Escolar Documentos
Profissional Documentos
Cultura Documentos
48
$04.000 1994 IEEE
0-8186-6660-9/94
Badre and Lewis present an empirical study of learning an This issue is addressed by the two laboratory formats used
algorithm using animations [19]. Their study involved an in the experiment.
advanced algorithm and examined students learning about The use of animation is also becoming increasingly
the algorithm using (1) a textual description or (2) a textual common in the area of Human Computer Interaction. It
description accompanied by an animation. The animation becomes crucial to formally investigate questions of how
subjects scored slightly higher, but the difference was not animation may best be incorporated into such areas as an-
statistically significant. The authors concluded that the imated help [231, animated interfaces [l], and animated
animation alone is not sufficient to improve understanding, presentation of concepts [21. Issues of design and use of
but must be included in a more active learning environment animations hinge on the careful study of how animations
for complete understanding. Their study afforded tight may most effectively be used.
controls on the two conditions, but it did not investigate
the use of animated algorithms in a classroom setting. Our
study was designed to mimic more closely a traditional 2 Empirical Study
classroom and laboratory use of animated algorithms.
In particular, our study was conducted in order to mea- This study was designed to investigate several presen-
sure the efficacy of using animated algorithms in varied tation issues conceming the use of animations to illustrate
teaching approaches. The approaches varied in the level algorithms. The subjects of our study were students at the
of control and active involvement of the student. The ex- Georgia Institute of Technology enrolled in CS1410, the
periment was conducted in a classroom setting. Algorithm first programming course for Computer Science majors at
animations utilized were created with the XTango algorithm the Institute. The students were volunteers who received
animation package [20] and the Polka algorithm animation class credit for their participation. Sixty-two students par-
system [21]. These two systems provide similar algorithm ticipated in the experiment.
animation environments: 2-D color graphics on top of the
X Window System. Several variables were studied, in- 2.1 Design
cluding style of presentation of an animated algorithm as a
lecture accompaniment (animation or transparencies), use The experiment was a 2 x 2 (nested 2) design as rep-
of a laboratory session to clarify algorithm concepts, and resented in Table 1. One variable was presentation of the
student interaction with the animation during the laboratory lecture example, using Animation or Prepared Slides. The
session, where one group received prepared data sets and second variable was Lecture Only or Lecture plus Labo-
the second group created their own data sets. ratory Session. This design also encompassed a nested 2
level factor under laboratory session where the variable of
This experiment was an in-depth look at how animated concern was Laboratory Type, either Active or Passive.
algorithms may be used in the teaching of computer al- The algorithm used was Kruskals Minimum Spanning
gorithms. As in actual teaching of such algorithms, group Tree Algorithm. Kruskals MST Algorithm finds a set of
sessions were used. Varied conditions allowed some groups edges of a graph that form a path to all vertices of the graph
to participate in a lecture with an extra laboratory session and that are also of minimum cost or weight. The first
while other groups participated in lecture sessions only. step in the algorithm is to sort the graph edges by their
All lecture sessions were accompanied by an example of weight. The next step is to iteratively add the edge of least
the algorithm which was either a series of transparencies weight that does not form a cycle. The problem of finding
prepared in advance or the same data set illustrated by an a minimum spanning tree is commonly presented in early
animation. Laboratory sessions were of two types, either computer science courses and is often solved by using either
using prepared data sets or allowing the student to create a Kruskals algorithm or another well-known variant, Prims
series of personalized data sets. algorithm.
The question of how best to present material is an age-
old question to the pedagogue. How, indeed, may one best 2.2 Materials
transfer to others the concepts which are so clear to the
teacher, yet such unknown territory to the learner? Felder A lecture describing the algorithm was presented to all
and Silverman [9] among others, stress that students have groups. The lecture was written in advance to ensure that
many different ways of learning and that the learning and each group would receive the same information. Students
teaching styles of both student and teacher affect the re- in the LecturdAnimation groups watched on individual
sults of the teaching process. One aspect of this is the workstations an animated example of the Kruskals Mini-
activdpassive dimension which reflects whether students mum Spanning Tree Algorithm created by the Polka Algo-
learn best by experimentation or by developing theories. rithm Animation software that allows step-by-step control
49
LABORATORY
CONDITION The example questions below are selected from the free-
response test.
Lecture PaSsiveLab
9 7 1. Under what conditions would the next shortest edge
not be added to the Minimum Spanning Tree?
