Creative Education
2012. Vol.3, No.1, 126-133
Published Online February 2012 in SciRes (
Copyright © 2012 SciR e s .
Trees and Nests: A Comparison between Two Hierarchical
Metaphors in Educational Applications
Adi Katz
Department of I nd us trial Engineering and Management, Shamoon College of Engineering, Ashdod, Israel
Received December 2nd, 2011; revised January 6th, 2012; accepted January 17th, 2012
The aim is to determine whether one of two hierarchical metaphors, the tree (parent-child) or the nested
(object-container), is more suitable for designing educational interfaces for children. To cope with this is-
sue an experimental educational application was designed with a prototype for each hierarchical metaphor.
The application was evaluated in a laboratory experiment, where children participants interacted with the
prototypes to find answers for questions that require searching for information. Task performance was
measured in terms of effectiveness, efficiency and subjective aspects such as user perception of ease of
use and user preference. The nested (object-container) metaphor was found to be preferred by users and
superior in several objective parameters of performance efficiency, but no significant differences were
found in the perceived ease of use and in the performance effectiveness. Implications for designing edu-
cational applications are discussed.
Keywords: Human Computer Interactions; Metaphoric Interfaces; Hierarchical Information Organization;
Educational Applications; Usability; Children Interfaces Design
Today, more than ever, children are exposed to computers in
their early years of development and have easy access to com-
puters and the Internet. Children use the Internet for school-
work, playing games and communicating with each other, and
are typically involved in activities that require searching and
browsing for information (Hutchinson et al., 2006). Creating
interfaces for young children presents particular challenges, and
the designers of such applications must take into consideration
that children are different than adults in the way they think and
learn. An important requirement for an effective learning proc-
ess is to foster a positive attitude and affect that nourish moti-
vation, arouse curiosity, engage creativity, and turn the brain
into an effective learning organism (Norman, 2004). It is most
important that applications for children are provided with suit-
able interfaces that make both their learning process and their
interaction with the applicatio n easy and enjoyable.
A successful user interface paradigm is to design the hu-
man-computer interaction based on metaphors already familiar
from real life objects, actions and situations. A metaphoric
system simply replicates the real life objects and portrays them
in the computerized environment, so that the user can intui-
tively manipulate objects and work in a familiar manner. A
suitable metaphor enables the user to function effectively in the
new system, and helps overcome cognitive limitations in com-
plex tasks (Te’eni et al., 2007). One of the implications of the
cognitive load theory (CLT) in the area of instructional design
is that the layout should be intuitive, so that the mind’s capacity
for processing information will not be overloaded and user
activities will be focused on the concepts to be learned (Sorden,
2005). Instead of learning the system and its features, meta-
phorical designs free some of the users’ limited cognitive re-
courses so they can be fully engaged and devoted to the data
and the activities at hand. When cognitive resources are unnec-
essarily devoted to mental activities such as retrieving from
memory and learning new objects and actions, it may be at the
expense of performance. Metaphors have a particularly impor-
tant function in interfaces for young children (Ellis & Blashki,
2001; Gilutz & Black, 2010), are highly preferred by them and
can induce curiosity and pleasure (De Angeli et al., 2006). A
useful metaphor must be suitable for the user population (Katz
& Vaserman, 2009), and therefore, when designing for young
users, the chosen metaphor must be borrowed from their eve-
ryday environment or conceptual world.
In order to cope with the overwhelming diversity of objects
and properties in the world, people mentally group objects,
treating them as instances of categories instead of as unique
individuals. Hierarchical organization is one type of a natural
evolution of a classification system in which people notice both
distinctions and similarities among objects, and organize cate-
gories into hierarchies in which more specific classes are in-
cluded in more general ones (Markman, 1989). In other words,
many natural categories are hierarchically organized into nested
class-inclusion relations, where some classes are super-ordinate
or subordinate to others. For example: organisms/plants/flow-
ers/Composite family/daisies/erbera daisies. Collins and Quil-
lian’s (1969) Hierarchical Network Model of semantic memory
states that concepts are stored and represented as nodes within a
hierarchical structure in our long-term memory (LTM), with
meaningful associations between concepts. The “Is A” link is
the most common link in this semantic network model.
