Diagram of study design decisions (calibration, recalibration, quality thresholds) feeding into validated webcam eye tracking accuracy

Design Considerations that Affect Webcam Eye Tracking Accuracy

Webcam-based eye tracking's accuracy depends heavily on the implementation behind it. This page covers the design decisions, calibration and screening thoroughness, stimulus positioning, and the recording environment, that determine what accuracy a specific study actually achieves.

Study Design Decisions That Determine Accuracy
Frequently Asked Questions
Published Research Using This Method

Study Design Decisions That Determine Accuracy

A study's webcam eye tracking accuracy comes from two layers: the baseline accuracy of the underlying technology, and the design decisions, calibration, stimulus positioning, and the recording environment, that determine how much of that baseline accuracy a specific study actually reaches.

Baseline webcam technology accuracy

Published accuracy figures vary widely across the field when it comes to the capabilities of webcam eye tracking technology. For example, WebGazer, the open-source library still used in the majority of published webcam eye tracking studies in psycholinguistics (Patterson, Nicklin & Vitta, 2025), reports about 4° of visual angle (Papoutsaki et al., 2016). More recent systems using newer gaze-estimation approaches reach well under 2°: Labvanced's eye tracking has the field's one peer-reviewed comparison against a research-grade hardware eye tracker published to date, reaching 1.3° at the center of the screen against an EyeLink 1000 (Kaduk et al., 2024). Its own further validation, not peer-reviewed but published with open data, puts real online participants at a median of about 1.8° on their own computers (Eye Tracking 2.0, 494 Prolific participants). Figures aren't always directly comparable across tools: some report in pixels rather than visual degrees, and not every vendor publishes a specific number at all.

None of these figures is a guarantee for your specific study, each is a ceiling reached under a matched setup. What determines whether a given study reaches it is the design decisions below, illustrated throughout with how Labvanced implements them.

Calibration

Any webcam eye tracking system estimates gaze position from a calibration mapping, and a more thorough calibration produces a more accurate one. Calibration is also the only point in a study where the participant has to actively engage with the system rather than simply attend to the task, so thoroughness is a real tradeoff against participant time and dropout, not a setting to maximize by default. The right amount depends on the AOI layout: many small, closely spaced AOIs need the accuracy a more thorough calibration provides, while a few large AOIs are often better served by a lighter one. Several decisions fall under this one tradeoff:

Length and point count. A longer calibration with more points produces a more accurate mapping.
Data-quality screening. How strictly to screen out participants whose calibration doesn't meet a quality bar. Without one, a poor calibration passes silently into the dataset indistinguishable from a good one; stricter screening buys more accuracy among the participants who remain, at the cost of turning more people away.
Recalibration cadence. Whether and how often to re-check calibration during a session, since a mapping taken once at the start can drift as the session goes on.
Head-position control. How far a participant's head is allowed to drift from the calibrated position before it's flagged.

In Labvanced's implementation: Eye Tracking 2.0 bundles all four into a guided accuracy tier and screening level, with live corrections for drift and lighting running by default; a Custom configuration or a study still on the older Eye Tracking 1.x sets each individually instead. How to Choose Eye Tracking Calibration Settings is the dedicated decision guide.

Stimulus positioning

Accuracy is not uniform across the screen: peripheral stimuli carry more error than central ones, which has two direct implications for study design.

Centering. Where a paradigm allows it, placing the critical stimulus or the region that matters most centrally yields the most reliable gaze data.
AOI spacing. Any two AOIs presented in the same trial need enough space between them to exceed the method's error margin, otherwise a fixation near the boundary between two closely-spaced AOIs cannot be reliably assigned to either one. The closer the accuracy figure being relied on, the more spacing a design needs to build in.

In Labvanced's implementation: in its peer-reviewed validation, accuracy improved for stimuli at the center of the screen (1.3° accuracy, 0.9° precision) relative to the overall figure (1.4° accuracy, 1.1° precision) (Kaduk et al., 2024).

Environmental and participant conditions

Choosing a thorough calibration and strict screening sets a ceiling; the recording environment decides whether a given session reaches it. Three conditions are worth planning for: lighting, eyewear, and webcam positioning.

Lighting

Backlighting, a bright light source such as a window or lamp positioned behind the participant, silhouettes their face against it, reducing the contrast that camera-based face tracking depends on to detect facial and eye landmarks precisely. Plan your participant instructions around this regardless of generation, it's cheaper to prevent than to correct.

In Labvanced's implementation: the calibration screen instructs participants to be in a stable, well-lit room with no bright light source behind them. On Eye Tracking 2.0 a lighting-change detector also triggers an extra check if the light on a participant's face changes mid-session, but this reduces the cost of a change, it doesn't remove it, so the instruction still matters.

Eyewear

What actually degrades tracking is light getting distorted before it reaches the webcam, glasses that cause reflections, or lenses with a tinted or blue-light-filtering coating, not eyewear as such. Glasses without those properties do not meaningfully affect calibration, making eyewear a narrower concern than it first appears.

In Labvanced's implementation: the default participant-facing calibration instructions currently tell participants not to wear glasses at all, a conservative default rather than a technical restriction. That instruction is editable: it lives in the eye tracking system messages under the Texts & Translate static strings, so a researcher whose participant population commonly wears glasses can revise the instruction rather than exclude those participants by default.

Webcam positioning

Calibration maps gaze to screen coordinates relative to the webcam's position, so an offset between the webcam and the screen center adds error beyond what correction alone fixes. This only applies to a webcam that isn't built into the device, since an integrated webcam's position relative to the screen is fixed.