Table 1: Number of participants in each of the conditions
of the experiment. 2. What, in your opinion, is the key part of the algorithm
which guarantees the Spanning Tree obtained will be
minimal?
of the animation. Students in the k t d s l i d e s group were 2.3 Procedure
shown the same example graph by means of a series of pre-
pared transparencies. Tbese transparencies were created The subjects were divided into 4 groups of approxi-
from window-dumps of the Polka example. Tbe examples mately sixteen studentseach. See Table 1fortheexad num-
differed in the dynamic nature of the animation. bers. The groups were Lecturelhimation, LectdSlides,
For those students in the active and passive laboratory Lecture and Lab/Animation,Lecture and Lab/Slides. Each
groups, two prepared sheets of instructions explained how laboratory section was further divided into Active and Pas-
to acoess an XTango animation of Kruskals Minimum sive subsections, four groups. Students were randomly
Spanning Tree Algorithm. The only difference in the two assigned to a particular sub-group. All students listened
handouts was that the Active group was instructed how to to the lecture presentation of the algorithm, accompanied
interactively create a graph to use as input to the animation, by either the Polka animation or by the prepared slides.
while the Passive group was given the names of prepared For those in the laboratory condition, interaction with the
data files to use as input. A sample window from the ani- XTango animation followed. Students in the Active group
mation appears in Figure 1. created graphs and observed the workings of the algorithm
The version of the animation used in this experiment on those graphs. Students in the Passive group were given
was based on previous experiments which indicated that a list of prepared file names and asked to observe the work-
a monochrome version of the algorithm with algorithmic ings of the Kruskal MST Algorithm in the XTango envi-
steps appearing as text was best, as measwed by perfor- ronment on those files. All students were allowed twelve
mance on a post-test [ll]. This type of animation was minutes for the laboratory session. The twelve minute
used for both the lecture example and for the animation time limit was derived from a previous experiment where
labwatory. it was determined that the average time a student spent
All groups completed a multiple-choidtrue-falseon- experimenting with the graphs was twelve minutes [lll.
line test requiring application or understanding of the al- Following this, the students completed the two post-tests.
gorithm. Groups also completed a free response test on
paper that was designed to require the students to articulate
concepts relating to understanding the algorithm.
Sample questions from the on-line test appear below:
3 Hypotheses
1. In Kruskal Algorithm, the first step in finding the Min. We hypothesized that the subjects who received the lec-
imum Spanning Tree is: ture accompanied by the animated example would perform
Sort the edges by weight better than those subjects who received the slides exam-
Select the two shortest edges ple. We also hypothesized that the addition of a laboratory
Select the shortest edge from node 1 session would lead to improved performance. Accuracy,
None of the above the dependent variable, was measured on two instruments,
the on-line fixed choice test and the paper and pencil short
answer test that was designed to measure behavioral objec-
2. In the given graph, if edges HG, IC, GF, CE and AB tives based on concepts and applications of the algorithm.
are already in the path, which edge will be added next? The tests differed in approach. Tbe fixed choice test was
IC designed to concentrate on questions of procedure and the
CD small steps of the algorithm. Questions were either mul-
HI tiple choice or trudfalse. The paper test was designed to
None of the above concentrate on conceptual issues including motivation as
50
1-
I' I
I
Figure 1: A frame from the XTango animation of Kruskal's minimum spanning tree algorithm.
4 Results
51
niques, recognition and guessing, which are not easily ap-
plied in free-answer style questions.
The two factors in the first analysis of variance on test
scores were Lecture Only versus Lecture Plus Lab and 22 I I I 1
Polka A n i o n versus Repared Slides.
In a comparison of groups that received the laboratory
session, results indicated that students completing a labo-
ratory session performed marginally better than those who
2o 1 Active Lab t
Passive Lab B-
No Lab h
On-Line
I
Free Response
I
The following results are for the paper post-test requir- Figure 2: Graphical representation of cell means for the
ing statement or application of concepts. This test had three lab conditions.
seven questions, each counted as three points for a maxi-
mum high score possible of twenty-one points. The three
points were necessary because each correct answer con-
sisted of a number of components. Tbe three point scale
allowed each component to be graded. The questions were
Passive
designed to address the basic concepts necessary for un-
Active
derstanding of the algorithm, in addition to requiring a
complete demonstration of the working of the algorithm on Table 3: Cell means for the three lab conditions. number
a provided graph. correct of twenty-one, on the free-response test.
