Netherlands

1

UNITED NATIONS ECONOMIC COMMISSION FOR EUROPE CONFERENCE OF EUROPEAN STATISTICIANS Expert Meeting on Statistical Data Editing 7-9 October 2024, Vienna Austria

12 July 2024

Current work on automatic multisource editing at Statistics Netherlands

Sander Scholtus, Arnout van Delden, Rob Willems, and Frank Aelen (Statistics Netherlands, The Netherlands) [email protected] Abstract As part of a new integrated uniform production system for business statistics, Statistics Netherlands aims to develop a system for simultaneous editing of variables that are observed across different data sources. In this way inconsistencies between statistics could be identified and resolved as early as possible. In addition to top- down manual editing of the most influential inconsistencies, we aim to introduce automatic editing of other inconsistencies where possible. An initial approach and results for automatic multisource editing from a pilot study were previously presented in Scholtus et al. (2022). Currently, there is a project at Statistics Netherlands to develop automatic multisource editing further for use in regular statistical production. This project is supported as part of a grant from Eurostat for timelier, more relevant and more integrated European business statistics. The project will run from January 2024 until June 2025. Regarding automatic editing across statistics, the main aims of the project are, first, to improve the quality of automatically edited data by introducing better edit rules and incorporating more unit-specific information, and second, to evaluate the effects of automatic editing on the quality of statistical output. As part of the project, a probability sample of units will be edited manually outside of regular production, to provide information both for the development of automatic editing logic and for the evaluation of the quality of automatically edited data. While the project is still ongoing, we will present some ideas and initial developments. Reference S. Scholtus, W. de Jong, A. Vaasen-Otten and F. Aelen (2022), Towards a new integrated uniform production system for business statistics at Statistics Netherlands: automatic data editing with multiple data sources. Paper prepared for the UNECE Expert Meeting on Statistical Data Editing, 3-7 October 2022 (virtual).

Edwin Horlings

8-9 July 2024

Developing a common language for sustainable and inclusive well-being in the Netherlands Presentation for the UNECE Seminar on Measurement of Well-being

The rise of “brede welvaart”

“brede welvaart”

The strength of GDP is that it has become a common language (based on the SNA)

No such language (or statistical standard) for inclusive and sustainable well-being

A common language

6

─ assessment offices (economic, social, environmental)

─ government departments

─ SDG community

─ regional policy analysis offices (national network)

─ public knowledge institutes and advisory councils

─ Court of Audit

─ civil society organisations

─ Parliamentary Analysis and Research Department

Strategic relations

Core network

Netherlands Institute for Social Research Netherlands Bureau for Economic Policy Analysis Netherlands Environmental Assessment Agency

Netherlands Court of AuditStatistics Netherlands

alignment of language and understanding of

the conceptual framework, examine and improve policy

practices in relation to “brede welvaart”

─ Rapidly growing community

─ Emergence of a common language for “brede welvaart” (inclusive and sustainable development)

─ Developing tools and practices: to integrate “brede welvaart” in policy development, budgeting, policy analysis, evaluation, and impact assessment

─ Training of policy professionals (Government Academy)

Stakeholder community of practice

─ Policy development tools

─ Integration into the budget cycle

─ Annual reporting by the Government and the Court of Audit (the state of well-being)

─ Evaluation (ex post and ex ante)

Tools and practices

13

Factsheets

SDG x, SDG y

SDG z

SDG x

SDG x, SDG y

SDG z

SDG x, SDG y

SDG z

SDG x

Hyperlinks to SDGs Archive - SDG Nederland

14

Internal network Data Database Methods Products

indicator data and metadata

trends

rankings

annual change

analysis database

Results MBW & SDG’s

rMBW

MBW CN

...

SDG’s

Factsheets

Network of experts

Methodology department

Dedicated IT-team Communications

department Dedicated data team

─ Main suppliers of indicators and data (teams and experts)

─ Communications department

─ Methodology department

─ Centre for Policy Statistics (for customised products)

─ Account managers (contacts with stakeholders)

─ Higher management (strategic positioning)

Collaboration within Statistics Netherlands

Edwin Horlings

8-9 July 2024

Monitoring inclusive and sustainable well-being in the Netherlands

Presentation for the UNECE Seminar on Measurement of Well-being

“brede welvaart”

International statistical framework

An international framework, based on scientific research, provides authority and guarantees neutrality

─ The CES Recommendations are the basic framework

─ Need to adapt and expand the framework - Reduction of themes to more manageable proportions

- Integration of the Sustainable Development Goals

- Interpretation of statistical outcomes

─ Technical and methodological challenges - Selection criteria for indicators

- Decision rules for observations

- Visualisations

Adapted and expanded framework

“Brede welvaart”

(inclusive and sustainable well-being)

Well-being concerns the quality of life here and now as well as the extent to which this quality is achieved at the expense of later generations or of people in elsewhere in the world.

