Rural Schools Hit By Test Errors
Justine Ferrari, Education writer, The Australian, February 03, 2010
THE national literacy and numeracy test results for small and rural schools fail to provide reliable information on their performance, skewing their results as either much better or worse than those of other schools.
The reliability of the national test results published on the My School website varies greatly according to the size of the school, with small schools having the most inexact results.
The variation is due to the measurement errors inherent in any form of test.
The error rates show that a school’s average score varies by as little as four points and as much as 38 points, depending on the size of the class and the subject tested.
Testing experts said yesterday the error rates were in line with expectations for the type of tests conducted, but said they highlighted the need to assess school performance on a broader range of measures than a single average score.
The margins of error also highlight that ranking schools based on the median score, as done by some newspapers after the My School launch, is meaningless. Education professor Stephen Lamb, who developed a reporting system for school results for the Victorian government, said the margins of error were critical in small schools, where a marked difference in performance by one student in one year could make a huge difference to the school’s average.
“The larger the number of students, the more reliable it is in representing school performance,” Professor Lamb said. “Small schools may appear to do much better or worse; it’s one of the difficulties with comparing rural schools, which tend to be small.”
