|
|
Emotions play a major role in the social interaction process with electronic commerce systems. This paper describes our attempts to design customer interfaces for cyber banking systems that can induce target emotions for cyber banking systems. Four experiments were conducted to identify the important emotive factors and design factors, and to establish and verify causal relations between the factors. Results indicate that it is possible to design customer interfaces that will elicit target emotions for the systems (e.g., trustworthiness).
emotional usability, customer interface, trustworthiness
© 1997 Copyright on this material is held by the authors.
Until recently, study of human computer interaction has focused on designing artifacts to help users overcome their cognitive limitations, with little regard for the emotive aspect of the interface [3]. However, human computer interaction is inherently a social process; people invariably behave as if the computer were a social actor [4]. Emotions play an important role in social interaction processes through its close relation to social reasoning and decision making [2]. By providing information as to the emotional desirability of the options available, they reduce and limit reasoning to those that induce positive feelings [1].
In a recent study on emotion, films and slides were found to elicit common target emotions [2]. Computer interfaces can basically be considered as a collection of slides composed of diverse visual and auditory stimuli. Accordingly, computer interfaces also have the potential to elicit common emotions in their users. This has important implications for the design of customer interfaces in electronic commerce systems, because emotions play an important role in the social interaction process involved in commercial transactions. This study focuses on cyber banking systems and investigates the possibility of designing customer interfaces that elicit desirbale emotions for the users of cyber banking systems.
This study consists of four experiments. In the first, the emotional factors that influence a user's overall impression of a visual interface is determined. The second experiment identifies the design factors that a user usually attends to in the visual interface. The third experiment establishes the relations between the emotional and design factors. Finally, the fourth verifies these causal relations.
Materials and Procedure. The aim of the first experiment is the selection of a comprehensive set of emotive scales for measuring the emotions elicited by cyber banking interfaces. A variety of sources related to cyber banking systems was combed for the 318 bipolar terms used in this first study. The experiment was conducted in three group sessions of five to ten subjects, with a total of 25 subjects. Subjects were shown one cyber banking interface at a time and asked to indicate the intensity and direction of the feelings elicited by the interface on 7 point Likert scales. Eight interfaces representative of the different types of cyber banking systems were used, the order of which was counter-balanced for each of the three group sessions so as to prevent an ordering effect.
Results and Discussion. From the 318 scales, we identified 10 groups of closely related emotive terms through cluster analysis. From each of the 10 clusters, a group of experts in cyber banking systems selected a few differential scales representative of the domain specific emotions that are most important in cyber banking systems. As a result, forty bipolar scales (e.g., reliable-unreliable) were selected to be included in the final self-report questionnaire used in the following experiments.
Materials and Procedure. Twenty-six interfaces through an extensive search of existing cyber banking systems in our attempts to include all different designs in the test material. The experiment was conducted in two group sessions of 30 subjects in total. Instructions for the task were the same as in the first experiment, with the exception of the free recall tests. Upon completion of the self-report questionnaire, the interface was hidden from view and subjects were asked to draw what they could remember of the interface. This was done at random intervals for a few selected interfaces during each session.
Results and Discussion. Subjects' drawings were analyzed to identify the most frequently recalled design items. The fourteen design factors that were identified from the drawings are shown in Table 1. A group of experts in cyber banking systems classified the identified design factors into four major groups : title, menu, main clipart, and color. These represent the major design factors from a user's perspective for the cyber banking systems.
Table 1. Design Factors of Cyber Banking Interfaces
| Categories |
Design Factors |
Values |
|---|---|---|
| Title | Format | Bar, Clipart ... |
| Graphics | Included, Not Included | |
| Position | Top, ... | |
| Menu | Content | Numbers & text in a box,... |
| Form | Square,... | |
| Size | Over 1/16 | |
| Main Clipart | Format | 3 Dimensional,... |
| Size | Over 1/2 total screen size,... | |
| Motion | None, Simple animation... | |
| Color | Color Tone | Cool shades... |
| Main color | Pastel Colors... | |
| Background | White covering over 1/2 ..... | |
| Brightness | High..... | |
| Symmetry | Stability of color tones.... |
Materials and Procedure. The experiment was conducted in three group sessions of 123 subjects in total. The procedure employed was identical to that of the first experiment. Twelve cyber banking interfaces were selected to cover all the fourteen design factors obtained from the second experiment. Consequently, each subject had to provide judgments for each of the 480 items (40 emotive differential scales x 12 cyber banking interfaces) on the self report questionnaire.
Results and Discussion. Factor analysis was conducted to identify the major emotional dimensions. Seven factors accounting for 80.6% of the variance were identified: trustworthiness, symmetry, attractiveness, sophistication, awkwardness, elegance, and simplicity. Due to space limitations in this paper we present the detailed analysis for only one of the seven factors, trustworthiness, because of its importance in the design of cyber banking interfaces. ANOVA tests were conducted on the factor scores computed for each observation to determine the relations between trustworthiness as the dependent variable and the design factors as independent variables. Table 2 displays the main effect of the design factors that were found to have significant impact on the extent of trustworthiness evoked by the interface.
Table 2. Effect of Design Factors on Trustworthiness
| Design Factor |
Main Clipart |
Color |
||||||
|---|---|---|---|---|---|---|---|---|
| Form | Size | Motion | Tone | Main | Background | Brightness | Symmetry | |
| F | 213.0 | 148.2 | 17.0 | 401.7 | 214.1 | 190.5 | 78.7 | 277.6 |
| P | .0001 | .0001 | .0001 | .0001 | .0001 | .0001 | .0001 | .0001 |
| Design Factor | 3 D (.41) | 1/2 size (.26) | anima (.17) | Cool (.28) | Pastel (.15) | None (.21) | Low (.19) | Similar (.15) |
The design factors that were most important in deciding the trustworthiness of a particular cyber banking interface was identified to be those related to the main clipart and color of the interface. The last row denotes the value of the design factor that is most effective in eliciting the feeling of trust, the numbers in parentheses indicating the mean factor scores of the respective design. The results indicate that interfaces that have an animated 3D clipart ½ the size of the full screen, and a cool, pastel tone color will elicit a greater extent of trustworthiness.
Materials and Procedure. Based on the results in Table 2, two cyber banking systems were built to verify the causal relations between the two groups of design categories and trustworthiness. The interfaces of the two systems were identical except with regard to the design factors in Table 2. Interface I was designed to enhance the extent of trustworthiness elicited by incorporating animated 3D clipart over a pastel tone color, whereas Interface II was designed so as to lower the extent of trustworthiness evoked. Two group sessions of a total of 55 subjects were conducted, in which subjects evaluated the two interfaces according to the same procedure as in the first experiment.
Results and Discussion. Factor scores for trustworthiness were computed for the observations using the scoring coefficients calculated in the third experiment. ANOVA tests reveal that the design factors manipulated in the design of the interfaces have a significant impact on the extent of trustworthiness elicited (F=50.32, p<0.0001). Interface I evoked a greater amount of trustworthiness (factor score=+0.36) than Interface II (factor score = -0.88).
This paper investigates the possibility of designing emotive interfaces which can involve the useres emotionally and thus enhance the quality of the decisions made while interacting with the system. The results indicate that it is possible to design customer interfaces of cyber banking systems that will elicit target emotions. We are currently conducting a similar study with different systems of electronic commerce : cyber shopping malls. Preliminary results from the study reveal that similar emotive and design factors can be identified in cyber shopping malls, an indication of the possibility of developing common emotive and design factors for electronic commerce systems.
|
|