Tennessee Journal of Law and Policy Tennessee Journal of Law and Policy
Volume 13 Issue 2 Article 3
July 2021
Tennessee's National Impact on Teacher Evaluation Law & Policy: Tennessee's National Impact on Teacher Evaluation Law & Policy:
An Assessment of Value-Added Model Litigation An Assessment of Value-Added Model Litigation
Mark Paige
University of Massachusetts - Dartmouth
Audrey Amrein Beardsley
Arizona State University
, audrey[email protected]
Kevin Close
Arizona State University
Follow this and additional works at: https://ir.law.utk.edu/tjlp
Part of the Law Commons
Recommended Citation Recommended Citation
Paige, Mark; Beardsley, Audrey Amrein; and Close, Kevin (2021) "Tennessee's National Impact on Teacher
Evaluation Law & Policy: An Assessment of Value-Added Model Litigation,"
Tennessee Journal of Law and
Policy
: Vol. 13 : Iss. 2 , Article 3.
Available at: https://ir.law.utk.edu/tjlp/vol13/iss2/3
This Article is brought to you for free and open access by Volunteer, Open Access, Library Journals (VOL Journals),
published in partnership with The University of Tennessee (UT) University Libraries. This article has been accepted
for inclusion in Tennessee Journal of Law and Policy by an authorized editor. For more information, please visit
https://ir.law.utk.edu/tjlp.
Tennessee Journal of Law and Policy Tennessee Journal of Law and Policy
Volume 13
Issue 2
(Winter 2019)
Article 4
February 2019
Tennessee's National Impact on Teacher Evaluation Law & Policy: Tennessee's National Impact on Teacher Evaluation Law & Policy:
An Assessment of Value-Added Model Litigation An Assessment of Value-Added Model Litigation
Mark A. Paige
University of Massachusetts - Dartmouth
Audrey Amrein-Beardsley
Mary Lou Fulton Teachers College
Kevin Close
Arizona State University
Follow this and additional works at: https://trace.tennessee.edu/tjlp
Part of the Education Law Commons, and the Law and Politics Commons
Recommended Citation Recommended Citation
Paige, Mark A.; Amrein-Beardsley, Audrey; and Close, Kevin (2019) "Tennessee's National Impact on
Teacher Evaluation Law & Policy: An Assessment of Value-Added Model Litigation,"
Tennessee Journal of
Law and Policy
: Vol. 13 : Iss. 2 , Article 4.
Available at: https://trace.tennessee.edu/tjlp/vol13/iss2/4
This Article is brought to you for free and open access by Volunteer, Open Access, Library Journals (VOL Journals),
published in partnership with The University of Tennessee (UT) University Libraries. This article has been accepted
for inclusion in Tennessee Journal of Law and Policy by an authorized editor. For more information, please visit
https://trace.tennessee.edu/tjlp.
1
[523]
TENNESSEE JOURNAL
OF LAW AND POLICY
VOLUME 13 WINTER 2019 ISSUE 2
ARTICLE
TENNESSEE’S NATIONAL IMPACT
ON TEACHER EVALUATION LAW &
POLICY
AN ASSESSMENT OF VALUE-ADDED MODEL
LITIGATION
Mark A. Paige
*
Audrey Amrein-Beardsley
**
Kevin Close
†
Abstract
Over the last decade or so, federal and state education
policymakers embraced the use of value-added models
(VAMs) to evaluate teachers’ performance and make high-
stakes employment decisions (e.g., tenure, merit pay,
*
Mark A. Paige, J.D., Ph.D., is an associate professor of public
policy at the University of Massachusetts-Dartmouth. Before
becoming a professor, he was an education law attorney
representing school districts.
**
Audrey Amrein-Beardsley, Ph.D., is currently a Professor in
Mary Lou Fulton Teachers College. Her research interests
include educational policy, educational measurement and
research methods.
†
Kevin Close, M.Ed., is a doctoral candidate and graduate
assistant researcher at the Arizona State University.
12
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[524]
termination of employment). VAMs are complicated
statistical models that attempt to estimate a teacher’s
contribution to student test scores, particularly those in
mathematics and reading. Educational researchers, as
well as many teachers and unions, however, have objected
to the use of VAMs noting that these models fail to
adequately account for variables outside of teachers’
control that contribute to a student’s education
performance. Subsequently, many teachers challenged the
use of VAMs through the courts. This article assesses those
challenges.
