Beyond Usability: Evaluating Emotional Response as an integral Part of the User Experience
Research on Emotions is important in the current HCI as they influence any person’s satisfaction and acceptance for any product. One important quote “People work more efficiently and are creative problem solvers when they are happy”. Even some studies have shown that emotions influence users’ purchase intention and determinant of the consumption behavior. This paper has reviewed existing emotion measurement tools in order to develop methods that combine both verbal and nonverbal factors for measuring emotional responses. Based on these
measurement factors – two usability studies have been conducted. The contribution of this paper is that “emotion assessment helps in better understanding the user
Definition of emotion
There are so many definitions for emotions and it is difficult to conclude one particular definition.
However there are two widely agreed constituents of the human emotions.
1. Emotion is a psychological reaction to events such as our needs, goals and concerns.
2. Emotion is a combination of physiological, affective, behavioral and cognitive components.
Emotion is complex subject of study due to several reasons – humans often have problems in describing how they feel and distinguishing between different emotions. It is also difficult to find the exact reason for any particular emotion because emotions are spontaneous. However there
are different existing emotion measurement tools which are verbal, nonverbal and physiological in nature. The study presented in this study is only makes use of verbal and nonverbal measurement tools as it was conducted remotely.
Verbal measurement tools
– Primarily used in the marketing and advertising sectors.
– Self report system where a participant them-self record their emotions using scales.
only captures conscious emotional states. For effective study => Comprehensive
studies are recommended but comprehensive studies often get lengthy. However, for
measuring emotions, lengthy studies are not recommended because emotions are
spontaneous and immediate so extensive time can disturb how users initially felt.
Verbal measures are language dependent so their application is limited to certain
population and people.
– Advantage: Widely used because they are easy to develop and use.
– Likert scales (strongly disagree to strongly agree) and semantic differential scales (bipolar adjective pairs at each end of the scale).
Non-verbal measurement tools
– Challenging to find some effective non verbal measure of emotion.
– Example: Visual representation of emotion and participant choose to characterize their
feelings. Moreover example, smiley faces or human like deceptions.
– Non verbal methods aims to capture unconscious emotional responses.
The aim of the study was to quickly and easily measure the emotions with low cost. The study
combined verbal scale with experimental non verbal emotion measure to improve existing emotion measurement methodologies.
Verbal PartPAD Semantic differential scale was used as a verbal component. The scale includes bipolar adjective pairs that were rated on nine point scale. PAD scale was used to measure three aspects of emotions –
Pleasure – positive affective state which is different from preference and reinforcement.
Arousal – emotional state from sleepy to very excited.
Dominance – when person feels unrestricted, free from outside.
PAD is extensively used validation tool for measuring emotional response. Its validity in the marketing and advertising is highly practiced but its use in application software is still new.
Non – Verbal Part
Emocard tool was used for non verbal component. It contains 16 cartoon like faces having equal number of male and female faces. Each represent different emotion. Each face represents a combination of two emotion dimensions example pleasure and arousal are identical to two of the dimensions included in the PAD scale.
Emocards can be divided into four quadrants – Calm-pleasant, Clam-unpleasant, ExcitedPleasant and Excited-Unpleasant. The results in Calm-pleasant and Excited-Pleasant are interpreted as positive.
Both PAD scale and Emocards were placed into the online survey in random order.
Different tasks were created as in any usability testing. Time for completing a given task and number of errors were collected. After completion of the every task, users were directed to the online survey where they select Emocard which represents initial emotional reaction to the task. Then PAD scale is filled by the user. Participants are also required to give qualitative feedback after this.
1. Factor Analysis: The motivation was to check the validity of PAD semantic differential scale for interfaces. All the three factors: pleasure, arousal and dominance identified by Mehrabian and Russel for marketing and advertising domain can also be extended for measuring the emotional response of interfaces as well.
2. T-test analysis: The motivation was to check if measuring the emotions in addition to traditional usability metrics gives any additional insight into the user experience. The analysis revealed differences in the emotional responses but no differences from the usability point of view.
1. Studying emotions is challenging as it is subjective, fluid and qualitative construct.
2. Usability metrics are widely used to indicate the user experience level however, usability is one small part of UX.
3. Strong reliance on traditional methods such as heuristics evaluations, usability testing, time taken for a task, errors made and so on. Usability is one of the easy component of UX to access UX. Qualitative emotional experiences can contribute to usability and helps in judging overall product UX.
4. The paper has contributed in studying “How measuring emotions can contribute in the existing user experience research”.
5. Usability of two interfaces can be identical but there was difference in the emotional experience of those interfaces.