Browsing for information is a methodical activity with a spe-
cific goal, which requires progressive filtering of results based
on visual scanning and searching. Since hierarchical organiza-
tion stems from greater accumulation of knowledge and ex-
perience (Markman, 1989), children have a difficulty dealing
with hierarchies when searching and browsing. Children do not
always navigate efficiently between categories, they may not
think hierarchically, and may have a difficulty understanding
abstract, top-level categories (Hutchinson et al., 2006). Meta-
phors may be a useful design paradigm to make it easier for
children to cope with browsing through hierarchical levels of
information. A metaphor based on a hierarchical organization
that children are already familiar with, actually utilizes knowl-
edge that has already been accumulated and this, as previously
described, decreases the cognitive overload. Clearly, the chosen
metaphor should also be appropriate for presenting the needed
relationships between categories. For example, an application
teaching organism hierarchies should present “Is-a” relation-
ships (A grizzly bear Is-a bear and a bear Is-a land mammal).
A well-known system for organization and management of
files and folders on our computing devices embodies two hier-
archical metaphoric-based interfaces; one is the file manager
interface, which reflects the tree metaphor, and the other is the
desktop interface, which is based on a nested object-container
metaphor. The tree metaphor layout that originates from the
very first instantiations of computer desktop environments, has
come out so pervasive that is firmly embedded into our intui-
tive ways of dealing with “items”, be it images, videos or gen-
erally speaking any type of information that can be defined as
“files” (Villegas, 2006). Users navigate by clicking on little
structural elements (plus and minus symbols) that open and
close various directories.
The nested (object-container) metaphor for hierarchical or-
ganization expresses the relationship of “object-within-object”
and allows recursion to be visualized in an intuitively nested
“Russian doll” fashion. The nested Russian doll is widely used
in recursive problem solving and in programming and mathe-
matics education to explain recursive models (Schiemenz,
2002). The most obvious example of the nested doll principle in
computers is Microsoft Windows, where “windows” is simply
windows inside windows inside windows.
The current research follows a previous study that compared
the tree and the object-container metaphors in the forms of a
family tree versus a cabinet of drawers—two alternative meta-
phoric designs to convey hierarchical learning of material. In
the previous study no superiority of one metaphor over the
other in the design of children’s interfaces was found (Katz and
Vaserman, 2009). Despite these results, we were still certain
that a cabinet metaphor was more suitable for children since it
inherits affordances that are the actions of organizing their be-
longings (toys, books, candy, etc.) in their natural environment
at playrooms in their homes and kindergartens. Therefore the
expectation was that this metaphor would be more intuitive and
easy to operate, more usable and more preferable for young
children than the tree metaphor. We believe that the previous
results were obtained due to specific design elements of the
metaphoric interface that caused an unwanted difference in
usability between the tree and the cabinet designs. In addition it
should be noted that the previous research was based on a small
sample of only 10 subjects. In this research we decided to re-
design the metaphors and test our hypotheses on a larger sam-
ple of subjects.
Problem Definition and Solution
Young children need encouragement, support and help in
learning about the world and developing cognitively and emo-
tionally. Pre-school children are at a very important stage of
development in life, in which they are exposed to activities of
the early learning of reading and writing. At this phase, most
children are constantly involved in confident exploring behav-
iors, interacting with their environment in an active way and
expanding their cognitive abilities on the basis of their own
activities. We believe that educational tools such as learning
applications for children should serve as a vehicle to promote
learning, curiosity, exploratory behavior and independence in
the learning process, and serve to flourish their knowledge. In
addition, successive interactions and positive learning experi-
ences at this stage in childhood can develop a high level of self
confidence, esteem and efficacy. A sense of self efficacy has a
positive influence on the success of the learning process, be-
cause previous successful and positive experiences develop
expectations to succeed in similar tasks (Bandura, 1997). Self
efficacy developing while interacting with an application used
for learning gives children the confidence to continue learning
with the same tool, and motivates them to explore additional
tools and features.
Creating interfaces for young children presents particular
challenges. Children have different cognitive capabilities than
adults; they think and learn differently. The variable of age en-
compasses many critical developmental differences between
childrens’ and adults’ ability to interact with technology, and
their unique characteristics play an important role in creating a
successful user experience for them. This is why usability test-
ing techniques should be applied with a user centered design
(UCD) approach, particularly when designing for children to
meet their capabilities, needs and expectations (Gilutz & Black,
The objectives for this research were to:
1) Design and implement the features of two hierarchical re-
cursive metaphor prototypes for a computer program that edu-
cates children about animals: the tree and the nested metaphors.
2) Compare the two prototypes by conducting an experiment
with pre-defined searching tasks, using subjective quantitative
measures of the perceived ease of use (perceived usability) and
user preference, as well as objective and quantitative measures
of effective and efficient performance. In addition, gather qua-
litative data by directly observing the child-application interact-
The Experiment
The two hierarchical metaphors, tree and nested, were tested
in a controlled experiment, in which children interacted with
two metaphoric-based educational systems to answer questions
regarding animals. The systems were aimed at the age group
between four and six since most children at this age range are
still unfamiliar with the file manager (tree) and nested-windows
previously mentioned, and are therefore not influenced by con-
ventional tree and nested metaphoric designs. This enabled
testing the suitability of the metaphors without excessive noise.
