Keywords

1 Introduction

User interaction with computers is now a constant in our lives, an active part of our daily routines. We not only use computers to perform tasks but also use them to manipulate remote Internet access and control devices such as home security devices, camera surveillance, and temperature or light controls. All users, from young to old, count on user interface to work in their favor, helping them successfully complete their essential tasks. One issue that has been observed concerns the way in which interface design reflects how users want to view the devices they are remotely managing. In other words, what kind of mental image should the design trigger when a user views the user interface? The knowledge as to how this mental image matches the real concept of user interface is the key to creating a capable design.

A variety of researchers have investigated the mental image, imagery, perception, and effects on how we interact with our world [1]. To just mention a few sources for example, Norman suggests that people form internal representations or mental models of themselves and the objects with which they interact to create predictive and explanatory powers for understanding the interaction [2]. Gentner and Stevens support the concept that mental models are based on the way people understand a specific knowledge domain [3]. Johnson-Laird believes that mental models play a central and unifying role in representing objects, states of affairs, sequences of events, the way the world is, and the social and psychological actions of daily life (397) [4].

Much research is focused on how the mental image differs among the designers and users. Overall, two mental models have been distinguished: a user’s mental model, referring to what an end user believes about a system [5], and a designer’s mental model, which refers to the conceptualization invented by a design [6]. Nielson believes that “what users believe they know about a UI strongly impacts how they use it. Mismatched mental models are common, especially with designs that try something new”. Several studies tend to investigate, understand, and use mental representations to analyze design interfaces that are based on users’ mental and propose frameworks [7].

Athavankar’s study illustrates that designer create virtual models in their ‘‘mind’s eye,’’ then manipulate and alter them, and make them behave according to their wishes during the development of their ideas [8]. Waren reports that users develop a mental model of the new system that is analogous to the old one [9].

In a previous study [10] we tried to understand what user interface would be easier for users to manipulate or control through UI, whether an abstract or concrete representation would be better received. For example, to turn a light on and off, would it be easier to click on a representation of a lighting fixture, such as a light bulb, or simply check a box? Another example is temperature regulation. Would it be better to manipulate a virtual representation of thermometer or enter a set temperature? We have shown that 71 % of the participants who created a UI for light control had a more abstract approach, using a nonfigurative image of the object, 4 % had a semi-concrete approach (somehow visualizing the light), and only 25 % had a very concrete view of a lighting fixture to illustrate the on and off functions. For the temperature control group, the breakdown was 52 % abstract, 32 % semi-concrete and 16 % concrete.

In this follow-up study, two user studies were conducted. In the first, the previous study was replicated with 33 graduate students. In the second study, we expanded the survey to the general population asking them to choose between the two options: Concrete and Abstract.

2 Method

Two user studies were conducted. In the first study, the previous study was replicated with 33 graduate students. The participants were a mixture of human factors and software engineering students (13 males and 20 female). The students were mid-way through the course and already had acquired fundamental knowledge of HCI and user interface design. The exercise is given during the class time where 10 min were given to provide a 1-page paper prototype (low fidelity) of the design case. No other instruction beside what was written on the exercise description was given (Table 1). Upon collection of the prototypes they were classified into following categories: Virtual (Concrete), and Abstract (See Figs. 1 and 2). After conducting this experiment we then gave the same group two design options: Abstract and Concrete (Figs. 3 and 4) asking them to select the one they prefer. We wanted to see whether participants were consistent in preferring their own design type when offered the alternate option.

Table 1. Design exercise description
Fig. 1.
figure 1

Classification of the design categories for light control

Fig. 2.
figure 2

Classification of the design categories for temperature control

Fig. 3.
figure 3

Concrete design option

Fig. 4.
figure 4

Abstract design option

For the second study we designed a survey that was administrated online and on paper to a population in Silicon Valley, California. The participants were asked to select their preference between two design options: Abstract and Concrete (Figs. 3 and 4). 130 people took this survey. The demographics were 48 % (58) female and 52 % (63) male. The age range of the participants was as follows: 44 % 18 to 25 years old, 22 % 25 to 35 years old, 7 % 36 to 45 years old, 11 % 46 to 55 years old, 14 % 56 to 65 years old, and 2 % over 65 years old. The level of education was 6 % high school, 63 % college, and 31 % graduate school.

3 Results

The results on the replication of the first experiment were consistent with the previous study [10]. 73 % of the participants’ designs were classified as abstract (compared to 71 % in the previous study) and 27 % concrete (29 % in the previous study). When the participants were asked to choose between abstract and concrete designs, 42 % preferred concrete designs versus the 58 % who chose abstract design. This illustrates that some participants preferred the concrete option when offered the two design options, despite the fact that their own design was considered to be abstract.

The results of the second survey asking general population preference between the two design options (Figs. 3 and 4) are shown in Charts 1, 2 and 3. The results show that 56 % of participants preferred an abstract design while 44 % preferred a concrete design (Chart 1).

Chart 1.
figure 5

Preference of the general population between abstract and concrete design options

Chart 2.
figure 6

Design preference by gender (Color figure online)

Chart 3.
figure 7

Design preference by age group (Color figure online)

We also tried to understand if the preference might change across gender and age groups. Chart 2 shows the design preference based on gender. 53 % of female participants prefer the abstract design compared to 59 % of the male participants (Chart 2).

47 % of female participants and 41 % of male participants prefer the concrete design (Chart 2). This illustrates that overall, female participants are more likely than men to prefer concrete designs.

The design preferences differ when we analyze the data based on age groups. Interestingly, the design preference for concrete design rises in correlation with participant age. The preference for concrete design among the 18–25 year age group is 38 %, while it is 29 % for ages 25 to 36, and 60 % for those over 36 years (Chart 3).

4 Discussion

The results of this study suggest the preference among the general population (male and female participants of all age groups) is for abstract UI design (59 % versus 47 %). However, female participants tend to prefer concrete design more than male participants (47 % female versus 41 % of the male participants).

The preference for concrete designs seems to rise with age, as older adult prefer the concrete design.

It is hard to draw more general conclusions due to the limited sample size, the fact that the participants were choosing the design on static page (screens shots) and a real application web application. Having said this, the study does seem to show a trend. For example, should one extrapolate Piaget’s theory of stages of intelligence: sensorimotor, preoperational, concrete operational and formal operational to this finding one might think that the accessibility of the concrete object might be more attainable to older adults who are utilizing more concrete thinking [11]. Again, the data set is too limited to expand in that direction.

Further studies should investigate children of younger ages and older adults to see if the results would support the results of this study.

This result would be applicable when designing user interfaces for and applications that target older adults.