Because the first research phase of the Research Unit CORE (2023 – 2027) is not yet complete, we continue updating and expanding the key findings.
During Phase 1, the nine CORE projects have established conceptual, methodological, and empirical foundations for investigating COR under authentic digital conditions. The findings available to date point to the following emerging results:
(1) Authentic, performance-based measurement. CORE has developed performance-based COR assessments in authentic open and closed online information spaces and combines them with process tracing and systematic analyses of the sources students access and use. This approach complements conventional self-report measures and assessments based on predefined information materials by capturing students’ information search, evaluation, and reasoning processes under more authentic conditions. Students’ written responses are scored with validated rating procedures; inter-rater reliability has exceeded .80 for the reported rating domains.
(2) Competence development with persistent challenges. Cross-sectional and longitudinal analyses indicate development in students’ COR skills, including their ability to construct more nuanced, well-reasoned, evidence-based arguments in domain-specific tasks. Development patterns vary across reasoning and disciplinary contexts. Challenges persist particularly in evaluating source quality and in developing well-founded arguments.
(3) Efficiency over systematic verification. Process data from log files, eye tracking, and think-aloud protocols indicate that students often prioritize efficient information seeking over systematic source evaluation and cross-source verification. They tend to engage selectively with online information rather than systematically interrogating sources in depth. Eye-tracking data also indicate limited attention to source credibility cues in relevant task contexts.
(4) Generative AI use: faster task completion, no automatic gain in reasoning quality. Initial CORE studies of generative AI use indicate that access to AI chatbots can reduce task completion time, but does not automatically improve the quality of students’ reasoning. Students who use AI more strategically, e.g., by critically evaluating AI-generated content and integrating it with information from other sources, show more differentiated patterns of source use and reasoning than students who rely more heavily on AI-generated responses. These findings highlight the importance of students’ ability to evaluate and regulate AI-supported information use critically.
