There’s a lot to be said for the Indiana Growth Model, the statistical method that Indiana uses to calculate year-to-year student growth on math and English test scores. It’s far from perfect, but it’s a much better measure of how schools are doing than looking at how many students pass the tests.
It may be a challenge, though, to use the model while complying with a state law that says Indiana must measure students’ growth in relation to their proficiency on state standards.
That’s especially true now that Indiana has adopted new “college and career ready” standards and will be giving a new version of the ISTEP+ exams, aligned with the new standards, in the spring of 2015. Is it possible to measure growth in proficiency when you give a test for the first time?
The law in question is House Enrolled Act 1427, adopted in 2013. It says accountability measures “must be based on measurement of individual student academic performance and growth to proficiency” and “may not be based on a measurement of student performance or growth compared with peers.”
But the State Board of Education voted this month to use the growth model in 2015. Indiana’s Center for Education and Career Innovation and the state Department of Education recommended the approval.
CECI and DOE staff cited an analysis by testing expert Damian Betebenner, who helped design the Indiana Growth Model and advises the state. He suggests using a statistical adjustment called “equi-percentile concordance” to correlate the 2014 test with the new, 2015 test. That, he says, will make it possible to keep using the growth model to measure students’ test-score gains.
But his report to the board also says that, with the move to new ISTEP+ exams, it won’t be possible to evaluate students’ gains or losses on a single test from one year to the next. “Without gains/losses,” he writes, “growth must be calculated using norm-based metrics that compare like students as they progress from the ISTEP+ to the Career and College Ready Assessment.” (Italics added).
Vic Smith and Derek Redelman, who don’t agree on much, both say that’s a problem.
Smith, with the Indiana Coalition for Public Education, says the legislature clearly called for an accountability system based on whether students make progress in relation to standards, not the peer-based Indiana Growth Model approach. “It is astonishing that the State Board is ignoring the call for fair comparisons in Indiana’s growth model,” he writes in his Statehouse Notes column.
Redelman, vice president for education and workforce of the Indiana Chamber of Commerce, told me via Twitter that he also thinks the “student growth percentiles” approach of the current model doesn’t comply with HEA 1427.
But State Board of Education general counsel Michelle McKeown says using the Indiana Growth Model will comply with the law. Under the model, she said, each student who takes ISTEP+ exams has a testing “target” that reflects whether the student achieves appropriate growth from the previous year.
She gave the hypothetical example of a girl who scored 405 on the fourth-grade math test in 2014. To set her target for 2015, the state looks back to the 2013 math exam, takes all the fourth-graders who scored 405, and plots the scores they recorded on the 2014 test. To demonstrate a year’s growth in 2015, the girl would be expected to achieve at least the median of those scores.
The key point, McKeown said, is that the student’s growth is measured against her target score, not against the performance of her fellow fifth-graders.
“Theoretically, every kid in Indiana could meet his or her growth target,” she said. “I think this is absolutely what’s required of HEA 1427. Only the student’s performance determines whether or not she meets her growth target.”
McKeown said the approach lets the state identify realistic and feasible growth targets for students at all developmental levels while shifting from one set of standards to another.
So who’s right? Maybe legislators can tell us what they meant. But it’s possible they won’t agree either.