Testing Our Children: Findings and Recommendations

Status: 
Archived
Subject: 
K-12 Testing

FairTest based its evaluation of state assessment programs on the Principles and Indicators for Student Assessment Systems. The Principles were developed by the National Forum on Assessment to help guide assessment reform and have been signed by over 80 education and civil rights groups.

 

The major findings of Testing Our Children: A Report Card on State Assessment Systems include:

 

1) While most states now have content standards, many state tests are not based on their standards, and many important areas in their standards are not assessed.

2) Most states rely far too heavily on multiple-choice testing and fail to provide an adequate range of methods for students to demonstrate their learning. The results include not assessing important areas, creating the likelihood that those areas will not be taught. (Twenty-six states use all or almost all multiple-choice, and 18 use a majority of multiple-choice items.)

3) A majority of states (33) use norm-referenced tests, which compare students to a reference group and not to achievement on state standards. These tests fail to assess important areas of the standards and encourage grouping and instructional practices that historically have failed to provide many students with a strong education.

4) Most writing assessments require students to respond to a prompt (34 states), fostering and reporting a limited conception of writing. Writing must serve many purposes and therefore take many styles. A major problem here is the potential reduction of writing instruction to fit the state exam.

5) The state testing burden is often too heavy, with students repeatedly tested in the same subjects. A few states test students in almost every grade. For accountability purposes, such extensive testing is not necessary.

6) Seventeen states use a single test as a requirement for high school graduation, violating widely recognized standards for good assessment practice, ensuring unfair treatment of many students, and increasing the likelihood that narrow tests will dictate curriculum and instruction. Five more states are planning to use such tests. Districts also may use state tests as graduation or grade promotion hurdles.

7) Rich assessment techniques, such as portfolios and performance events, are rarely used by states. Thus, important areas of learning are not assessed and important signals are not sent to schools about what students should be learning and how assessment can support that learning.

8) Very few states use sampling for accountability, public reporting, and program improvement, even though this proven technique can provide accurate data, is less expensive and less intrusive, and allows greater use of portfolios and performance events.

9) Most states (32 of the 44 which responded to the FairTest survey) have bias review panels, often with significant authority to delete or revise items on state-made tests, though some still do not. These panels with strong authority are a positive development.

10) States tend not to adequately assess or include in state reports students with Individual Education Plans (IEP, e.g., "special education") and students with Limited English Proficiency (LEP). Inclusion of all students, using appropriate accommodations or alternate assessments as needed, is necessary for proper program evaluation and education for these students. The recently reauthorized federal Individuals with Disabilities Education Act (IDEA) will require all students with disabilities to be assessed appropriately, but such provisions do not exist for LEP students.

11) States are generally quite weak in providing adequate professional development in all aspects of assessment to teachers and other educators. Such teacher education, particularly in classroom assessment, is fundamental to assessment and broader school reform.

12) Few states evaluate teacher competence in assessment or study district, school and classroom assessment practices or their impacts. Thus, educators lack information to help improve the quality of assessment at all levels and to halt harmful practices.

13) Student and parent rights, such as the ability to review tests after completion, to challenge flawed items or to appeal scores, exist unevenly across the states. Such rights support fairness and help parents to better understand assessment and education in general and to view themselves as important partners in their children's education.

14) Public reporting and education about assessments are often limited. Furthermore, few states report in languages other than English, even if they have sizeable populations for whom English is a second language.

15) State reviews of their assessment systems need substantial improvement. Most do not study the impact of testing on curriculum, instruction, or graduation rates; and most do not review whether their assessments measure the ability of students to think critically or in complex ways in the various subject areas. In an era in which testing is proposed as a fundamental tool for school reform, states often cannot even be sure whether increasing scores are based on real learning gains or teaching to the test.

 

Trends

Among the observed trends in state testing, based on Testing Our Children and other reports, were:

 

1) The amount of state testing has remained fairly steady over the past decade.

2) The number of states requiring or planning to require graduation tests declined in the early 1990s but is rising again. The figure had declined to 17 from a high of 23. Now, if the states planning such tests implement them, by early in the next decade the number will rise back to 22. Most such states are in the south and northeast.

3) States with graduation exit exams appear less likely to make use of performance assessments.

4) There is a heavy "southern effect" which includes high stakes testing, a heavy testing load, use of an NRT, and relatively less use of constructed-response and performance assessments. As a group, the southern states still comprise the nation's poorest region, so this is also a "poverty effect."

 

If it is true that using performance assessments signals or spurs a shift toward teaching and assessing more challenging, cognitively complex material, then the southern states will be left behind. As teaching to narrow tests also has been found to most powerfully affect schools with large proportions of minority-group and low-income children, such students in these states are particularly at risk of continuing to receive a low-level education that will not prepare them well for their adult lives. Students in large cities that also emphasize teaching to traditional tests face the same risk.

 

Recommendations

Despite these problems, the opportunity for reform is not over. A great deal has been learned, some of it from pioneering efforts in a few states, some of it in districts, most of it in schools and classrooms. What is lacking is not the technical know-how, though certainly problems remain, but the political and social will to recreate assessment as part of reinventing education.

 

If large-scale assessments are to support excellence and equity in education, the underlying conceptions and basic practices in most states need to be fundamentally changed to be brought into alignment with the Principles and Indicators for Student Assessment Systems.

 

Based on these findings, FairTest makes the following recommendations:

 

1) Base all state and district assessments of student achievement on clear standards.

2) Employ multiple methods of assessment, limiting multiple-choice to no more than one quarter of test-takers' scores.

3) Rely on methods that allow students to demonstrate understanding by applying knowledge and constructing responses while ensuring assessment of complex and critical thinking in and across subject areas.

4) Do not use norm-referenced tests, or limit their use to very light sampling.

5) Do not make high-stakes decisions, such as high school graduation, using single exams as a hurdle. Rely on multiple sources of information instead.

6) Employ sampling procedures to collect information on large populations, using performance and portfolio assessments, including sampling from classroom-based measures. This should include work which allows individual choices and expressions of knowledge and provides students the opportunity to evaluate their own work.

8) Enhance efforts to appropriately include all IEP and LEP students in assessments and reporting, along with reporting disaggregated data by important population groups.

9) Ensure adequate professional development in assessment, particularly in classroom and performance assessment, for in-service teachers and students in education schools.

10) Systematically involve teachers and other educators in developing and scoring performance assessments and portfolios.

11) Institute comprehensive reviews of assessment systems and use the results to improve assessments.