Methods & Data
In order to achieve the aims of the project (to understand the implications of harmonizing definitions of the 'household' for survey and census data to represent the realities of intergenerational relationships in Anglophone and Francophone settings in Europe and Africa), we employed three methods across each of the five project countries.
1) Key Informant Interviews (approx. n = 30 per country). Identifying Key Informants all along the chain of data production (from survey designers and commissioners, to interviewers, to statisticians, to policy makers and academics) we interviewed people about their roles in the collection and/or use of household survey data: the sorts of data sets that they used, the way in which these data sets were helpful or not, the way in which these data sets could be used and any difficulties and challenges that respondents had using these datasets. Discussions particularly focussed on the way in which the "household" is defined and used in these surveys/census and the extent to which survey households are able to capture and represent intergenerational relationships. We spoke to people in government ministries, local and national statistics offices, universities, commercial research organisations and NGOs and charities. All interviews were anonymised, transcribed verbatim and entered into an NVivo database. The UK and Burkina transcripts will be available on the ESRC Data Archive in 2013.
The NVIVO codebook was developed during a three week workshop and built on the codebook developed for the pilot study in Tanzania. The whole team participated in the first week of the workshop and then the country level research assistants worked together for two weeks developing and refining the code book, standardising their use and understanding of different codes and jointing coding a number of interviews.
2) Household Interview Case Studies: Qualitative interviews with households in a number of settings about their household (how they subjectively understood this), their relationships of care/support across the generations and inter- and intra- exchanges of care. With:
a. 12 households in Senegal, Uganda and Burkina Faso in each of 4 sites (a rural village, a small town, an established district area of the capital city and a poor unplanned area of the capital). N=48/country
b. 8 households in UK and France in each of 4 sites (a rural site, a suburban site, Ugandan diaspora in UK/Malian, Burkinabe, Senegalese diaspora in France, non-diasporic households living in the city). N= 32/country
We adapted household interview schedules commonly used in surveys in order to record the information on household structure by hand. Data from grids was then anonymised, input into Excel, and coded inductively using Nvivo. The data from the UK and Burkina Faso household interviews will be available on the UK Data Archive in 2013.
Household interviews in the UK were undertaken in English, interviews in France were in French, interviews in Uganda were in Lugandan, English, interviews in Senegal were in French, Wolof and interviews in Burkina Faso were in French, Moore, Dioula, Dagara, Birifor, Bwa. Where interviews were not in English or French, they were translated into these languages by an interpreter during the course of the interviews.
For each household a summary profile was produced highlighting the characteristics of that household and particular difficulties or ambiguities (if any) would exist if specific household definitions were applied. These profiles were entered onto the NVivo data base and anonymised.
3) Review of household definitions Analysis of all available censuses and household surveys from 1960 to 2011 across all 5 countries. Using surveys and interview guides found in university and national libraries and in statistics offices in each of the 5 countries, we recorded the way in which the household has been defined.
These definitions were then analysed longitudinally to examine how they have evolved since 1960 and comparatively to examine similarities and differences between different sorts of surveys, between Anglophone and Francophone Africa, between France and the UK, between the UK and Anglophone Africa and France and Francophone Africa in order to enable us to obtain a rich understanding of how these definitions have evolved, how they have been used and the extent to which they inform each other.