Article: Implicit measurement of college students' values driven by multimodal affective computing Journal: International Journal of Information and Communication Technology (IJICT) 2026 Vol.27 No.80 pp.51 - 73 Abstract: Traditional measures of college students' values rely on self-report questionnaires, which are susceptible to social desirability effects and introspective bias, making it difficult to capture genuine, unconscious tendencies. To address this challenge, this paper proposes a multimodal affective computing-driven implicit measurement paradigm. It collects students' microexpressions, speech prosody, and eye-tracking data in response to value-conflicting stimuli, and constructs a temporal attention network to uncover the mapping patterns between emotional responses and deep-seated value orientations. The experiment utilised both publicly available emotional datasets and a custom-built database. Results showed that the multimodal fusion model achieved an area under the curve of 0.87 in classification tasks - a 14% improvement over single-visual-modal approaches - and a normalised cumulative gain of 0.91 in ranking prediction, validating the validity advantages of implicit measurement. This study provides a feasible pathway to overcome the limitations of self-reporting and achieve objective assessment of values. Inderscience Publishers - linking academia, business and industry through research

Title: Implicit measurement of college students' values driven by multimodal affective computing

Authors: Wenjun Hu; Donghui Wang

Addresses: School of Marxism, Xinjiang Normal University, Urumqi, 830000, China; School of Marxism, Xinjiang Communications Polytechnic University, Urumqi, 830000, China ' College of General Education, Xinjiang Communications Polytechnic University, Urumqi, 830000, China

Abstract: Traditional measures of college students' values rely on self-report questionnaires, which are susceptible to social desirability effects and introspective bias, making it difficult to capture genuine, unconscious tendencies. To address this challenge, this paper proposes a multimodal affective computing-driven implicit measurement paradigm. It collects students' microexpressions, speech prosody, and eye-tracking data in response to value-conflicting stimuli, and constructs a temporal attention network to uncover the mapping patterns between emotional responses and deep-seated value orientations. The experiment utilised both publicly available emotional datasets and a custom-built database. Results showed that the multimodal fusion model achieved an area under the curve of 0.87 in classification tasks - a 14% improvement over single-visual-modal approaches - and a normalised cumulative gain of 0.91 in ranking prediction, validating the validity advantages of implicit measurement. This study provides a feasible pathway to overcome the limitations of self-reporting and achieve objective assessment of values.

Keywords: multimodal affective computing; latent measures; college students' values; physiological and behavioural responses.

DOI: 10.1504/IJICT.2026.155109

International Journal of Information and Communication Technology, 2026 Vol.27 No.80, pp.51 - 73

Received: 27 Mar 2026
Accepted: 29 Apr 2026
Published online: 27 Jul 2026 *

Title: Implicit measurement of college students' values driven by multimodal affective computing

Keep up-to-date