farquhar, M.C., Ewing, G. & Booth, S. (2011). Using mixed methods to develop and evaluate complex interventions in palliative care research. Palliative Medicine, 25(8), 748-757. doi: 10.1177/0269216311417919.
Carole completed the OU's postgraduate conversion diploma in psychology in 2011 and is currently working as a Research Assistant at the Primary Care Unit, Department of Public Health at the University of Cambridge supporting projects taking a mixed methods approach to living with breathlessness http://www.phpc.cam.ac.uk/pcu/research/research-projects-list/living-with-breathlessness-study/.
This article first appeared in News & Views December 2013.