Mixed Methods and convergence analyses

Home Forums Default Forum Mixed Methods and convergence analyses

Viewing 8 posts - 1 through 8 (of 8 total)
  • Author
  • #4493

    Does anybody has some experience using multimethod-multitrait matrix for to analyze data produced under a mixed method design?

    Thanks in advance.


    Jason Taylor

    Not entirely sure what your asking about here. Can you describe a little bit further? I Have lot of experience in mixed method analysis and visualization and can probably help or guide a little better if I have a clearer understanding of what your talking about. One thing that seems to work well is a 3 or 4 dimensional mixed trait visualization such as a bubble chart as in

    More info at http://www.dedoose.com


    Hi Jason, thanks for answer me!

    I will try to explain it in a better way. Look, I have, obviously, a lot of data gathered using a questionnaire with closes-ended questions and, in addition, data gathered through focal groups and open-ended questions. So, I want go to the next step, integrate these kinds of information. I can apply a design called multitrait-multimethod for to try to set the validity of my methods, usually a set of measures. This design was thought to be used with traditional methods (tests) and not with qualitative methods, however, Creswell (2009) proposed it as a possibility for get an unified results of our quan-qual data. Have you ever worked with this design? Please, look this link:

    Oh, Dedoose! I never heard about it!, thanks friend! Great!!. Well, maybe you can help me because I have another question for you, I’m evaluating some software programs, dedoose it seems a nice possibility, but I played with Qualrus, MaxQDA and QDA miner, do you know something about QDA miner, I like it because it has qualitative and quantitative functions, it is perfect for work in mixed methods. Any advice?


    Upps! Jason, I Dedoose your company? I’m sorry friend, but I think you are the perfect person to set a comparative between your web application and those other softwares that I mentioned.

    Jason Taylor

    Lol, yes I am one of the developers involved in the application, and EthnoNotes before that, so please excuse me if I do try to show off our new tools 😛 From my understanding of what you are doing, personally the way I would tend to go about it is to make a trait / tag / code system with weighted answers. You would then need to develop a rubric to accurately weight the open ended answers, where the closed ended answers would likely be much easier to weight. At that point all of your quantitative data has been marked with quantitative tag/weight information allowing you to then analyze the combined data in the same space. I however am no real expert on true mixed method theory, but I would be happy to put you in touch with our resident Expert Eli Lieber who should be able to assist further as well as fully answer your comparative questions regarding similar products.
    ~ JT


    Hi Jason.

    No worry, I liked to see your web app. Indeed is a great idea (I regret I never thought to do something like that :oP ) and it is not expensive, so I would try it before to buy a more expensive software like QDA miner (that is my choice at the first place). Thanks a lot for your suggestions, I like the idea about weighted answers. Tell me, are there some books/articles that you can recommend me?

    Sorry by my latency, I have a lot of work these days.


    Jason Taylor

    Hi Pablo, I am one the development side of the team. Eli Lieber, our support specialist as well as a professor in the field for quite some time should be assisting more on this topic in a few.

    Eli Lieber

    Hi Pablo. Been watching your conversation with Jason and he pulled me in to offer a bit more from the perspective of my experience. The software you are looking at among a set of very useful tools for qualitative data analysis (I particularly like MaxQDA for some of it’s features). That said, Dedoose and much of my work for the past decade in academia are really focused on mixed methods research. We built Dedoose because we couldn’t find a solution for really advanced level research that truly integrated methods. I still use stats programs for the higher-level quant analysis, but then bring these results into Dedoose to move much deeper into qualitative and mixed methods work. Your reference to Creswell’s suggestions for how to bring methods together in a multi-method multi-trait matrix analysis is well taken. While this approach is right on from a theoretical basis, I think it is less straight-forward from an analysis point of view unless the qualitative data can be converted to quant and then used in the traditional Campbell and Fiske sense. It’s complicated stuff and ‘mixed methods’ are really a big (and growing) family of strategies to capitalize on the strengths of different traditional disciplines. It’s all good, but really open territory and we work with people who apply mixed methods in lots of ways. I think the ultimate goal is to choose the combination of methods that best serve the research problem and then figure out how to implement as effectively and efficiently as possible (and do it well and within budget and time constraints)e. Dedoose was designed to be as flexible and intuitive as possible so it can accommodate anything a mixed methods researcher throws at it. From what you describe, it sounds like you can do some great stuff by integrating as much of your demographic and scale score (quant) data as possible, linking these data to your qual interviews and focus groups, and using a code system and code weighting in thoughtful and creative ways. Glad you decided to check out Dedoose to see if it can server your needs and I’d be happy to respond to more of your questions. A few other things to be aware of with Dedoose besides its low-cost is how easy it is to work collaboratively with others and how easy (essentially automatically) it is to get great presentation graphics that can tell the visual story and serve as filters to take you deeper into your findings. We’ll look forward to hearing what you think.

Viewing 8 posts - 1 through 8 (of 8 total)
  • The forum ‘Default Forum’ is closed to new topics and replies.