Survey rates Ontario schools
August 18, 2009
Comments on this story
(37)
Louise Brown
You can't judge a school fairly without looking at the background of its students – and an unusual Ontario survey tries to do both.
In its latest education report, the C.D. Howe Institute compares the test scores of schools where children face the same demographic privilege or penalty – from how rich they are to how long they have lived in Canada and the amount of formal education of their parents – in order to measure how well schools are serving students once the demographic differences are removed.
"About half the difference in test scores between schools comes from the socio-economic differences between students – but the other half comes from the school itself, and that's what we're measuring when we remove the other factors," says economics professor David Johnson of Wilfrid Laurier University, author of the institute's latest report card on Ontario schools.
Unlike more simple surveys that rank schools by test scores alone, Johnson has designed an intricate formula that factors in a demographic picture of a school based on nearly a dozen socio-economic factors – and then compares schools with similar profile, like apples to apples.
"It's one thing to see how well your child's school does on provincial tests, but it's more meaningful to know how it does compared to schools with the same type of students," said Johnson, whose sweeping research on Ontario's schools is being released today by the non-profit think tank.
"If you care about the quality of a school, this is how you do it – you see how well a school does compared to others with a similar type of student body."
Here's how the rating chart works.
Each school carries a "percentile" number, which shows whether it does better or worse on Ontario's standardized tests than other schools in the same demographic bracket. A percentile rating of 50 plants it right in the middle, where you might expect a school with such students to score, says Johnson. He calculates the percentile by averaging a school's Grade 3 and 6 test scores over the past three years.
The higher the school percentile is above 50, the more schools in its category it beats. A school with a "70" percentile, therefore, does better than 70 per cent of schools with children of similar economic background. Likewise, a school with a "10" percentile scores higher than only 10 per cent of schools in like neighbourhood.
"We often hear that all differences between school scores can be chalked up to social and economic factors – but this data proves that's wrong. You have schools in the same demographic bracket landing scores that are more than 40 points apart. And they can't blame it all on the students' background – because we've already taken that into consideration."
thestar.com