Sticky and Magnetic: Evaluating Error Correction and User Adaptation in Gaze-and-Pinch Interaction Conference: Extended Abstracts of the 2026 CHI Conference on Human Factors in Computing Systems; April 13–17, 2026; Barcelona, Spain Extended Abstracts of the 2026 CHI Conference on Human Factors in Computing Systems (CHI EA ’26), April 13–17, 2026, Barcelona, Spain DOI: 10.1145/3772363.3798839 ISBN: 979-8-4007-2281-3/2026/04

Jazmin Collins, Prasanthi Gurumurthy, Eric J. Gonzalez, Mar Gonzalez-Franco · 2026

Read Sticky and Magnetic: Evaluating Error Correction and User Adaptation in Gaze-and-Pinch Interaction Conference: Extended Abstracts of the 2026 CHI Conference on Human Factors in Computing Systems; April 13–17, 2026; Barcelona, Spain Extended Abstracts of the 2026 CHI Conference on Human Factors in Computing Systems (CHI EA ’26), April 13–17, 2026, Barcelona, Spain DOI: 10.1145/3772363.3798839 ISBN: 979-8-4007-2281-3/2026/04 by Jazmin Collins, Prasanthi Gurumurthy, Eric J. Gonzalez, Mar Gonzalez-Franco (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-26608.txt.

Introduction

Collins and co-authors test two fixes for timing errors in gaze-and-pinch control for virtual reality. Users often pinch just before their gaze lands, or after it leaves a target. A sticky rule adds a time buffer, while a magnetic rule pulls selection toward nearby targets. Nine users tried both against a baseline on a Samsung Galaxy XR headset. Speed and throughput held steady, yet errors fell and user habits shifted in clear ways. Magnetic selection led users to trade precision for speed, an offloading effect.

Catalog record

CREATORJazmin Collins, Prasanthi Gurumurthy, Eric J. Gonzalez, Mar Gonzalez-Franco
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE26219
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.26608

