taken into consideration Once again, the goal for this experiment was
to show how a role-play illustrating the functionalities and behaviors of
threads would increase the students’ understandings of the concept.
Execution
Execution was exactly the same as in Experiment 1, but with a different role-playing game to illustrate the concept of threads.
The Role-Play
The second role-play was conducted in the manner of a live
Guess-Who? [ 2] game. One person would hunt for a specific person by asking questions and eliminating students based on physical traits. One
student (call him Jeff) was chosen to be the leader of the exercise and
was told to leave the room while the game was set up.
The instructor then chose one student at random and gave that student a treasure to hide on a person. Once the student had hidden the
treasure, the instructor asked the entire class to stand up. Jeff was
asked to come back in, and was told he needed to find the student
with the treasure by process of elimination. The process was to be iterated until only one student remained:
1. Ask a question about the person’s appearance. For example, “Is the
treasure held by a student with blond hair?”
2. Go through the class, person by person, and tell each person that did not
fit the criteria to sit down and each person that did to remain standing.
Once the treasure was found, Jeff was asked to leave the room once
again. The instructor chose three people who would act as Jeff’s
threads, and they were told to join him outside the room. Again the
treasure was hidden on a random student and the entire class was asked
to stand up. When the treasure was again hidden, Jeff and his threads
were asked to re-enter the room. He was told to do the same thing,
except this time he could dispatch an idle “thread” to sort through the
class. This added the option to have multiple people working at once.
The class could now clearly see the benefits of multitasking and thus,
the benefits of having multiple threads in a computer program.
Initial Instructor Reaction
After the role-play class, Kinley explained that the students really enjoyed
the exercise. He said that the entire class was energized, awake, and
engaged in the learning of threads. The change of pace was refreshing
for the students as well as the instructor and everyone was more interested than usual. Once again, the instructor said he believed the experiment was effective because the students seemed to learn a large amount
from the exercise. Not only did it interest them and get them involved,
but the analogy of dispatching additional aid seemed to help the students
understand the purposes, benefits, and functionality of threads.
Statistical Results
Percent change (Table 4) was calculated for each student. The data
reflects the mean, median, and standard deviation of the changes, not
the change in mean, median, and standard deviations. At first glance
it seems that, on average, the experimental section improved more
(28% mean increase versus 5%), but at the end of the exercise the control section still had a slightly higher mean score and thus a deeper
understanding of threads. In both cases, the standard deviation shrunk
to 13%, which indicates that the spread of scores lessened after the
class on threads.
Control Section
Mean Accuracy
Median Score
Standard Deviation
Experiment Section
Mean Accuracy
Median Score
Standard Deviation
Pre-Quiz
78%
83%
16%
Pre-Quiz
53%
50%
23%
Post-Quiz
83%
83%
13%
Post-Quiz
82%
83%
13%
Change
5%
3%
17%
Change
28%
33%
28%
Table 4: Results of the experiment.
The experimental class, however, started with a bigger spread than
the control class and ultimately coalesced more. This would indicate
that more of the students moved closer to the mean. The median score
in the control section did not change at all, which implies that the lower
scores moved an insignificant amount, and that the scores that started
above the median remained the same as before. For the students in the
class above the median this session did not affect their knowledge of
threads. The median in changes for the control section is very low,
which indicates that most of the students had little score change, if any.
The median score in the experimental section followed the same
upward trend as the mean, which indicates that the increase happened
class-wide and not just with a select few students. The fact that the
standard deviation dropped by about 10% indicates that the benefit had
the biggest impact on people below the average, because their scores
grew much closer to the mean. Overall, it seems the experimental section made the most progress over one day with the lecture on threads.
C ontrol
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%
E xpe rimental
*
*
*
*
p r e
po st
pr e
po st
Figure 2: Threads Test Results.
This time two different instructors were teaching the classes, which
might have lead to the students in the control and experimental sections having different levels of understanding. The experimental section, however, showed a 28% mean gain in one session that may have
taken the other session two. A significant change occurred in the experimental section (Figure 2): the wide distribution shrunk to what appears
to be a line instead of a box for the post test. Also visible is the climb in
low scores. Most of the scores in the experimental section grew by more
than 25%, whereas in the control section the lower scores stayed low.
Further Analysis
There are many different types of atypical techniques. Role-playing is
just one example. Each technique targets different learning styles and
