Abstract

Abstract: The field of survey statistics has undergone significant transformations over the years, shaped by evolving methodologies, technological advancements, and the increasing availability of diverse data sources. This presentation explores the historical development of survey statistics, from early reliance on exhaustive censuses to the gradual acceptance of sampling techniques. It critically examines the impact of emerging data sources, such as administrative records and internet-generated data, highlighting their limitations in extrapolation. The discussion also addresses challenges faced by official statistics, including declining survey response rates and data privacy concerns.
Additionally, the role of new statistical techniques, including machine learning and calibration methods, is analysed in the context of improving data integration and inference. The presentation concludes by emphasizing the need for continuous innovation, methodological rigor, and institutional adaptation to maintain the relevance and credibility of survey statistics in an increasingly data-driven world.