An affective interface for conveying student feedback

Abeyratna, Savandie An affective interface for conveying student feedback. Master of Philosophy thesis, University of Liverpool, Liverpool Hope University.

PDF AbeyratnaSav_Apr2012_6513.pdf - Author Accepted Manuscript Available under License Creative Commons Attribution No Derivatives. Download (9MB)

Abstract

In the present information age, decision-makers and modern society in general are challenged by the need to effectively handle large amounts of interrelated data obtained via electronic means. This thesis attempts to addresses the need for more effective data analysis and interpretation for decision-making. In particular, the study investigates whether virtual facial expressions (FEs) can be effectively applied as a non-verbal means to convey student feedback ‘at-a-glance’ and accurately with regard to affective content. This research has a threefold aim: (i) to handle the complex nature of multi-criteria type feedback data; (ii) map the feedback data into appropriate FEs and (iii) represent the data using a non-verbal affective interface. The approach adapted is such that the two-dimensional Kano model of satisfaction is established to evaluate feedback data in accord with multiple criteria; based on this, an aggregate score is generated that best represents the student feedback. Facial expressions of emotion are mapped to one-dimensional scales and the two-dimensional satisfaction space using psychophysical methods; mappings used to convert multi-criteria based student satisfaction ratings onto a pictorial representation in the form of cartoon facial expressions. A proof-of-concept prototype of an affective interface is developed and evaluated in terms of accuracy of the proposed non-verbal feedback analysis approach. The main findings of this study are that multi-criteria evaluation that takes into account two-dimensional quality can produce measures of satisfaction significantly correlated with manual rating. Student feedback can be conveyed accurately using virtual FEs provided that the multi-criteria analysis has been successful. Use of FEs to convey student feedback is faster than conventional feedback display modes.

Item Type:	Thesis (Master of Philosophy)
Additional Information:	Date: 2012-04 (completed)
Divisions:	Faculty of Health and Life Sciences > Institute of Population Health
Depositing User:	Symplectic Admin
Date Deposited:	06 Aug 2012 09:15
Last Modified:	16 Dec 2022 04:36
DOI:	10.17638/00006513
Supervisors:	Paramei, Galina Tawfik, Hissam
URI:	https://livrepository.liverpool.ac.uk/id/eprint/6513