The metaphor design, tree versus closet, was manipulated to
result in a natural tree design for the tree metaphor and a play-
room closet design for the nested metaphor.
The Dependent Variables
The subjects’ performance was observed in terms of efficiency
and effectiveness, user usability perceptions, and preference.
Table 1 presents the variables and how they were measured.
Copyright © 2012 SciRe s . 127
Table 1.
The dependent varia bles and their measures.
Dependent variables Measure
Effectiveness Overall number of correct answers
Overall task time
Total number of clicks
Total number of false clicks
Perceived ease of use
(usability) “How easy was it to search for answers in the
application?” (4-point Likert scale)
Preference “Which application wo uld you prefer having a t
home: the tree or the closet?”
The Independent and the Control Variables
The independent variable was the metaphoric design of the
interface: tree versus closet. The two designs are described in
detail in Section “Apparatus—Metaphoric Interface Design”.
The controlled variables were the participant, the tasks and
the order of task performance. The participants were thirty kin-
dergarten children (twelve girls and eighteen boys), all within
the target age group of four to six years old. The participants
were classified as novice users in terms of Shneiderman’s ge-
neric classification of users (Shneiderman, 1987) since they did
not possess significant semantic knowledge of objects and ac-
tions in computers such as the hierarchical organization of files
and folders, and lacked the syntactic knowledge of the meta-
phoric applications designed for this research. All participants
interacted with both metaphoric educational applications at the
Each child had to carry out a series of search tasks to enable
answering a set of questions, such as: “what is produced from
tuna fish liver?”, and “how does a gorilla react to threat?” Overall
there were fourteen questions regarding animals, with seven
different tasks for each application. Two different but structur-
ally and cognitively similar sets of tasks were created, with cor-
responding questions in each application. For example, a ques-
tion appearing in the tree metaphor was: “what can be made of
goose-feathers” and the corresponding question in the closet
metaphor was: “what is produced from tuna fish liver”.
In order to avoid possible bias due to the order of exposure of
the subjects to each metaphor, the children were randomly di-
vided into two groups of 15 children each, so that one group
interacted with the tree prototype first and continued with the
closet, and the second interacted in the opposite order (hereafter,
Tf and Cf will be used as abbreviations for Tree first and Closet
first, respectively). After having conducted seven search tasks
in the first application, the applications were switched to per-
form the additional seven search tasks in the second applica-
Pilot Test
After having implemented the two prototypes to meet the de-
sign criteria of the two hierarchical metaphors, both applica-
tions were pilot tested on six children as representatives of the
potential users, to obtain feedback regarding the design, to
identify potential usability problems, and to make sure that the
experiment was not too difficult. The pilot participants were
three girls and three boys between the ages 4.5 - 6. Each used
both prototypes where three started with the tree metaphor and
the other three started with the closet metaphor. The pilot was
conducted so as to follow the user centered design (UCD) ap-
proach, to ensure the prototypes were compatible with the end-
users’ characteristics , such as cognitive development, cognitive
limitations, preferences and other factors that may have af-
fected the understandability and usability of the applications.
The pilot participants’ interaction with the applications was
observed and the design was adjusted accordingly. UCD is
extremely important when designing for children, and is widely
used in the HCI literature in various variations such as coopera-
tive inquiry and participatory and informant design (Baek &
Lee, 2003; Druin, 2005; Mazzone et al., 2008). The pilot re-
sulted in some improvements in design features that were not
easy or intuitive or caused confusion, and in several experi-
mental changes.
Apparatus—Metaphoric Interface Design
The research was conducted in Israel, and therefore the in-
terfaces’ language was Hebrew. The two educational programs
were identical in the type and amount of information, in the
hierarchical categorization and levels (with “Is-a” relations be-
tween categories), and in icons and labels representing object
Figure 1 presents the opening screens of the hierarchical met a-
phors: the tree metaphor on the left side, and the closet meta-
phor on the right. Figure 2 presents two compatible screens of
the tree and the closet metaphors. The hierarchical classes that
are displayed in both screens represent the following hierarchi-
cal levels: Mammals/Land Mammals/Ungulates/Horse; Deer;
Cow; Donkey.
