This is the final version of the letter, which I submitted today.
July 22, 2016
The Honorable John King
Secretary of the Education Department
400 Maryland Avenue, SW
Washington, D.C. 20202
Dear Mr. Secretary:
The Every Student Succeeds Act (ESSA) marks a great opportunity for states to advance accountability systems beyond those from the No Child Left Behind (NCLB) era. The Act (Section 1111(c)(4)(B)(i)(I)) requires states to use an indicator of academic achievement that “measures proficiency on the statewide assessments in reading/language arts and mathematics.” The proposed rulemaking (§ 200.14) would clarify this statutory provision to say that the academic achievement indicator must “equally measure grade-level proficiency on the reading/language arts and mathematics assessments.”
We write this letter to argue that the Department of Education should not mandate the use of proficiency rates as a metric of school performance under ESSA. That is, states should not be limited to measuring academic achievement using performance metrics that focus only on the proportion of students who are grade-level proficient—rather, they should be encouraged, or at a minimum allowed, to use performance metrics that account for student achievement at all levels, provided the state defines what performance level represents grade level proficiency on its reading/language arts and mathematics assessments.
Moving beyond proficiency rates as the sole or primary measure of school performance has many advantages. For example, a narrow focus on proficiency rates incentivizes schools to focus on those students near the proficiency cut score, while an approach that takes into account all levels of performance incentivizes a focus on all students. Furthermore, measuring performance using the full range of achievement provides additional and useful information for parents, practitioners, researchers, and policymakers for the purposes of decisionmaking and accountability, including more accurate information about the differences among schools.
Reporting performance in terms of the percentage above proficient is problematic in several important ways. Percent proficient:
- Incentivizes schools to focus only on students around the proficiency cutoff rather than all students in a school (Booher-Jennings, 2005; Neal & Schanzenbach, 2010). This can divert resources from students who are at lower or higher points in the achievement distribution, some of whom may need as much or more support than students just around the proficiency cut score (Schwartz, Hamilton, Stecher, & Steele, 2011). This has been shown to influence which students in a state benefit (i.e., experience gains in their academic achievement) from accountability regulations (Neal & Schanzenbach, 2010).
- Encourages teachers to focus on bringing students to a minimum level of proficiency rather than continuing to advance student learning to higher levels of performance beyond proficiency.
- Is not a reliable measure of school performance. For example, percent proficient is an inappropriate measure of progress over time because changes in proficiency rates are unstable and measured with error (Ho, 2008; Linn, 2003; Kane & Staiger, 2002). The percent proficient is also dependent upon the state-determined cut score for proficiency on annual assessments (Ho, 2008), which varies from state to state and over time. Percent proficient further depends on details of the testing program that shouldn’t matter, such as the composition of the items on the state test or the type of method used to set performance standards. These problems are compounded in small schools or in subgroups that are small in size.
- Is a very poor measure of performance gaps between subgroups, because percent proficient will be affected by how a proficiency cut score on the state assessments is chosen (Ho, 2008; Holland, 2002). Indeed, prior research suggests that using percent proficient can even reverse the sign of changes in achievement gaps over time relative to if a more accurate method is used (Linn, 2007).
- Penalizes schools that serve larger proportions of low-achieving students (Kober & Riddle, 2012) as schools are not given credit for improvements in performance other than the move to proficiency from not-proficient.
We suggest two practices for measuring achievement that lessen or avoid these problems. Importantly, some of these practices were utilized by states in ESEA Flexibility Waivers and are improvements to NCLB practices (Polikoff, McEachin, Wrabel, & Duque, 2014).
Average Scale Scores
The best approach for measuring student achievement levels for accountability purposes under ESSA is to use average scale scores. Rather than presenting performance as the proportion of students who have met the minimum-proficiency cut score, states could present the average (mean) score of students within the school and the average performance of each subgroup of students. If the Department believes percent proficient is also important for reporting purposes, these values could be reported alongside the average scale scores.