Parallel analysis of variance was performed on the free-
response test results. Results of this analysis indicated that
students who completed a laboratory session performed
significantly better on the free-response post-test (F4.36, the factor in a second analysis, we discovered cell means
d.f. 1,58, p< 0.05) than those who did not. The amount indicating that those students in the active condition had
of difference in this result indicates that student laboratory the highest scores on the free-response test. These results
participation is more effective for questions which require appear in Table 3. Analysis of Variance for the three possi-
more conceptual knowledge than questions which require ble lab conditions indicated that laboratory condition was
recognition of the individual steps of the algorithm. No significant (F = 2.83, d.f. 2,59, p< 0.07). Pairwise t-test
significant difference was discovered between groups view- were performed to determine where the difference in con-
ing the algorithm demonstration as an animation and those dition actually lay. The significant difference (p= 0.05)
viewing the algorithm as a series of prepared slides. This was discovered between the active and the no laboratory
result also seemed to indicate that the type of visual pre- condition.
sentation is not as strong a differentiating factor as the type
of reinforcing experience.
5 Discussion
4 3 Laboratory Style, Active, Passive
Our experiment was interesting in several different as-
The results described above led to further study of the pects. First, it appeared that the example used (animation
differences among the conditions based upon the type of vs. slides) did not make a significant difference in teaching
laboratory session - active, passive, or none. Using this as the algorithm. In fact, the animation in lecture group did
52
slightly worse on the free response test when no lab was Study of the results for the two laboratory conditions
involved. Even though it is valuable to have a visual aid indicated that students who are in the active condition and
to concept formation, the animation, while enjoyable to the create their own data sets for the algorithm achieve higher
student, did not provide added clarity over the transparen- scoresthan those who observe prepared data sets. The result
cies in our study. Certainly, this result could be a factor of suggests that a study on the effect of student involvement
the particular algorithm we utilized. Other algorithms may in the learning process should be conducted.
benefit more from animation.
A second aspect of interest was that the advantageof the 5.2 Implications
interactive laboratory session was confirmed. We found
that these students excelled when compared to the students With the new curriculum for Computer Science (ACM
who did not participate in a laboratory session as well as to IEFZ91) comes a focus on breadth of learning and expo-
the students who were in the passive laboratory condition. sure of the beginning computer science student to many
Of special interest is the fact that the intuitive advantage of varied concepts and areas of computer science. M i -
the laboratory group was not statisticallysupported. Simply tions may well be used to enhance this breadth of exposure
having a laboratory session was not enough to improve and to enable the student to grasp an understanding of the
performance; the issue of control and interaction also was field through an understandingof its underlying algorithmic
necessary. A possible implication of this result is that one processes.
valuable use of algorithm animations will be to make them Additionally, one emphasis of the new curriculum is to
available to the studentsoutside the classroom setting. Such include more closed laboratory sessions during the courses
availabilitymay be provided in either a closed laboratory or of the computer science major. Animated algorithms could
open laboratory setting where studentswouldcreate sample well be employed to serve as a portion of these closed
data sets and observe the workings of the algorithms on laboratory sessions and to provide enhancement and rein-
these sample data sets. Our result suggests that active forcement to lecture and textbook material.
student participation is a key issue in this design process. Strong implications exist for application of these results
A third item of interest was that while those in the ac- in general areas of Human Computer Interaction. The de-
tive laboratory performed at a slightly higher level for both sign of animated help, animated interfaces, and other uses
portions of the test, the difference was larger for the free of animations can be guided by empirical results such as
response test than for the on-line test. The nature of the two these which indicate the most effective use of animations.
tests is important in understanding this result. In general,
the questions on the on-line test required recognition of the 5 3 Future Work
correct response rather than generation of aresponse. These
questions might be described as being more on a procedu-
This study is simply a beginning step in a longer trek
ral and operational level than on a conceptual level. This
toward uncovering the true instructionalvalue of algorithm
speaks to the issue of what types of learning are most af-
animations, and identifying the most valuable visual and
fected by the use of the animations. A previous hypothesis
pedagogical techniques that optimize their utility. Many
was that these animations may aid in concept formation.
other conditions remain to be studied: level of algorithm
The results support that hypothesis.
(beginning through advanced); animation presentation is-
sues such as color, dynamism, dimensionality, and sound;
5.1 Conclusions
how the animation is used as an instructional accompani-
The results of this experiment indicate that an advantage ment, and so on. Establishing solid empirical evidence
was shown for students who interacted with the algorithm that animations can help teach algorithms is critical to their
animation in the laboratory session. The advantage was continued use. The goal of future work in this area is to
more marked for those questions that required knowledge examine these questionsfurther and either confirm or refute
at a deeper level (the free-response test). Questions on the benefits of these visual presentations.
this test required drawing conclusions from the questions
asked, as well as demonstrating a holistic version of the
algorithm. Students who received the laboratory session References
also performed better on the on-line trudfalse or multiple
choice questions, but not at significant levels. This finding [l] Ronald Baecker and Ian Small. Animation at the
indicates that the animation session is more crucial for those interface. In Brenda Laurel, editor, The Art of
conceptual questions than for the more basic operational Human-ComputerInterface Design,pages 251-267.
level of question. Addison-Wesley,Reading, MA, 1990.
