Figure 9: Dragonmeter example.
add more details to our understanding of children’s
view of their experience with the evaluation software,
we have involved their parents in a focus group. We
asked the parents to describe what the children told
them about the activity. Three parents agreed to par-
ticipate to the focus group. All the parents reported
that the children have mentioned the fact that there
was a game with a red panda. The children have
mentioned different objects (geometric shapes, ani-
mals, fruits) and some of the tasks they had to per-
form. They were very enthusiastic about the stick-
ers and diplomas they have received after the session.
The parents have mentioned that the children were
very enthusiastic about their experience. They have
also mentioned they would like to participate in fur-
ther similar activities. Their only concern was about
the period of time the children were exposed to the
computer. Although we have ensured them that from
our experience they did not sit in front of the com-
puter more than 10 minutes, they were sceptic, as
from children enthusiasm they believed the children
would never stop playing. From all these assessment
methods, we have drawn the conclusion that children
had a great time interacting with our assessment ap-
plication. They did not mention anywhere that they
had been evaluated or tested.
4 DISCUSSION AND LESSONS
LEARNED
Designing for children is different than designing for
other stakeholders, as there are additional constraints
brought by their physical and psychological develop-
ment. In the following we will present how the usual
software design phases need adaptation such that the
clients and the final users of the system participate in
the design process. Also, implementation of function-
ality that requires no supplementary effort when the
users are adults, brings the necessity of design and im-
plementation decisions when the users are small chil-
dren who cannot read or write.
4.1 Considerations on the Software
Engineering Process
Requirements. The process of gathering information
about their characteristics and needs must be adapted.
Spending time with the children in their familiar envi-
ronment provides useful information about their inter-
ests, their skills, and their knowledge. The entire ac-
tivity of requirements gathering should be organized
as a play activity, to encourage children participation
and to help them connect with the design team mem-
bers. When the goal of the designed product is educa-
tional, further assistance from the educational experts
needs to be integrated in the process. They can pro-
vide information about the developmental stages of
children, their knowledge on a specific domain and
further educational goals. Parents are valuable stake-
holders in the design process, as they can provide their
view on children’ knowledge, interests and interac-
tion skills.
Alternative Designs. In order to evaluate the alter-
native designs, there are two possible options: cre-
ating abstract representations of the design solutions
and involving an educational expert only to provide
feedback, or implementing executable prototypes of
the designed solutions such that the preschoolers are
able to give feedback. Although it is more comfort-
able to interact with the educational experts (adults)
to identify possible interaction problems, there are
some aspects that cannot be predicted by the educa-
tion experts. Task formulation accepted by the adult
users might be misinterpreted by the children. For
example, if a task required the children to count the
number of objects on the screen, the children always
answered verbally, without interacting with the inter-
face. Such situations cannot be identified without ob-
serving a child interacting with an executable proto-
type.
Prototyping. As we have previously mentioned, the
most appropriate approach when working with such
small children is to merge the design alternatives and
prototyping step. This means that more development
effort is involved in the early project steps, but the
children can participate to a larger degree in the de-
sign process. Involving children in the process is es-
sential, as we consider that the acceptance and en-
gagement of children in interaction is determinant on
their task performance.
Evaluation. Evaluation of computer assisted assess-
ment evaluation tools requires multiple aspects to be
taken into consideration. An evaluation from the ed-
ucation experts is needed to validate the content, nav-
igation, and task sequences. The evaluation with the
children is required to provide information about their
A User Centered Approach in Designing Computer Aided Assessment Applications for Preschoolers
511