Dimensions of well-being

Distribution

─ an international agenda and framework, co- signed by Dutch government

─ provide statistics on policy-relevant themes without having to interpret government policy

Integrating the SDGs

─ Resources and opportunities: rights, freedoms and other possibilities that people have; capacity of infrastructures and systems; access to and affordability of services

─ Use: number of people, households, companies who use resources and opportunities, including aspects of behavior

─ Outcomes: positive and negative effects of use

─ Subjective assessment: people's subjective assessment of outcomes, including satisfaction, concerns, expectations

Framework for SDG indicators in context

Resources and opportunities What resources are used to maintain and improve the healthcare system? Is healthcare affordable and accessible?

Use What behavior do people exhibit that affects their health? How do they use health care?

Subjective assessment Are people satisfied with their own health and with the Dutch healthcare system?

Outcomes What is the current physical and psychological health of the population and the quality of care?

2024

Monitor

& de Sustainable Development Goals

Brede Welvaart

Visualisations

Dimensions of the CES Recommendations

“here and now” “later” “elsewhere”

─ Sparse: as little information as possible to make understanding easier for users

─ Colours to signal significant observations

─ No composite index

Dashboards

─ Individual indicators by sex, age, education level, country of origin/birth

─ Cumulative differences for c. 7,000 individuals

Distribution of well-being

12 provinces

40 regions

345 municipalities

3 Caribbean islands

Regional well-being

regionale Monitor Brede Welvaart 2023 [regional Monitor of Well-being] (https://www.cbs.nl/nl-nl/visualisaties/regionale-monitor-brede-welvaart)

From indicators to a narrative

Number of indicators

“here and now” 29

“later” 22

“elsewhere” 17

Everything included (SDGs, resilience) around 300 indicators

Layering

conclusions

FULL TEXT detailed description of indicators

expert knowledge, nuance, complicated

MAIN OBSERVATIONS bullets, shorter, less complicated

indicators linked to concepts

SUMMARY short, easier to read

indicators linked to concepts

MBW&SDGS IN A NUTSHELL 3-pager

conclusions about concepts

The Netherlands is a social and prosperous society 'here and now'

- A close-knit society with high subjective well-being, high material prosperity and a lot of work

- Not everyone benefits from high material prosperity

With current developments, the next generation will achieve lower well-being 'here and now'

- Natural capital is deteriorating

- The economy is only slowly becoming more sustainable

- Essential institutions appear to be at a standstill

- Young adults have been lagging behind for two years in a row

Main conclusions in 2024

Critical evaluation (the seven-year itch)

Publications

17 September “Budget Day” ➔ Factsheets Well-being for each department (since 2023)

International SDG Day ➔ SDGs International (since 2016)

30 November

25 September

“Accountability Day” ➔MBW & SDG’s (since 2018) (since 2019)15 May

7 December

MBW-SDG Caribbean Netherlands (since 2023)

Regional MBW (since 2020)

─ Usefulness and accessibility in relation to usage

─ ‘Indicator creep’

─ Match with major societal challenges and themes

─ Communicating about recognisable groups

─ Incomparable international frameworks

First observations

─ Economic capital, Natural capital, Human capital, Social capital are ubiquitous

─ Hard to explain what they mean in real life

─ Appear to have a strong economic bias

─ Social capital is a reservoir for everything not included in economic, human or natural capital

Four capitals of well-being ‘later’

─ Pay more attention to deprivation

─ Measure well-being of recognisable groups (e.g. old- age pensioners or households living on social security)

‘Top-down’ bias

Dimensions of well-being

Distribution

Distribution

Statistical framework of the future

UNITED NATIONS ECONOMIC COMMISSION FOR EUROPE

CONFERENCE OF EUROPEAN STATISTICIANS

ModernStats World Workshop

21-22 October 2024, Geneva

Modernization and agility powered by Communities of Practice

Speaker: Matjaž Jug And Olav Ten Bosch, Statistics Netherlands

Author(s):

Abstract

With recent changes in society and looming challenges, such as climate change and political instability, there is

increased demand on National Statistical Organizations (NSOs) to timely deliver new insights to policymakers,

researchers and general public. The official statistics community is expected to respond faster than ever before,

which poses a challenge to identify and secure new data sources, to invent new agile but reliable methods, to

apply new AI techniques in a responsible way, to effectively use open source libraries, tools and models and to

provide outputs in modern forms for many different types of users. Hence NSOs are faced with an increased

demand for capability development to build up and maintain the latest skills and develop practices needed.

