Reaching the community with intelligent tutoring

	Glenn Blank

Each year, the increasing demand for innovative technology is reflected by the emergence of such new products as car navigation systems and iPhones. But fewer students are choosing to pursue careers in technology fields.
“There is a real lack of women and minorities in information technology,” said Glenn Blank, associate professor of computer science and engineering. “In the 1980s, women made up about 25 percent of computer science majors in research universities, but today it’s down to 10 to 15 percent.”
Fewer people are going into information technology (IT) careers, Blank says, partly because of stereotypes that the field is dominated by “anti-social nerds.” In reality, he says, software development and other IT endeavors require diverse teams of people to solve problems for real customers.
Another possible turnoff is that the IT classroom is often filled with male students writing code on their own—not a very supportive environment for novices with different backgrounds or learning styles.
“If we can introduce computing in a more engaging way, helping students when they get stuck, then students would be more interested in pursuing computing careers,” says Blank. “And if you catch students early—in middle schools before many of them are turned away by peer pressure or stereotypes—then you can have more success.”
Blank helps oversee two grants from the National Science Foundation (NSF) that support programs that encourage struggling middle and high school students to prepare for IT careers.
Launch-IT, encourages minority and female students to pursue college and careers in information technology (IT).
The Lehigh Valley Partnership for STEM Education, or LV STEM (Science, Technology, Engineering and Mathematics), sends eight teams of Lehigh faculty and graduate students into K-12 schools to develop and help teach innovative curricula in STEM classes.
The programs are directed by Blank and Henry Odi, executive director of academic outreach and special projects at Lehigh.
Blank advises several graduate students in the two programs. The students are developing and implementing intelligent tutoring systems, which help educators learn about a child’s strengths and weaknesses in a subject. The systems also analyze students’ errors by comparing what they do with how an expert might solve the same problem.
Computer tutors
“Computer tutors can cause students to move as much as a letter grade up,” says Blank, adding that the systems, once they are created, are less expensive than human tutors.
Shahida Parvez and Sally Moritz, Ph.D. candidates in computer science, have developed DesignFirst-ITS, an intelligent tutoring system for high school students that emphasizes the design of software systems in the first few weeks of a programming course. After learning basic concepts with multimedia created in Flash, students begin constructing designs in Unified Modeling Language (UML). DesignFirst-ITS helps them, step by step, with advice or lessons. The system has been used successfully at Dieruff and Whitehall High Schools.
“This kind of approach involves not just coding, but rather catering to customers and finding solutions,” Blank says.
Moritz has also created a tool that translates an instructor’s English-language description of a problem into an expert solution in UML. Her expert evaluator lets students design UML in their own language and matches them with the expert solution.
Nick Moukhine, Mike Sands, Isaac Rieksts and Yu Yang, all graduate students in computer science, and Tamara Peffer, a Ph.D. candidate in the College of Education’s
Teaching, Learning and Technology program, are also designing new intelligent tutoring systems.
Moukhine works at Liberty High School and Rieksts at Harrison-Morton Middle School as LV STEM fellows. Both spend about 10 hours a week in the classroom and five hours in preparation. Sands and Peffer help Launch-IT students learn Flash and Java once a month during the academic year and for three weeks during the summer.
Moukhine and Sands are also developing an intelligent tutoring system that helps students learn how to debug programs.
A student model then estimates how well a student understands the relevant concepts and strategies. Finally, a pedagogical advisor decides how to provide tutoring advice or lessons to the student. Parvez’s research focuses on how to adapt feedback to a student’s learning style.
Peffer is working on an intelligent tutoring system that helps middle school students with an online game, First in Math, that works in tandem with the “24” game.
“When students get stuck or make a number of errors, the program provides guidance and cognitive hints to help solve fractions,” says Peffer. The program then gives students additional problems to solve that are similar to problems the students answered incorrectly.
--Rita Shankar ’08