BACK TO SCHEDULE REGULAR

Session 14

Improving Quality of Social Statistics

3 June 2026
16:30 – 18:00
ŠIBENIK V SHOW ON MAP

Session chair
Vesna Buterin
Associate Professor and Researcher in Macroeconomics, Vice-Dean for Cooperation with Business Community and Affairs, Faculty of Economics and Business (EFRI), University of Rijeka, Croatia

Read more Read less

Vesna Buterin, PhD, is an Associate Professor and Vice Dean for Cooperation with Business and Affairs at the Faculty of Economics and Business, University of Rijeka. She teaches courses in macroeconomics, institutional development, economic demography, and the philosophy of economic science, and is actively involved in doctoral-level teaching. She has extensive academic, scientific, and project experience, with numerous research engagements and leadership roles. She is currently leading a scientific project on the macroeconomic effects of demographic changes, as well as the EU-funded project EkonInfoChecker under the National Recovery and Resilience Plan, and participates in several international projects funded through Horizon and other European programs. Throughout her career, she has held numerous positions in scientific, professional, and governance bodies. She has completed a wide range of professional training in higher education, statistics, leadership, international cooperation, and economic policy. She publishes scientific papers, serves as a reviewer for international journals, and participates in the organization of scientific and professional conferences.

Presentation title
Quality Monitoring of the Monthly Unemployment Rate
PRESENTATION PAPER

From its onset, the unemployment data presented by Eurostat has followed the definitions of the International Labour Organisation, applied in a time consistent and harmonized manner by the EU Labour Force Survey (for quarterly and annual data).

Read more Read less

For the higher frequency monthly data, additional compilation techniques were introduced, which merit particular attention, especially when seen in a time series context. Various strategies are in place at the level of EU Member States to deal with the challenge of compiling and presenting the monthly data series, revealing the need for a systematic overview of the quality the monthly statistics.

This paper presents in detail the first quality monitoring exercise for the monthly unemployment rate, carried out by Eurostat during the course of 2025. After explaining the legal framework set in the Commission Implementing Regulation (EU) 2019/2241, the paper describes the indicators used to measure two quality dimensions, namely the volatility of the series and the magnitude of data revisions. Following this, the paper illustrates chronologically all the steps taken by Eurostat in the quality monitoring exercise and discusses its results as well as how selected elements were presented in a structured and understandable way.

The work done in this area shows the high overall quality of monthly unemployment rate data, while also identifying opportunities for improvement. The paper also outlines the quality challenges involved and outlines possible interlinks between the qualitative descriptions already published and the time series features monitored herein.

Main author / Presenter
Nevena Cholakova
Eurostat

Read more Read less

Nevena Cholakova joined Eurostat in 2021, first working in the unit for Population and Demography. Since 2022, she has been a Statistical Officer in the unit dealing with Labour Market and Skills, where she leads the production and dissemination of the EU monthly unemployment rate and also contributes to quarterly ЕU Labour Force Survey (EU LFS) products. In this context, in 2025 Nevena was responsible for carrying out the first quality monitoring exercise of the EU monthly unemployment rate.

Presentation title
Time Slot Model in the Swedish LFS
PRESENTATION PAPER

The Swedish Labour Force Survey (LFS) suffers from an increasingly challenging response climate.

Read more Read less

This has led to a decrease in response rate and an increase in the cost of the data collection. The nonresponse rate in the Swedish LFS has increased from around 20 to 57 percent between 2008 and 2024, mostly due to an increase in non contacts.

Today's data collection encounters several challenges, which makes it essential to optimize resources to achieve a cost-effective data collection process. Therefore, Statistics Sweden initiated a project with aim to achieve a more cost-effective data collection, while also taking quality into account. The project was partly funded by EU funds and was carried in 2024 and 2025.

One part of the project included analysis on contact strategies and how available information from the survey can be used to improve the efficiency of the data collection with focus on time for contact attempts, referred to as time slots.

