24.01.2020 Friday 11:00-12:00
Notes
Lecture 2: Household Surveys and Empirical Methods I
Examples of some modern household surveys:
- Indian National Sample Survey (NSS) 1950
- Malaysian Family Life Survey 1976-1977
- World Bank Living Standards Measurement Study (LSMS) Surveys
- Work began under Robert McNamara in 1979
The first LSMS surveys were in Peru and Cote d’Ivoire in 1985. The purpose of these surveys is to measure living conditions, compare the poverty rates across countries and to investigate the relationship between poverty and growth.
LSMS Surveys
Living Standards Measurement Study (LSMS) Surveys that are conducted by the World Bank or other comparable surveys that are conducted by national governments are presently available for a wide range of countries.
The surveys generate relevant data for policy makers and the research community. It is a response to a perceived need for policy relevant data that would allow policy makers to move beyond simply measuring rates of unemployment, poverty, and health care use, for example, to understanding the determinants of these observed social sector outcomes.
The Content of Household Surveys
The main components are the household questionnaire, the price questionnaire, and the community questionnaire.
The household questionnaire consists of:
- Household composition, housing, fertility
- Education, health, economic activities
- Food expenditures, durable goods expenditures
- Other income sources including remittances, saving, assets, credit markets
- Anthropometric measures
Vocab: Anthropometric –
The sampling uses a two stage sample design. Clusters are selected at the first stage of the design. A cluster can correspond to a community or village. In the second stage, households are picked from each selected cluster. This two stage sample design lowers the costs of collecting data, it facilitates revisits, and makes it a lot more worthwhile to collect community data.
Stratification is the sampling within specifically targeted groups of interest to the researcher.
Vocab: Stratification –
Sample weights are used to ensure that the sample is ‘representative’ of the population.
Household Surveys versus Census
A population census covers the entire population by definition. The data obtained in a census is much less detailed than the data obtained from the household survey. Because conducting a census is much more costly, it is conducted much less frequently than household surveys. For example a census can be conducted once every 10 years. The data collected from the census can be of use to a large variety of studies and still be effective compared to household survey data. The data from the census can also be used to cross-validate the data gathered from the household survey.
Home 