53
[2] Ronald M. B&m, Ian Small, and Richard Mander. [15] Susan Palmiter and Jay Elkerton. An evaluation of
Bringing icons to life. In Proceedings of the ACM animated demonstrationsfor learningcomputer-based
SIGCHI '91 Conference on Human Factors in Com- tasks.In Proceedings of theACMSIGCHI '91 Confer-
puting Systems, pages 1 4 , New Orleans, LA, May ence on Human Factors in Computing Systems, pages
1991. 257-263, New Orleans, LA, M a y 1991.
[3] Frederick P.Brooks. No silverbullet,Essence and ac- [16] Rajeev K. Pandey and Margaret M. Burnett. Is it
cidents of softwareengineering. Computer,20(4):10- easier to write matrix manipulation programs visually
19, April 1987. or textually? an empirical study. In proceedings of the
1993 IEEE Symposium on Ksual Languages, pages
[4] Marc Brown. Exploring algorithms using Balsa-11. 344-351, Bergen, Norway, August 1993.
Computer,21(5):14-36,1988.
[171 S.K. Reed. Wects of computer graphics on improv-
[5] Marc H. Brown. Zeus: A system for algorithm ani- ing estimates to algebra word problems. Jouml of
mation and multi-view editing. In Proc. I991 IEEE Educational Psychology, 77:285-298,1985.
Workshop on Ksualhnguages. IEEE,October 1991.
[181 John Stasko. TANGO: A Eramework and system for
[6] Marc H. Brown and John Hershberger. Color and algorithm animation. Computer,23(9):39-44,1990.
sound in algorithm animation. Computer,25( 12):52-
63, December 1992. [19] John Stasko, Albert Badre, and Clayton Lewis. Do
algorithm animations assist learning? An empirical
[7] Marc H. Brown and Marc A. Najork. Animating study and analysis. In Proceedings of the INTERCHI
algorithms in 3D. In Proceedings of the UIST 1993, '93 Conferenceon Human Factors in Computing Sys-
pages 93-100, Atlanta, GA, November 1993. tems, pages 61-66, Amsterdam, Netherlands, April
1993.
[8] Edsger W. Dijkstm. On the cruelty of really teach-
ing computer science. Communicationsof the ACM, [20] John T. Stasko. Animating algorithms with
32(12):139&1404, December 1989. XTANGO. SIGACTNews,23(2):67-71, Spring 1992.
[9] Richard M. Felder and Linda K. Silverman. Learning [21] John T. Stasko and Eileen Kraemer. A methodology
and teaching styles in engineering education. Engi- for building application-specific visualizations of par-
neering Education, 78(7):674-681,1988. allel programs. Journal of hrallel and Distributed
Computing, 18(2):258-264, June 1993.
[lo] Peter A. Gloor. AACE - Algorithm animation for
computer science education. In Proceedings of the [22] John T. Stasko and Joseph E Wehrli. Three-
I992 IEEE Workshop on Msual Languages, pages dimensional computation visualization. In Proceed-
25-31, Seattle, WA, September 1992. ings of the I993 IEEE Symposium on Ksual Lan-
guages, pages 100-107, Bergen, Norway, August
[ 1 11 Andrea W. Lawrence. Empirical Studies of the lralue 1993.
of AlgorithmAnimation in Algorithm Understanding.
PhD thesis, Georgia Instituteof Technology, 1993. [23] Piyawadee Sukaviriya. Dynamic construction of ani-
mated help from application context. In Proceedings
[12] Richard E. Mayer and Richard B. Anderson. Ani- of the '88ACM SIGGRAPH Symposiumon UserInter-
mations need narrations: An experimental test of a face Sofhvare and Technology,pages 190-202, Banff,
dual-coding hypothesis. Journal of Educational Psy- Alberta, Canada, October 1988.
chology, 83(4):484-490,1991.
[13] Richard E. Mayer and Richard B. Anderson. The in-
structive animation: Helping students build connec-
tions between words and pictues in multimedia learn-
ing. Journal of Educational Psychology, 84(4):444-
442, 1992.
[14] Richard E. Mayer and Joan K. Gallini. When is an
illustration worth ten thousand words? Journal of
Educational Psychology, 82(4):715-726.1990.
54