Pigida A. Evaluation and consideration of design-effect in multistage stratified sampling for national social survey in Ukraine.

Українська версія

Thesis for the degree of Candidate of Sciences (CSc)

State registration number

0415U005360

Applicant for

Specialization

  • 22.00.02 - Методологія та методи соціологічних досліджень

14-09-2015

Specialized Academic Board

Д 26.001.30

Taras Shevchenko National University of Kyiv

Essay

The dissertation covers the phenomenon of design-effect, which occurs in sampling surveys, and also covers the development of techniques designed for design-effect's reduction. The dissertation examines at what stages design-effect took place in construction of sampling for national survey in Ukraine, how it can be measured, and what approaches should be used to reduce it. Detailed analysis is applied for investigation of stratified selection stage and for impact of grouping elements of population into strata on the sample design-effect. There is offering of the technique that allows breaking of general population into strata, optimal from the point of impact on survey accuracy. Stage of cluster selection is analyzed as well as its impact on survey design-effect. The problems of constructing optimal cluster sampling in terms of consumption of resources are considered. The method for constructing of such optimal sampling units is proposed.

Files

Similar theses