In Labvanced's implementation: for participants using an external webcam, the calibration screen instructs them to position it as close to the center of the screen as possible.

Summary Table

Decision	General principle	Labvanced's implementation
Calibration	More AOIs, smaller and closer together, need a more thorough calibration, stricter screening, more frequent recalibration, and tighter head-position control to resolve reliably	A guided accuracy tier and screening level on Eye Tracking 2.0; `Custom` or Eye Tracking 1.x set each individually. See the calibration settings guide
Stimulus positioning	Accuracy is higher center-screen than peripherally; closely-spaced AOIs risk misclassification	Validated center-screen accuracy figure (1.3° in the laboratory) available to plan AOI spacing against
Environmental and participant conditions	Lighting affects tracking quality; glasses are fine unless they cause reflections or a tinted coating	Participant instruction against backlighting, backed by a lighting-change detector on Eye Tracking 2.0; default "no glasses" instruction is an editable static string, not a restriction

None of these are unusual requirements. They are the same category of decision a lab makes about a hardware eye tracker's setup, applied to a setting where the researcher cannot physically supervise the participant.

Whether a given accuracy level fits your research question at all, before any of these decisions come into play, is covered in How to Run a Webcam Eye Tracking Study's first step.

Frequently Asked Questions

How accurate is webcam eye tracking?

It depends heavily on the implementation: published figures range from about 4° for widely-used open-source libraries like WebGazer up to well under 2° for the most rigorously validated systems. Labvanced's peer-reviewed laboratory comparison against an EyeLink 1000 reached 1.3° at the center of the screen (Kaduk et al., 2024), the most rigorously benchmarked webcam eye tracking figure published to date. Its own further online validation (Eye Tracking 2.0, not peer-reviewed but published with open data) puts real participants at a median of about 1.8° at the standard setting, down to about 1.4° at the most thorough one. See the Eye Tracking 2.0 Lab Note for the full validation.

Is webcam eye tracking accurate enough for reading studies?

Yes, at the most thorough calibration and screening setting, which delivers about 1.5°, enough to resolve word-level AOIs with normal spacing. Larger AOIs, such as faces, scenes, or page regions, work at a lighter setting.

How long does webcam eye tracking calibration take?

It scales with how thorough your design needs it to be: roughly 4 minutes for a standard setup and about 7 minutes for the most demanding one, with lighter setups taking under a minute or requiring no calibration at all. Pick based on your AOI spacing first, calibration time follows from that choice. See the calibration settings guide for the exact options.

Does webcam eye tracking work with infants?

Yes. The lightest setting tracks gaze with no calibration at all, at an estimated error of about 4°, enough to tell left from right in preferential-looking studies. For designs that can tolerate a brief, infant-friendly calibration, accuracy improves further. See the calibration settings guide for the infant-specific options.

Do I still need to schedule recalibration during a session?

Not on Eye Tracking 2.0: live corrections for head drift and lighting changes run by default on every calibrating setting, so there's no manual recalibration cadence to configure. On Eye Tracking 1.x, recalibration had to be triggered manually at defined points, for example between blocks, to guard against drift.

How do I know if a participant's calibration is good enough to include in my data?

On Eye Tracking 2.0, this is what the screening level decides: participants who don't meet the bar for your chosen level are excluded or asked to retry before data collection counts, rather than something you check case by case afterward. On Eye Tracking 1.x, a maximum calibration error threshold did the same job as a setting you configured directly.

How much space should there be between AOIs in a webcam eye tracking study?

Enough to exceed the accuracy margin of the calibration and screening setting chosen: as tight as about 1.8° at the standard setting, down to about 1.5° at the most thorough one, or 1.3 to 1.4° in peer-reviewed laboratory validation (Kaduk et al., 2024). Two AOIs placed closer together than that margin risk having a fixation near the boundary misassigned to the wrong one. Placing critical stimuli centrally and building in spacing proportional to the accuracy figure being relied on reduces this risk.

Labvanced's calibration instructions tell participants not to wear glasses. Does that mean glasses aren't supported?

No. What actually affects tracking is light getting distorted before it reaches the webcam, glasses that cause reflections or have a tinted or blue-light-filtering coating, not eyewear as such. Glasses without those properties don't meaningfully affect calibration. The default "do not wear glasses" instruction shown to participants is a conservative default text, not a technical restriction, and it's editable: it's one of the eye tracking system messages under the `Texts & Translate` static strings, so a study with a participant population that commonly wears glasses can revise it rather than exclude those participants outright.

Published Research Using This Method

Cassano-Coleman, R. Y., Izen, S. C., & Piazza, E. A. (2026). Listeners Systematically Integrate Hierarchical Tonal Context, Regardless of Musical Training. Psychological Science. (Includes gaze-contingent elements).
Zhang, P., & Zhang, S. (2025). Attention and learning in L2 multimodality: A webcam-based eye-tracking study. Language Learning & Technology. https://doi.org/10.64152/10125/73626
Serrano-Carot, M., Angele, B., Xu, H., & Vasilev, M. R. (2025). Webcams Can Be Used to Study Eye Movements during Reading. PsyArXiv (OSF Preprints). https://doi.org/10.31234/osf.io/bzt2h_v1
Lester, C., et al. (2025). Effect of uncertainty-aware AI models on pharmacists' reaction time and decision-making in a web-based mock medication verification task: Randomized controlled trial. JMIR Medical Informatics. https://doi.org/10.2196/64902
Wies, S., Bleier, A., & Edeling, A. (2022). Journal of Marketing. Webcam-based eye tracking in a marketing research context.
Banki, A., de Eccher, M., et al. (2022). Frontiers in Psychology. Developmental and infant eye tracking.