In the Swedish LFS, time slots are assigned based on contact questions or background information. Contact questions are asked at the end of the interview and refers to questions on the most suitable time to contact the sample individual. Based on answers to these questions, the individual is placed in a time slot in which they are most likely to respond. If the individual has not responded to the contact questions, they are allocated a time slot based on background information. The time slots are used only for those who responded in the previous or the one before last survey occasion, and only during the first two days of the data collection.

In the project, analyses were conducted related to both contact questions and time slots in relation to when the interview was carried out. Analyses related to the contact questions included examining the distribution of answers to the contact questions and whether the sample individuals tend to give similarly responses on these questions over time. Analyses of time slots included examining whether the sample individuals respond within the allocated time slot and if the interview tend to take place in the same time slot over time. These analyses show that there is room for improvements within the current time slot model, especially when it comes to using information from previous survey occasions.

Main author / Presenter
Frida Videll
Statistics Sweden (SCB)

Read more Read less

The authors work at Statistics Sweden with the Swedish Labour Force Survey (LFS). Frida has a background as a methodologist working with methodological questions concerning the Swedish LFS and Sara work as a survey manager, thereby working closely with the data collection in the Swedish LFS.

CO-AUTHOR:

Presentation title
Evaluating Sources of Selection Bias in a Mixed-Mode Labor Market Survey
PRESENTATION

Telephone surveys have historically been a popular form of data collection in labor market research and continue to be used to this day.

Read more Read less

Yet, telephone surveys are confronted with many challenges, including imperfect coverage of the target population, low response rates, risk of nonresponse bias, and rising data collection costs. To address these challenges, many telephone surveys have shifted to online and mixed-mode data collection to reduce costs and minimize the risk of coverage and nonresponse biases. However, empirical evaluations of the intended effects of introducing online and mixed-mode data collection in ongoing telephone surveys are lacking. We address this research gap by analyzing a telephone employee survey in Germany, the Linked Personnel Panel (LPP), which experimentally introduced a sequential web-to-telephone mixed-mode design in the refreshment samples of the 4th and 5th waves of the panel. By utilizing administrative data available for the sampled individuals with and without known telephone numbers, we estimate the before-and-after effects of introducing the web mode on coverage and nonresponse rates and biases. We show that the LPP was affected by known telephone number coverage bias for various employee subgroups prior to introducing the web mode, though many of these biases were partially offset by nonresponse bias. Introducing the web-to-telephone design improved the response rate but increased total selection bias, on average, compared to the standard telephone single-mode design. This result was driven by larger nonresponse bias in the web-to-telephone design and partial offsetting of coverage and nonresponse biases in the telephone single-mode design.

Main author / Presenter
Joseph Sakshaug
Institute for Employment Research (IAB), Ludwig Maximilian University of Munich (LMU), Germany

Read more Read less

Joseph Sakshaug is Professor of Statistics at the University of Munich (LMU) and Distinguished Researcher at the Institude for Employment Research (IAB), Nuremberg.

CO-AUTHOR:

Presentation title
Ready, Set, Go! Implementing a Smart Survey App for the Household Budget Survey
PRESENTATION

Relevance and research question - Household Budget Surveys (HBS) are a key source for official statistics in Europe but are also known for their high response burden and risk of underreporting.

