What Do ‘Better Tests’ Look Like?

As U.S. schools prepare to leave bubble tests behind to enter a new era in assessments, educators everywhere hope and expect that lawmakers and districts get it right. After a decade of punitive, narrow high-stakes testing, educators and parents have had just about enough. So what are the lessons we have learned from testing? Probably too many to count, most educators might tell you. Rand Education, however, recently waded into years of research to identify and analyze the key issues and answer two fundamental questions: How has testing influenced instructional practice, and what conditions and policies have will make the impact of new assessments aligned with the Common Core State Standards more positive for teachers and students?

In the report, “New Assessments, Better Instruction?”, Rand identified major ways testing has affected classroom instruction, including changes in curriculum content and emphasis, allocation of time and resources across different pedagogical activities, and teacher-student interactions.

The curriculum has been altered in a variety of ways – a narrower content focus, favoring one specific skill-set over another, and changes in the sequence of topics. “While reduction in emphasis on social studies, art, and other subjects that are frequently omitted from high-stakes testing programs typically receives the bulk of attention from critics of testing, both forms of narrowing—across and within subjects—raise concerns about what students are missing,” the report states.

Of course, high stakes testing has affected not just the what but also the how. The Rand report goes into great detail describing the drift away from real instruction educators know all too well. Over the past decade, teachers have spent more and more time preparing students for tests by focusing less on the content and skills assessed and more on the format and structure of the test. This in moderation can be helpful, but not at the great expense of classroom time on higher learning and student engagement.

The relentless focus on test scores has had a significant impact on how teachers interact with their students. One of the most troubling aspects identified in the report is the focus on the so-called “bubble kids” – those students on the threshold of passing the test who are given a disproportionate share of individual attention – at the expense of other students who may be in greater need of extra instruction.

So what kinds of assessments would be most useful to educators and actually gauge and improve real student learning? If the Common Core-alighned assessments are to succeed, what should they look like? Rand identifies the following conditions:

  • Tests must reflect high-quality instruction. Both in format and content, test should demonstrate goals of deeper learning.
  • Use tests only for original purpose. Understanding what the test is intended to accomplish is critical. Piling on “unintended purposes” can obviously lead to faulty interpretations and lousy decisions.
  • Teachers should receive proper training and support to interpret test scores and use them effectively. This is a crucial responsibility that requires ongoing professional development, not a single “drive-by” workshop.  Educators need the time to analyze the data accurately and collaborate with peers.
  • Less high-stakes consequences. Punitive measures based on test scores alone   are corruptive and counterproductive. Instead, assessment systems must build in multiple measures that “emphasize important outcomes so that test is not the sole incentive or source of guidance,” the Rand report states.
  • Tests should be part of an integrated assessment system that includes formative components. High-stakes summative assessments cannot alone be used to evaluate student learning.  More timely formative assessments can track student progress and provide educators with the flexibility to adjust their instruction during the school year.
  • Real change depends not only on tests. “In isolation, tests can send strong signals to educators about what they should focus on and how they should teach,” according to the Rand report. But new and better assessments should only be one tool to improve student learning. Rand recommends “a coherent system of reforms” that includes curriculum and instructional materials, professional development and support for teachers, data systems, and strategies to engage parents and other community members.
  • Meg

    So, what exactly do the new CC tests do to alleviate these issues? Oh yeah…nothing. If anything, they are more high stakes then ever. More and more pressure on students, parents, administration and teachers, only with a fancier test and no funding for the technology they require.

  • These characteristics all have one thing in common– they are not about helping a teacher do his/her job better, but about helping the teacher prove to someone in a bureaucratic management role that the teacher is doing his/her job. This is not about helping to better educate students– it’s about helping students insure that classroom and school scores are kept up.

    Of course, even if we could hit these barely-adequate marks, it would be better than what we are promised under CCSS. CCSS is bringing us higher stakes, more tests, worse tests, and testing “data” that is thinly sliced and of little use.

  • This is what I think about the current data mania….


  • David Zeeman

    Dear NEA, what are you doing? What have you become? I can’t believe my union sends dues to you!

    There needs to be change in the leadership of the NEA. They are blatantly selling out teachers and public education.

    I’m going to start making noise in my union about how we can stop sending our money to the NEA.

  • Does anyone else feel like they just read through an advertisement for Smarter Balanced and/or PARCC?

  • Cheryl Binkley

    Less testing, not different testing.

  • New Teacher

    BARF! Here’s an idea, NEAToday: Let’s scrap this whole system and bring back the Trivium: Grammar, Latin, and Logic. As it stands now, your support of subpar, cheap government mind control has bred a servile population unable to read their own founding documents. While I’ve taken this course in my secondary classroom, the sad truth is that the youth are systemically dumbed down for a purpose.

    As a sidenote, at least I sleep well knowing that union vampires live well off my dues. I can barely afford to eat meat this week thanks to my student loans–i.e., living on a budget–but at least the Union sleeps well.


  • Louisiana Purchase

    Standardized tests used to stack and rank students and teachers can NEVER be better. Testing should be used only for teachers to assess students and help them, NOT to punish or label students.

  • Jim

    NEA, why ask this question? This was decided after you accepted money from Bill Gates. Stop taunting the membership with this kind of disingenuous inquiry.