Background & History
The Affinity Learning System is similar to, but
distinguishable from, other electronic tutor development efforts. Many
educational simulation, electronic tutor or virtual classroom programs, such as
the IMMEX system (www.immex.ucla.edu) developed at UCLA, provide opportunities
for students to participate in active learning. Affinity can accommodate these
types of presentations and assessments. In addition, Affinity adds customized
support to help students overcome misconceptions, missing knowledge, or missing
skills. Like IMMEX, Affinity provides data allowing instructors to understand
how students used the system and how well they learned. Affinity-based lessons
in mathematics, geosciences, and medical education have shown positive learning
achievement (Zygielbaum and Feese, in press). The system can also distinguish
student learning strategies – such as those used by male and female students
(Zygielbaum and Grandgenett, 2001).
Affinity is based on predicting the likely misconceptions
of students. Misconceptions are detected through embedded assessments and
resolved through remediation or augmentation learning nodes. Lessons are
prepared by breaking down the topic to be taught into small learning
experiences or “learning nodes.” Lessons appear as a hierarchical network of
interconnected nodes. Each node consists of a web-based presentation of
material and an assessment of student understanding of the material. As a
student completes one learning node, he or she is directed to a subsequent node
based on the outcome of the assessment. The student is either directed to
continue in the learning progression or diverted to an augmentation node. If
the outcome is unanticipated or not covered by a succeeding node, the system
notifies the faculty member and the student. The faculty member then enhances
the lesson by adding new node(s) and interconnections to better serve the
student (Bruning, Zygielbaum and Grandgenett, 2001).
Students’ progress through the network of learning
nodes is monitored by the web-serving computer. Each node outcome is assigned a
weight by the author that indicates the educational value of that response. A
large weight denotes an element critical to understanding the overall topic. A
small weight indicates ancillary information. If the instructor wishes to track
learning efficiency (progress versus time), he or she may assign negative
weights to particular outcomes. Data collected include the order of node
visitation, the related sequence of outcome weights, time spent on each node,
and the results of assessments. These data are presented graphically to the
instructor.
Understanding misconceptions is the basis of Affinity
and relates to the work of James Minstrell work at FACET Innovations
(www.facetinnovations.com). Under NSF and other funding, Minstrell has
developed methods of identifying and collecting misconceptions and create the
DIAGNOSER software system.
Why Pursue
Research on Affinity
Affinity Learning has provided insight into student
behavior and learning strategy. It offers the potential aid in educational
research and to develop more effective online teaching. For example, in the
past, the PI and his colleagues have implemented response to each learning node
assessment individually. In the future, building upon techniques developed to
identify matching
DNA sequences, sequences of learning nodes will be
compared to develop templates indicating student mastery and correct
identification of misconception. Applied to large sets of students, such
techniques could provide guidance in how to present lessons and what knowledge
or skills are required as prerequisites. Based on such sequence information,
the Affinity Learning nodes could be reconfigured dynamically in response to
observed student behavior by techniques such as the use of intelligent software
agents.
Most assessment in education, whether formative or
summative, is done at the level of the answers to problems. The Affinity
Learning System offers the potential to assess student actions while they are
in the process of solving problems (what NCITE has entitled “microassessment”).
The ability to identify and correct misconceptions at
this level will provide targeted, immediate help to students. Micro-assessment
will prove useful in many fields. For example, in mathematics, most instructors
are less interested in the answer to a problem then they are in the way the
student got to the answer. In computer science, the concept of recursion is
very difficult for students. The essence of the difficulty appears as students
attempt to create recursive solutions to programming problems. Affinity
software will identify and help correct student misconceptions as they attempt
to employ recursion in solving problems.