Nathalie Bick ^1,2* Laura Froehlich ^1,2 Jan-Bennet Voltmer ² Jennifer Raimann ^1,2 Natalia Reich-Stiebert ^1,2 Niels Seidel ^1,3 Marc Burchart ^1,4 Sarah E. Martiny ⁵ Jana Nikitin ⁶ Stefan Stürmer ^1,2 Andreas Martin ^7,8

• ¹ CATALPA, FernUniversität in Hagen, Hagen, North Rhine-Westphalia, Germany
• ² Department of Psychology, FernUniversität in Hagen, Hagen, Germany
• ³ Department of Mathematics and Computer Science, FernUniversität in Hagen, Hagen, Germany
• ⁴ Research Institute for Telecommunications and Cooperation e.V., Dortmund, Germany
• ⁵ UiT The Arctic University of Norway, Tromsø, Troms, Norway
• ⁶ Department of Developmental and Educational Psychology, University of Vienna, Vienna, Austria
• ⁷ Department of Humanities and Social Sciences, FernUniversität in Hagen, Hagen, Germany
• ⁸ German Institute for Adult Education – Leibniz Centre for Lifelong Learning, Bonn, North Rhine-Westphalia, Germany

Collaboration improves multiple academic and social outcomes. Accordingly, computer-supported collaborative learning (CSCL) can be beneficial in distance education contexts to overcome the issues specific to online learning (e.g., underperformance, low identification with university). Distance universities often attract a substantial number of non-traditional students (e.g., students with disability, students with migration background). Despite their representation, non-traditional students face negative stereotypes and associated social consequences, including social identity threat, diminished sense of belonging, and less motivation for social interactions. In the context of online learning, where there is little individuating information, social categories like socio-demographic group memberships become salient, activating stereotypes. Consequently, socio-demographic group memberships can have detrimental consequences for the integration of non-traditional students. The purpose of the present study was to a) determine the extent of social identity threat for students in higher distance education, b) explore the social consequences of this threat in the same context, c) validate these findings through longitudinal analyses embedded in a CSCL task, and d) use learning analytics to test behavioral outcomes. In a longitudinal study with three measurement occasions over eight weeks (N = 1,210), we conducted path analyses for cross-sectional associations and Random Intercept Cross-Lagged Panel Models for longitudinal predictions. The results showed that nontraditional students mostly reported higher social identity threat than traditional students. While the expected longitudinal within-person effects could not be demonstrated, we found stable betweenperson effects: students who reported higher levels of social identity threat reported also lower social approach motivation mediated by lower sense of belonging. Exploratory analyses of actual online collaboration during CSCL offer potential avenues for future research. We conclude that social identity threat and its social consequences play an important role in higher distance education and should therefore be considered for successful CSCL.