|Ms. Ieva Burakauskaitė
|Non-probability sample integration in the survey of Lithuanian census
|The Statistical survey on population by ethnicity, native language and religion 2021 differs from its previous round in 2011 as it consists of both the voluntary sample and the probability sample drawn from the rest of the census population. Two approaches to the evaluation of sample estimates were considered – a natural post-stratified calibrated estimator and its combination with corrected estimates. As it turned out, the post-stratified calibrated estimator tends to underestimate small proportions of interest (e.g., minor religions). Since the inclusion of the voluntary sample results in a significant voluntary response bias, the propensity scores for individuals using the demographic and the previous census data were evaluated. As it will be presented at the conference, the combination of the corrected estimates with the calibrated ones improves the estimation accuracy for small proportions of interest such as minor religions.
|Dr. Andrius Čiginas
|Small area estimation in the survey of Lithuanian census
|The sample size in the Statistical survey on population by ethnicity, native language and religion 2021 is not sufficient to derive accurate direct estimates of parameters in small domains like municipalities. Estimation becomes even more complicated due to the small proportions of interest. We apply robust design-based composite estimators which exploit domain-level information available from the previous full census. We will present also an estimation of the accuracy of these shrinkage estimators at the conference.
|Mr. Albert Motivans
|Considering demand-driven approaches to address gender gaps in official statistics
|In relation to the Sustainable Development Goals and national/regional development programmes, more attention has been paid to data that can support greater understanding of and inform policy efforts to promote gender equality and inclusion. However, official statisticians around the world face challenges in meeting these demands, due to a variety of reasons (specialist knowledge, resources, tools, etc.) This paper examines the demand-side for gender data, highlighting the use cases and value propositions for national statistical offices collaboration with stakeholders in civil society and academia as a way forward to filling data gaps about women, girls and inclusion, by drawing on examples of collaboration for different entry points in the data value chain (e.g., providing contexts for difficult to reach or small population groups, providing gender perspectives to instrument design, communicating results and putting official statistics into action, etc.) from collaborative projects in Africa and Latin America.