Try the SAGE handbook of social research methods to help you clarify the issues and to learn how to work with your data. If you can not find the book try the Journal of Mixed Methods Research. Lots of interesting articles that could be of use. If that does not work try googling Julia Brannen, she is the grand ole lady of mixed methods research in Europe. Good luck!

TQ Linda, I am trying to find them.

There is a dedicated chapter to mixed method research in "RESEARCH DESIGN" by John Cresswell. I have used the book and find it to be an excellent reference for doctoral researchers. I can provide you with a copy in PDF format if you let me have your email address.

Thank you very much sir, my email is [email protected]

Very interesting discussion on mixed-methods research with "3 different conclusions". In the book by Vicki L. Plano Clark and John W. Creswell (2008) titled "The Mixed Methods Readers" pages 288 - 294, may help answer your concerns Gabriel. Especially with Meta-inference. Page 893 provides some good information on "Multiple validities legitimation".

Thank you Lynne, but i did find some clues after reading some journals where cross case analysis is usable in my concern. Any comment on that?

Ms Lynne, can you please provide me the book that you mentioned/ suggested earlier i.e., The mixed methods readers in pdf form? Probably you can email to me if you have the book. My email add is [email protected].

Thank you in advance.

The first issue is how your various data sets are related to your research question. In particular, are they all designed to answer the same question? Or, does each one serve a specific purpose within the overall project?

This is directly related to your choice of research design. For example, if the quantitative data from the survey is your main data source, and if the two smaller qualitative data sets are intended to add different kinds of content to that core data, then your design could be represented as: QUANT + qual.

But you may be worrying more about this issue than you need to. The current standard for integration in most substantive fields would simply be to have 3 separate Results sections and then a Discussion section that considers how those results relate to each other.

Thank you Prof David. Fyi, Yes, generally speaking all these three responses are used to answer the same research question (i.e., problems/ issues of the current institutions). This is some sort of ONE BIG/ Main research question. In order to do that, i have to come out 3 different sub-questions that will each of them pursues specific purpose (meaning to say, i have to gain three responses-from three different perceptions i.e., public, government, and private sectors' perspectives) as you mentioned above BUT all of them later on account for the same aforesaid main research question. The responses among 3 sectors are somewhat interrelated, so i believe after producing three separate results (although separated, they are related), in discussion part, i will do the 'mixing' or integrating in particular (according to some literatures, this can be referred as 'cross-case analysis'). One more thing of my research strategy is employing mixed-method concurrent triangulation strategy (both of them given same priority and executed simultaneously on my data collection & data analysis).

I suggest a chapter with the integration of the results. You mention 3 stakeholder groups as 3 sets of results but you also have those for the survey ie 4 sets of results. Make a matrix table and compare against each set of results. That way you can see where findings are similar, different or perhaps add extension to the question being explored. Leaving the integration to the discussion chapter reduces the potential of presenting fully the integration of the data.

It sounds like the best approach is to treat the data as three different perspectives on the same basic topic. That would indeed point to three sections of results, with the integration in a separate discussion section.

I'm uncertain about using the term triangulation. The key issue is how well the population for the original 200 responses matches the subgroups who participated in the qualitative interviews. Most people interpret triangulation to mean making a direct comparison between two sets of results where the only thing that differs is the method. So, you should avoid the term triangulation if there are differences in both the methods you used and subgroups you studied with those methods. Alternatively, if the survey respondents also consisted of officers in private companies and government officials, like the qualitative interviews, then you would be in a better position to make a direct comparison.

Thank you for your suggestion Prof Sharon, You are indeed pointing out a good direction. Expansion of 'mixing' or integration is vital e.g., before integration,, transformation of the data type (i.e., case description on the quantitative data) and make, say one or few sentences, a paragraph to facilitate the comparison/ integration.

Prof David, you are correct. Thanks for alerting me about this. and i believe it is more suitable for me to use the embedded/ nested design of mixed method.

One of the best and highest quality sources for methodological questions is:

http://www.qualitative-research.net

At the "searche" section simply type your key word and you will get plenty of interesting articles.

I don't think you want to call your research "embedded" because Creswell and Plano-Clark (who defined this term) use it to mean a separate qualitative study that is included within a larger quantitative project, and which pursues a different set of goals from the quantitative study as a whole.

If you look at almost all their examples of embedded studies, they turn out to be QUANT --> qual sequential studies, where something about the core quantitative study leads to a follow-up with a smaller qualitative study.

But I'm not sure why you want to have a technical label for your research design -- especially since you are trying to apply that label after the fact. Instead, I think that all you need is a brief description of your goals and why the led you to collect these three kinds of data.

Hm~Thank you Prof David. However, there were some studies also indicate Qual(quan) design which quantitative approach is embedded (as supplemental, or supporting role or as enhancement)- realising that my quantitative method is less prioritised in this objective (secondary method) and most importantly, according to some sources, this nested or embedded design besides one of the methods become less prioritised, this is useful when we are studying different groups' perceptions on something (e.g., students, client, private etc) Qual(quan) in particular (Hanson et al., 2005). Yes i do agree that the questions for this embedded design may mean that, the secondary method is for entirely separate question but it can also be meant to get a broader perspectives from different groups (Creswell 3rd edition, 2009) another source i gain to support this is from Dr Terell on mixed method design (The Qualitative Report Volume 17 Number 1 January 2012 254-280 http://www.nova.edu/ssss/QR/QR17-1/terrell.pdf). However i do agree with you that technical labelling is not so important rather i should put emphasis on the description on how my methodology process flows based on the aim, research questions and objectives. TQ

Dear,

here you go

I think it will be valuable for you