The use of mean scores places the focus on improving the academic achievement of all students within a school and not just those whose performance is around the state proficiency cut score (Center for Education Policy, 2011). Such a practice also increases the amount of variation in school performance measures each year, providing for improved differentiation between schools that may have otherwise similar proficiency rates. In fact Ho (2008) argues if a single rating is going to be used for reporting on performance, it should be a measure of the average performance because such measures incorporate the value of every score (student) into the calculation and the average can be used for more advanced analyses. The measurement of gaps between key demographic groups of students, a key goal of the ESSA law, is dramatically improved with the use of average scores rather than the proportion of proficient students (Holland, 2002; Linn, 2007).
If average scale scores cannot be used, a weaker alternative that is still superior to percent proficient would be to allow states to use proficiency indexes. Schools under this policy would be allocated points based on multiple levels of performance. For example, a state could identify four levels of performance on annual assessments: Well Below Proficient, Below Proficient, Proficient, and Advanced Proficient. Schools receive no credit for students Well Below Proficient, partial credit for students who are Below Proficient, full credit for students reaching Proficiency, and additional credit for students reaching Advanced Proficiency. Here we present an example using School A and School B.
|Proficiency Index Example|
|School A||School B|
Points Per Student
# of Students
Points Per Student
# of Students
|Well Below Proficient||0.0||27||0.0||0.0||18||0.0|
|NCLB Proficiency Rate: 55%
ESSA Proficiency Index: 75
|NCLB Proficiency Rate: 55%
ESSA Proficiency Index: 83
Under NCLB proficiency rate regulations, both School A and School B would have received a 55% proficiency rate score. Using a Proficiency Index, the performance of these schools would no longer be identical. A state would be able to compare the two schools while simultaneously identifying annual meaningful differentiation in the performance of School A from that of School B. The hypothetical case presented here is not the only way a proficiency index can be used. Massachusetts is one example of a state that has used a proficiency index for the purposes of identifying low-performing schools and gaps between subgroup of students (see: ESEA Flexibility Request: Massachusetts, page 32). These indexes are understandable for practitioners, family members, and administrators while also providing additional information regarding the performance of students who are not grade-level proficient.
The benefits of using such an index, relative to using the proportion of proficient students in a school, is that it incentivizes a focus on all students, not just those around an assessment’s proficiency cut score (Linn, Baker, & Betebenner, 2002). Moreover, schools with large proportions of students well-below the proficiency cut score are given credit for moving students to higher levels of performance even if still below the cut score (Linn, 2003). The use of a proficiency index or providing schools credit for students at different points in the achievement distribution improves the construct validity of the accountability measures over the NCLB proficiency rate measures (Polikoff et al., 2014). In other words, the inferences made about schools (e.g., low-performing or bottom 5%) using the proposed measures are more appropriate than those made using proficiency rates alone.
What We Recommend
Given the findings cited above, we believe the Department of Education should revise its regulations to one of two positions:
- Explicitly endorsing or encouraging states to use one of the two above-mentioned approaches as an alternative to proficiency rates as the primary measure of school performance. Average scale scores is the superior method.
- Failing that, clarifying that the law is neutral about the use of proficiency rates versus one of the two above-mentioned alternatives to proficiency rates as the primary measure of school performance.
With the preponderance of evidence showing that schools and teachers respond to incentives embedded in accountability systems, we believe option 1 is the best choice. This option leaves states the authority to determine school performance how they see fit but encourages them to incorporate what we have learned through research about the most accurate and appropriate way to measure school performance levels.
Our Recommendation is Consistent with ESSA
Section 1111(c)(4)(A)) of ESEA, as amended by ESSA, requires each state to establish long-term goals:
“(i) for all students and separately for each sub- group of students in the State—
(I) for, at a minimum, improved—
(aa) academic achievement, as measured by proficiency on the annual assessments required under subsection (b)(2)(B)(v)(I);”
And Section 1111(c)(4)(B) of ESEA requires the State accountability system to have indicators that are used to differentiate all public schools in the State, including—(i) “academic achievement—(I) as measured by proficiency on the annual assessments required [under other provisions of ESSA].”
