IASS Webinar 45: A nested error regression model with high-dimensional parameter for small area estimation

Map Unavailable

Date/Time
Date(s) - 30/10/2024
2:00 pm - 3:30 pm

Category(ies) No Categories


IASS Webinar 45:   A nested error regression model with high-dimensional parameter for small area estimation

 

Speaker:  Nicola Salvati, Hukum Chandra Memorial Prize Winner

The University of Pisa

 

30 October 2024  at 2pm – 3:30pm (CET)

 

All are invited to the webinar, organised by the International Association of Survey Statisticians.

 

Please register for the IASS Webinar at:

https://us06web.zoom.us/webinar/register/9317199219179/WN_9dppQxuoQMi4i8mh6r8pGQ#/registration

 

 

After registering, you will receive a confirmation email containing information about joining the webinar. There will be time for questions. The webinar will be recorded and made available on the IASS and ISI web site. See below for the abstract and biography of the speakers.

 

Webinar Abstract

Planning and evaluation of government programs often require reliable statistics across various socio-economic, environmental, and health measures at national and sub-national levels. Due to the sparseness of data in smaller geographical areas, standard estimation methods are inadequate. The primary goal of small area estimation (SAE) is to leverage strength from related sources via statistical models that integrate different data sets. Mixed models are predominantly used in SAE literature due to their ability to borrow strength from larger datasets.

Small area estimation faces significant challenges due to heterogeneity in the data and the need to provide reliable estimates for areas with limited data. Traditional methods often fail to account for this heterogeneity adequately, leading to biased estimates and unreliable uncertainty measures. Our objective is to develop a model that incorporates heterogeneity in regression coefficients and variance components, providing a more accurate and robust framework for SAE.

We propose a nested error regression model with high-dimensional parameters that allows for heterogeneity in both regression coefficients and variance components across different areas. The model uses small area-specific estimating equations that facilitate appropriate pooling of information from a large number of areas. We develop parametric bootstrap and jackknife methods to estimate mean squared errors, standard errors, and coefficients of variation, providing comprehensive measures of uncertainty. Both model-based and design-based simulation experiments are conducted to validate the performance of our methodology. Our approach outperforms traditional methods, offering significant improvements in accuracy and robustness, making it a valuable tool for policymakers, researchers, and analysts.

 

References

Lahiri, P., & Salvati, N. (2023). A nested error regression model with high-dimensional parameter for small area estimation. Journal of the Royal Statistical Society: Series B.

 

 

Biography

Nicola Salvati  is Full Professor in Statistics in the Department of Economics and Management, University of Pisa, Italy. He is director of  the Tuscan Universities Research Centre – Camilo Dagum on  Advanced   Statistics for the Equitable and Sustainable Development – ASESD. He is Associate Editor for the Biometrical Journal and the Journal of the Royal Statistical Society (Series A). His research is focused on small area estimation, and particularly its use to estimate poverty measures based on M-quantile and latent variable models. His research interests also include survey sampling, robust regression, spatial statistics, new technologies in survey methodology, Big Data analysis and statistical machine learning. His most recent area of research involves development of new statistical methods based on latent variable models for estimating parameters from non-deterministically linked data.