Emocard

Recently, research on evaluating emotions has received attention in different HCI conferences. Humans are driven by emotions and it is also often said that emotions is a fundamental component of human being. Emotions can influence users’ satisfaction and acceptance towards any product or service.

Erevelles [17] stated that emotions can even influence the purchase intention and is an important determinant of consumption behavior. So far most of the research on emotions is focused on emotion metrics but on theoretical level [12]. Only few case studies exist on examining emotions in context to evaluating UX.

Before actually discussing emotion evaluation I need to answer first, what are emotions? In practice, there exist many definitions of emotion so it is both hard and challenging to find one precise definition that is also scientifically

Figure 3.10: Modified SUXES Evaluation Metrics [64]

correct. However, there exists an agreement on “what actually constitutes human emotions” [12]. First, emotion is a psychological reaction to events such as our needs, goals and concerns. Second, emotion is a combination of physiological, affective, behavioral and cognitive components.

Due to inherent relationship between emotions and product’s acceptabil-ity, they are considered key in determining UX. Andrew and Agarwal [6]

claimed that “emotion assessment helps in better understanding the UX.

However, evaluating emotions is not only complex but also challenging as users often face trouble in explaining as how they feel. Moreover, users can-not even differentiate between different emotions due to the instantaneous nature of emotions [15]

There are several emotion evaluation tools that help to capture and in-terpret emotions. These tools are broadly classified as verbal and non verbal evaluation tools. Verbal emotion measures include self reports where test participants use a scale to record their emotions. Non-verbal evaluation uses different visual representations of the emotions from which the test partici-pants choose how they feel. [6]

In this thesis, I used only non verbal tool for emotional reponse evalua-tion. Non verbal tools involves human like deceptions that have been vali-dated cross-culturally so they are reliable in their interpretation. Non-verbal tools attract a wider audience compared to verbal techniques because they capture conscious state of human mind unlike verbal measures [6].

Evaluat-ing emotions can enable product designer in better understandEvaluat-ing the UX. In order to attain this goal, I used Emocard [15] which is a well known non ver-bal evaluation tool. Emocard is used in evaluating users’ emotional response against MMR prototype and proofs of concept created in Chapter 5.

There are numerous emotional response evaluation tools available in ex-isting literature but I did not found single emotional evaluation technique that alone is reliable for the assessment of emotional response. I was aiming at finding a quick and easy method that can help in evaluating the users’

emotional response against MMR prototype and proof of concept in a cost effective way. Andrew and Agarwal [6] stated that most of existing emo-tional response evaluation tools are very much experimental and of unknown validity. Due to these reasons, I first decided to use a combined verbal and nonverbal emotional evaluation as done by Andrew and Agarwal [6] but later I decided to perform emotional evaluation by using Emocard only. There are two main supporting reasons for my decision namely - first, verbal emotional response techniques like PAD differential scale [48] has several overlapping statements to AttrakDiff questionnaire so making use of PAD or similar ver-bal emotional scale will result in a double effort from test participants’ end.

Second, my study already contains variety of UX evaluation methods such as SUXES, AttrakDiff and Emocard so having one more additional metric will result in a heavy and lengthy study that can make test participants’

experience dull and boring.

Emocard is widely used and empirically supported technique but its va-lidity in evaluating UX of mobile applications such as MMR is unknown.

Emocard helps in refining the experimental technique to boost my emotional response evaluation methodology.

Emocard consists of 16 different cartoons like faces (half male and half female faces)(see Figure 3.11). These faces represent different emotion and each face shows combination of two emotion dimensions such as pleasure and arousal. Now based on these dimensions Emocard is divided into four quadrants - calm-pleasant, clam-unpleasant, pleasant and excited-unpleasant. In Emocard study, if the user reaction is more pleasant and higher in the arousal then it is considered desirable. Moreover, if the users’

responses are more towards the calm-pleasant and excited-pleasant then it is understood as positive results. [6]

Emocard faces can be confusing to some users and it is also possible that users cannot interpret the meaning of different faces if used in users study.

This was found with two pilot tests performed by me during UX evaluation (see Chapter 6). In order to bridge this challenge faced by Emocard, I took three preventive measures namely 1) I added description and explanation for each pair of faces based on T¨ahti et al. [63]. 2) I gave training to participants

on how to interpret Emocard and how to give responses based on different combination of faces. 3) All test participants are asked to first perform two dummy tasks using Emocard before performing the actual test. For example,“what is your emotional response when I ask about your workplace?”

and “What is your emotional response When I ask about your current living place?”

Figure 3.11: Overview of Emocard having added explanations [63]

I modified the simple emocard method used in the earlier studies [6]. The method is not changed as such but further explanations and interview com-ponent is added to gather qualitative feedback of the participants. This kind of modification is essential for getting rich user data during UX evaluation.

Emocard alone is not suffice and participants’ reponses can be hard to un-derstand if qualitative explanations are not gathered. This was found during the pilot test. The Emocard study can be performed in three phases if only one prototype/concept is available for UX evaluation. However if the number of prototype and concept is two then Emocard study can be completed in six phases (see Figure 3.12). In the first phase, participant is introduced with the study setup and given training on using Emocard. Participant is

intro-duced with concept/prototype in phase second for which emotional response needs to be evaluated. In third phase, participant is instructed to answer Emocard based on their experience with the prototype/concept in phase sec-ond. Later, participant is interviewed for getting qualitative opinion on the shown concept. This process is repeated for any number of prototypes and concepts.

Figure 3.12: Overview of Emocard study process