Our suggested approach is supportable under these provisions based on the following analysis. The above-quoted provisions in the law that mandate long-term goals and indictors of student achievement based on proficiency on annual assessments do not prescribe how a state specifically uses the concept of proficient performance on the state assessments. The statute does not prescribe that “proficiency” be interpreted to compel differentiation of schools based exclusively on “proficiency rates.” Proficiency is commonly taken to mean “knowledge” or “skill” (Merriam Webster defines it as “advancement in knowledge or skill” or “the quality or state of being proficient”, where “proficient” is defined as “well advanced in an art, occupation, or branch of knowledge”). Under either of these definitions, an aggregate performance measure such as the two options described above would clearly qualify as involving a measure of proficiency. Both of the above-mentioned options provide more information about the average proficiency level of a school than an aggregate proficiency rate. Moreover, they address far more effectively than proficiency rates the core purposes of ESSA, including incentivizing more effective efforts to educate all children and providing broad discretion to states in designing their accountability systems.
We would be happy to provide more information on these recommendations at your pleasure.
Morgan Polikoff, Ph.D., Associate Professor of Education, USC Rossier School of Education
Educational Researchers and Experts
Alice Huguet, Ph.D., Postdoctoral Fellow, School of Education and Social Policy, Northwestern University
Andrew Ho, Ph.D., Professor of Education, Harvard Graduate School of Education
Andrew Saultz, Ph.D., Assistant Professor, Miami University (Ohio)
Andrew Schaper, Ph.D., Senior Associate, Basis Policy Research
Anna Egalite, Ph.D., Assistant Professor of Education, North Carolina State University
Arie van der Ploeg, Ph.D., retired Principal Researcher, American Institutes for Research
Cara Jackson, Ph.D., Assistant Director of Research & Evaluation, Urban Teachers
Christopher A. Candelaria, Ph.D., Assistant Professor of Public Policy and Education, Vanderbilt University
Cory Koedel, Ph.D., Associate Professor of Economics and Public Policy, University of Missouri
Dan Goldhaber, Ph. D., Director, Center for Education Data & Research, University of Washington Bothell
Danielle Dennis, Ph.D., Associate Professor of Literacy Studies, University of South Florida
Daniel Koretz, Ph.D., Henry Lee Shattuck Professor of Education, Harvard Graduate School of Education
David Hersh, Ph.D. Candidate, Rutgers University Bloustein School of Planning and Public Policy
David M. Rochman, Research and Program Analyst, Moose Analytics
Edward J. Fuller, Ph.D., Associate Professor of Education Policy, The Pennsylvania State University
Eric A. Houck, Associate Professor of Educational Leadership and Policy, University of North Carolina at Chapel Hill
Eric Parsons, Ph.D., Assistant Research Professor, University of Missouri
Erin O’Hara, former Assistant Commissioner for Data & Research, Tennessee Department of Education
Ethan Hutt, Ph.D., Assistant Professor of Education, University of Maryland College Park
Eva Baker, Ed.D., Distinguished Research Professor, UCLA Graduate School of Education and Information Studies, Director, Center for Research on Evaluation, Standards, and Student Testing, Past President, American Educational Research Association
Greg Palardy, Ph.D., Associate Professor, University of California, Riverside
Heather J. Hough, Ph.D., Executive Director, CORE-PACE Research Partnership
Jason A. Grissom, Ph.D., Associate Professor of Public Policy and Education, Vanderbilt University
