Standards keep shifting, and some changes may make exams easier to game
As back-to-school season looms, grades are due out any day now for the 2017 Math and English Language Arts exams that measure proficiency across grades 3-8. In a new report for the Manhattan Institute evaluating 2016 scores in traditional (non-charter) public schools, I find that ELA proficiency shot up by 7.8 percentage points, while math proficiency rose only 1.2 points. The broad ELA gains across ethnicity, gender and even disability status are striking in light of the relative stagnation in math — and in fact likely reflect major changes to the testing setup that juiced ELA performance.
The results we’re about to see will mark the second year since the state education department made big changes to the testing process. Specifically: Tests were made shorter; 75% of the preceding year’s (similar) exam questions were released in advance; and students get unlimited time to complete the exams. The State Education Department itself warns that exam changes render scores “not comparable” on the web infographics that, um, compare 2016 and 2015.
In the state Education Department’s defense, there is an intuitive explanation for removing time limits. These exams test proficiency in subject matter, so speed of completion ought not be relevant (as it might be when assessing, say, IQ). In the real world, people generally have at least several hours to write and revise their work, and that’s the kind of substantive skill demonstration that matters.
Alex Armlovich is a fellow at the Manhattan Institute and author of the new report, Poverty and Progress in New York: English and Math Proficiency in NYC Schools, 2015–2016. Follow him on Twitter here.