A Comparison of test score gains, 2006-2007 by students who exceeded in West Sylvan, Portland and George Middle Schools
In the course of putting together these graphs, I discovered something interesting. Conventional wisdom holds that students who "exceed" will always underperform students who "meet" benchmarks. The assumption is that they are closer to the ceiling of the tests and so don't have as "far to go". In addition, the margin of error is greater at the top and there is also the issue of regression to the mean.
So I compared the "exceeds" score gains for reading between West Sylvan with the "meets" score gains for the overall PPS average and found, sure enough, that West Sylvan students who exceeded benchmarks in the previous year made lower gains than the average PPS student who just met the benchmarks the previous year.
Math is supposed to be much more susceptible to ceiling effects for Middle School students, because advanced students usually start geometry and algebra in middle school and these are not well represented on the state tests for this age group. Nevertheless, West Sylvan students who exceeded outperformed PPS students who just met benchmarks at all three grade levels. This isn't because West Sylvan keeps advanced students in lower-level classes--they move on to algebra and geometry. Is it simply because they are doing a good job at math instruction? Can they achieve the same results in reading in the future? At George, students who exceeded not only failed to keep pace with students who met benchmarks, they lost ground every single year they were in middle school.
Research, Evaluation & Assessment - 09/07/07ttj(2007-0135)
This report includes students who have a valid score from both 2005-06 and 2006-07 and were enrolled at the same school Oct. 1, 2006 and May 1
2007.
NOTE: Because of problems with the state’s computer-based assessment system (TESA) in early 2007, testing on it was discontinued and moststudents took paper-pencil tests in May. Students who met state standards on TESA between January and March were not retested in May. Many of those students might have exceeded state standards and had higher test scores had they been retested. Readers should keep this situation in mind when comparing 2006-07 results with results from other years, or when viewing reports that show test score gains.