# Toughen the tests

New York needs tougher tests that measure student progress, writes Diane Ravitch. In response to No Child Left Behind, New York made it much easier for weak students to be classified as “proficient,” she writes.

In 2006, a seventh-grade student needed to get 59.6 percent of the points on the state math test to become proficient (Level 3); by 2009, it was just 44 percent. Remember the old days when 44 percent was a failing mark? Not any more.

. . . In 2006, third-grade students had to get 43.6 percent of the points on the math test to earn a Level 2 — but by 2009, they needed to get only 28.2 percent of the points. On the English language-arts test, the cutoff to earn a Level 2 in sixth grade dropped from 41 percent of the points in 2006 to just 17.9 percent in 2009.

New York City wants to end social promotion by requiring students to reach Level 2 to move to the next grade. But students who guess blindly can do well enough to reach Level 2.

The Regents exam also has been downgraded, Ravitch writes.

To get a diploma, students must get a 65 on each of five Regents exams. Sounds tough — but it’s not anymore, thanks to the State Education Department’s statistical magic.

On the algebra Regents, a student collects a passing score of 65 if he or she earns only 34.5 percent of the possible points. On the biology exam, a “pass” requires earning only 46 percent.

Ravitch suggests giving honors, college-ready and work-ready diplomas that  reflect “realistic goals for everyone, rather than a low hurdle that almost everyone can step over.”

I think this makes a lot of sense.

1. I think it makes a lot of sense too. Sadly, a situation like the one Ravitch describes would preclude the sort of test scores that self-serving cynical politicians demand.

2. tim-10-ber says:

I wish him all the best…I hope he has the courage to raise the bar…

3. FuzzyRider says:

You can’t be serious about making the tests more difficult…someone might fail and their self-esteem would be permanently crushed!
(sarcasm off…)

4. hardlyb says:

When I was a kid, the NY State Regents Exams were substantial. Not really hard – the courses for which they were the finals, like Geometry, Trig, and English, were sort of a pain, because the coursework didn’t match the tests that closely (at least in the honors classes I took), so we had to study for the test with the special paperback textbook (with the REALLY cheap paper) that went with the test. So I didn’t like the books or the tests (this high school had a fairly fierce competition to be valedictorian, because being in the top five students got you lots of privileges), but the tests were decent – infinitely better than that exit exams that I’ve seen recently. I suppose that they are as dumbed down as everything else these days.

5. Does the test really need changed, or does the score necessary for “proficient” just need to be revised up? It’s probably not the instrument, it is how the instrument is being used.

6. Cal says:

I’ve been calling for tiered diplomas for years. It’s a great idea–and politically unworkable.

Rather than create new tests, you could do the same thing with existing tests–AP, SAT, SAT Subject, IB, whetever.

7. CharterMom says:

I knew NCLB was in trouble when I heard it was modelled after NC’s ABC tests and that states would be able to set their own standards. Not enough kids passing? Just change the measures!

This year in NC was pretty comical — everyone is just bursting with how improved things were this year. However this year the standards changed — schools are now allowed to count both kids who passed the initial test and kids who pass the first retest in their totals. Previously only kids who passed on the first try were counted. Now I don’t necessarily think it is a bad idea to include the first retest (everybody has a bad day from time to time) but I really have to laugh at all the press releases and announcements celebrating the “improvement” this year. The news articles dutifully mention somewhere down toward the bottom that the measure changed this year but I have yet to see a single report that compares last year to this year on an apples to apples comparison. Surely that data is available!

8. There should be tiered diplomas, absolutely. You can’t demand everyoneb pass the test on one hand and that standards be held to college readiness on the other at the same time. It just isn’t possible. Have you ever run across adults who are just plain Not Bright? News flash: they weren’t bright as kids, either. It’s pretty impolitic to say that, so I only do so in my super-secret anonymous blog persona.

The data I would like to see is longitudinal. The cohort we have in high school now taking the last of the state exams is the very first to have taken them as elementary school students. I’d like to see if there has been any improvement within the cohort (even though I’m not sure the tests have been well designed to show that). Have any students in the class of ’10 jumped from below basic or basic to proficient at some point, or is once basic always basic? That’s what I would like to know.

9. Charlie says:

I know that there are cases where lower scores on the same or comparable tests are used to make it easier to attain a given category score; I know that in some cases the cut score on a multiple choice test that is needed to pass is hardly over the chance level; I am not sufficiently familiar with the details of this case to know what is going on here.

That said, I do want to make the more general point that a percent-correct score doesn’t tell the whole story. Take a typical test with items that range from easy to hard, and split it into the easy half of the items and the hard half. A score of 40% on the hard half might represent the same level of performance/knowledge as 90% on the easy half, and it is easy enough to estimate this by comparing the scores of the same people on their scores on the two separate half-tests. Similarly, just by picking which items are on a test, we could make one history test so hard that Dr. Ravitch would only get 30% right, and a second test with different items that a sixth grader would get 80% right. To understand the level of knowledge that a test and cut score require to be classified as proficient, you really need to see the items (and ideally the item statistics) as well as the cut score. This also means that if you really do create harder tests, you could have a cut score at a lower percent yet represents better performance.

And it means too that item selection is one of the places for administrators to play games with. For example, make the items easier and make the cut score is higher, and it seems to people who only see the cut score that you are raising standards.

In one project my colleagues worked on, the legislature passed a law that required the cut score on some licensure test (might have been plumbing) to be 70. The testing agency developed tasks that were important for plumbers to know, did studies and worked with master plumbers to determine about how many right they would expect a just-qualified plumber to get right, and noted this percentage. Then they rescaled the scores so that this score was a 70 on the resulting scale. Not only was the cut score set this way legally-defensible, I would argue it is better justified than setting a percent-correct cut point without knowledge of the items or what someone needs to know to be a competent plumber.

10. Hardlyb, I’m there with you. I went to high school when the NY Regents exams required extra studying and most of us purchased or borrowed red Regents study guides. Our high school had tiered diplomas: Academic, Regents, and two others that I can’t remember right now. To get the Academic diploma you have to meet all the state requirements (which meant doing well on the Regents exams) as well as some extra stuff. The Regents exams were usually our final exams depending on how much work the teachers wanted us to do. As a senior, all the AP classes required three exams: AP exam, Regents exam, and a separate final.

The Regents exams started changing when my sisters were in high school. Intead of taking a separate test each for algebra, geometry, and trig for example, they took math year I, year II, and so on. It just went downhill from there.

I believe in a tiered diploma system and I believe in real exams that test real things. I am not impressed by the watered-down tests I’ve helped students study for in Indiana and Ohio. Those things are jokes and only prove how bad education is these days.