IEEE Consumer Electronics Magazine - July 2018 - 23

PROCEDURE
All participants were brought into the lab for a pretest of their
dance skills. The assessment was made by two professional
Tujia folk dance teachers with an evaluation scale on dance.
During the following four weeks, participants in the control
group were brought back to the lab twice a week for a 2-h
practice following the video. Participants in the experimental
group were brought back to the lab twice a week for a 2-h-long
practice with the PAC. The professional dance teachers then
assessed all of the participants' dance skills after corresponding interventions of the two groups. They were blind to the
practice groups of the participants.
During the intervention, the participants were asked to
pretend that they were learning the waving dance in class by
following the choreographer. The experimental process is
shown in Figure 8. The participants in the two groups were
learning in different places and at different times to avoid
mutual influence, but their experiment environments were
similar. We can see the setup for the experimental group in
Figure 9, in which the PAC is run on a computer and shown
on the screen in front of the user. The experimenter left the
room while the participants were learning the dance, and the
participants were asked to leave the room once their time was
up. Participants were told to practice in their own way but
that they should interact with all of the modes in the PAC.
Camera 1 recorded the process of the experiment, while camera 2, built into the Kinect, recognized and captured the
dance movements of players to help the PAC evaluate the
movement of the player instantly. Camera 1 recorded the dance
process with video, and camera 2 recognized movements of
the participants.

Because the device in the PAC is
handy and inexpensive, it can be
used to research and develop a
personal dance instructor.
ed into four aspects. Furthermore, the code of rating levels
on hand-waving dance performance was designed and
revised by the professional folk dance professors with
reference to the five-point Likert scale. The results of the
consistency analysis on the code turn out to be good
^kappa = 0.45, p = 0.000 1 0.05 h, indicating that the two
teachers were largely consistent in the understanding of the
code and the evaluation of the dancers.
Emotional acceptance is shown as the outcome of the satisfaction scale, which has been revised on the basis of the
online learning satisfaction scale made by Wang, combined
with the features of the game-based learning system PAC [18].
The satisfaction scale is shown in [19]. The reliability was
evaluated by the internal consistency of items in the scale representing each factor using Cronbach alpha. The 12-item satisfaction scale had a reliability (Cronbach alpha) of 0.808,
exceeding the minimum standard of 0.60, which showed good
reliability and internal consistency.
According to the single variable control principle, the
experiment tools in this study include the dance learning system, PAC, teaching video containing the same content as the
PAC, pretest and posttest of the dance performance evaluation
scale, and self-evaluation of the learning satisfaction scale.

MEASUREMENTS
The participants were pretested after grouping to make sure
their dance foundations would make no significant difference. The dance used for pretesting was different, however,
from the dance used for the formal test. The professional
teachers used the evaluation scale to grade participants' performance during the pretest. Then participants learned the
Tujia hand-waving dance in different ways following the
methods of Figure 8. In addition, the measurement of dance
performance and satisfaction was carried out using the dance
evaluation scale at posttest as well as the satisfaction scale at
the end of the practice of the two groups.
The effect of learning is mainly reflected by the dance
score measured before and after the intervention. As a traditional sport of the Tujia people, the moves of the handwaving dance come from the labor of production, daily
life, and wars. The evaluation rubric has been revised on
the basis of the dance performance evaluation rubric made
by Hsia et al. according to the features of the hand-waving
dance, which mainly focused on the choreography, technical skills, performance skills, and rhythm [16]. The revision includes the description of choreography and technical
skills. As shown in [17], the description of the code of rating levels on hand-waving dance performance can be divid-

RESULTS
In this section, we describe the different influences of the
PAC and traditional video on the waving-dance learning performance and learning satisfaction of the participants.

20
18
16
14
12
10
8
6
4
2
0

17.3

16.3
10.6

10.3

Experimental Group

Control Group

Average Score of Pretest
Average Score of Posttest
FIGURE 10. A comparison of the average scores of the learning
performance test.

july 2018

IEEE Consumer Electronics Magazine

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - July 2018