EdVANTAGE Blog

We’ve reported on comments by State Board of Regents Chancellor Merryl Tisch suggesting that state tests should be made more difficult.  Last week the New York Daily News explored the issue in more depth.

The paper argues, “Soaring scores on the state math test don’t necessarily add up to better schools or smarter kids.”  It concludes that it has become easier for teachers to “teach to the test” as it has become easier for teachers to predict what questions will appear on the tests.

The Daily News characterizes this as a defect in the tests.  It could also be seen as a natural consequences of accumulating experience with the tests, first given in 2006.

It may also be an outgrowth of the state’s decision to issue grade-by-grade curricula, to help schools and teachers identify what topics in math will be tested when.

One part of the paper’s analysis may be flawed.  It notes that only 54.7 percent of specific math skills expected for 7th graders were ever tested in the four years the assessment has been given.

Chancellor Tisch told the Daily News that the fact that only a limited number of math skills are tested is a “substantial” and “important” finding, adding, “That’s what we need to fix.”

But the 7th grade assessment, as well as those in 3rd and 5th grades are designed as shorter tests.  The longer 4th, 6th , and 8th grade assessments would be expected to cover more ground.

The paper acknowledges that, “The tests getting potentially more predictable is not a conspiracy to boost scores and the reputation of schools statewide. It’s not even unusual.”  The article explains, “Experts said that developing a wider range of test questions year-to-year requires paying the test preparation companies more money.”

Accordingly, the analysis highlights some of the limitations or complexities of test-based school accountability.

Still, the Chancellor has raised issues worth debating as the state seeks proposals from test-making companies to develop the next round of state tests.

One superintendent observed to me that the tests may measure proficiency with reasonable accuracy, but are less and less meaningful above that. As such the tests are, “Not so much easy as irrelevant.”

Lengthening the tests would be necessary to make finer valid measurements of student achievement across all levels of performance.  This is a required ingredient in developing a strong growth or value-added school accountability model.  Developing more questions for each topic may be needed to ensure the assessments are statistically reliable measures of student and school performance.