Multimodal text types hold promise for promoting classroom discourse. Because of the ways that language combines with various other modes such as sound and movement, video-based texts can offer access points for meaning-making and engagement that print-only text cannot. Through the lens of New Literacies, this mixed methods study examines one small group of 5th grade multilingual students identified as struggling with engagement in reading, and their interactions and engagements with one another and their teacher as they participated in text-based discussions using three kinds of text types: print-only, print-plus image, and video-only. Lesson observations, teacher interviews, and surveys of students' engagement with each text shed light on differing talk distributions, interactional moves, and levels of engagement between print-only, print-plus image, and video-only texts. This study finds talk distributions to be consistent with the literature on student and teacher talk with no statistically significant difference between text types. Student and teacher talk moves were found to vary little between text types, though higher rates of instructing moves were statistically different from print-only and print-plus image texts. Video-only engagement for students was rated overwhelmingly higher as compared to other text types. Together these findings suggest video to be a meaningful text type to use during text-based discussions.