Overall validity of eye-tracking measures of learning in IVR is limited, averaging 1.93 of 4, with data quality the weakest component¶

Claim · All claims
Evidence · 1 study (2 entries) · quant-synthesis r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? Across 38 reviewed studies, eye-tracking measures of learning constructs in IVR averaged 1.93 out of 4 on the validity scoring scheme (SD = 0.77), i.e., about 48%. → Lee 2026 q2 i? Data quality received the lowest average component score (0.32) while consequences scored highest (0.66), indicating insufficient control of data quality as a main driver of limited validity. → Lee 2026 (2)

Evidence¶

Lee 2026¶

Lee, J. Y., Moon, J., Paas, F., Cho, M., Darejeh, A., Holmer, S., & Ayres, P. (2026). Validity of Eye-Tracking Measures in Immersive Virtual Reality: Using Cognitive Theories of Learning. Educational Psychology Review. https://doi.org/10.1007/s10648-026-10186-9

q2 · i? · quant-synthesis · r2

Systematic review of 38 empirical IVR eye-tracking studies (2014–2024) scoring validity of measures on a 0–4 scheme. The review reports "The average score 1.93 out of 4 (SD = 0.77)" with only 2 measures (2%) reaching the maximum score of 4.

"The average score 1.93 out of 4 (SD = 0.77), ranging from 0.5 to 4. In terms of distribution, 12 measures (13%) received as a score of 0 or 0.5; 15 (17%) scored 1 or 1.5; 36 (39%) scored 2 or 2.5; 26 (29%) scored 3 or 3.5 point; and 2 (2%) achieved the maximum score of 4."

Lee 2026 (2)¶

Lee, J. Y., Moon, J., Paas, F., Cho, M., Darejeh, A., Holmer, S., & Ayres, P. (2026). Validity of Eye-Tracking Measures in Immersive Virtual Reality: Using Cognitive Theories of Learning. Educational Psychology Review. https://doi.org/10.1007/s10648-026-10186-9

q2 · i? · quant-synthesis · r2

Component-level analysis in the review's Results (Table 4) shows data quality averaged 0.32, the lowest of the four components. The review states "Scores for relations and data quality were below 0.5 on average," attributing low data quality to underreported methodological details.

"Scores for relations and data quality were below 0.5 on average, while content and consequences scored above 0.5, indicating variation in the strength of evidence across validity components."

Discussion¶

Related Claims¶

• Attention is the most frequently assessed learning construct in IVR eye-tracking research, while motivation, self-regulation, and immersion are unexamined — related
• Eye-tracking measures of cognitive load show significantly higher validity than measures of other learning constructs in IVR — related
• IVR eye-tracking studies commonly use small samples, adult participants, professional-domain tasks, and default vendor hardware and software configurations — related
• Neither self-report nor eye tracking can reliably distinguish strategic from habituated behavior in empirical practice — related