Experiences with statistical communities of practice (CoPs) can show the way forward but need to be extended

to align with external partners in communities that work not only within and across NSOs but also with

Academia and private sector.

In this presentation we will present our ideas how communities of practice may help to unlock the innovation

potential of our organizations based on pooling skills and experiences which can ultimately make our

organizations more agile. We use examples from different fields of expertise such as data science, privacy

technologies, smart metadata, (linked) open data, data dissemination, geospatial modeling and insights into

open source communities. We will spot criteria of success, basic operating models and the power of

interlinking communities of practice, such as for example data science and privacy protection. Finally, we will

come up with some recommendations what can NSOs do to leverage the potential of CoPs and better connect to

wider community networks.

Submitted by the experts from the Netherlands Informal document GRVA-19-11 19th GRVA, 25 June 2024

(For review at the Troy meeting 20-24 May 2024)

Provisional agenda item 8(c)

1

Proposal for a new supplement to UN Regulation No. 13

The text below was prepared by the experts from the Netherlands. The modifications to the existing text of the Regulation are marked in bold for new or strikethrough for deleted characters.

I. Proposal

Paragraph 4.5 of Annex 15., amend to read:

“4.5. Type II test (downhill behaviour test):

4.5.1. This test is required only if, on the vehicle-type in question, the friction brakes are used for the Type-II test as required by Annex 4 paragraph 1.6 (i.e. with a 6% slope) or 1.8.2.5 (b) (i.e. with a 7% slope).

4.5.2. Brake linings for power-driven vehicles of Category M3 (except for those vehicles required to undergo a Type-IIA test according to paragraph 1.6.4. of Annex 4 to this Regulation) and Category N3, and trailers of Category O4 shall be tested according to the procedure set out in paragraph 1.6.1. of Annex 4 to this Regulation.

Brake linings for power-driven vehicles of Category M3 and Category N3 required to undergo a Type-IIA test according to paragraph 1.6.4. of Annex 4 to this Regulation and only complying with this requirement by application of provisions of paragraph 1.8.2.5 of Annex 4 to this Regulation, shall be tested according to the procedure set out in paragraph 1.8.2.5 (b) of Annex 4 to this Regulation.”

II. Justification

1. This amendment is to clarify the need to apply, when relevant, the Type II test as required by Annex 4 paragraph 1.8.2.5 (b) in Annex 15 for alternative brake lining purposes.

2. During the September 2020 GRVA session document ECE/TRANS/WP.29/GRVA/2020/36 as modified by informal document GRVA-07-73/Rev.1 was adopted, essentially reflecting the possibility to use, in case of vehicles with electric regenerative braking possibilities, the Type II endurance test as required by Annex 4 paragraph 1.8.2.5 (b) as alternative to a regular endurance brake as required by Annex 1.6. In this case, the friction brake is used when storing energy in the traction battery is not possible only because the maximum state of charge of the battery has been reached.

3. At the time, the procedure and requirements of Annex 15 on alternative brake lining have not been amended and refer only to the Type 0, I, II as requested by Annex 4 paragraph 1.6 and the Type III test.

4. In case of a manufacturer making use of the above standing Type II test as required by Annex 4 paragraph 1.8.2.5 (b), applying a certain (‘standard’) brake lining and this manufacturer later on wants to apply an alternative brake lining, it can make use of the existing procedure as described in Annex 15. However, this Annex does not refer to the Type II test as required by Annex 4 paragraph 1.8.2.5 (b).

Based on paragraph 1.3 of the same Annex 15 it is possible to demand additional tests according to Annex 4 – so also the Type II test as required by Annex 4 paragraph 1.8.2.5 (b) - however in such case the test needs to be executed twice so with and standard an alternative brake lining which is double work:

1.3. The Technical Service responsible for conducting approval tests may at its discretion require comparison of the performance of the brake linings to be carried out in accordance with the relevant provisions contained in Annex 4 to this Regulation.

5. To overcome such double work and to ensure a level playing field, it is proposed to clarify the need to use, when relevant, the Type II test as required by Annex 4 paragraph 1.8.2.5 (b) in Annex 15.