Jeffrey Nellhaus, Ed.M., Chief of Assessment, Parcc Inc., former Deputy Commissioner, Massachusetts Department of Elementary and Secondary Education
Jeffrey W. Snyder, Ph.D., Assistant Professor, Cleveland State University
Jennifer Vranek, Founding Partner, Education First
John A. Epstein, Ed.D., Education Associate Mathematics, Delaware Department of Education
John Q. Easton, Ph.D., Vice President, Programs, Spencer Foundation, former Director, Institute of Education Sciences
John Ritzler, Ph.D., Executive Director, Research & Evaluation Services, South Bend Community School Corporation
Jonathan Plucker, Ph.D., Julian C. Stanley Professor of Talent Development, Johns Hopkins University
Joshua Cowen, Ph.D., Associate Professor of Education Policy, Michigan State University
Katherine Glenn-Applegate, Ph.D., Assistant Professor of Education, Ohio Wesleyan University
Linda Darling-Hammond, Ed.D., President, Learning Policy Institute, Charles E. Ducommun Professor of Education Emeritus, Stanford University, Past President, American Educational Research Association
Lindsay Bell Weixler, Ph.D., Senior Research Fellow, Education Research Alliance for New Orleans
Madeline Mavrogordato, Ph.D., Assistant Professor, K-12 Educational Administration, Michigan State University
Martin R. West, Ph.D., Associate Professor, Harvard Graduate School of Education
Matt Chingos, Ph.D., Senior Fellow, Urban Institute
Matthew Di Carlo, Ph.D., Senior Fellow, Albert Shanker Institute
Matthew Duque, Ph.D., Data Strategist, Baltimore County Public Schools
Matthew A. Kraft, Ed.D., Assistant Professor of Education and Economics, Brown University
Michael H. Little, Royster Fellow and Doctoral Student, University of North Carolina at Chapel Hill
Michael Hansen, Ph.D., Senior Fellow and Director, Brown Center on Education Policy, Brookings Institution
Michael J. Petrilli, President, Thomas B. Fordham Institute
Nathan Trenholm, Director of Accountability and Research, Clark County (NV) School District
- Tiên Lê, Doctoral Fellow, USC Rossier School of Education
Raegen T. Miller, Ed.D., Research Fellow, Georgetown University
Russell Brown, Ph.D., Chief Accountability Officer, Baltimore County Public Schools
Russell Clement, Ph.D., Research Specialist, Broward County Public Schools
Sarah Reckhow, Ph.D., Assistant Professor of Political Science, Michigan State University
Sean P. “Jack” Buckley, Ph.D., Senior Vice President, Research, The College Board, former Commissioner of the National Center for Education Statistics
Sherman Dorn, Ph.D., Professor, Mary Lou Fulton Teachers College, Arizona State University
Stephani L. Wrabel, Ph.D., USC Rossier School of Education
Thomas Toch, Georgetown University
Tom Loveless, Ph.D., Non-resident Senior Fellow, Brookings Institution
Alexander McNaughton, History Teacher, YES Prep Charter School, Houston, TX
Andrea Wood Reynolds, District Testing Coordinator, Northside ISD, TX
Angela Atkinson Duina, Ed.D., Title I School Improvement Coordinator, Portland Public Schools, ME
Ashley Baquero, J.D., English/Language Arts Teacher, Durham, NC
Brett Coffman, Ed.S., Assistant Principal, Liberty High School, MO
Callie Lowenstein, Bilingual Teacher, Washington Heights Expeditionary Learning School, NY
Candace Burckhardt, Special Education Coordinator, Indigo Education
Daniel Gohl, Chief Academic Officer, Broward County Public Schools, FL
Danielle Blue, M.Ed., Director of Preschool Programming, South Kingstown Parks and Recreation, RI
Jacquline D. Price, M.Ed., County School Superintendent, La Paz County, AZ
Jennifer Taubenheim, Elementary Special Education Teacher, Idaho Falls, ID
Jillian Haring, Staff Assistant, Broward County Public Schools, FL
