Effects of Instructional Typography on Procedural Performance in Virtual Reality

Joanna Lewis¹, Joanna Pugilese¹, Rebecca Pharmer², Lucas Plabst², Jalyn Nicoly³, Ben Clegg⁴, Chris Wickens², Leanne Hirshfield³, Francisco Ortega²; ¹University of Northern Colorado, ²Colorado State University, ³University of Colorado, ⁴Montana State University

Virtual reality (VR) training relies heavily on visual instruction, yet the visual design of instructional text itself is rarely examined. Typography is often treated as a formatting choice, yet poor design may result in reduced training efficiency by competing for attention. We examined whether typographic/readability influences procedural performance in VR and whether these effects depend on task complexity. Participants completed a VR assembly task in which multi-component shapes were constructed repeatedly. Framing our study within Cognitive Load Theory and prior VR assembly work (Pharmer et al., 2025; Spencer et al., 2025), we manipulated intrinsic task demands by varying the visual composition of the assembly components: either uniformly white or multicolored. Instructional text also varied across extraneous load conditions, manipulating font contrast and spatial density to alter the text's legibility which influences visual processing. Performance was evaluated using assembly training time and error rates, and perceived cognitive workload was assessed with NASA-TLX. Increased intrinsic load led to longer training completion times and higher error rates. Extraneous load predominantly affected completion time, with low-readability instructions slowed performance without increasing errors or reported workload. These results indicate that instructional typographic properties influence information extraction, impacting the completion time without impacting overall task difficulty. Participants viewed a completed version of each shape before assembly, which may have reduced their reliance on instructional text. A follow-up study removed this preview to increase dependence on visual instructions. Initial data show the same general pattern that typography continues to affect speed but not accuracy. Combined results emphasize how task demands and poor design impact VR procedural learning. Ongoing work incorporates eye tracking and pupillometry to examine how typographic design shapes gaze behavior during VR, assessing whether performance differences are associated with changes in fixations or attentional engagement as task complexity and typographic design vary for instructional text.

Acknowledgements: ONR N00014-23-1-2298