In 1999, semi-structured interviews were carried out with a subsample of the MDICP-1 respondents to extensively investigate the role of informal conversational networks in rural Malawi.
Eligible respondents were a random sample taken from the MDICP-1 survey. The sample size was determined by estimating the sample size that would have provided variation across respondents similar to the the distribution of responses to attitudinal questions about family planning in the MDICP-1. Given the pronounced regional differences in responses in the MDICP-1, it was estimated that it was needed a sample of 20 wives and 20 husbands in each of the three regions and that a non-response rate of about 17% should have been expected. However, it was immediately clear that this estimate of non-response was too low, so that the sample size for the semi-structured interviews was increased. In the South, 38 women and 41 of their husbands were selected, and 23 and 28, respectively, were interviewed. In the Center, 37 women and 37 of their husbands adn 26 and 27 were interviewed. In the North, 41 women and 41 men were selected, and 27 and 25 of them were interviewed.
The interviewers were the best of the local interviewers who had participated in the MDICP-1, and were thus familiar with the overall emphasis of the project on informal conversational interactions. In order to encourage the interviewers to shift from the structured approach of the MDICP-1 household survey to a more conversational approach in these interviews, they were given a guide that listed the four major areas to be covered in the interviews, with a small set of specific questions in each area. The interviews were conducted in the mother tongue of the respondent, and usually transcribed and translated by the interviewer. An a priori coding scheme was developed by Susan Watkins and Eliya Zulu, and the interviews were coded for analysis.