Small area estimation

Silvia Pacei

  • Date:

    01 JANUARY
    -
    30 MAY 2026
     from 11:00 to 13:00
  • Event location: Aula IV, Department of Statistical Sciences, Via Belle Arti 41 (STAT PhD Classes Virtual Room on need)

  • Type: Cycle 40 - Elective Courses

Aims: to provide statistical methods for the small area estimation (SAE) of population parameters.

Learning outcomes: at the end of the course students will be able to recognize when small area estimation is necessary and choose the most appropriate strategy. Moreover, they will be able to evaluate the improvement in estimates due to small area methods.

Course contents

-        Basic concept and problems in SAE. Examples of need for reliable socio-economic statistics for small areas/domains.

-        Statistical approaches suggested in the literature: direct estimators and model-based estimators derived from implicit or explicit small area models (area level models and unit level models).

-        Three possible approaches to estimation: EBLUP, EB and HB estimators.

-        Hints at possible extensions including non normality assumptions, multivariate models, estimation of ranks or percentiles or histograms (Triple-goal estimation), spatial models, time series models and small area estimation in presence of non-sampling errors.

-        Practical applications in laboratory using R.