I agree with all of the above. I would suggest pragmatically that the size of the survey and number of respondents might guide you. If you have few qualitative questions and few subjects it is probably not worth doing a laborious content analysis by hand or by software. Instead you could extract the main themes and quantify them. It is often good to get a colleague to test the validity of the themes you have extracted, if you have not many response then this is possible. Otherwise I would suggest content analysis is the best way forward as already mentioned. Often it is the richness and quality of your analysis that will count - what do you make of the survey data? Non-parametric analysis of a poorish content analysis or of a small data set can lead to spurious conclusions so I do encourage people to look deeply at the raw data rather than jump too quickly to statistical analysis. It is also worth seeing what is different and what is the same about your findings from the literature and what is special about the population you sampled from? There is no substitute for rational analysis and triangulation of results in social science. After all at the end of the day the views you collect are only the views of the sample!

Dear Jing,

It is probably better to separate your qualitative responses from your quantitative ones at least initially. The appropriate qualitative technique might depend upon the length of the responses. If they are generally short phrases then perhaps a synonym frequency analysis would be appropriate. If they are generally longer then you can carry out a content analysis.

...Or, my dear,  you can code open-ended answers for the variables/codes of your interest and treat them as quantitative data and run statistical analysis; However, at the same time you can content analyze them and combine qualitative-quantitative data. This is what I do in my current study.

Dear Peter,

Thanks so much for your information! The qualitative responses in the survey are quite short without long content, in this case, synonym frequency analysis sounds suitable. I am just wondering if initially I separate qualitative analysis from quantitative, later on how to integrate the results (qualitative & quantitative) into the paper as a whole.  

Dear Lili, long time no see!! Hope everything is super fine with you and your work!-:) I take great interest in learning how to combine qualitative-quantitative data in the paper. Could you recommend me some of your own publications or other literatures on this topic? Then I will have much clearer clues of how to operate these techniques in analysis. 

it can be sort first in different group then it can be coded and analyze...

If you have access to one of the major qualitative data analysis programs (e.g., ATLAS, Dedoose, MAXQDA, or NVivo), they will all let you both code your qualitative data and merge it with you quantitative data. Most of these programs refer to this as "mixed methods" analysis, so you could check out the tutorials on this subject on their websites.

You can use the qualitative data through content analysis software or can also use it to supplement the findings of quantitative analysis during discussion.

I agree with all of the above. I would suggest pragmatically that the size of the survey and number of respondents might guide you. If you have few qualitative questions and few subjects it is probably not worth doing a laborious content analysis by hand or by software. Instead you could extract the main themes and quantify them. It is often good to get a colleague to test the validity of the themes you have extracted, if you have not many response then this is possible. Otherwise I would suggest content analysis is the best way forward as already mentioned. Often it is the richness and quality of your analysis that will count - what do you make of the survey data? Non-parametric analysis of a poorish content analysis or of a small data set can lead to spurious conclusions so I do encourage people to look deeply at the raw data rather than jump too quickly to statistical analysis. It is also worth seeing what is different and what is the same about your findings from the literature and what is special about the population you sampled from? There is no substitute for rational analysis and triangulation of results in social science. After all at the end of the day the views you collect are only the views of the sample!

Jin, you asked about how to combine the qualitative and quantitative data in your work. I would recommend thinking in terms of the results from analyzing the qualitative and quantitative data. One possibility is what is known in mixed methods research as an "explanatory sequential" design or QUANT --> qual. In this case, you begin by analyzing the larger quantitative data set and then use the supplementary qualitative data to gain a better understanding of those results. 

That approach is particularly useful when you have strong research questions that are not fully answered by the quantitative results, so a further analysis of the qualitative data will help you "generate hypotheses" about why your quantitative results came out the way they did.

In contrast, if your quantitative results are more or less as predicted, one of the primary uses for the qualitative data would be "illustrate" those results with brief quotations. This is sometimes known as "giving the data a human face."

How to deal with qualitative data of the open-ended questions in the quantitative surveys?

Suggesting before you structure your survey questionnaire, you need to ascertain your research objective, research question & conceptual framework / research model with the rigorous literature review also on the research method including the survey questionnaire.  

In the event you can't adopt / adapt others' questionnaire & you need to develop your own, try to avoid yes/no questions as they can't offer detailed information from respondent, can't represent their deeper opinion, limited analysis can be done & lower explanation power etc.  Instead of asking yes / no question then followed by open ended question - why not ask categorical, ordinal, interval / ratio type of questions especially based on literature you'd reviewed..

Also don't combine both closed questions (for quantitative analysis) & open questions (for qualitative analysis) in the same survey questionnaire unless you intend to perform a mixed method research e.g. concurrent triangulation, transformative etc. designs.

Alternatively, you can also consider like what other researchers / scholars had stated - to perform mixed method sequential explanatory / exploratory designs.  You can refer to an example of sequential explanatory in the following RG link:

https://www.researchgate.net/publication/259478672_Relationship_between_Team_Satisfaction_and_Project_Performance_as_Perceived_by_Project_Managers_in_Malaysia_-_A_Mixed_Methods_Study

Article Relationship between Team Satisfaction and Project Performan...

In my experience on the treatment of surveys, I suppose that there would have been a previous selection of the respondents.

To perform the evaluation of the results I propose to identify the number of identical words that the group has used to respond, use Chi2 for all of them and discriminate with yes / no after, applying the same statistic.

Afterwards, I would observe the obtained results and apply qualitative and quantitative analyzes. The answers (yes / no) can be used as demo varialbles.

All of this presupposes the use of the words of the answers as well as quantitative or qualitative variables, which allow us to analyze the human coexistence both in positive (yes) and negative (negative).

I do not know if this strategy can help you, but I would also consider the guidance given by our colleagues earlier.

Dear colleagues, thanks so much for sharing your invaluable experience on the subject! I appreciate your contributions a great deal! I have very carefully read your advice and suggestions. I will take a try according to my research questions and the type of information the survey respondents have offered. Many thanks from Jing

Not sure if this will help now. I am using a thematic approach. This involves reading and reading again, coding words and phrases, compressing those and identifying themes. These themes emerge from groupings of factors, ideas and thoughts. The key aspect then is to write up the findings in a way that describes the process and the outcomes so that it adds value to the research. 

I did the same as Professor Sushil has recommended in a similar study of mine.