Read more Read less

Within earlier Eurostat Grant projects (@HBS, @HBS2 and Smart Surveys Implementation (SSI)), Statistics Netherlands developed and field-tested an app-based approach for HBS data collection, demonstrating the potential of smart survey solutions. Building on these results, Statistics Netherlands initiated a renewed development trajectory aimed at scaling up the app from a project-based innovation to a sustainable solution for regular statistical production. This transition aligns with the Eurostat Modernization Maturity Model, towards production-ready smart survey solutions. The central question addressed is how smart survey principles can be operationalized in an app-based HBS while enhancing data quality, respondent inclusion and organizational feasibility. Methods and data In 2025, a testing program consisting of six complementary tests was carried out using the HBS app. Two usability tests employed cognitive interviewing and think-aloud methods to assess the full respondent journey, from invitation to completion. A field test with a fresh sample evaluated the app in a realistic survey setting, including experimentation with different interviewer roles and support strategies. Quantitative survey outcomes were enriched with evaluation questionnaires and follow-up telephone interviews. Three internal tests supported the transition towards production use. One focused on collecting receipt data for training and validating machine-learning algorithms for automated expenditure classification. The other two enabled rapid iterative testing rounds to finalize design choices. From 2026 onwards, the app is implemented in regular HBS fieldwork, complemented by a web-based version to support mixed-mode participation. Results The presentation outlines the development trajectory from the Eurostat-funded projects to full-scale implementation and presents empirical findings from the usability, field and internal tests conducted in 2025. Results provide evidence on usability issues, respondent burden and reporting behavior, interviewer support strategies and automated expenditure classification performance. Findings reveal substantial heterogeneity in respondents’ digital skills, underlining the importance of inclusive design and mixed-mode strategies. Iterative testing informed concrete design and algorithmic improvements. First results from the 2026 fieldwork are presented. Added value This contribution illustrates how smart survey innovations developed under Eurostat Grants can be embedded in regular statistical production. It highlights practical lessons on multidisciplinary collaboration, aligning innovation with survey requirements, and integrating new tools into existing systems and workflows. The presentation shows how testing results informed incremental design decisions and trade-offs between functionality, usability and production constraints.

Main author / Presenter
Jelmer de Groot
Statistics Netherlands (CBS)

Read more Read less

Jelmer de Groot has a Masters degree in Communication Science and has been working at statistics Netherlands since 2010, first as a survey methodologist, followed by a job as a survey designer in the data collection department. Since 2017, he is a project manager at the data collection department, responsible for the fieldwork and its preparations in a wide variety of household surveys. His field of interest lies in new ways of data collection and inplementation. From that perspective, he currently is smart surveys project manager within Statistics Netherlands and will tell about the experiences he had within the smart Household Budget Survey.

CO-AUTHORS:

Presentation title
Integrating Administrative Data Sources to Improve Small Area Estimation of Municipal At-Risk-of-Poverty Rates
PRESENTATION PAPER

Direct estimates from the EU Statistics on Income and Living Conditions are often too unstable at the municipality level.

Read more Read less

We estimate municipal at‑risk‑of‑poverty (AROP) rates using an area‑level Fay–Herriot model that combines calibrated direct estimates with auxiliary information derived from administrative income data (wages and benefits) linked through the Statistical Population Register. The key covariate is a municipality-level proxy AROP rate, computed from administrative sources and used as the regressor in the Fay–Herriot model. We present the resulting empirical best linear unbiased predictions and discuss their stability relative to direct estimates. We also briefly discuss exploratory work on using retail scanner indicators as additional covariates for local poverty monitoring.

Main author / Presenter
Andrius Čiginas
Statistics Lithuania

Read more Read less

Dr. Andrius Čiginas is a methodology specialist working on sample surveys at the State Data Agency (Statistics Lithuania). He is also a Senior Researcher at Vilnius University.

Presentation title
Estimating the Size of the Resident Population of Poland Based on Administrative Records
PRESENTATION

In this study, we discuss the methodological and practical issues involved in deriving the resident population based solely on integrated administrative records and estimating its size.

Read more Read less

Using information from multiple administrative registers—such as the population register (PESEL), the social insurance register, and tax registers, among others—spanning multiple years, we derive the resident population for both Polish and non-Polish populations.

The main goal of this study is to present our approach to deriving the resident population from integrated administrative datasets. This approach requires not only linking data using identifiers but also probabilistic record linkage. Additionally, we impute length of stay based on information from registers as well as from documents that permit non-Polish residents to stay in Poland.

Initial results suggest that this is a promising approach, and simulation studies support our claims. However, there is a need to integrate administrative data sources with sample surveys to validate this methodology.

Main author / Presenter
Zofia Danek
Statistical Office in Poznan, Adam Mickiewicz University (AMU), Poland

Read more Read less

We work at the Centre for Methodology of Population Studies at Statistics Poland, Poznań. We specialise in linking administrative data to derive population estimates. Zofia and Aniela are mathematicians, while Maciej serves as the head of the Centre.

CO-AUTHORS: