School Information System
Newsletter Sign Up |

Subscribe to this site via RSS: | Newsletter signup | Send us your ideas

November 1, 2006

Grading a School's Grades

Alan Finder:

Mr. Hartranft, a nuclear engineer who had been forced to retire early because of Parkinson’s disease, came up with what he thinks is a rigorous mathematical model to compare the school’s demanding grading system with more lenient grading in other schools. The model, he and some local school administrators say, is a bold new way to think about grades.

“I’m giving you a G.P.S. navigation system, as opposed to scraps of maps,” Mr. Hartranft said. “If all you have are scraps of maps, which is all that admissions offices get in the existing protocol, then this gives you an overall orientation.”

Mr. Hartranft created an analytical method he calls the g.p.a. plot; it uses national data on grade-point averages and SAT scores to compare national grading norms with those at the local high school. The purpose, he said, is to reduce the variability and subjectivity of grades — and to make it absolutely clear to college admissions offices that a B or B-plus at Simsbury may be the equivalent of an A at most high schools.

Simsbury has included his statistical comparison in its admissions submissions for the last four years. In the suburb just to the north, Granby Memorial High School is using the g.p.a. plot for the first time this fall.

Here in Simsbury, administrators and parents appear satisfied with the results of the model, even though it is unclear whether it has helped increase the number of Simsbury students admitted to elite colleges. Neil Sullivan, the high school’s principal, said the proportion of students admitted by the most selective universities had increased somewhat over the last four years, after dipping slightly when the number of A’s dropped sharply between 1998 and 2001. But the number of A’s given out by Simsbury teachers has also increased in recent years.

He took the scores of 1.5 million students and graphed them against the students’ grade-point averages, as reported by the students on their SAT exams. In a given year, for instance, the analysis might show that on average nationally, students with an A average had a combined SAT score of 1,150, under the old two-part aptitude test. Then he would perform the same comparison for students at Simsbury, where, on average, a student with an A average might have a combined score of 1,220.

Details at hartranft.org.

Posted by Jim Zellmer at November 1, 2006 6:19 AM
Subscribe to this site via RSS/Atom: Newsletter signup | Send us your ideas