It looks like your design involves at least 3 steps, with a clear sequential explanatory design (QUANT --> qual) in the middle -- where you follow-up your survey with qualitative methods to help understand the quantitative results. In case you are wondering why I labelled the follow up as "qual" rather than "QUAL," the reason is because the qualitative methods are being driven by the original quantitative results. In my view, whether a study is primary or supplemental depends more on its purpose than on its size, and here the purpose and design of the qual methods is determined by the prior QUANT study.

Following up on the results of this section with a quantitative pilot study is also fairly straightforward, leading to QUANT --> qual --> quant.

What is least clear to me, however, is beginning with a concurrent design. Ordinarily, the input to a QUANT study is an exploratory sequential design (qual --> QUANT), so you would need to be very clear about the purpose of the quantitative study at this point. In particular, "triangulation" is no longer a favored term because it means too many different things. For instance, it can mean examining the convergence between studies that are intended to produce similar results. Or it can mean producing complementary results where each method makes a different contribution to the overall outcome.

The clearest message I would offer, however, is that you are trying to do entirely too much. Very often, I see studies that begin as a simple, two-stage sequential design collapse into a mono-method study because of the time and effort required to complete both stages. In your case, you are proposing to use at least six methods over the course of the project, which would almost certainly require an expert team with several years of time and a large budget.

An excellent response from David. I fully agree. Your approach looks overly complex Jonathon - but nice diagram/schematic!! That said - are we seeing it all? Your image looks chopped at the ends. Your 'pilot' at the end confuses me. We normally pilot to 'pre-test'. What are you pre-resting for. It gives the impression that there will be at least one more phase to your study. My other query would be 'are you not triangulating data?' Your triangulation appears to be 'front-loaded'. If that is the case - is it a triangulation of the literature? Your qual phase has two data collection techniques - single interviews and focus groups. Is there a comparative element to your study - hence the focus groups? That would add even more complexity.

Hi David L Morgan and Dean Whitehead

I have reflected over your responses and can see that my current process is overly complicated!

It is for a PhD study, and so I do have a reasonable amount of time on my hands, but even so, I can see that I need to simply considerably.

I believe the way to go, based on your feedback, is to split this up into a series of smaller studies, as follows:

1. exploratory survey leading to quantitative data analysis and final 'themes'

'qual --> QUANT'

2. case studies to test 'themes'

(within-group analysis for each case study, via 'qual --> QUANT')

3. Focus group to discuss findings of the case studies

'QUAL'

Thank you for taking the time to respond to my query - I definitely feel that this is now a more direct and straightforward approach.

Best regards,

Johnathan

I could use some help understanding your first step:

1. exploratory survey leading to quantitative data analysis and final 'themes'

First, survey research is seldom exploratory because you need a clear idea about which topics you want to address in order to write the questions that will "operationalize" the variables that are necessary to measure the underlying concepts. Second, quantitative analysis seldom produces themes, which are usually the result of interpreting qualitative data.

Hi David L Morgan

I think I could be misleading things slightly by my use of the word 'themes'.

The initial exploratory survey (large scale qualitative survey to the general public) will be combined with secondary research findings to develop a series of 'factors' or 'attributes' (I'm not quite sure what is the best term to use here).

These factors/attributes will then feed into a MaxDiff/Best-worst to work out the order of importance of these factors/attributes.

This will then allow me to take those important factors/attributes and build scale questions around them for use within a standard experiment (I originally labelled this as a case study, as it is based on a 'case study building'). This will be a 'within-group' design, structured as: 'Participants --> Pre --> Post --> Post Test'.

I hope that makes (some) sense! And thank you for your feedback, it is helping me develop the study.

Kind regards,

Johnathan

Could you give us more detail about what you mean by a "large scale qualitative survey"? The main reason most qualitative studies do in-depth interviews with a relatively small number of participants (typically 20 or less) is because of the difficulty of doing qualitative analysis. If you don't already have experience with this kind of analysis, I would warn you agains collecting a large amount of qualitative data.

Hi David,

Sorry, my incorrect wording again - I was imagining somewhere in the region of 30-35. I have some experience with qual analysis from my Masters so I am hoping this approx. amount should be manageable. Also going to make sure I don’t have too many questions as well!

Best regards,

Johnathan