This investigation reports on a conceptual replication of Goldman et al. (2012) which sought to determine processing patterns of better and poorer learners as they attempted to comprehend and learn from seven textual sources on the Internet in their native language English. The aim of the current investigation was to explore how reading multiple textual sources in the L2 interacts with sense making, monitoring and evaluation processes during reading and researching on the Internet. Nineteen intermediate and advanced learners of German were asked to outline the position of the German government regarding immigration. L2 readers received a selection of five Internet sources which varied in reliability and appropriateness for the research assignment. Students' reading behavior was recorded through think-aloud protocols. The findings only partially replicated Goldman et al. Better learners used significantly more varied reading strategies overall than poorer learners. Their reading behavior was marked by greater use of self-explanation and monitoring. Better learners also skipped more unfamiliar words. Nevertheless, neither group of L2 readers employed reading and research skills that lead to a coherent intertext model fulfilling the goal of the research task.