Background: When learning about complex topics using the Internet, students commonly encounter a multitude of textual, non-textual (e.g., images and graphs), and multimedia (e.g., videos) resources. Yet students' learning from multiple texts and multiple (non-textual) resources (MT-MR learning) has received insufficient consideration in the literature. Aims: We examine the associations among (1) undergraduates' conceptions of reasons for multiple resource access, (2) log-data of resource use when completing a MT-MR task, and (3) writing performance. Sample: Participants were 72 undergraduate students in the United States. Methods: Undergraduates were provided with a library of five texts and one video, with the option of accessing supplemental data (e.g., graphs and maps) in association with each resource. Log-data (e.g., time and supplemental data access) of undergraduates' resource use were collected. Undergraduates were then asked to compose a research report and to describe what they considered the purpose of multiple resource access to be. Results: Four types of conceptions were identified, reflecting a desire to (1) access a lot of information, (2) understand multiple perspectives, (3) corroborate and evaluate information, and (4) develop a personal understanding of a given topic. Undergraduates who considered corroboration and evaluation to be the purpose of multiple resource access were more likely to access more supplemental data sources and performed better on a multiple resource learning task. Conclusions: Undergraduates in our sample held conceptions largely similar to, but in some aspects distinct from, those identified by Barzilai and Zohar (Cognit Instruct, 30, 2012, 39). Conceptions were associated with resource access during task completion and with writing performance.