Modern Sampling Algorithms for Surveys with Applications in R                              

Level of instruction: Advanced   

1-Day course

Instructors : Maria Dickson & Yves Tillé  

About this short course

The aim of the course is to introduce modern sampling methods such as balanced sampling, spatially spread sampling and temporally coordinated sampling. It will also show how to put them into practice using different R packages (sampling, BalancedSampling, Wave, StratifiedSampling). The question of determining an optimal plan based on a superpopulation model will also be addressed.

About the instructors:

Maria Michela Dickson

Maria Michela Dickson is full professor at the Department of Statistical Sciences, University of Padua (Italy). Her studies and research are mainly focused on spatial sampling, probability and non-probability sampling methods for socio-economic surveys, measurement errors, panel rotation, quantitative social criminology, spatial statistics methods, and spatial diffusion processes of social, economic, health and environmental phenomena. She is committed in scientific dissemination both at high school, university, and research level, participating to many conferences and seminars. During her career she has taken part as statistical expert to several EU and national projects concerning social interactions, urban crime, econometrical high frequency data, COVID-19 diffusion, and forestry surveys.

Yves Tillé

Yves Tillé is full professor at the University of Neuchatel, Switzerland. He obtained his PhD. in statistics at the Free University of Brussels, Belgium, before becoming a lecturer at the National School of Statistics and Information Analysis in Rennes, France, and then a professor at the University of Neuchatel, Switzerland. His research contributed to the development of the cube method, a balanced sampling technique that ensures that the estimators of the totals of selected samples are equal or nearly equal to the totals of the population variables. Throughout his career, Yves Tillé has published numerous articles and books in peer-reviewed statistical journals. His work has also led to the creation of several algorithms and methods that are now widely adopted in official statistics.

Plan:

• Population, sample and sampling design

• Algorithms for simple random sampling
• Unequal probability sampling
• Balanced sampling
• Spread-sampling (GRTS, local pivotal, WAVE)
• Methods for sample coordination
• Sampling in a stream
• Discussion of design optimality under one model

Learning outcomes

• Select an appropriate sample design for your survey objectives
• Define the sample design
• Implement sample design and select sample in R language
• Estimate parameters and make inferences from the sample

Course is perfect for: PhD students and advanced researchers in the field of sampling, statisticians from statistical institutes or international organizations, statisticians in ecology and forestry using spatial sampling methods.