I am wondering if using group discussion to support statistical results of quantitative research method can be an adequate technique in sensitive topics such as fraud, corruption, etc. ?
In mixed methods research, what are known as "sequential explanatory designs" use qualitative methods to follow up on the results from earlier quantitative studies (QUANT --> qual). The basic purposes for these designs is to get more insight into how and why the quantitative study produced the results that it did, which often includes getting illustrative quotations from the participants.
Sure- Record/count the rounds, structure the agreement/disagreement objectives, and solicit formal/individual opinions / votes. Quant can proeuce agreement / consensus indices also, rank-ordering an option in groups.
Mattew Kerry has told you how to generate quantitative data from group discussion.
David L Morgan has told you how to use group discussion to "support" the interpretation of quantitative data by using qualitative data.
However, if you consider the topics are sensitive and people may not give you the honest answers in "quantitative research method" such as questionnaire, you may use group discussion to "check/ challenge" the statistical results.
One the other hand, if you suspect the topics are sensitive and people may not give you the honest answers in a "quantitative research method", it is no use for you to "support" the statistical results by group discussion.
You may need to give more details for people to give deeper suggestions.