EESW23 Short Courses

On Tuesday, September 19, 2023, two one-day short courses are offered adjacent to EESW23:

For fees and registration to a short course, please follow this link.

To those course participants who so require, certificates of course attendance will be issued.

New developments in business data collection methodology

Part I: The Respondent Centered Design approach in developing business surveys

Presented by: Chris Andrews

The term Respondent Centred Design (RCD) refers to qualitative research practices which focus on the needs of the respondent during the development of new and transformation of existing surveys.

Practices such as the creation of end-to-end user journeys, the use of mental modals and the implementation of iterative qualitative testing help create a positive experience for the user of the service. RCD also promotes the use of AGILE principles where the aim is to 'fail fast' and to keep testing and re-testing until we as researchers meet the needs of the respondent whilst still maintaining data quality. The benefits of such an approach include a reduction in respondent burden, increases in data quality and a long-term reduction in costs.

This short course provides an insight into RCD, and - using examples from business surveys - looks at how we can maintain data quality whilst also placing the respondent at the heart of the process.

Suggested reading:

About the lecturer: Chris Andrews has worked for the Office for National Statistics for five years, within the Data Collection Methodology team. As a Methodologist he has worked on the census but now focuses on business surveys contributing to the transformation of economic statistics.


Part II: Innovating the business statistics data collection

Presented by: Ger Snijkers and Paulo Saraiva

In the 20th century, sample surveys have proven to be a cost-efficient method to produce accurate statistics, although they come with a high cost both for the National Statistical Institutes (NSIs) and businesses, who may experience high response burden. This burden is caused by retrieving the data from internal business data sources and calculating the answers to questions in a questionnaire.

In this course, we’ll discuss how this manual process can be automated, resulting in System-to-System (S2S) data communication, also known as automated data collection. Nowadays in the information age, there are a lot of new digital data sources in smart industries. In some cases, these data sources allow for data communication with other computer systems without human intervention e.g. via Application Programming Interfaces (APIs) or web services. In this way human intervention in the data collection may be reduced to a minimum, e.g. only checking the pre-filled data to be submitted.

In addition, we’ll discuss the conditions for S2S as well as the consequences of such approach for the production processes and survey designs (questionnaire, sample). Examples from Statistics Portugal and Statistics Netherlands will be presented.

Suggested readings:

About the lecturers:

Ger Snijkers, PhD, is a senior methodologist and advisor in business data collection at Statistics Netherlands. Previously, he held the position of professor of business survey methodology at Utrecht University. He is the lead author of the textbook Designing and Conducting Business Surveys (Wiley, 2013) and the editor-in-chief of the edited volume Advances in Business Statistics, Methods and Data Collection (Wiley, 2023). 

Paulo Saraiva is Director of Data Collection and Management at Statistics Portugal. In this position, his task is to manage all data collection activities through surveys, census, and administrative data. He has extensive experience in the production of official statistics, information systems and dissemination. He is an author or co-author of numerous papers, books and presentations on data collection for business and household surveys.

Quality of multisource statistics

Presented by: Arnout van Delden and Sander Scholtus

In the course we will give an overview of various typical situations that can arise when producing statistics based on multiple data sources for which one wants to measure the quality of the output. The course will provide the participants with knowledge of specific quality measures, their calculation methods and when and how to apply them, and hands on experiences of the quality measures. We will also discuss examples from the participants’ experience with the production of multisource statistics and the quality measurement of these statistics. Although the methods can be applied to both social and business statistics, nearly all treated examples involve business statistics.

After the course, participants will be able to identify several practical situations where the quality of multisource statistics can be measured, and will be able to apply appropriate quality measures and methods to calculate these measures.

The topics covered during the course include: general introduction on multisource statistics and its quality; bootstrapping; using latent class modelling for correcting for classification error; macro-integration techniques; case studies based on input from the participants; etc.

Participants are asked to bring their own laptop with R installed (www.r-project.org/) since the course includes some hands-on exercises; RStudio is recommended.

Suggested readings:

About the lecturers:

Arnout van Delden, PhD, is a senior methodologist at Statistics Netherlands with a focus on business statistics. His main fields of expertise are processing of administrative data, methods for data integration, estimating output quality especially for non-sampling errors, and use of text mining and machine learning in official statistics. He is a co-editor of the edited volume Advances in Business Statistics, Methods and Data Collection (Wiley, 2023).

Sander Scholtus, PhD, is a methodologist at Statistics Netherlands. His main areas of expertise are data editing and imputation, sampling theory, measurement error models, and estimating output quality including non-sampling errors. Together with Ton de Waal and Jeroen Pannekoek, he is a co-author of Handbook of Statistical Data Editing and Imputation (Wiley, 2011).