In 1974 Phillip Uri Treisman, a calculus teacher at Berkeley, received the visit of two graduate students. The purpose of their visit was to videotape Prof. Treisman's class.
He had been rated as either one of the ten best or ten worst teachers in Berkeley, and he didn't know which. They wanted to study if other teachers could identify good teachers from the bad ones. Prof. Treisman couldn't know if he was in the top or bottom 10.
Prof. Treisman had been had been trying to get his students interested in Math through collaboration and creative problem solving, innovative techniques at the time. The students had met his experiments with certain reluctance: "is this going to be on the test?" they would ask. He had no idea if he was a good or a bad teacher.
Turns out he was one of the good ones, as he was given a grant to study why there was a performance gap in Black and Hispanic students. It was a problem that had been bothering some time, and he was willing to tackle it head on.
He would start by surveying his colleagues for possible hypothesis, and he found four widely-held beliefs:
- Minority students are not as motivated as other groups
- They come unprepared, as they often enter university with fewer credit hours of science and math.
- Their families lack a strong cultural and intellectual background, and thus lack understanding of the importance of higher education.
- That the income gap reflects on the educational gap, and if those variables are controlled there would be no performance gap.
All these beliefs sound reasonable, however, when they started looking at actual data, they noticed they were entirely wrong. Being an elite university, minority students, especially black students, had to make huge social sacrifices to be able to get accepted into Berkeley. Motivation was not the problem, it was disorientation.
They were also surprised to find that, among Black students, calculus grades correlated negatively with good math grades in high school. Black men with high SATs often faced academic dismissal. It was the students in the middle range who were the best math students in university.
They interviewed the families of the students to find if that was the cause of their poor performance. What they found was a strong network of support, that they had made a conscious effort into getting their kids into college, and quite a few of them were college graduates too.
Perhaps the most unexpected finding was that family income was negatively correlated with performance. This is because many of the poorest students come from families where the parents would work in public schools and don't earn much, but had a long standing tradition of education.
So, after going through all the data and interviews, they were at a blank state again. They decided to do some ethnographic research: follow the students around, videotape them in the least conspicuous way possible, let them grow to their presence.
So they decided to follow a group of Chinese students and a group of Black students. Black students would go to class, take notes, study eight hours per week, and then fail. What was surprising was the difference they saw with the Chinese students:
They studied calculus for about 14 hours a week. They would put in 8 to 10 hours working alone. In the evenings, they would get together. They might make a meal together and then sit and eat or go over the homework assignment. They would check each others' answers and each others' English. If one student got an answer of "pi" and all the other got an an answer of "82", the first student knew that he or she was probably wrong.
It was interesting to see how the Chinese students learned from each other. The would edit one another's solutions. A cousin or an older brother would come in and test them. They would regularly work problems from old exams, which are kept in public file in the library.
Based on these findings, Prof. Treisman obtained funding for creating what was essentially a "learning group". Berkeley had previously tried to enroll minority students in optional preparation courses, but he knew that minority students dismissed them as something for underperforming students.
His learning group was—in essence—a place to tackle problems to together.
Most visitors to the program thought that the heart of our project was group learning. They were impressed by the enthusiasm of the students and the intensity of their interactions as they collectively attacked challenging problems. But the real core was the problem sets which drove group interaction. One of the greatest challenges that we faced and still face today was figuring out suitable mathematical tasks for the students that not only would help them to crystalize their emerging understanding of the calculus, but that would show them the beauty of the subject.
We were able to convoke the students in our orientation that success in college would require them to work with their peers, to create for themselves a community based on shared intellectual interests and common professional aims. However, it took some doing to teach them how to work together. After that, it was really rather elementary pedagogy.
The results of this program were surprising: Black and Latino participants outperformed not only their minority peers, but their White and Asian classmates as well. Black students with Math SAT scores in the low-600s were performing comparably to White and Asian students whose Math SATs where in the mid-700s.
Here is Prof. Treisman's first hand ccount (PDF) if you're interested in learning more.