1

UNITED NATIONS ECONOMIC COMMISSION FOR EUROPE CONFERENCE OF EUROPEAN STATISTICIANS Expert Meeting on Statistical Data Collection and Sources 22-24 May 2024, Geneva, Switzerland

19 April 2024

Applying Workforce Management principles to personal interview modes

Jack Mommers and Martijn van de Riet (Statistics Netherlands, Netherlands) [email protected] Abstract Statistics Netherlands incorporates a range of modes in its data collection. It employs a sequential mixed-mode strategy for most of its social surveys, commencing with CAWI for all sampled individuals, followed by CATI and CAPI for those who do not respond. In the CATI mode, individuals without a response are surveyed if a telephone number is available. Over time, acquiring telephone numbers for sampled individuals has become increasingly challenging. Year after year, Statistics Netherlands observes a decline in the number of located telephone numbers and a decrease in their quality. In the CAPI mode, a targeted approach is employed to include non-respondent groups that exhibit lower response rates in CAWI. This targeted approach has led to a concentration of CAPI interview addresses in urban areas and a reduction in rural addresses. Consequently, travel costs in rural areas have soared. Additionally, persuading individuals for an interview requires greater effort due to the targeted approach for non-respondent groups. Statistics Netherlands’ interviewer corps comprises individuals with both permanent and temporary contracts. Currently, the organization faces challenges in recruiting (and planning) interviewers due to a tight labor market. This session explores the topics of contractual arrangements and the planning of interviewers at Statistics Netherlands. Within the contact center industry, Workforce Management (WFM) principles are widely adopted to allocate resources, enhance efficiency, forecast workloads, and manage schedules. By implementing the WFM cycle within Statistics Netherlands, we can effectively analyze, forecast, schedule, monitor, and evaluate the performance of CATI and CAPI, thereby improving employee satisfaction, reducing the need for short-term temporary contracts, and enhancing efficiency. In the near future, Statistics Netherlands plans to merge the CATI and CAPI departments and their interviewers. By enabling CAPI interviewers in rural areas to conduct CATI interviews, we aim to maintain a national presence of CAPI interviewers and reduce travel costs. Additionally, efforts will be made to further decrease the reliance on temporary workers by offering more permanent contracts. Furthermore, interviewing during peak periods may be outsourced to minimize the need for temporary contracts.

1

UNITED NATIONS ECONOMIC COMMISSION FOR EUROPE CONFERENCE OF EUROPEAN STATISTICIANS Expert Meeting on Statistical Data Collection and Sources 22-24 May 2024, Geneva, Switzerland

19 April 2024

Smart surveys: how to implement smart data collection in official statistics?

Jelmer de Groot (Statistics Netherlands, Netherlands) [email protected] Abstract Smart surveys face an important question in designing and analyzing the combination of survey questions, sensor measurements and data donation: what is the best way of implementation of these smart surveys in the regular data collection strategies? A variety of surveys contain elements that make them very well fit for smart data collection. Especially surveys that tend to be burdensome, both in time and in cognitive effort, and when the information requested may not always be readily available for respondents. Smart surveys introduce features of smart devices such as the use of internal sensors, internal storage and computing and linkage to external sensor systems. Some of these features, such as GPS, receipt scanning and advanced product search algorithms are very promising. They are also challenging in terms of user interface design, recruiting respondents and their motivation and involvement. In the presentation, two smart survey case studies on consumption and mobility are discussed. In all, the extent of respondents’ recruitment and involvement plays a prominent role: receipt scanning, OCR and automatic classification in household budget surveys and location tracking in the mobility survey.

1

UNITED NATIONS ECONOMIC COMMISSION FOR EUROPE CONFERENCE OF EUROPEAN STATISTICIANS Expert Meeting on Statistical Data Collection and Sources 22-24 May 2024, Geneva, Switzerland

19 April 2024

Developments in interviewing at Statistics Netherlands: The challenges for personal interviewing in a targeted approach

