Test more, not less

Testing is under attack, but the solution is more testing, not less, argue Russ Whitehurst and Katharine Lindquist on Brookings’ Chalkboard.

New Common Core tests demand more of students, pushing down proficiency rates even in suburban schools, they write. That’s accelerated the push to test only in a few grade levels.

Linda Darling-Hammond, who advised one the consortia developing the new tests, recommends three tests — once in elementary, middle and high school — while still breaking out scores for “vulnerable” groups.

Her proposal would end value-added evaluation of teachers and schools, they write.

. . .  value-added calculations at the teacher level depend on the difference between the test scores of a teacher’s students at the end of the school year and the test scores of those same students at the end of the previous school year.  The annual gain in test scores of the teacher’s students, with some additional statistical information, is the teacher’s value-added.  

. . . Value-added can’t be calculated for many teachers as it is, but in the tested grades and subjects in which it can be estimated, it provides an important point of validation for other more widely deployed measures such as classroom observations. It is also the basis for calculating the school’s value-added, e.g., the test score gains between 3rd and 5th grade for all the students attending a particular elementary school relative to the gains in other elementary schools.

The fewer kids who are tested the more unreliable the scores: Testing at only one grade level “makes results nearly meaningless,” especially for the “vulnerable” subgroups, Whitehurst and Lindquist conclude.

Testing more is the answer, they write.

Consider what would happen to the pervasive test-prep sessions that consume weeks of class time in many schools leading up to the end-of-the-year test if students, instead, spent an hour or so monthly being tested on content drawn from their lessons in the previous few weeks.   Under this scenario the high stakes tests blend into the tests and quizzes that good teachers have always given their students regularly, and that research shows without a doubt increase student learning.

Monthly tests (on computers, I assume) would give teachers feedback in time to adapt their teaching.

About Joanne


  1. I’ve never understood a lot of the pro-and-anti-testing hysteria. It’s seemed to me that standardized testing is terribly implemented, but teachers have been using low stakes ‘standardized tests’ for years. Weekly quizzes, spelling tests, multiplication table drills, etc, are somewhat ‘standardized’ after the teachers have taught for a few years (especially if all 4th grade teachers in a school or district use the same book), as are the more cumulative ‘end of unit’ tests. Teachers obviously teach to the test for these – nobody gives a test about adding fractions without teaching students how to add fractions.

    It seems that very frequent low stakes testing would help everybody – teachers and students would get feedback, no time would be spent on test prep, since after the first few weeks students should be familiar with the format, and teaching to the test would be the goal. The level of student stress should decrease, since ‘you have to do extra fractions homework until you can complete this’ has a very different tone that ‘you’re going to fail 3rd grade if you don’t take this seriously’. It would seem like these tests could be standardized fairly easily – maybe something like 80% is standard, 20% is whatever twist the teacher wants to add.

    I know that when I taught at a CC, students did a lot better if there were frequent small quizzes – it encouraged them to keep on top of their work, and it also alerted us to any problems before they spent too much time working with their incorrect understandings.

  2. Roger Sweeny says:

    At least at the high school level, most textbooks come with “auxiliaries”–including tests and test banks. My impression is that most teachers use them.

  3. Richard Aubrey says:

    Suggest dropping the bottom, say, 20pct of scores to exclude the intentional non-learners, the test-shy, and the mainstream kids who should not be there.
    Those kids’ scores do not or cannot reflect conventional teaching.

  4. The teacher’s job description has to be changed so that the teacher is responsible for student learning. Currently the classroom teacher is only responsible for presenting information orally. Two years later, those who didn’t learn it can be sent to a specialist for a re-try. Somewhere between 0 and 2 years, selected students can be sent to RtI resources. Currently no one has the responsibility for re-teaching failed lessons…a student could not comprehend the division lesson, and that’s it . S.O.L. for the next two years.

  5. In an ideal world, maybe. In this world? I think “Take This Test Please” rather covers it.

    Do we fund schools to educate children, or to provide education reformers/researchers with “good data?” Is Value Added more important than teaching and learning?

    Do education reformers believe they understand education better than teachers? I believe they are willfully naive about the influence of more frequent high stakes testing on classroom instruction. I watched test prep infiltrate homework for my first grader in public school–two years before NCLB testing began in third grade. I can’t imagine increasing test prep will improve results in any way.

    I listened to my children complain about the “stupid, boring” and repetitive test-prep. Increasing test prep diminishes even further the attention given to the kids who already “get it.”

    The private schools my children have attended do administer standardized tests to track student growth. They’re about 2-4 hours in total, and no one preps the student for them, other than normal classroom instruction. That classroom instruction is heavily weighted toward lots of homework, lots of writing, and classroom discussions. There are no multiple-choice tests.

    In comparison, our local public school devotes 3 weeks to testing. Students are not assigned homework the night before tests; homework before the testing period is heavy test prepping, and much of classroom instruction is devoted to completing and discussing such tests.

    Increasing scores on standardized tests after such a regimen does not mean you’ve learned more.

    • Richard Aubrey says:

      It means they learned more of what was going to be on the test.
      The actual question should be whether what is tested for reflects a useful academic goal.
      I’m sure the military isn’t a great example of something or other, although, given the competition, it’s hard to think of what that may be.
      However, when they want you to learn something, they teach it to you. And then they test you to see if you’ve learned it. If you do well, you probably have. If not, you haven’t. They don’t spend much time teaching or testing anything that isn’t likely to be useful. Which is the question; is it useful or a valuable goal to know this stuff?
      If so, why not test for it?
      I suppose the response will be that the stuff the tests test for aren’t useful in education.

      • Richard Aubrey

        I would feel more optimistic about the Common Core if the Armed Forces were in charge of it. The Armed Forces would be more pragmatic.

        The larger the domain of knowledge you’re testing, the more uncertain the outcomes. The smaller the domain, the easier it is to prep it to death.

        Consider if you had to design a history test (no holds barred.) If your subject were the history of the world, any country any era, you’d expect even good students to miss questions.

        They’d do much better if you set the domain of the test as “Atlanta, Georgia in the 19th century,” simply because there would be a smaller array of facts and themes. They’d do tremendously in comparison to others if the test were supposedly on the history of the world, but you figured out that the test emphasizes Atlanta, Georgia.

        More frequent high stakes tests will mean more test prep. How can it not, when the people pushing the tests want to use the tests to determine teacher employment? It’s setting up a bad incentive system. Such a system chooses to ignore the current bad outcomes of entire school districts condoning and encouraging outright cheating.