The No Child Left Behind Act requires public schools to demonstrate yearly improvement as measured by students' scores on standardized tests. The Assistment system aims to solve a dilemma for teachers posed by the law: how to prepare students for tests without sacrificing quality instruction time.
The system is designed to quickly predict a student's score on a standardized test, provide feedback to teachers about how they can adapt their lessons to address gaps in students' knowledge and provide individualized tutoring to suit each student's needs. In developing the Assistment system, researchers have drawn upon the proven success of Carnegie Mellon's popular Cognitive Tutor®, a comprehensive secondary mathematics curricula and computer-based tutoring program that has been commended by the Education Department and is in use in 1,500 schools nationwide.
"Unlike the traditional test prep, we will build a system that is focused on helping students learn as opposed to being strictly an assessment system," said principal investigator Kenneth R. Koedinger, an associate professor of human-computer interaction in Carnegie Mellon's School of Computer Science. "It's not just going to say 'These kids aren't learning fractions.' It will teach them fractions. There's a real challenge of providing instruction that adapts to an individual student's needs."
The investigators are planning to have eighth-grade teachers in the Worcester, Mass., Public School District begin to test the system in the spring of 2004 as they prepare students for the Massachusetts Comprehensive Assessment System (MCAS). According to Koedinger, Worcester was selected in part because it is an urban school district with a high percentage of minority and low-income students, who, on average, perform poorly on standardized tests. Sixty-four percent of African American students failed the eighth-grade MCAS test last year, while the failure rate of Hispanics was 74 percent. Researchers hope to improve students' performance before they reach high school, when students must pass the MCAS in order to receive their diplomas. About one-fifth of all Worcester seniors are expected to fail the test this year.
The four-year grant, which comes from the Education Department's Institute for Education Sciences, will allow researchers to compare the performance of students who use the Assistment system to those who do not. Although the system is being tested in Massachusetts, Koedinger said it has been designed so that it can be readily adapted for use in other states.
In addition to Koedinger, other researchers on the Assistment project include Brian Junker, a professor of statistics at Carnegie Mellon; Neil T. Heffernan, an assistant professor of computer science at Worcester Polytechnic Institute; and Steven Ritter, a senior cognitive scientist at Carnegie Learning.
Carnegie Learning, which markets the Cognitive Tutor curricula, was founded in 1998 by Carnegie Mellon researchers to apply and extend more than 20 years of award-winning research in cognitive science to mathematics instruction.