I wish things were as clear as Senthilvel states, but if you look at most of the examples that Creswell & Plan-clark give for embedded designs, they turn out to have a sequential component where the a dominant QUANT design is the source of the qual component. this amounts to: QUANT --> qual.
Personally, I think the term embedded design leads to more trouble than it is worth, since you can get more clarity from determining whether there is a primary method and whether the combination of methods is sequential or not.
FYI, what Senthilvel describes would amount to QUANT + qual or QUANT --> qual, then why not just address that issue directly?
Things become even more confusing when you recognize that "nested design" is also a special case of ANOVAs which refers to special measurement scenarios: In a nested design-ANOVA, certain levels of one factor can only be realized at certan levels of another factor. Example: 2 teaching methods, 4 teachers at 2 schools with teacher A and B at school X and teacher C and D at school Y. Then the factor "teacher" is nested into the factor "school"
I agree with David that one should better name the design eplicitly than use a category like "nested" or "embedded" which is nearly meanningless without further specification
I am also very interested in the mixed method designs.My understanding is pretty much the same as Senthilvel, the embedded/nested design is less dominant within the primary design. the primary design carries more weight or is more dominant. I am planning to use a mixed method design, initially I was looking at sequential explanatory but then I felt concurrent nested/embedded will be more ideal for my study. The more I read the more I got myself confused, what David and Werner are describing is actually what or where I am, confused .
Part of the reason Nolundi and others are confused by this terminology is that these terms largely come from lists, especially those compiled by John Creswell, where the items are not necessarily mutually exclusive.
For example, as I said earlier, nearly all "embedded" designs are actually of the form QUANT --> qual where the QUANT is dominant and the qual is used as a follow up.
In Creswell's terms, this would be "sequential explanatory" design (where the origin of that label is from an article by Jennifer Greene et al., 1989). So, in practice, most embedded designs are also explanatory sequential designs.
My advice would be to begin by deciding which method is your primary one. That means determining what your overall goals are and thus which kinds of methods have the core strengths that you need to meet those goals. After that, you should choose a supplementary method that makes an important contribution to your method.
I disagree with David, because nested design involve collecting data at the same time while the term sequential as use to be attached to either of the designs mentioned by creswel signifies collection of data at different time.
So to easily get it right, you can use quan or qual as your primary method and the left one as secondary or embedded method while collecting data at same time, with embedded data supporting your primary data after analyzing them differently
In their seminal book, Creswell and Clark (2010) propose four major types of mixed methods research designs: convergent (parallel or concurrent) design, embedded (or nested) design, sequential (explanatory sequential or exploratory sequential) design, and multiphase designs. In this paper, we illustrate the use of an embedded mixed method (EMM) research design. Embedded mixed method designs are described by Creswell and Clark.
Reference: Using Embedded Mixed Methods in Studying IS Phenomena: Risks and Practical Remedies with an Illustration