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Sticky and Magnetic: Evaluating Error Correction and User Adaptation in Gaze-and-Pinch Interaction Conference: Extended Abstracts of the 2026 CHI Conference on Human Factors in Computing Systems; April 13–17, 2026; Barcelona, Spain Extended Abstracts of the 2026 CHI Conference on Human Factors in Computing Systems (CHI EA ’26), April 13–17, 2026, Barcelona, Spain DOI: 10.1145/3772363.3798839 ISBN: 979-8-4007-2281-3/2026/04 Jazmin Collins, Prasanthi Gurumurthy, Eric J. Gonzalez, Mar Gonzalez-Franco Abstract . The gaze-and-pinch framework offers a high-fidelity interaction modality for spatial computing in virtual reality (VR), yet it remains vulnerable to coordination errors—timing misalignments between gaze fixation and pinch gestures. These errors are categorized into two types: late triggers (gaze leaves a target before pinch) and early triggers (pinch before gaze arrival on target). While late triggers are well-studied, early triggers lack robust solutions. We investigate two heuristics— Sticky selection (temporal buffer) and Magnetic selection (spatial field)—to mitigate these errors. A within-subjects study ( N = 9 N=9 ) on the Samsung Galaxy XR evaluated these heuristics against a baseline. Findings indicate that while throughput and selection time remained stable, the heuristics fundamentally shifted user behavior and significantly reduced errors during selection. Notably, Magnetic selection induced an “offloading” effect where users traded precision for speed. Additionally, the heuristics reclassified ambiguous failures as explainable coordination errors. We provide recommendations for selection heuristics that enhance interaction speed and cognitive agency in virtual reality. Figure 1. We investigate two types of coordination errors common to gaze-and-pinch interactions (A), Early and Late Trigger errors, along with two correction heuristics (B), Magnetic and Sticky selection. † † cc-license: by-nc-nd 1. Introduction Virtual reality (VR) is growing in popularity across sectors, from healthcare to education. As major firms continue to invest in the ecosystem, VR is transitioning from entertainment technology into a platform for spatial computation. In this new paradigm, a 3D digital workspace promises immersion and utility that could supersede traditional computing. However, realizing the potential of spatial computing requires input methods that match the fidelity of traditional interfaces. In particular, making selections in VR remains a key usability bottleneck. Users find that selecting objects in 3D space is slower, less accurate, and more fatiguing than using standard devices such as a mouse or trackpad ( 8 ; 1 ; 20 ) . To address this, the HCI community has examined multimodal interactions in VR ( 21 ; 10 ; 13 ; 9 ) , including the gaze-and-pinch framework ( 18 ) , which combines the speed of eye tracking for targeting with the ease of pinch gestures for confirmation. However, gaze-and-pinch interaction still poses challenges to users. In particular, prior work has noted that because gaze is a highly rapid input—often moving faster than manual motor responses—users frequently experience coordination errors when performing gaze-and-pinch selection ( 17 ; 15 ; 16 ) . Such coordination errors (i.e., when the timing of the gaze fixation and the pinch do not align) are particularly common due to the asynchrony between high-speed gaze and the slower, manual pinch. These errors fall into two categories, shown in Figure 1 A: Late Trigger Errors , which occur when a user fixates on a target but shifts their gaze away before they pinch, and Early Trigger Errors , which occur when a user initiates the pinch before their gaze lands on the target. Past work has attempted to address these errors, exploring solutions to late trigger errors like machine learning (ML) models trained to identify late pinches and correct system output ( 16 ) . However, early trigger errors remain underexplored, despite major differences between the error types. With late trigger errors, a system can review past gaze data to identify and correct for an intended target; in contrast, early triggers require a predictive approach to determine where gaze was going when the pinch occurred. Solutions for late and early triggers are fundamentally incompatible, leaving a gap in how we address them both. We aimed to address this gap by investigating two solutions for coordination errors (Figure 1 B): • (1) Sticky selections : A temporal adjustment that keeps a gaze reticle artificially “stuck” to a target for a brief window of time (designed for late trigger errors). • (2) Magnetic selections : A spatial adjustment that adds a larger “magnetic field” around an intended target, snapping a gaze reticle to the center of the target once it touches the field (designed for early trigger errors). We evaluated these two heuristics in a within-subjects study with nine expert users of gaze-and-pinch. We had our participants perform repeated target selections in VR with a baseline and each of our heuristics. Finally, we collected log data of their performances to evaluate the effects of our heuristics via statistical analysis of key metrics. Ultimately, we discovered that our heuristics significantly reduced the errors that participants would have experienced during selection, without impacting performance metrics such as throughput, accuracy, or selection time. We also identified behavioral shifts among our participants as they adapted to the heuristics, such as performing less precise selections on targets with a "magnetic field." By sharing these results, we provide recommendations for mitigation of coordination errors to create a truly seamless spatial computing experience in VR. 2. Related Work 2.1. Gaze-and-Pinch Challenges The concept of combining gaze for coarse targeting with manual input for fine selection was first popularized by Zhai et al. with MAGIC pointing ( 23 ) . Since then,HCI researchers have explored various interaction problems with gaze-and-pinch selections in VR ( 16 ; 17 ; 15 ; 18 ) . Notably, Pfeuffer et al. ( 17 ) presented five principles of gaze-and-pinch interactions and challenges that must be addressed, such as improving drag and drop mechanics, expanding how users manipulate objects, and addressing coordination errors. In recent years, the HCI community has begun to explore one of these challenges, coordination errors, in more depth ( 16 ; 15 ) . Park et al. trained ML models to identify late triggers, incorporating dynamic timing adjustments to pinches that were flagged as late ( 16 ) . These models resulted in an overall decrease in late triggers among their participant group, demonstrating a promising approach for these errors. While previous work has explored solutions for late trigger errors, there remains a gap in the study of early trigger errors. Although less frequent, early triggers pose a distinct usability barrier in gaze-and-pinch interaction. Specifically, late trigger errors are addressed through retroactive adjustment, leveraging past gaze data to identify the intended target. In contrast, early triggers occur before the user’s gaze has reached the target, necessitating predictive action. Thus, there is a need to explore unique solutions for early trigger errors. Our work has taken a first step to this goal, introducing two heuristics designed to compensate for early and late triggers, respectively. 2.2. Cognitive Impact of Coordination Errors Within the neuroscience community, the concept of intentional binding, or action-binding, provides a critical framework for understanding coordination errors. Action-binding describes the process of mentally linking an action that one intended to perform to its result ( 5 ; 6 ; 4 ; 2 ; 14 ; 19 ) . Past work in this area by Haggard et al. has shown that when results following intentional

Read the complete text