As aforementioned, specific design elements of the metaphoric
interface have caused an unwanted difference in usability be-
tween the prototypes in the previous study comparing between
the tree and cabinet metaphors (Katz & Vaserman, 2009). The
main problem was that the cabinet metaphor was inferior in
usability characteristics not directly derived from the metaphor,
such as a relative lack of information regarding the hierarchy in
the visualization of hierarchical organization, and the use of a
confusing navigational aid. Drawing conclusions from that
research, great emphasis was placed on a unified transitioning
between various hierarchical levels in both prototypes. This
goal was accomplished by requiring an equal number of clicks
to move from one level of the hierarchy to another, and design-
ing each prototype to manifest the complete hierarchical path
for each choice of the user.
Figure 1.
The opening screens of each metaphoric interface.
Copyright © 2012 SciRe s .
In the tree metaphor (Figure 2(a)), the four main (general)
categories were displayed at the bottom part of the screen,
closest to the tree’s roots, and the highest branch displayed the
most specific (leaf level) categories. This is obviously a natural
way to present a physical tree that has a “bottom-up” develop-
ment growing upwards from the roots in the ground. An alter-
native metaphoric design that utilizes the tree metaphor is the
family tree (Katz & Vaserman, 2009; Park & Park, 2010), but it
is typically drawn “top-down”, presenting ancestors above their
descendants. Note that both versions of the tree metaphor con-
ceptually convey a “top-down” relation, in which categories
start with the most general class, until reaching the lowest level
of elementary subclass. The claim is that while adults distin-
guish between “top-down” and “bottom-up” in many areas that
differentiate levels of abstraction, children are not aware of this
“vertical flip” of real-life trees and hierarchical category de-
velopment and therefore will find it easier to handle a hierar-
chical tree metaphor organizing information in a manner that
imitates the natural growth of trees.
The tree metaphor endows the affordance of following paths
from the root, and from one branch to another. Paths are re-
vealed and concealed by unfolding and folding tree branches,
respectively, by clicking on categorical nodes. Clicking on a
node where the continuous branch seems hidden unfolds it
above that node (in other words, expands the tree upwards),
while clicking on an internal node folds the continuous branch
back by hiding it. When clicking on another node, at the same
or at an upper level in the hierarchy, the current node with an
open continuous branch folds back to the level of the clicked
node, and a new continuous branch unfolds at the clicked node.
The reason for folding a branch when the user wishes to unfold
another is to prevent a cognitive load that is likely to occur
when too much and irrelevant information is exposed. We used
navigation aids to highlight the hierarchical path, coloring the
chosen nodes (categories) light blue, while leaving the other
nodes (categories not on the path) in light green, and also by
presenting the branches connecting chosen nodes in sharp col-
ors, while the other branches in faded colors.
Following the idea of Katz and Vaserman (2009) of meta-
phoric designs for hierarchical organization for children, the
cabinet metaphor used in the previous study was redesigned to
a different variation of an object-container metaphor, which
was named the closet metaphor (Figure 2(b)). The closet meta-
phor matches the typical organization of objects (such as books
and toys) in children’s playrooms. While the Katz and Vaser-
man (2009) cabinet metaphor strictly followed the recursive
“Russian doll” fashion relations of doll within doll and so forth ,
by a cabinet of drawer within drawer and so forth, the newer
closet metaphor contains different objects at different levels of
the hierarchy, that can be placed one inside the other. The four
main categories (see Figure 2(b)) are displayed as the closet’s
four doors. Opening each door reveals its nested drawers that
contain boxes that contain animal cards representing the most
specific (leaf level) categories. We believe that the current
nested variation of the closet metaphor is more interesting and
fun than the previous, and is a better imitation of a playroom
closet that children are familiar with. Familiarity is an impor-
tant design factor that has a strong impact on children: the more
familiar is the interface metaphor, the better the comprehension
achieved (Gilutz & Black, 2010).
In Figure 2, one door—the mammals—is open to reveal two
drawers, land ma mmals and sea mammal s. The land mammals’
drawer is also open to reveal four boxes nested inside it: carni-
vores, ungulates, rodents and haplorhini. The ungulates’ box is
open and exposes four cards for the following animals: horse,
deer, cow and donkey. The closet-doors-drawers-boxes-cards
relationships are in a nested “Russian doll” fashion: the biggest
doll opens to reveal the next biggest doll, and so on until the
tiniest doll is revealed, nested in all dolls together. Opening
doors, drawers, and boxes (closed categories) is accomplished
by intuitively clicking on their handles, and closing these ob-
jects is accomplished by clicking on them again. Clicking on an
object that is closed (door, drawer or box) opens it and closes
the object that was open so far. For example, in the current path
presented in Figure 2, if one clicks on the sea mammal’s
drawer, it opens to reveal its boxes, and at the same time the
land mammal’s drawer closes. A click on the fish door reveals
its drawers, and causes the mammal door to close. The closet
metaphor endows a set of affordances that are extremely intui-
tive for young children: searching and revealing by opening and
closing doors, drawers, and boxes.
