SC971-7-__ : Survey Sampling, Non-Response and Inference


Applied Sampling is an applied statistical methods course that focuses on the design of probability samples to be used for data collection. Sample designs are driven by analytic goals of an investigator, but this is not an analysis course. In particular, this course focuses on the principles of designing and selecting samples of individuals. This course will focus on sampling populations of humans because of the unique challenges posed by such populations. This will include both methods for sampling the general population and methods for sampling specialist or minority populations, including screening methods and two-phase sampling. Sampling techniques covered will include simple random sampling, stratification, cluster sampling, systematic sampling, multistage sampling, probability proportional to size sampling, and flow sampling. Applications of these methods for face-to-face interview surveys, telephone surveys and self-completion surveys will be examined. Cost, sampling frames, and sampling error estimation techniques will also be addressed.

Aims

To provide a practical introduction to methods of survey sampling and estimation.


Learning Outcomes

By the end of the module students should:

      1. Understand fundamental characteristics of different sample designs.

      2. Recognise the strengths and weaknesses of alternative proposed designs for a particular survey

      3. Be able to apply techniques for sample design and sample selection.

      4. Be familiar with statistical notation and formulas for different sample designs.

      5. Understand and be able to estimate the effects of sample designs on survey estimates.

      6. Be able to design and describe, in writing, a multistage probability sample of a population.




Syllabus


    • Principles in sample selection

    • Sampling frames and frame problems

    • The methodology to draw samples and estimate correct measures of central tendency and variation using the following sampling techniques:

    • Simple random sampling

    • Stratified sampling

    • Cluster sampling

    • Complex (stratified, multi-cluster) sampling

    • Systematic sampling

    • Probability proportionate to size sampling

    • The calculation of design effects, the variance inflation factor

    • Introduction to advanced variance estimation techniques