Individual differences in crowding predict reading speed in rapid serial visual presentation
Poster Presentation 23.406: Saturday, May 16, 2026, 8:30 am – 12:30 pm, Pavilion
Session: Object Recognition: Reading
Schedule of Events | Search Abstracts | Symposia | Talk Sessions | Poster Sessions
Fengping Hu1 (), Anna Bruns1, Najib J. Majaj1, Jonathan Winawer1, Denis G. Pelli1; 1New York University
Visual crowding—defined as failure to recognize in clutter—reduces the visual span, which reduces reading speed. Here, complementing the reported stimulus effects that relate crowding to reading, we ask whether there are also individual difference effects: Do individual differences in crowding predict differences in reading speed across a large online sample? 168 online participants completed three tasks: (1) Visual crowding: identifying the middle letter of a trigram at ±8° eccentricity (QUEST threshold spacing); (2) RSVP reading: identifying three sequentially presented nouns at fixation (QUEST threshold word duration); (3) Ordinary reading: speed in word/min based on time taken to read a short IReST story (Trauzettel-Klosinski et al., 2012), with Flesch-Kincaid grade level 7.2±1.2, with subsequent questions to verify comprehension. Results confirmed large individual differences: all three measures varied several-fold, with individual differences much larger than within-subject measurement noise. Combining the measures revealed a moderate correlation (r = –0.38) between crowding distance and RSVP reading thresholds, indicating that individuals with more crowding (larger spacing thresholds) read more slowly. In contrast, crowding did not correlate with ordinary reading speed. We speculate that for some participants, letter spacing is the bottleneck that limits the rate of RSVP reading, whereas there are other, narrower bottlenecks that limit the rate of reading the 7th-grade-level passages presented to our adult participants.
Acknowledgements: Supported by: Meta Inc. contract to D. Pelli.