Juan Gomez, Middle School Math Instructional Coach Carmel High School, Carmel, CA
Mahnaz R. Charania, Ph.D., GA
Mary F. Johnson, MLS, Ed.D., Retired school librarian
MaryEllen Falvey, M.Ed, NBCT, Office of Academics, Broward County Public Schools, FL
Meredith Heikes, 6th grade STEM teacher, Quincy School District, WA
Mike Musialowski, M.S., Math/Science Teacher, Taos, NM
Misty Pier, Special Education Teacher, Eagle Mountain Saginaw ISD, TX
Nell L. Forgacs, Ed.M., Educator, MA
Oscar Garcia, Social Studies Teacher, El Paso Academy East, TX
Patricia K. Hadley, Elementary School Teacher, Retired, Twin Falls, ID
Samantha Arce, Elementary Teacher, Phoenix, AZ
Theodore A. Hadley, High School/Middle School Teacher, Retired, Twin Falls, ID
Tim Larrabee, M.Ed., MAT, Upper Elementary Teacher, American International School of Utah
Troy Frystak, 5/6 Teacher, Springwater Environmental Sciences School, OR
Other Interested Parties
Arnold F. Shober, Ph.D., Associate Professor of Government, Lawrence University
Celine Coggins, Ph. D., Founder and CEO, Teach Plus
David Weingartner, Co-Chair Minneapolis Public Schools 2020 Advisory Committee
Joanne Weiss, former chief of staff to U.S. Secretary of Education Arne Duncan
Justin Reich, EdD, Executive Director, Teaching Systems Lab, MIT
Karl Rectanus, CEO, Lea(R)n, Inc.
Kenneth R. DeNisco, Ph.D., Associate Professor, Physics & Astronomy, Harrisburg Area Community College
Kimberly L. Glass, Ph.D., Pediatric Neuropsychologist, The Stixrud Group
Mark Otter, COO, VIF International Education
Patrick Dunn, Ph.D., Biomedical Research Curator, Northrop Grumman TS
Robert Rothman, Education Writer, Washington, DC
Steven Gorman, Ph.D., Program Manager, Academy for Lifelong Learning, LSC-Montgomery
Torrance Robinson, CEO, trovvit
Booher-Jennings, J. (2005). Below the bubble: “Educational triage” and the Texas accountability system. American Educational Research Journal, 42(1), 231–268.
Center on Education Policy. (2011, May 3). An open letter from the Center on Education Policy to the SMARTER Balanced Assessment Consortium and the Partnership for Assessment of Readiness for College and Career. Retrieved from http://cep-dc.org/displayDocument.cfm?DocumentID=359
Ho, A. D. (2008). The problem with “proficiency”: Limitations of statistics and policy under No Child Left Behind. Educational Researcher, 37(6), 351–360.
Holland, P. W. (2002). Two measures of change in the gaps between the CDFs of test-score distributions. Journal of Educational Behavioral Statistics, 27(1), 3–17.
Kober, N., & Riddle, W. (2012). Accountability issues to watch under NCLB waivers. Washington, DC: Center on Education Policy.
Linn, R. L. (2003). Accountability: Responsibility and reasonable expectations. Educational Researcher, 32(7), 3–13.
Linn, R. L. (2007). Educational accountability systems. Paper presented at the The CRESST Conference: The Future of Test-Based Educational Accountability.
Linn, R. L., Baker, E. L., & Betebenner, D. W. (2002). Accountability systems: Implications of requirements of the No Child Left Behind Act of 2001. Educational Researcher, 31(6), 3–16.
Neal, D., & Schanzenbach, D. W. (2010). Left behind by design: Proficiency counts and test-based accountability. Review of Economics and Statistics, 92, 263–283.
Ng, H. L., & Koretz, D. (2015). Sensitivity of school-performance ratings to scaling decisions. Applied Measurement in Education, 28(4), 330–349.
Polikoff, M. S., McEachin, A., Wrabel, S. L., & Duque, M. (2014). The waive of the future? School accountability in the waiver era. Educational Researcher, 43(1), 45–54. http://doi.org/10.3102/0013189X13517137
Schwartz, H. L., Hamilton, L. S., Stecher, B. M., & Steele, J. L. (2011). Expanded measures of school performance. Santa Monica, CA: The RAND Corporation.