Abstract
The collection of data through surveys is a costly and time-consuming process, particularly when complex economic data are involved.The paper presents an efficient approach, based on Gaussian quadrature, to survey sampling when some information is available about the target population. Using household data from Mozambique, we demonstrate that Gaussian quadrature subsamples, based on relatively easy to observe household characteristics such as size and educational attainment of members, generate better estimates of the moments of household expenditure than random samples of equal size.
Original language | English (US) |
---|---|
Pages (from-to) | 355-364 |
Number of pages | 10 |
Journal | Journal of the Royal Statistical Society. Series C: Applied Statistics |
Volume | 55 |
Issue number | 3 |
DOIs | |
State | Published - May 2006 |
Keywords
- Gaussian quadrature
- Sampling
- Survey methods
ASJC Scopus subject areas
- Statistics and Probability
- Statistics, Probability and Uncertainty