MultiplEYE Data Collection Guidelines

Eye-tracking is a gold-standard method for studying reading and language comprehension, yet the field lacks large-scale, multilingual datasets collected under standardized and FAIR-compliant conditions. This preregistration describes a large-scale, international eye-tracking-while-reading study conducted across multiple testing sites as part of the COST Action MultiplEYE (CA21131). Participants from diverse linguistic backgrounds read short naturalistic texts while their eye movements are recorded using a harmonized experimental protocol. The stimulus materials consist of parallel texts across languages and genres, enabling systematic cross-linguistic comparisons of reading behavior as well as comparison across different text types and levels of complexity. In addition to the reading task and comprehension questions, demographic information is collected for all participants, and a subset of sites administers standardized psychometric tests assessing a range of cognitive and linguistic abilities. Data collection, preprocessing, quality control, and documentation follow jointly defined standards to ensure comparability and reproducibility across sites. The resulting multilingual corpus will be openly shared via EyeStore, a FAIR-compliant repository hosted by the Research Data Center at the Leibniz Institute for Psychology (RDC at ZPID), providing a sustainable resource for research in psychology, linguistics and machine learning.

This document describes the data management plan (DMP) for COST Action "MultiplEYE" (CA21131). We outline the project details and provide a formal description of the data. Furthermore, we address data privacy & protection principles, as well as intellectual property rights & copyright issues. The DMP defines criteria for data storage, backup & extermination, data organization, data quality, and data documentation. Finally, we describe the agreement for data sharing & licensing.

Data and materials for the 39 language versions of the MultiplEYE Text Corpus pertaining to Kaspere, Bondar, Nisioi, Stegenwallner-Schütz et al. (2026). Text Corpus: Towards a Diverse and Ever-Expanding Multilingual Text Corpus. Proceedings of the 15th Language Resources and Evaluation Conference (LREC 2026). European Language Resources Association. For each language version, the repository includes three file types: (1) a stimuli-experiment file containing paginated stimulus texts, (2) a metadata file containing bibliographic and provenance information for each text, and (3) token-level linguistic annotation files. In addition, the repository provides (4) a pagination correspondence table documenting the alignment between each language version's pagination and the English reference pagination, and (5) a language coordinator list identifying the individuals coordinating the compilation of texts for each language version of the MultiplEYE Text Corpus.