|
This
session was organized around the Genetics Cognitive Tutor (GCT),
a computer-based teaching tool designed to promote problem-based
teaching and learning of genetics. The session had two goals: 1)
to educate participants about the cognitive tutor in genetics and
the potential advantages it has over other methods of teaching genetics
and 2) to explore the efficacy and merits of developing similar
computer based tutors in teaching other subjects.
Presentation:
Session leader
Jones began by explaining the impetus for her involvement in the
development of the GCT. First, in her experience, students do not
like taking notes and prefer technology-based methods of learning;
second, previous attempts to incorporate genetics software in her
lessons have failed because of the lack of top quality computer-based
teaching programs; and, finally, Dr. Jones felt that participating
in this effort would provide a great opportunity for her to be part
of a team of professionals that included experts from the Human-Computer
Interaction Institute, other biologists at her own institution,
Carnegie Mellon University (CMU), and biologists from several other
institutions, including Harvard University and the National Science
Foundation, who share her desire to change how genetics is taught.
The project
team had several goals in creating the GCT. The primary goal was
to speed the students’ learning of genetics and improve their
command of the subject. A second goal was to find a modality that
is neutral in teaching students of diverse racial, ethnic, and socioeconomic
status. Math cognitive tutors for middle and high school students,
on which the GCT is based, have this capacty. A third goal was to
improve the teaching of genetics nationally by disseminating the
software widely to colleges and universities and distributing simpler
models to middle and high schools. A fourth goal was to eventually
make the Cognitive Tutor commercially available.
The Cognitive
Tutor has great potential to enhance student interest and involvement
in a subject – and thereby improve learning – because
it enables students to learn via a modality that to them is as natural
as breathing, namely through computers. The approach it uses is
to identify challenging problems and for each problem do a task
analysis and develop a cognitive model, which is an expert system
that mimics the ways students solve problems. Software is then written
based on this model to support the students’ learning. An
advantage over less sophisticated genetics software is that errors
are flagged “just-in-time” to provide help in the form
of “hints” that allow students to succeed in solving
complex, authentic problems. Students are required to answer each
question correctly and show their work before they can proceed.
The interfaces are designed to “make thinking visible.”
As the Cognitive Tutor accepts answers, the program interprets student
behavior, and, when completed, will customize the lesson for the
individual and predict how a student will perform on future exams.
Studies have shown an average gain of 16 points, or 36% improvement,
on exams among students who have used the Cognitive Tutor. The GCT
program, which integrates principles of artificial intelligence,
cognitive psychology, human computer interaction, and genetics,
is based on earlier Cognitive Tutors used for teaching algebra,
geometry, statistics, and computer programming.
Discussion:
The discussion
began with several questions about the advantages and disadvantages
of the
GCT: Has the non-biased nature of the program been tested? Is it
possible to measure whether this approach alleviates misconceptions?
Do problems exist with this program that mimic those encountered
in tradition methods of teaching? What new problems does the GCT
create? Can educational software effectively foster meta-cognitive
skills and if so, will students become better learners as a result?
These are important issues that must be addressed, but, because
the GCT is still in the testing phase, sufficient data has not yet
been collected to provide the answers.
The discussion
then turned toward the challenges in creating a sophisticated computer-based
learning tool. These include the various intellectual and administrative
problems that accompany new developments, the need to test and demonstrate
the efficacy and value of the tool, and the need for widespread
dissemination. One major difficulty is the time and effort it takes
to pull together a project of this magnitude and find professionals
with appropriate experience willing to become involved. The extent
of effort required of the disciplinary professionals, computer programmers
and cognitive experts as well as the cost of the materials necessary
to develop the GCT is enormous; it took, for example, two years
to develop the eleven modules of the GCT that are currently being
tested at colleges and universities across the country and is expected
to cost $1,000,000 by the time it is completed.
Next, the session
leader stressed the significance of the test institutions in creating
superior and effective computer-based teaching tools and emphasized
the importance of allowing the test groups enough time to incorporate
the tutors into their lesson plans. Testing is carried out in two
phases. The first phase consists of instructors in the test groups
using the software for sufficient time to develop some expertise,
involving enough students to determine merits and deficiencies and
gauge the effectiveness of the software, and, following this, bringing
the teachers who have used the Cognitive Tutors together with the
disciplinary experts and the programmers to discuss software glitches
and make suggestions for improving the program. The second phase
involves refining the software based on the feedback and suggestions
put forward. Testing can be expensive, particularly because the
institutions testing the programs do not pay for the materials.
The final challenge
is the difficulty in disseminating programs like the Cognitive Tutor
widely. Broad dissemination is likely to occur only after the benefits
of using the software have been documented. The Cognitive Tutor
Algebra (CTA) course, for example, the most successful of the Tutor
initiatives, is currently being used by 200,000 students in 1800
schools. Studies have shown remarkable gains by students in CTA
classes in comparison to their counterparts in control classes,
and the U.S. Department of Education has designated the CTA course
as one of five exemplary curricula for K-12 mathematics education.
The group agreed that the first step in acquiring the resources
necessary to develop, test, and disseminate more computer-based
programs is to test and document the benefits of the Cognitive Tutor
programs that have already been developed and share this information
with funding agencies.
Other computer-based
teaching and learning programs are already widely available, but
they do not offer the immediate feedback and the individualized
active learning environments of the Cognitive Tutors. Examples of
web-based teaching and learning tools include the Multimedia Educational
Resource for Learning and Online Teaching (MERLOT), a catalog of
online, peer reviewed learning and teaching materials; and Dyann
Schmidel’s interactive educational websites which provide
a myriad of games, puzzles, and quizzes for all education levels
in a variety of disciplines.
Recommendations:
- Development
of cognitive tutors requires a myriad of experts, including disciplinary
and cognitive specialists and computer programmers. One suggestion
was to bring these experts together in the hope that a joint effort
would help increase resources and speed dissemination.
- The group
expressed an interest in the development of web tutorial templates
with essential characteristics that can readily be adapted to
many disciplines. These templates could be valuable teaching tools
that teachers can tailor the programs to their own classrooms.
References/Resources:
Websites
- The PACT
Center website: www.pact.cs.cmu.edu
This site provides information about the Center’s goals,
completed and current research, biographies of the people involved,
and links to sites with detailed explanations of these projects.
The site also provides an extensive list of publications on tutor
development and pedagogy research, cognitive tutor evaluation
and implication, and other relevant PACT publications.
- The Human-Computer
Interaction Institute website: www.hcii.cmu.edu
HCII is an interdisciplinary group of faculty and students at
Carnegie Mellon University dedicated to research and education
in topics related to computer technology in support of human activity
and society. From the site’s homepage, click the “research”
link to find information about interesting research projects
and computer-based learning tools.
- The MERLOT
website: www.merlot.org
- Dyann Schmidel’s
interactive educational websites: http://schmidel.com/dyann.cfm
|
