High Stakes Testing

High Stakes Testing


The bill maintains and expands Ohio’s proficiency testing “accountability” system. In spite of the name changes, what are now referred to as “diagnostic tests” and “achievement tests” are likely to be very similar to the current criterion-referenced “proficiency” tests used now. In the proposed revised testing system, test questions will be written based on “standards” rather than “learning competencies.” Even curriculum experts are having difficulty identifying how a “standard” differs substantively from a “competency.” Both terms address some desirable end state of learning for students. The real educational challenge is how the identified “standard” or “competency” becomes firmly embedded in the child’s brain as a functioning concept. Listing doesn’t equate to learning. Making an intended learning outcome more specific doesn’t change the fact that children differ in the ease in which they can learn and retain information. Some children need much more time, support, reinforcement, and personal attention than other children do to learn essentially the same things.


Once the new testing program is adopted, the same large testing companies who had contracts with the state to produce the current proficiency tests will also submit proposals to develop the new “diagnostic” and “achievement” tests. The same accelerated time line for test development will be used. The demand to create a bank of questions to be changed from test to test often resulted in proficiency test questions (what teachers and parents call “bad questions”) that had not undergone adequate field-testing.

The highstakes quality of the current testing program will transfer to the new testing system. “Diagnostic” and “achievement” tests will assume the same importance in the school and community as the current “proficiency” tests, which will result in the tests themselves becoming the curriculum teachers need to emphasize to get test results that will be acceptable to their communities.

There will be confusion as to what actually constitutes improved test score performance. An arbitrary level of acceptable performance or cut score will have to be set for each test. Even if test makers manage to write better, fairer questions for the next generation of tests, if more students pass these “better” tests, the general public will accuse the state of lowering its standards. If the present high levels of performance continue to be required, students will continue to experience a level of stress that is unhealthy for some children. In summary, unless careful planning is undertaken in the development of the new testing system, any improvements to the current much-criticized state testing program will be cosmetic at best.



1.Just how effective is high-stakes testing?  

There are 13 states in the US which have instituted high stakes testing and while Texas is doing admirable well on its own TASS, its students have not done well on national tests such as the SAT, ACT, and TIMSS.

2. What is actually being taught and learned?

The University of Colorado retested students in that state last year a few weeks after the students had taken their achievements tests.  By changing just a few simple words on the questions on the  test, but without changing the  essence of the question themselves, Professor Lorrie Shepard was able to produce a sharp decline in test scores.  Again, the question was essentially the same and asked for the same information.  Conclusion: students are learning by rote memorization, they are not learning higher thinking skills.  Even Robert Schwartz, president of Achieve Inc. has acknowledged that exams linked to state standards tend to test the lowest level skills reflected in the standards.

3. Who is being helped and hurt by these tests?

Testing doesn’t tell us about students and the ways they will learn, the things they will learn, and what they will do in life.  Testing discriminates against any student who is not a good test taker.  



The third TIMSS test reported that 8th graders in well-funded American school districts score nearly as well as the top nations worldwide, beaten only by Finland in math and science, but students in poor urban areas are near the bottom.  This is not acceptable.

One Nevada high school reported over 100 drop-outs last year because of the stress of taking a graduation exam.  Two high schools in Cincinnati were reported last week (May 3) on NPR as having a drop out rate of nearly 60% because of the proficiency tests.  Texas is reporting a whopping 40% dropout rate for African-American and Latino students.  This is not acceptable

The last NAEP report suggested no substantial gain in reading scores for 4th graders since 1992 in spite of the push for high stakes testing that has been “reforming” education throughout the country over the last ten years.  


  1. In reality, testing means nothing.  it’s like judging Sammy Sosa’s baseball year on one day’s at-bats.  Tests don’t predict recognition in life, they don’t predict supervisor’s job ratings, and they don’t predict wages.  Even the SAT, designed to predict college grades, doesn’t work well.  ETS admits that the SAT accounts for only 20% of what goes into freshmen grades.  That means that 80% of what determines who makes the dean’s list and who gets academic probation comes from qualities that are not measured by tests. And, ETS admits that after the freshmen year, the SAT means nothing at all.
  2. What determines a citizen’s worth in the US includes absolutely nothing that these tests test—creativity, critical thinking, motivation, ambition, persistence, humor, attitude, reliability, politeness, enthusiasm, civic mindedness, artistry, self-awareness, self-discipline, empathy, and leadership to name just a few among many.  An April 2001 survey of Fortune 100 companies showed that 4% of the CEO’s valued high stakes testing for students—they were more interested in who had taken a lot of business courses or liberal arts courses in school, who had gotten good grades—not the test scores.
  3. What matters more in the educational life of a student is not the high stakes test he or she takes.  It’s the teacher.  Richard Rothstein of the Economic Policy Institute and Henry Levin of Columbia University have both released independent studies of the relationship between tests and employability in the US marketplace.  They have both concluded that that “the link between jobs and test scores is so small that when test scores were used for employment selection, there was a high rate of workers selected for jobs they could not perform and a high rate of workers rejected for jobs they were found later able to perform.
  4. Multiple assessments based on long-term performance are much better for determining such things as promotion/graduation of students, that end-of-course exams are difficult to design, and that criterion-referenced tests are spurious at best.
  5. Follow Maine’s and Maryland’s examples.  They are crafting stronger tests that, while linked to state standards, focus on writing tasks and other “performance” measures that push students beyond low-level skills.
  6. Genuine assessment of a student’s knowledge, employability, and civic mindedness takes time.  It considers the entirety of a child’s skills and abilities and seeks to build on strengths, while recognizing weaknesses.


Keys to New Testing:

The keys to making the new testing system work include the following: allowing adequate time for test-item development; selecting items or types of tests that fit the age of the children being tested (e.g., young children are more accurately assessed through developmental checklists rather than answering multiple-choice questions); using a test format that permits teachers to show parents exactly what a student has mastered and what the student needs to learn next; reconciling the new testing program with the state’s definition of giftedness so that gifted students do not fail state achievement tests (a phenomenon that happens all too frequently with the current proficiency  testing system).

Plea for Time:

Please  avoid quick fixes and changes that use new terminology but merely replicate the existing flawed system. Please help legislators understand that the primary effort needs to be on diagnosing what each child needs to learn well and making certain the child has the learning situation that will guarantee success. Ultimately the educational system improves by increasing what children are able to learn, retain, and use in the future, not merely by producing numbers tied to test scores. A good testing system documents real growth in learning.

Ohio Department of Education:

The current legislature is unfamiliar with ODE and its leadership role in developing curriculum, so we need to build up the department in the minds of the House and Senate. If you are so inclined, please add ideas such as new leadership, increased accountability to professional organizations (OCTELA and OCIRA), willingness to involve classroom teachers, increased staffing to match curriculum and tests to standards, and so on.


Core Curriculum:

We can develop a core curriculum for Ohio, but it should be developed by Ohio’s educators, parents. and community members.