Semi-Structured Interviews

(Adapted from FAO 1990; WEST LOTHIAN COUNCIL 2004)

The semi-structured interview is the most common form of interviewing people and is a common and useful tool in the exploring phase of a planned SSWM intervention. Semi-structured interviews are conducted with a fairly open framework, which allow for focused, conversational, two-way communication. They can be used both to give and receive information.

By using this type of data collection, the interviewer has worked out a set of questions beforehand, but intends the interview to be conversational. Therefore, the interviewer can change the order of the questions or the way they are worded. The interviewer can give explanations or leave out questions that may appear redundant. Not all questions are designed and phrased ahead of time. The majority of the questions are created during the interview, allowing both the interviewer and the person being interviewed the flexibility to probe for details or discuss issues. The interviewer follows the guide, but is able to follow topical trajectories in the conversation that may stray from the guide when he or she feels this is appropriate.

The main job is to get the interviewee to talk freely and openly while making sure you get the in-depth information on what you are researching.

(Adapted from LAFOREST 2009)

Conducting a good semi-structured interview, requires a thoughtful planning which includes: identifying respondents, deciding on the number of interviews and preparing the interviews.

Identifying Respondents

Key informants from a community under the study regarding water and sanitation issues are “privileged witnesses, or people who, because of their position, activities or responsibilities have a good understanding of the problem to be explored”.

These witnesses are not necessarily members of the targeted location, but they all have a major interest and knowledge about the location. In addition to being privileged witnesses of specific problems, they may represent specific client groups and areas, have administrative responsibilities in a municipality or community organisation, be experts in the filed of water, sanitation or agriculture, etc.

It is recommended that the stakeholder table (see also stakeholder identification) or the list of community organisations be consulted to identify key informants. If necessary, members of the relevant committees or organisations can also be asked to suggest respondents or even to act as key informants themselves. It can be useful to draw up a list of potential respondents before launching the recruitment process.

Number of Interviews

It is hard to determine the exact number of interviews that have to be done in order to get enough information about your topic. However, several questions must be taken into account in deciding on how many interviews to conduct:

When semi-structured interviews may be viewed as a way of supplementing other data collection methods, it may be sufficient to conduct only a few interviews with key informants from the study community. However, if semi-structured interviews are the sole source of information, more interviews should be conducted.

Semi-structured interviews can be a way of capturing concerns and perceptions of groups that have not been contacted with other data collection methods.

Time is needed to prepare, conduct, transcribe and analyse a semi-structured interview for scientific research. The number of interviews scheduled must take into account available time and resources. It is also important to consider the fact that key informants are often stakeholders in high demand and therefore not always available.

Under optimal conditions, data collection from key informants should end once data saturation is achieved, i.e. when interviews do not provide any new or additional insights because the information gathered is repetitive.

Even though semi-structured interviews are flexible, they require rigorous preparation. It is essential to define their objectives, devise an interview plan and draw up a consent form. Some of the main steps are as follows:

• Contact the respondent in advance and explain them the goal of the interview,
• obtain his/her permission,
• schedule an appointment and agree on where the interview will be held. The place selected should be neutral, confidential, comfortable, quiet, free of distractions, and easily accessible for the respondent. If necessary, send the consent form and the interview plan to the respondent.
• Furthermore, if required, prepare equipment for recording the interview.
• It is always good to contact the respondent again to confirm the date and location of the interview.

If you are doing this for the first time, it can be helpful to do it quite exactly – however, you must remember that you may not be able to stick to all your questions exactly how you have written them down. It’s best to make your interview guideline on a sheet of paper that you can take along, even when you are planning to record the interview on tape or video:

• Write down what you want to say during your introduction – don’t forget to include a sentence about yourself and why you are conducting this interview.
• Write down a number of general questions that you would like to discuss, and think of the more detailed sub-questions that can be attributed to the general questions. This will help you to be prepared – you can never predict in which direction an interview will go.
• Maybe mark key questions hat you really need to have an answer to, and “second priority” questions that are just “nice to know”.
• Don’t forget to think about a concluding statement, e.g. what you are going to do with the data collected.
• Take along this interview sheet to every interview. Make a new copy for every interview, and leave space on every sheet to write down name, contact number and date of the interview. This will help you to get back to the interviewee, should you have some questions afterwards.
• If you are working with a translator, don’t forget to discuss how key terms are going to be translated!

(Adapted from LAFOREST 2009)

The duration of an interview depends of course on the questions that need to be discussed – however, make sure you do not prepare an interview that will last for too long, so that the interviewee gets annoyed:

Initiating Interviews

An easy way to start an interview is to introduce yourself to the respondent and then remind him/her of his/her goals and projected length and the topics to be discussed. It is important to tell the respondent that he/she will be interviewed as an expert or as a representative of a group of people or an organisation. If the interview is to be recorded, the respondent must be asked for his/her written or verbal consent and reminded that his/her statements will be kept confidential at all times. This is a good time to have the respondent sign the consent form, if necessary. Note that taping can only be done with the prior approval of the respondent. If he/she refuses, notes must be taken instead.

If necessary, formulate questions so that informants answer on behalf of the people they represent. Listen carefully to all answers and ask more questions to obtain additional information. Ensure that key informants thoroughly understand each question.

Conducting Interviews

Start the interview with a general, open-ended question. Then ask as few questions as possible because the respondent should do most of the talking. Making reference (anonymously, of course) to statements made in other interviews or to findings based on other data sources can a good way to encourage respondents to express themselves. It is also useful for validating information already gathered. Respect the respondent’s pace and do not be afraid of pauses or silences. Interviewers should not judge what respondents say. They must keep the interview focused on the topics previously defined, refrain from suggesting answers and be careful not to ask leading questions. Be careful not to ask closed questions that leave respondents no room to elaborate and that can slow the interview’s pace. Be sure to cover all of the pertinent topics included in the interview plan. Ask clear and direct questions such as How? Where? When? Who? What? Why? How much? How many?

Concluding Interviews

Since semi-structured interviews do not consist of closed questions, it may be hard to end them. When interviewers feel that all topics have been discussed and that the time set aside for the interview is up, they can ask the respondent if he/she has anything to add. Interviewers must then thank the respondent for participating, explain how the rest of the study project will proceed and mention that the results will be sent to him/her once the study is complete. Soon after the interview, it is a good idea to summarise what the respondent said and to write down your impressions and any things that stood out. This will make it easier to transcribe and analyse the interview.

(Adapted from LAFOREST 2009)

Ideally, transcribe and reread the tape recordings or notes made during the interviews. Since transcribing recordings in their entirety can be a fairly cumbersome task, a compromise solution is to simply identify and write down the main themes that emerge as you listen to the tapes if the interview has been recorded.

Classify the information gathered using an analytical framework based on the topics discussed by the respondents during the interviews. It can be useful to work with specialised software which makes it easy to categories the interview. Identify the main ideas expressed for each topic and identify the most important points and classify them. Validate the findings among the members of the project team and with other designated authorities. Be careful not to construe too much, and try to remain unbiased.