By Greg Russell
It almost sounds like the making of a sci-fi novel: animated agents wresting mind
control from a young audience with the help of an everyday computer screen. But this
is not Stephen King drivel — it is an important research project in the Department
of Psychology in the U of M’s College of Arts & Sciences that is enhancing learning
in area middle and high schools.
Kyle Cheney and other researchers use computer software that features animated agents
to optimize learning in classroom settings.
Psychology professor Dr. Barry Gholson, researcher Dr. Scotty Craig, undergraduate
research assistant Kyle Cheney and others are using computer software that makes use
of animated agents (talking heads) that appear on computer monitors with an ultimate
goal of improving classroom performance.
“Our major goal is to optimize learning in classroom settings by incorporating computers
that present dynamic conversations between animated agents,” said Craig, who works
in the U of M’s Institute for Intelligent Systems. “These agents seek to focus students’
attention on central information that provides the conceptual foundation for course
content that is then given expression by the classroom teacher.”
Cheney said the researchers enter a classroom, discuss course content and then have
the class watch a video.
“We have an overall pretest to gauge their base knowledge — where they are when we
start. Then on each day we have a pretest. The pretest is only six questions and it
pertains to only specifically what we are teaching them. They watch the video, the
teacher teaches a little bit and then we test them afterwards to see how much better
they did in comparison to how they did in the beginning.”
Cheney said the researchers follow that model for six days and on the seventh day
conduct a post-test. The results from each day as well as the overall pre- and post-test
results are put into a computer to see what learning gains are made and if there is
anything on a particular day where the students did really well or poorly on.
The video on the screen features two animated talking heads — a teacher and a student
— who discuss the content of a particular course as students observe.
“Within the video are deep-level reasoning questions, which are questions that are
intended to make a student want to know how and why something works — what makes a
particular concept work. We want them to know more about a subject rather than just
be able to fill in the blanks on a test. When they get an explanation after they are
thinking about a concept, it gets further into their memory and they retain more of
it,” said Cheney.
The videos last between 13 to 20 minutes and focus on one subject at a time.
“For example, in physics classes the conversational agents discuss the conceptual
content of Newton’s first three laws,” said Craig. “Following these presentations,
the teacher is then able to focus the students’ attention on equations that give quantitative
expression to those concepts and their interrelationships. The exchanges between conversational
agents, which include animations displaying dynamic relationships among concepts,
are built on deep-level reasoning questions and explanations.
“These encourage students to engage in deep reasoning about the conceptual content
and how to give it quantitative expression,” Craig said.
The researchers say the project has included biology, physics and information technology,
but they expect future work to expand into broader areas of the curriculum, and to
eventually include elementary schools.
The research is being funded by the Cognition and Student Learning research program
in the Institute of Education Sciences, which is a division of the Department of Education.>