I. Introduction 525
II. VAMS: Promise and Controversy 530
A. A Brief History of VAMs in Educational Policy 530
1. The Rise of VAMs in National Education Policy:
Race to the Top 532
B. Statistical and Practical Controversies 535
1. Reliability 535
2. Validity 537
3. Bias 538
4. Transparency 540
5. Fairness 542
6. Consequential Use 543
7. Intended Consequences 544
8. Unintended Consequences 545
III. The Cases 547
A. Federal Substantive Due Process Rights & Equal
Protection Arguments: VAMs May Be Unwise But
Still Constitutional 547
1. Cook v. Bennett 547
2. Trout v. Knox County Board of Education 551
3. Wagner v. Haslam 553
4. Matter of Lederman v. King 554
B. Legislative State Agency Authority Questioned 555
1. Leff v. Clark County School District 555
2. Stapleton v. Skandera 557
3. Louisiana Federation of Teachers v. State 559
23
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[525]
4. Robinson v. Stewart 560
5. Filed but not Adjudicated 561
C. Process & “Fundamental Fairness” Cases 563
1. Houston Federation of Teachers 563
2. Washington Teachers’ Union v. D.C. Public
Schools 566
IV. Current Policy Landscape in Wake of ESSA 569
A. ESSA Reauthorization 570
B. State Plans 572
V. Conclusions 573
I. Introduction
In March of 2017, William “Bill” Sanders passed
away in Tennessee.
1
To most policymakers outside of
education (and many within it) he was a relatively
unknown statistician. His work in education policy
started far away from schoolhouses. Indeed, after he
received his degree in statistics at the University of
Tennessee, he began assessing the impact of radiation on
farm animals.
2
But his career trajectory changed markedly. In
1982, after reading a newspaper article about how
Tennessee Governor Lamar Alexander sought a model of
teacher compensation that would pay teachers for
performance, Mr. Sanders concluded he had the answer.
3
He wrote to Alexander explaining that he developed a
statistical model that could determine who the “best”
teachers were—a so-called “value-added” model (e.g., the
Tennessee Value-Added Assessment System (TVAAS)
1
Kevin Carey, The Little-Known Statistician Who Taught Us
to Measure Teachers, N.Y. TIMES (May 19, 2017),
https://www.nytimes.com/2017/05/19/upshot/the-little-known-
statistician-who-transformed-education.html [https://perma.
cc/2VBF-CZWY].
2
Id.
3
Id.
34
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[526]
which is more generally known as the Education Value-
Added Assessment (EVAAS)).
4
This model estimates a
teacher’s contribution to student achievement on
standardized tests,
5
and it formed the basis for his
private company that developed algorithms for the
models.
6
Tennessee ultimately incorporated value added
models into policies and laws, linking high-stakes
employment decisions and evaluation to student test
scores.
7
Mr. Sanders’s models—sparked by this random
collision of events—has had profound impact on national
educational policy. In 2009, President Obama’s Race to
the Top (RttT) program conditioned state receipt of
federal education dollars on states’ use of VAMs to
evaluate and make employment decisions for teachers.
States seeking much-needed
federal money during the
4
Id. VAMs have a policy history that precede Mr. Sanders’s
adoption of the term in education. They had been used in
economics since the 1960s. See, e.g, Douglas Harris, Would
Accountability Based on Teacher Value Added Be Smart
Policy? An Examination of the Statistical Properties and Policy
Alternatives, 4 J. EDUC. FIN. & POL’Y 319, 321 (2009). Yet
Sanders is widely credited as the one who popularized the use
of VAMs for educational accountability. E.g., Carey supra note
1.
5
E.g., EDWARD WILEY, A PRACTITIONER’S GUIDE TO VALUE
ADDED ASSESSMENT 5 (2006) https://nepc.colorado.edu/
publication/a-practitioners-guide-value-added-assessment-
educational-policy-studies-laboratory-resea [https://perma.cc/
EH6R-S7QN].
6
SAS® EVAAS® FOR K-12, https://www.sas.com/en_si/
software/evaas.html [https://perma.cc/65TE-VEFG] (crediting
the development of this particular model sold by a private
company to Mr. Sanders).
7
TENN. CODE ANN. §§ 49-1-302(a)(2)(C), 49-5-503(4) (2016);
TENN. STATE BD. OF EDUC., TEACHER AND ADMINISTRATOR
POLICY § 5.201 (2017) (statutory and regulatory framework
delegating authority to state department of education to
develop policy for evaluation and further linking that
evaluation to tenure determinations).
