Course teached as: B029809 - DESIGN AND ANALYSIS OF SAMPLE SURVEYS Second Cycle Degree in STATISTICS AND DATA SCIENCE Curriculum STATISTICA UFFICIALE
Teaching Language
English
Course Content
The overall aim of this course is to provide participants with the knowledge required: a) to project and realize a sample survey with attention to the most common sources of non-sampling errors; b) to use the collected survey data in order to estimate different parameters of a population by choosing in each case the more correct approach.
Yves Tillé (2020) Sampling and Estimation from Finite Populations, John Wiley & Sons Ltd
Paul S. Levy, Stanley Lemeshow (2008) Sampling of Populations 4th Edition, JOHN WILEY & SONS
Steven G. Heeringa, Brady T. West, Patricia A. Berglund (2010) Applied Survey Data Analysis, Chapman & Hall/CRC
Sarndal, Swensson and Wretman (1992) Model assisted survey sampling. New York, Springer Verlag
Hedayat and Sinha (1991) Design and inference in finite population sampling. New York, Wiley
S. K. Thompson (2012) Sampling, 3rd Edition. New York, Wiley
G. Nicolini; D. Marasini; G.E. Montanari; M. Pratesi; M.G. Ranalli; E. Rocco (2013). Metodi di stima in presenza di errori non campionari. Milano: Springer-Verlag Italia
Class notes and slides.
Learning Objectives
After completing the course the students should be able to:
- correctly design and analyze simple and complex sampling strategies for the study of specific phenomena;
- estimate different descriptive parameters of a finite population, through a correct chose/use of estimators belonging to different types and provide for each estimate a measure/estimate of its error;
- analyze the problems associated with the presence and statistical treatment of non-sampling errors, with particular attention to the missing-data problems.
Prerequisites
Preliminary teachings: "Statistical Inference and Computational Methods" ("Inferenza statistica e metodi computazionali") and "Probability and Mathematics for Statistics" ("Probabilità e matematica per la statistica")
Teaching Methods
Lectures and classrooom exercises
Type of Assessment
The evaluation is based on three elements:
- Written exam including both practical exercises and questions on theory;
- Project based on the analyses of data
- Oral exam: the student who received a sufficient evaluation in both the written exam and project will be admitted to the oral exam; the oral exam covers the theory and interpretation and includes the discussion of the project; at the end of the oral exam, a final mark will be given
Course program
The basic concepts in sampling theory for finite populations and the Horvitz-Thompson's estimator.
Most commonly used probabilistic sampling designs: simple random sampling with and without replacement, stratified sampling, cluster sampling, two-stage sampling, systematic sampling, probability-proportional-to-size sampling, complex sampling designs.
Types of non-probability sampling.
Sample Size and Sample Allocation.
Estimation using known auxiliary variables: the ratio estimator, the difference estimator, the regression estimator, the post-stratified estimator, calibration and weighting estimators.
Designing samples for repeated surveys.
Non-sampling errors; frame imperfections; unit and item nonsponse. Missing data: assumptions and treatment. Reweighting and imputation methods. Multiple imputation.
Variance estimation for complex sampling design. Bootstrap methods.
Estimation for domains. Small area estimation methods: model assisted and model based methods.
Some method will be illustrated using empirical analyses from various fields. Practical analysis will be done with R and/or Stata software.