Any ideas about exploratory mixed methods? Creswell (2014) suggests moving from the qualitative data analysis to scale development. Is it better to develop a scale, or questionnaire, or survey?
A lot depends on how much survey based content you already have. For example, I worked on a project where the innovative aspect was the creation of a new dependent variable. We had everything else that we needed in terms of background (demographic) variables and pre-existing measures for the independent variables. So, we concentrated on doing a series of focus groups that produced a 15-item scale to measure the dependent variable.
Alternatively, you may be in a much more "discovery" oriented mode, where you are unsure about what topics your survey should cover. In that case, you might need to start by generating hypotheses, followed by further qualitative work to create the measures to operationalize the key concepts you discovered in the earlier phase of your work.
Overall, the key point is that a qual --> QUANT exploratory sequential design is a general template that you have to apply to a specific set of circumstances. Once you are clear about your research goals, then you can work out the specific implementation of the design.
Good response from David. Yes - it is correct that there are variations depending on what you want to achieve, what the existing data suggests and where it can lead you to, and what approach allows you to answer the over-arching question that links the two paradigms and facilitates a relatively singular outcome that encompasses and answers the whole study. My other question Tatyana is 'how do you differentiate questionnaire and survey?'
I'd like to make one correction in David's response in that a sequential exploratory design weighs the data in QUAL-->quan form since the focus is to try to define the variables and qualities of an unknown phenomena. If the topic under study is already known, then a explanatory design is usually chosen and a typically calls for qual -->QUAN weighing in most cases.
If you need solid examples or in need of a more robust explanation of designs beyond Creswell, I'd suggest looking at the text: Tashakkori & Teddlie (2010). Sage handbook of mixed methods in social & behavioral research. There are also several other well established researchers in mixed methods and emergent designs, such as Nagy Hesse-Biber, Onwuegbuzie, and Mertens.
I'd also highlight Dean's question in 'how do you differentiate between questionnaire and survey?', as from my point of view the quantitative data collection can incorporate both open ended answers, as well as scale-based questions. In this particular design, the questions are derived from the qualitative data and use the language of the participants to create the questions. I'm sure in your qualitative results you discovered certain emergent themes and characteristics that define the topic under study and provide deeper understanding and meaning. Using these themes and qualities you can then develop questions to test the prevalence and variation within a second larger sample. The idea of this particular design, as I'm sure you are already aware of, is that the quantitative portion is used to support and expand the results of the qualitative findings. Hence, the qual - quan sequencing. Coming at your question from a phenomenological stance, you could frame the questionnaire (or survey) to test the emergent and contextual aspects of your topic question.
Hope this helps!
A good response Angeline,
However, while Qual-Quant is the most typical of sequential designs; there are plenty of examples of the Quant-Qual variation.
Absolutely Dean. I was just defining it in terms of Exploratory design where Qual-Quant sequencing is typically used. For explanatory design the sequencing would would be flipped. Go Kiwis!
Hi Angeline,
Glad we are agreed. I take it that you like the Kiwi thing. I do too - although I'm English - so very nervous about the All Blacks versus England this weekend; as well as the football World Cup!!
- Badges
- Science topic
- Similar topics
- Statistics
- Survey