In both designs, clicking on last level nodes (leaf level cate-
gories) opens an information window displaying an educational
description of animal items, e.g. information about the deer or
donkey. The children were required to access the information
windows relevant to each task, to answer correctly. The number
of levels and categories at each level of the hierarchy is ad-
justed according to a recommendation referring to 5 ± 2 catego-
ries in children’s applications (Baek & Lee, 2003).
Figure 2.
screen shots of two compatible hierarchical paths for the two metaphors.
(a) Tree metaphor ; (b) Closet metaphor.
Copyright © 2012 SciRe s . 129
Copyright © 2012 SciRe s .
Children need positive reinforcement and praise, which help
them to experiment successfully (Yahaya & Salam, 2009). To
motivate the participants’ involvement and arousal, we used
positive reinforcement in the form of a glass candy jar that is
gradually filled with a different candy for each task. Each time
a child chooses the right question; a candy appears at the upper
side of the screen and slowly falls down into the jar while
playing a funny sound of applause and cries of glee (“correct!”,
“well done!”). Animation and sound effects are positive ele-
ments in designing for children (Gilutz & Nielsen, 2007).
Whenever a wrong answer is chosen, a frowning “smiley” ap-
pears in an animated entrance with an audio encouraging the
child to try again. Audit output is crucial in interfaces designed
for young children, to over come the lack of a reading ability.
All text (displayed on buttons, information pages, instructions,
etc.) is presented with punctuations as customary in Hebrew
texts designated for young children, and is followed by a hu-
man voice that reads the text.
focus on gestures, facial cues and voice intonations. In addition
the “think aloud” technique was used, by encouraging the chil-
dren to speak out their thoughts, problems, indecisions and
impressions. The information gathered from these observations
added rich and useful information on the comprehensibility of
the designs and user experience, as later discussed.
Experimental Results
We used T tests for paired (dependent) samples to test per-
formance differences between the two prototypes: the tree ver-
sus the closet metaphor. The results are presented in Table 2. It
can be seen that there is significant difference in efficiency in
terms of overall task time, total number of false clicks, and total
number of clicks, in favor of the Closet metaphor application;
that is, it took considerably less time to answer the task ques-
tions, there were less false clicks and a smaller total number of
clicks when using the Closet application. However, there is no
significant difference in effectiveness in terms of the number of
correct answers. Note that almost all children answered all the
questions correctly with both applications.
Conducting the Experiment
Each of the children participants interacted separately with
both metaphors at a computer station in a quiet room at the
kindergarten, under our supervision. As said, 15 children were
randomly allocated to a group that interacted with the tree ap-
plication first and the closet application second, and the other
15 children interacted with the closet application first and the
tree application second. At the beginning of the interaction with
each application, the children received a short tutorial about it,
after which they practiced the use of the application by per-
forming two preliminary search tasks prior to the actual ex-
periment. The experiment itself was composed of fourteen tasks,
seven in each application. After having conducted seven search
tasks in the first application, the child performed the additional
seven search tasks in the second application.
A Wilcoxon signed ranks test was used to test the difference
in user perceptions of the ease of use of the metaphors (per-
ceived usability). The results indicate that there is no statistical
difference in the perception of ease of use between the tree
(Median = 2) and the closet (Median = 1) metaphors (Z =
–0.808, p = 0.419). For both prototypes, most users stated that
it was very easy or easy to use (26 and 28 for the tree and closet
prototypes, respectively) Figure 3 presents user rankings of
ease of use for each metaphor.
4- very ha rd3- hard2- easy1- very ea sy
Tr ee
During the experimental phase, the applications’ logs re-
corded the dependent variable measures of performance and at
the end of the interaction with each application, a question re-
flecting the child’s perceived ease of use of the application
appeared on the screen (“How easy was it to search for answers
in the application?”; the response was recorded on a 4 point
Likert scale). At the end of the experiment, after interacting
with both applications, a question regarding the child's applica-
tion preference appeared (“Which application would you prefer
to have at home: the tree or the closet?”).
During the experimental phase qualitative data was gathered
by direct observation of the child-application interaction, with a Figure 3.
Perceived ease of use of each metaphor.
Table 2.
Performance differences between the two interfaces.
