Dear Houda Boumediene

One aspect that is relevant to all your questions is the threat to validity. The threat might have varied origin:

• - the interviewer effect (how might the age, ethnicity, gender or affiliation of the interviewer influence the responses?)
• -sources of bias (such as social desirability bias, where the participants provide the responses that they perceive as desirable)
• - the techniques used to interview (such as use of leading questions)

However there are multiple other factors to consider, this text may be helpful:

http://discovery.ucl.ac.uk/1436174/2/semi-structured_qualitative_studies.pdf

data summarizing for each interview and cross analysis are the major issue here. please see the example in my paper

Article The Current Status of Customer Relationship Management (CRM)...

What I find most important to me is to think how and whether the interview schedule answer the purpose and objectives of your research. Without that clarity and linkages, you may collect a lot of data but it will be challenging to establish the linkages.

I would say before you ask a question, it must be in your mind how that question is going to contribute or address your objectives. Sometimes we go with the flow and do not think about it. But if you think before you ask, it may save the time of the interviewer and the interviewees.

There are already a lot of paper on the strengths and weaknesses of qualitative research. Depending on the context and aim of your research, you choose what is best for you. Just want to mention that generalisation of research findings and universalisation of application is not possible as it is very much context specific.

My comments are as follows:

1. Remember you are working with a human being, a person. To gather rich qualitative data in semi-structured interviews you must attend to this fact. This means attending to creating a relationship, a context, that allows the participant to reflect upon their experience in relation to the phenomena under study. Trust is a major component of what must be considered particularly related to the sensitivity of the experience to be investigated.

2. In relation to "thinking" prior to the interview. I would say being sensitized to the conceptual framework that supports your study and informs your questions is vital. You must be aware of what potential concepts and ideas may surface in the interview while simultaneously not holding onto these ideas too tightly- making sure to be open to new ideas.

3. The strengths of semi-structured interviews 9if a safe, warm, trusting environment is created and the interviewer has good interviewing skills) is a reflexive conversation leading to rich understanding of the phenomenon.

the weaknesses: these have been elaborated upon above from what I would say is a predominately post-positivist perspective. From my perspective there's a tension of overly controlling the interview w/ too many questions asked during the interview. Space must be allowed, from my perspective, to allow the interviewee take the question where they wish to go, to the meaning they wish to make. The interviewers job is to follow, to inquire, and deepen description.

Much more could be said, but I'll stop here for now. David

The following publications may further help:

• Adams, W. C. (2010) Conducting Semi-Structured Interviews, in Wholey, J.S., Hatry, H.P. and Newcomer, K.E. (eds.) Handbook of Practical Program Evaluation, Chapter16, pp. 365-377. (See: Disadvantages and Advantages of semi-structured interviews, pp. 366-368).
• Qu, S. Q. and Dumay, J. (2011) The qualitative research interview, Qualitative Research in Accounting and Management, 8, 3, pp. 238-264.
• Schmidt, C. (2004) The analysis of semi-structured interviews, in Flick, U., von Kardorff, E. and Steinke (eds.) A companion to qualitative research. London: SAGE Publications Ltd, pp. 253-258.
• Whiting, L. S. (2008) Semi-structured interviews: guidance for novice researchers, Nursing Standard, 22, 23, pp. 35-40.

You say that you're going to do semi-structured interviews---so you are not doing completely unstructured interviews; therefore, you do need to think about eventual analysis; but even before this, you need a clear purpose----what are you trying to find out? Answering this question will help in deciding how structured your interviews should be.

Do you need demographic data, for instance? Age? Sex? Gender (not quite the same)? Employment?

Will you state the parameters of the topic to respondents?

The answers to these type of questions will depend on your research question---what do you want to find out?

And thinking about how you will eventually analyse the data collected can help in constructing actual questions to ask respondents, as well as enabling you to 'visualise' your overall project.

So thinking about analysis can only help rather than hinder interviewing practice.

As a research tool, the strength of the semi-structured interview is essentially in the degree of 'researcher control' over data collection towards answering a research question; however, this pre-determination can be its weakness---say from the point-of-view of unstructured interviews in which respondents can 'talk themselves out' on a topic, potentially introducing aspects which the researcher may not have considered.

I would like to add to the answer from Philip Adams, about the semi structured interview and its strengths and weaknesses. When you make the questions for a structured interview, your pre-conceptions and imagniation might hinder what can come forth in the interview when compared to an unstructured interview where it is the interviewee that delimits and sets the boarders for the subject.

Before deciding omn unstructured, semi structured or structured interviewe you need to have thought through what knowledge you are interested in, since the way questions are posed sets the boarders for the answers. Things you have not thought to be of importance will probably not come forth since no questions are asked about things that you do not know or think about.

In my opinion,

1) Before doing qualitative research, you should know what you want to know or study and what is your aims of study. If we know our aims, semi-structured interview will be determined to answer research questions. Before collecting data, you can take semi-structred interview to consult with research team or expertist in that field to get the comments. It will help you see variouse perspectives. Semi-structered interview needs to improve to get rich data from interviwing.

2) During interview, research is best tool to use semi-structered interview in ask, interview, discuss with participants. It is only a guide and to prevent we will not forget about questions. If participants do not understand your pre-determined questions, you might change or revise the question again after first interview for participants understanding as general language that participants can understand without techincal terms/ vercabulary. Because some participants can not understand our questions.

Pros:

- Helps in getting insights.

- Saves a lot of time

Cons:

- Unstructured

- Difficult to interpret

- Bias may result in leading questions