45
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[527]
“Great Recession” eagerly complied.
8
As a consequence,
VAMs became codified in state teacher evaluation and
employment laws across the country.
9
Despite their widespread adoption, the use of
these statistical models in improving public schools is a
source of considerable debate in law and policy. Some
scholars applaud their use, arguing that they provide a
clear measure of a teacher’s worth and address a
persistent policy dilemma: How to improve the quality of
our public school teachers.
10
Detractors insist that a
teacher’s value is much more than the measure of test
scores and, more importantly, that VAMs are statistically
flawed.
11
Critics note that VAMs fail to account for the
complexity of teaching and cannot accurately control for
the impact of other variables (e.g., students’ individual
8
See generally Rhoda Freelon et al., Overburdened and
Underfunded: California Public Schools Amidst the Great
Recession, 2 MULTIDISCIPLINARY J. EDUC. RES., 152 (2012)
(documenting the impact of the Great Recession on public
schools in California, but also noting the broader impact of the
recession on schools and institutions beyond California).
9
KATHRYN M. DOHERTY & SANDI JACOBS, STATE OF THE STATES
2013: CONNECT THE DOTS: USING EVALUATIONS OF TEACHER
EFFECTIVENESS TO INFORM POLICY AND PRACTICE 10 (2013)
(noting that in 2013 at least 31 states had adopted the use of
standardized test in their teacher evaluation protocols); see
also MARK A. PAIGE, BUILDING A BETTER TEACHER:
UNDERSTANDING VALUE-ADDED MODELS IN THE LAW OF
TEACHER EVALUATION 15, 16 (2016) (describing the links
between teacher evaluation systems and teacher employment
statutes, such as tenure, and warning against such use for
high-stakes decisions).
10
See, e.g., Eric A. Hanushek, Conceptual and Empirical Issues
in the Estimation of Educational Production Functions, 14 J.
HUM. RESOURCES 351, 353 (arguing for the adoption of
production function models to evaluate teachers).
11
E.g., Linda Darling-Hammond, Can Value-Added Add Value
to Teacher Evaluation?, 44 EDUC. RESEARCHER 132, 133
(placing the use of value added models in the larger policy
debate about how to improve teacher quality).
56
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[528]
motivation) that impact student achievement.
12
Because
of these issues, commentators cautioned against the use
of VAMs in high-stakes employment decisions (e.g.
termination), noting such use would invite legal action.
13
Notwithstanding these warnings, many states
embraced VAMs. Florida, for example, amended their
teacher evaluation statutes to ensure that VAMs played
a controlling role in teacher employment status,
including tenure decisions.
14
Teachers and unions almost
immediately challenged the use of VAMs through legal
means. Lawsuits ranged from violations of the Federal
Constitution
15
to assertions that requirements to use
VAMs violated the non-delegability doctrine.
16
Many of
these received widespread attention in the popular
press.
17
12
Id.; see also SEAN P. CORCORAN, CAN TEACHERS BE
EVALUATED BY THEIR STUDENTS’ TEST SCORES? SHOULD THEY
BE? THE USE OF VALUE-ADDED MEASURES OF TEACHER
EFFECTIVENESS IN POLICY AND PRACTICE 22 (2010).
13
PAIGE, supra note 9, at 22 n.28; see also Preston C. Green III
et al., The Legal and Policy Implication of Value-Added
Teacher Assessment Policies, 2012 BYU EDUC. & L.J. 1, 15–16
(2012).
14
E.g., FLA. STAT. ANN. § 1012.22(1)(c)(5) (West 2013)
(connecting teacher salary to an evaluation system that
requires use of VAMs).
15
E.g., Cook v. Bennett, 792 F.3d 1294, 1298 (11th Cir. 2015)
(alleging use of VAMs violated substantive and procedural due
process clauses, as well as the Equal Protection Clause of the
14th Amendment).
16
E.g., State ex rel. Stapleton v. Skandera, 346 P.3d 1191, 1194
(N.M. App. 2015).
17
E.g., Peter Greene, Over a Year Ago a Federal Court Struck
Down VAM: Why Are We Still Using it to Evaluate Teachers?,
FORBES (June 25, 2018, 08:23 PM), https://www.forbes.com/
sites/petergreene/2018/06/25/over-a-year-ago-a-federal-court-
struck-down-vam-why-are-we-still-using-it-to-evaluate-teachers/
[https://perma.cc/AA4M-NRQ5]; