'Watching kids learn'
Teachers College researchers got a few surprises when they looked at the data from their Affinity Learning project.
Researchers at the National Center for Information Technology in Education discovered they actually were watching kids learn.
    In the project, sponsored by the National Science Foundation, 10th grade science students in an Omaha school were guided through online activities and assessments in accord with their skills and rates of learning. Freshmen and graduate students at the University of Nebraska at Omaha also were involved in the project.
    The Web-based learning application, designed to teach the mathematical concept of acceleration, included a series of examples and an assessment of how much a student learned from each example. If the student answered the example correctly, the student moved on to another activity. If the answer was incorrect, a remediating step - designed for each predicted wrong answer - was provided. Sometimes, the wrong answer hadn't been predicted and the student was told to contact the teacher. A new remediating approach then was developed and added.

Fuzzy nodes
    "The module consists of multiple small lessons, each of which is a fuzzy node consisting of an activity, an assessment and an identified number of outcomes. Outcomes might be correct or erroneous. Correct outcomes lead to subsequent activities, incorrect ones to remediating. Unexpected outcomes lead to new fuzzy nodes," Arthur Zygielbaum and Roger Bruning of UNL and Neal Grandgenett of UNO explained in a paper presented in 2001 at a symposium in Monterey, Calif.
    "The Affinity Learning tool demonstrates the use of an electronic tutoring technology that interacts with both teacher and student and grows to provide a comprehensive and personalized learning experience," according to their abstract.
    Researchers found what they suspected about computer learning in the Affinity project - a 10 percent better performance by students using the computer activities compared with students in the traditional paper-and-pencil model. But Zygielbaum said they were in for a few surprises as well.
    The boys hurried through the activities as fast as they could to get to the end, spending little time on the explanation notes and most of their time in the node labeled "Contact a teacher," he said. The girls, on the other hand, took their time and were careful not to make errors. As the researchers pondered the results, the students' 10th grade math teacher, who had developed the nodes and node structure, explained "these are characteristics of real students learning."

'Eureka moment'
    The computer assessment also showed, what the researchers labeled the "Eureka moment," the point at which the students began to comprehend the topic. A fluctuation at the end of the assessment indicated that students had difficulty applying the topic more broadly indicating either "that the final assessment of the student's equation is intrinsically hard or that we did a poor job in aiding that testing online," the researchers wrote in their abstract.
    "The assessment not only evaluated learning but defined it," Zygielbaum said. "Basically, without sufficient motivation, the kids didn't want to learn, they wanted to get to the end of the exercise." Educators' role, he said, "is to make people more excited about learning."
    Ultimately, teachers could use similar Web-based presentations to help create a more individualized learning environment, Zygielbaum said. In a classroom of 30 students, the activities provide students with individualized support without a teacher having to be at the student's side.
    "These figures give a sense of the possibilities for an intimate look at student learning as they tackle a complex topic," Zygielbaum, Bruning and Grandgenett wrote. "Many questions can be raised. Can we predict from early node performance how a student will perform on later nodes? If we can, then we can tailor the nodes to fit that student's learning characteristics. Can we discriminate poor presentation and assessment methods from student performance? If we can, then we can improve the presentation and assessment virtually continuously as the tool is used."
    NCITE is using the technology for work it is now doing with the University of Nebraska Medical Center as it develops computer examples on the diagnosis of cardiovascular problems in the leg.

NSF proposals
    Two other multimillion-dollar proposals to the National Science Foundation are being developed by NCITE. One involving "intelligent learning objects" uses intelligent agents to respond to students and modify their experiences. Intelligent agents are computer programs that "watch how you do your work and make suggestions to you," Zygielbaum said. They reconfigure the information to best suit the student without the student intervening.
    The other proposal calls for creation of a Center for Learning and Teaching at Teachers College. The National Science Foundation currently sponsors 10 centers around the country to help schools and teachers, kindergarten through 12th grade, in the areas of science, technology, engineering and math.
    NCITE also is jointly sponsoring a graduate-level seminar with the Department of Computer Science and Engineering on "reinventing computer science," Bruning said. "We will look at how to use technology to provide a significant part to beginning instruction in the computer science curriculum. It is a major effort to revamp the beginning curriculum in computer science."
    In addition, NCITE continues to support the development and refinement of two online learning tools developed by the Center for Instructional Innovation. ThinkAboutIt! and InfoGather are used in undergraduate teacher education classes.