Paired diffe rences
Performance Mean Std. D t df Sig. 2-tailed
Tree 11:43 3.11
overall task
time Closet 10:41 2.57 –2.742 29 .010
Tree 12.63 8.12
# of false
clicks Closet 9.13 6.14 –3.249 29 .003
Tree 42.77 9.66
total # of
clicks Closet 39.46 7.77 –2.452 29 .020
Tree 6.60 0.62
Effectiveness total # of
correct answers Closet 6.77 0.50 1.153 29 0.258
Of 30 children participants, 12 (40%) preferred the tree pro-
totype and 18 (60%) preferred the closet. Six children said they
preferred the closet because it was easier to use for finding
information, 7 children said it was more fun to use, 3 said it
was prettier and 5 subjects said they preferred the closet be-
cause they had a similar closet at home (some gave more than
one explanation for their choice). Only one child said she pre-
ferred the tree because it was easier to use for finding informa-
tion, 2 children said it was more fun to use, and 2 said it was
prettier. Two subjects said they preferred the tree metaphor
because they liked trees and nature. These preference explana-
tions show that the metaphor evaluations were not only prag-
matic but also hedonic in nature. This point is elaborated in the
The observations revealed that all children found interest in
both applications, and their visible behavior and verbal expres-
sions showed they were positively and happily engaged with
them. There were no signs of discomfort, frustration or dissat-
isfaction, on the contrary: the children were highly motivated to
find answers for the questions using both the tree and the closet
metaphors. The conclusion is that both metaphors were com-
prehensible and satisfying.
We were interested in determining whether one of two hier-
archical metaphors, the tree or the closet was more suitable for
the design of children’s interfaces.
The results regarding interface efficiency, i.e., overall time to
perform tasks, number of false clicks, and total number of
clicks, show significant differences in favour of the closet
metaphor interface. The overall task completion time and total
number of clicks are related to Nielsen’s “efficiency” dimen-
sion of usability, and the number of false clicks is related to the
“few errors” dimension (1993). Therefore we can conclude that
in efficiency performance parameters (expressing the resources
consumed in the processes of achieving the task), the closet
metaphor is more usable than the tree metaphor. However, in
terms of performance effectiveness (i.e., the success of the task
in terms of the number of correct answers to the questions),
there was no difference between the two metaphoric designs.
Based on the results of T tests for independent samples to
test the differences between the performances of the groups in
the first and second exposure to the prototype, we can conclude
that the order of exposure to the prototypes does not matter.
Referring to the test of user perception of the interfaces’ ease
of use, the results do not show a significant difference: both
metaphorical interfaces were evaluated as easy or very easy to
use. We expected the closet metaphor to be perceived as more
usable for young children than the tree metaphor since it re-
sembles the way they organize their belongings in their play-
rooms and therefore would be more intuitive. Overall, the re-
sults show that both metaphors are perceived usable for kin-
dergarten children. Additional data, obtained from observations
of the children while they interacted with the applications,
showed that both metaphors were comprehensible, and there
was no significant difference in the difficulty expressed in the
action of transitioning between hierarchical levels.
The closet metaphor was designed to resemble any closet in
the natural environment of children’s playrooms. Choosing a
metaphor that matches between the system and the real world
of the user population is an important usability heuristic pro-
posed by Nielsen (1994). The system should speak the user’s
language, with words and concepts that are familiar to the user,
and should follow real-world conventions, presenting informa-
tion in a natural and logical order. Most children have indeed
preferred the closet metaphor. We believe that the current
nested variation of the closet metaphor is more interesting and
fun than the previous cabinet variation (Katz & Vaserman,
2009), and is a better imitation of a playroom closet that chil-
dren are familiar with. On the other hand, the tree metaphor
seemed less familiar and was also more monotonous; using the
same objects at each hierarchical level (nodes on paths that are
made of branches). We refer to more heterogeneous design
ideas for the tree metaphor later.
In summary, the results of the current study show superiority
of the closet metaphor in user preference and in several objec-
tive parameters of performance efficiency, but not in the per-
ceived usability or in performance effectiveness. Both meta-
phoric designs were effective in achieving the users’ goals, and
were similarly perceived as easy to use, but the closet metaphor
stimulated a better user experience, leading the children to pre-
fer it over the tree metaphor. User experience (UX) emphasizes
the importance of understanding how the user feels about the
system, in opposed to considering only the mere usability, with
its emphasis on user performance. Although the main focus was
to test the suitability of the metaphors in terms of usability, the
results are in line with the shift from a cognitive and functional
focus on computer applications towards an experiential affec-
tive view (Tractinsky et al., 2000; Dillon, 2001; De Angeli et
al., 2002; Hassenzahl, 2003), and show that UX is extremely
important . There are two distinct design objectives in HCI: one
is optimizing user performance and the other is optimizing user
satisfaction in achieving both pragmatic and hedonic goals
(Bevan, 2009). We claim that UX is most essential in systems
designed to foster a positive attitude in the learning process of
young children. De Angeli et al. (2006) found that participants
preferred an interface they evaluated as more engaging and fun,
despite an acknowledged inferior usability. We agree that while
it is indeed important to optimize user performance and to ful-
fill the pragmatic goals of the user, it is most important to opti-
mize user hedonic goals as well.
