The project collected three main types of data: (1)A second wave of a 2002 household surveys in Brazil and South Africa, which sampled just over 1000 households each in the Western and Eastern Cape region of South Africa, and Rio and Ilheus in Brazil, was collected in 2008. The sample frame in both countries was provided by census areas, the sample was proportionate to size, and stratified by urban and rural areas in Brazil and South Africa, and additionally by race in South Africa. The survey instruments included (i) a household questionnaire with a section collecting information on all household members; and (ii) and a supplement collected on all household members aged 60 and over, with questions on their entitlements, relations to other household members, health status, goals, social networks and community participation, and income decisions. The instruments were similar in all respects across the two countries. In South Africa, the questionnaires were translated into Xhosa and Afrikaans. (2)Semi-structured interviews with older respondents, stratified by age, race, and location in South Africa, and by age and location in Brazil. 30 in each country. The qualitative component complemented the information on households, especially as regards major events and household arrangements and circumstances in period intervening between the two survey points. (3)A mapping of institutions, policies, and programmes supporting older people and their households in the two countries. This study examines the impacts of individual ageing on the well-being of older people and their households in Brazil and South Africa, with a view to informing appropriate policies to address the challenges of accelerated population ageing in developing countries. The main hypothesis organizing the research is that the contribution of older people to social and economic development is a primary factor determining the dynamics of their, and their households', well-being. A core feature of the research will be to undertake a longitudinal comparative study of well-being among older persons and their households in South Africa and Brazil, by re-visiting a sample of households with older people from these two countries which participated in a 2002 household survey. The longitudinal dataset combined with in depth interviews will enable the analysis of the dynamics of household income and livelihoods in the two countries, including inter-generational relations and care. Comparison of South Africa and Brazil, two countries with rich social policy environments and recent innovations, will help identify the impact of old age support and existing anti-poverty programmes on the well-being of older people and their households.

The 2002 survey data was collected using a proportionate to size methodology, through face-to-face and telephone interviews; one-follow up to cross sectional. The same methodology was employed in Brazil, but pensioners receiving a non-contributory pension were over sampled using administrative data in the census areas selected for the study. The 2008 survey revisited 2002 households, plus replacements. Brazil 2002: 3252 cases; Brazil 2008: 2836 cases; South Africa 2002: 5560 cases; South Africa 2008:5769 cases; Qualitative data consists of transcripts of 30 interviews each for Brazil and South Africa. The Brazil interviews are in Portuguese, the South Africa transcripts are in English. The units under observation were households(1006 households in Brazil and 973 households in South Africa (note this is a second round of a 2002 survey; the 2002 survey was collected on 1006 households in Brazil and 1117 households in South Africa), and within households, persons aged 55 and over. The household survey was implemented on the most knowledgeable member of the household and a supplement was implemented on all persons aged 55 and over. A small sample of respondents was selected for semi-structured interviews.