I4NG Evaluation of test equating and linking methods for data harmonisation
Summary
The talk will help researchers navigate the advanced statistical rigour required to quantify errors introduced by methodologies used in the data harmonisation process, specifically test equating and linking methods.
Description
Register via Zoom
This specialised technical session is to support the interactive Infra4NextGen training course: Creating and working with ex-post harmonised survey data.
The validity of inferences in the social sciences depends on the comparability of scores obtained from harmonised data. Therefore, the methods used for harmonisation must be rigorously evaluated to improve the precision of results.
This session explores how test equating and linking methodologies can be applied to survey harmonisation. These methodologies can be used when multiple survey questions differing in content are used to measure the same construct, and the results obtained from these combined.
Following an introductory talk from Dr. Rabia Karatoprak Ersen (GESIS – Leibniz Institute for the Social Sciences), participants can ask questions about the presentation, the online training course, and the I4NG Harmonised Data Gateway.
Agenda
- Evaluation of Test Equating & Linking Methods for Harmonisation – Dr. Rabia Karatoprak Ersen
- Q&A: Attendees can ask questions on the course, harmonised data, harmonisation strategies, and the methodologies presented
Key Learning Objectives
After this session, participants will understand:
- The equating and linking workflow: Decisions regarding data collection, operational definitions, statistical estimation methods, conducting linking, and evaluating the results of linking.
- Application of methods: Application of the different methods in a harmonisation context: identity linking, mean linking, linear linking, equipercentile linking.
- Evaluation of results: How to set up an evaluation study to understand the performance of linking methods, and to assess group invariance under various conditions.
Further Support Slots
In addition to this session, there will be two further opportunities to ask questions about data harmonisation:
- 28 October 2026, 11-11.30am (CET)
- 1 December 2026, 11-11.30am (CET)
A kick-off session for the course took place on 9 June 2026.
About the speaker
Dr. Rabia Karatoprak Ersen is a postdoctoral researcher at GESIS – Leibniz Institute for the Social Sciences. Within the Infra4NextGen project, she works across all five NextGenerationEU pillars, focusing on the harmonisation of cross-national survey data and the construction of the NextGen Harmonised Data Gateway. Her research interests centre on psychometric properties of item or test scores. Specifically, test equating and linking, item response theory linking, reliability estimation, and application of these methods to harmonised data.