Referring to the results that show superiority of the closet
metaphor in efficiency but not in effectiveness, and also the
result of superiority in user preference for the closet, we argue
that while it is indeed important that the user accomplishes the
end goal using a system, namely performance effectiveness (i.e.,
the quality of the task solution), the system should also address
the process, namely performance efficiency (i.e., the resources
consumed in the process of achieving the goal). The effi-
ciency-effectiveness distinction is well known and respectively
matches the process-outcome levels of user experience with
technology (see for example Dillon, 2001). Yet, many studies
measure only objective aspects of the process, overlooking the
affective and emotional aspects and more attention should be
given to user choice and preference. Dillon suggests the POA
approach for viewing user experience, which introduces three
levels of user interaction: Process (action), Outcome (result),
and Affect (emotion), which deal respectively with what the
user does, what he attains and how he feels (Dillon, 2001). It is
reasonable to argue that users’ evaluations of the system are
influenced by the process of pursuing their goal, in terms of
Copyright © 2012 SciRe s . 131
their enjoyment of the task and not only in terms of the time
consumed or the level of task accomplished.
As described, this study is a follow up to a previous study that
compared the tree and the nested metaphors (Katz & Vaser man,
2009), but is not an exact replication of the comparison, espe-
cially because the sample size, age range and also the meta-
phoric designs have changed. Therefore we can not compare
the results of the two researches and claim that one metaphor is
more appropriate than the other for a given age range. A follow
up study will be conducted to compare the current metaphoric
designs of the natural tree versus the closet with children in the
age range of the previous study (ages 7 - 11).
The study was limited to the comparison between two spe-
cific metaphoric designs. Although the closet metaphoric de-
sign was found to be superior in several important parameters,
one should not conclude from the results that the tree metaphor
is inappropriate. We stress that it is important not only to choose
a good metaphor, but also to properly design its characteristics.
The tree metaphor may be more successful if designed differ-
ently, and more creatively, perhaps with special interactive
effects that are playful or humoristic (for example see De An-
geli et al., 2006) or in the form of an online game that makes
the learning experience more fun (see Park & Park, 2010). Also,
a tree metaphor that contains different objects at different hier-
archy levels (such as branches, leafs, flowers and fruits) or
alternatively having a gradual change of branch thickness (from
the thickest to thinnest branches, imitating a real-world tree
growth) may be more interesting and fun variations of the cur-
rent. In addition, the tree metaphor may be more suitable than
the closet for other tasks or for children at ages that we did not
We conclude that applying metaphors that are familiar to
children from their real life environments is a successful user
interface paradigm. We recommend that when designing an in-
terface that can exploit alternative metaphoric ideas, one should
test and compare them, by using the user centered design (UCD)
approach to find out which design is better, which characteris-
tics and features are more easy and intuitive to use, which are
difficult, and how children react to each metaphor as whole. In
addition, creating artifacts such as user interfaces is a very crea-
tive process, and since children are known as very creative,
collaborating with them can inspire and empower adults with
new insights to generate new ideas they would otherwise never
have thought of. Working closely with children that represent
the target audience as design partners in the application devel-
opment process from the earliest stage of the design is a good
strategy to create interfaces that are particularly suitable for
them. Involving children in the design process, conducting user
testing techniques that combine objective usability measures
with subjective measures (such as user satisfaction, attitudes,
feelings and preferences) and collecting qualitative data from
observations are highly recommended as ways for optimizing
both pragmatic and hedonic goals of the user.
We than k Ofir Marco and R avid Kahalani for their contribu-
tions in designing the metaphors, planning of the experimental
tasks, and running the pilot and the laboratory experiment.
Bandura, A. (1997). Self-efficacy: The exercise of control. New York:
W.H. Freeman.
Baek, J. S., & Lee, K. P. (2003). A study of cognitive characteristics of
children’s information architecture using participatory design tech-
nique. Research Papers of Ergonomics for Children and Educational
Environment, IEA Technical Committee.
Bevan, N. (2009). What is the difference between the purpose of us-
ability and user experience evaluation methods? UXEM’09 Work-
shop, INTERACT 2009, Uppsala.