Jack Mommers and Jacky Deneer (Statistics Netherlands, The Netherlands) [email protected] Abstract Developments in interviewing at Statistics Netherlands: 2. The challenges for personal interviewing in a targeted approach Statistics Netherlands incorporates a range of modes in its data collection. It employs a sequential mixed-mode strategy for most of its social surveys, commencing with CAWI for all sampled individuals, followed by CATI and CAPI for those who do not respond. The use of CATI and CAPI ensures that all target groups are sufficiently included in the statistics. In the CAPI mode, a targeted approach is employed to include non- respondent groups that exhibit lower response rates in CAWI. This approach (called the target group approach) entails a number of challenges for CAPI. The targeted approach has led to a concentration of CAPI interview addresses in urban areas and a reduction in rural addresses. This automatically leads to a greater distribution of respondents in rural areas, making the distances for the interviewers relatively large. Consequently, travel costs in rural areas have soared. Additionally, persuading individuals for an interview requires greater effort due to the targeted approach for non-respondent groups. The number of non-contacts requires a high degree of persistence from the interviewers. This can also lead to a reduction in job satisfaction. The questionnaires are drawn up in Dutch. In the CAPI target groups, the number of non-Dutch speaking respondents is relatively high and increasing year on year. We currently don’t have a way to cater to this development. Based on the experiences gained over several years, we are now looking for appropriate measures to address the challenges described above. We are thinking of combining and merging the CATI and CAPI modes. By enabling CAPI interviewers in rural areas to conduct CATI interviews also, we aim to maintain a national presence of CAPI interviewers and reduce travel costs. Furthermore we are introducing multilingual questionnaires, both at CAWI and CAPI and making adjustments in the approach strategy such as clustering addresses and limiting the number of visits to respondents in the event of no contact.

1

UNITED NATIONS ECONOMIC COMMISSION FOR EUROPE CONFERENCE OF EUROPEAN STATISTICIANS Expert Meeting on Statistical Data Collection and Sources 22-24 May 2024, Geneva, Switzerland

19 April 2024

10 years of communication experiments at Statistics Netherlands

Jelmer de Groot (Statistics Netherlands, Netherlands) [email protected] Abstract Within Statistics Netherlands, web data collection is being used more frequently to replace traditional interviewer modes. After a series of experiments and parallel data collection with other modes as CAPI and CATI, the CAWI-only approach became common practice. For the LFS, web data collection was implemented in 2012. This CAWI-only approach brings challenges when it comes to quantity and quality of collected data. Increasing demand for self-completion surveys means we cannot rely on skilled interviewers to secure response rates. Reaching out to our potential respondents can only be done by traditional mail. In a serial mixed-mode or a non-CAPI/CATI design, letters are the only way of communicating and thus are of great importance. Where CAWI as a data collection method was extensively tested before it was implemented, there was barely any attention for the advance letter design. This paper describes how advance letters of Statistics Netherlands were composed in the first place, the design and view on advance letters and on the testing process over the past years. Up to now, we are still improving both our materials and testing methods in order to be as efficient as possible when it comes to communication.

1

UNITED NATIONS ECONOMIC COMMISSION FOR EUROPE CONFERENCE OF EUROPEAN STATISTICIANS Expert Meeting on Statistical Data Collection and Sources 22-24 May 2024, Geneva, Switzerland

19 April 2024

Use the Blaise 5 system to implement multi-mode surveys

Gina Cheung (CBS, Netherlands) [email protected] Abstract Employing multi-mode or mixed-mode data collection strategies can significantly enhance the quality and representativeness of survey data. These strategies accommodate diverse respondent preferences, minimize bias, and improve accessibility. In this presentation, I will demonstrate leveraging the Blaise 5 system to offer respondents multiple participation options. We will explore critical elements of multi-mode survey design and system processes, including:

1. Increased Response Rates: By offering various modes such as online surveys, phone interviews, or paper questionnaires, researchers can reach a broader spectrum of respondents, ultimately leading to higher response rates.

2. Reduced Non-response Bias: Non-response bias occurs when certain groups are more likely to respond, skewing results. Multiple modes of data collection help mitigate this bias by reaching out to individuals who might not participate through a single mode.

3. Enhanced Accessibility: Different populations have varying access to modes of communication. Multiple modes ensure the survey is accessible to a broader audience, including those with disabilities or limited technological resources.

4. Improved Data Quality: Different modes may elicit varied responses or levels of detail. Utilizing multiple modes can improve the depth and accuracy of the data collected.

5. Cost Efficiency: While using multiple modes can incur additional costs, it can sometimes be more cost- effective than relying solely on one method, especially considering the potential increase in response rates and data quality.

6. Flexibility: Certain populations may be more reachable or responsive through specific modes of communication. Flexibility in data collection methods allows researchers to adapt their approach based on the target population and research objectives.

By considering these elements and leveraging the Blaise 5 system, researchers can design and implement multi- mode surveys that yield more representative and reliable data.