In most of the research, we select mixed method based on our research questions. The objective of the mixed method is to make best possible use of qualitative and quantitative

Quantitative method is objective in nature while qualitative focuses on subjective elements. it is possible that the thing supported by quantitative approach does not go along with qualitative. you as a researcher has to provide the justification through literature and with supportive arguments that why there is a discrepancy.

 it can also become your contribution as your research will challenge the existing tested questionnaire designed to measure technology adoption

Hi Kashif,

I'm not quite sure what you mean by "match." If you're really doing mixed methods (vs. dual methods), then matching isn't really an outcome of the methods. For a real mixed methods approach, you employ methods that should inform one another. If you are collecting data separately and attempting to match them, then this sounds like a dual methods study rather than a mixed method approach. In mixed methods, you effectively have either sequential or concurrent designs, in which your data collection is done in an effort to achieve something in tandem that a single method would be unable to do. Commonly, researchers apply it to explain the results. It's also used to achieve other ends, like triangulating results or contextualizing your inferences. 

So, as I said, I'm a bit confused as to how they do not match... as the mixed data are collected in an effort to mutually inform one another. They are not collected to have different forms of data that say the same thing.

In your case, I would also caution you against using the term "significant" for qualitative research. It's a reserved term and has an operational definition for statisticians.

I would also consider the types of questions that are appropriate to each method and how those questions inform one another. The qualitative analyses can tell you why something was important. 

What I might suggest for you, is to explore what tests, error, or noise in the analyses you ran might have lead you to conclude that there not significant when the qualitative analyses suggest that they are important. From a quantitative perspective, there is a tremendous amount of intervening variables that may have influence over your statistical outcomes that you simply have not, or cannot, account for in your model. While technology adopters may perceive one thing, the mathematical model may not be robust enough to show the specific influence of that factor.

So, I suppose... when all is said and done... it's not crucial for the results to align. But thinking of it in terms of two separate data collections is problematic and not strictly indicative of mixed-methods research or the utility that they present.

PG

In mix methods we focus two important components:

Quantitative Part: Cover the  adequate sample size to generalize the findings as far as possible. For this,  standard survey questionnaire is prepared, and sometimes validated tools are used. But this method has difficulty to answer the 'cause of cause'. 

Qualitative Part: In qualitative part, the researchers try to find out the answers of 'causes of causes'.  This is called 'Thick Description'. 

Regarding your question, if the results does not match to each other methods, it is important to revisit the research protocol, and especially the design part, make sure that both the protocol are focusing on same issue and methodologically correct. 

I agree with the feedback above. Quantitative to achieve generalizable data and qualitative to explore and probe into interesting aspects from the quantitative' statistical analysis.

There should be a link between the two with the quantitative being used to inform the direction of the qualitative investigation.

Therefore, in relation to your question - My view is that they do not necessarily have to match, but one should inform the the other.

I think it depends on the purpose that you are pursing with your two data sources, and there a number of different goals that can be met with mixed methods. One common goal is data triangulation, which is also known as convergent results. For that purpose, you do want to see equivalent results. If not, you have what is known as divergence rather than convergence, and you would usually need to collect more data to understand the source of the difference.

FYI, an example of a research design that does not assume equivalent results would be what is known as a sequential explanatory design (QUANT --> qual), where the purpose of the qualitative study is to better understand the results from the quantitative study.

"Significance" in scientific way can be tested only with quantitative analisys

In addition to the above valuable comments, please also refer to the following papers which may be helpful. As Jick (1979, p. 607) has pointed out that: “……divergence can often turn out to be an opportunity for enriching the explanation.” 

• Jick, T. D. (1979) Mixing Qualitative and Quantitative Methods: Triangulation in Action, Administrative Science Quarterly, 24, 4, pp. 602-611.
• Onwuegbuzie, A. J. and Leech, N. L. (2006) Linking Research Questions to Mixed Methods Data Analysis Procedures, The Qualitative Report, 11, 3, pp. 474-498. 

As if I understand it is not the matter of matching of result in terms of both approaches, rather one approach (say quantitative) should be used to support or reject the result obtained by another approach (say qualitative).

I believed the mixed method approach is used in a sequential manner where you use the qualitative to further probe/explain the result in quantitative approach. One issue here....how you get 'Significant' result in qualitative phase?

I highly appreciate all comments. I did use sequential explanatory and it was the qualitative part that did not match with the quantitative one. 

@Schrader; data is collected separately but from the same sampling frame. First quantitatively and later qualitatively.

@ Bano; if discrepancy would be treated as contribution then I'd love to have it

@ Others; I am trying to support the discrepancy from literature

Most of the work on discrepancy falls under the heading of "divergence" in triangulation. For triangulation, you need to conduct two independent studies on the same research question, where the classic goal is to have them agree (convergence), but it is a very different situation if they disagree (divergence).

In contrast to triangulation, sequential explanatory designs usually incorporate the results from the original QUANT study into the design of the subsequent qual study. So, if you simply studied the same thing twice, without letting the first set of results influence the second study, that would be closer to a triangulation design rather than an explanatory sequential design.