Schools in Norfolk, Virginia are closing the achievement gap between black and white students, reports the St. Pete Times. Overall, the district is two-thirds black; 60 percent of students come from low-income families.
In 1998, 67 percent of Norfolk’s white third-graders passed the state English exam. Only 41 percent of the district’s black third-graders met that standard.
Five years later, the passing rate for black students had jumped to 61 percent.
A black superintendent named John Simpson took over in 1998 with a mandate to improve achievement. He stressed using data to improve teaching.
Teachers were ordered to test their students — over and over, if necessary — until they could determine what they knew. That information was used to tweak lesson plans and identify children who needed extra instruction.
It sounds like schools have found a way to group students by performance, discreetly. A class of 36 students and two teachers is split into two groups: One serves students who are doing well; the other is for children with learning problems. Teachers use different approaches but students are expected to learn the same material.
At a school with very low-income students and high transiency, fourth grade teachers analyze test scores.
They weren’t happy with what they saw.
The teachers wondered whether the students were tired by the time they got to a question with a high percentage of wrong answers. They noticed an even distribution of answers to another question, an indication that none of the students knew how to solve the problem.
In Norfolk, knowing why students answered incorrectly is as important as the overall test score. It allows teachers to identify weaknesses and adjust their instruction accordingly.
There is a maxim in education circles. Testing at the end of the school year is an autopsy. Testing during the year is a check-up.
At one middle school, the principal enlarged PE classes so she could cut PE teachers and hire more English and math teachers. Students now spend 85 minutes a day learning math. And it shows.