Semi-structured dialogue

How do we Evaluate Programmes?

Data collection methods

used in the PPA  approach

Semi-structured dialogue uses a checklist of questions related to each topic of interest. Questions can be added or omitted as appropriate. It is a flexible tool which must be used systematically in order to produce valid results. Always remember that whatever exercise that you will be involved in as researchers is NOT an interview but a DIALOGUE. A key objective is to make the process as interactive as possible.

An example of SSD:

Topic: How do people in this area describe vulnerability?

Key issues and questions:

•  What are the local definitions of vulnerability?

•  Who are the vulnerable?

•  How do people cope with vulnerability?

Questions can be added or omitted as appropriate.

Key steps in SSD

During the actual session the following steps may be useful:

•    Always ensure that you have prepared and read through your checklist of questions, and that you are familiar with the topic/s that you are going to conduct an SSD on.

•    Begin with the traditional greeting and state that the interview team is here to learn.

•    Begin the questioning by referring to someone or something visible.

•    Conduct the interview informally and mix the questions with discussion.

•    Be open-minded and objective.

•    Let each team member finish their line of questioning (do not interrupt) unnecessarily.

•    Do not begin with sensitive questions - carefully lead up to these (use your best judgement);

•    Assign clear roles for team members, and rotate these;

•    Be aware of non-verbal signals – for example, are your respondents yawning?

•    Avoid leading questions ad value judgements.

•    Avoid questions that can be answered with ‘yes’ or ‘no’.

•    Do not carry on for too long – individuals and groups get tired even if they my be enthusiastic about what you are discussing with them.

•    Make sure that you have a list of topics and key questions written down in order in a notebook.

Common mistakes in SSD

As a researcher you should always be conscious of the common mistakes that are

usually made while carrying out semi-structured dialogue. These include:

•    failing to listen closely

•    repeating questions (sometimes because you were not attentive when a respondent was talking to you).

•    helping the interviewee give an answer (often done by completing sentences for respondents).

•    asking vague or insensitive questions.

•    failing to cross-check a topic, or failing to judge answers (believing everything)

•    asking leading questions

•    allowing the interview to go on too long

•    over-generalising findings.

•    relying too much on information from the well-off, more articulate members of the community, the educated and men

•    ignoring anything that does not fit your ideas and preconceptions.

•    giving too much weight to answers that contain quantitative data; and,

•    incomplete note-taking.

See aslo : Analysing findings from semi-structured interviews