Collins A. M., & Quillian M. R., (1969). Retrieval time from semantic
memory. Journal of Verbal Learning and Verbal Behavior, 8, 240-
248. doi:10.1016/S0022-5371(69)80069-1
De Angeli, A., Lynch, P., & Johnson, G. I. (2002) Pleasure versus
efficiency in user interfaces: Towards an involvement framework. In
W. S. Green, & P. W. Jord an (Eds.) , Pleasure with products: Beyond
usability (pp. 97-111). London: Taylor & Francis.
De Angeli, A., Sutcliffe A., & Hartmann, J. (2006). Interaction, usabil-
ity and aesthetics: What influences users’ preferences? Proceedings
of the 6th Conference on Designing Interactive systems (DIS’06),
New York, 271-280.
Dillon, A. (2001). Beyond usability: Process, outcome and affect in hu-
man-computer interactions. Canadian Journal of Library and Infor-
mation Science, 26, 57-69. doi:10.1086/428691
Druin, A. (2005). What children can teach us: Developing digital li-
braries for children. Library Quarterly, 75, 20-41.
Ellis, K., & Blashki. K. (2001). Producing an artifact as research: Mul-
timedia for young children, 2001. Proceedings of the National Advi-
sory Committee on Computing Qualifications, Napier, 259-264.
Gilutz, S., & Black, J. B. (2010) Child and design factors interacting in
children’s HCI helping children focus on the content, not the inter-
face. Designing for Children, 2-6 February 2010, Mumbai.
Gilutz, S., & Nielsen, J. (2007). Usability of websites for children. USA:
Nielsen Norman Group. URL (last checked 3 October 2011).
Hassenzahl, M. (2003). The thing and I: Understanding the relationship
between user and product. In M. Blythe, C. Overbeeke, A. Monk, &
P. C. Wright (Eds.), Funology: From usability to enjoyment (pp. 31-
42). Dordrecht: Kluwer.
Hutchinson, B. H., Bederson, B. B., & Druin, A. (2006). The evolution
of the international children’s digital library searching and browsing
interface. Proceedings of the 2006 Conference on Interaction Design
and Children (IDC’06) , (pp. 105-112). New York, NY: ACM.
Katz, A., & Vaserman, D. (2009). A case study of two hierarchical
recursive metaphors for information organization: The closet meta-
phor vs. the tree metaphor. Computer Modelling and New Technolo-
gies, 13, 7-17.
Markman, E. M. (1989). Categorization and naming in children: Prob-
lems of induction. Cambridge, MA: The MIT Press.
Mazzone, E., Read, C. J., Beale, R., Abuelmaatti, O., & England, D.
(2008). Understanding children’s contributions during informant de-
sign. Proceedings of the 22nd British HCI Group Annual Conference
(HCI 2008), Li ve rpoo l, 1-5 September 2008, 61-64.
Nielsen, J. (1993). Usability engineering. Boston: Academic Press.
Nielsen, J. (1994). Ten usability heuristics. URL (last checked 3 Octo-
ber 2011).
Norman, D. (2004). Emotional design: Why we love (or hate) everyday
things. New York: Basic Books.
Park, E. Y., & Park, Y. H. (2010). A hierarchical interface design of a
puzzle game for elementary education. International Journal of u-
and e-Service, Scien ce and Technology, 3, 43-50.
Schiemenz, B. (2002). Managing complexity by recursion. In R. Trappl
(Ed.), Cybernetics and Systems (pp. 475-479). Wien: Austrian Soci-
ety for Cybernetic Studies.
Shneiderman, B. (1987). Designing the user interface: Strategies for
effective human-computer interaction. Boston: Addison-Wesley, Inc.
Sorden S. (2005). A cognitive approach to instructional design for mul-
timedia learn in g . Informing Science Journal, 8, 263-279.
Te’eni, D., Carey, J., & Zhang, P., (2007). Human computer interaction :
Copyright © 2012 SciRe s .
Copyright © 2012 SciRe s . 133
Developing effective organizational information systems. Hoboken:
John Wiley & Sons, Inc.
Tractinsky, N., Shoval-Katz, A., & Ikar, D. (2000). What is beautiful is
usable. Interacti n g with Computers, 13, 127-145.
Villegas, P., Concejero, P., Pérez, S., Prieto, J., Aragón, L., & Diego, S.
(2006). Human factors issues in 3D visualization module for multi-
media collections in AceMedia. The 20th International Symposium
on Human Factors in Telecommunication, Sophia-Antipolis, 20-23
March 2006.
Yahaya, W. A. J., & Salam, S. N. A. (2009). Usability design strategies
for children: Developing children learning and knowledge in de-
creasing children dental anxiety. Proceedings of the International
Conference on Primary Education, Hong Kong, 25-27 November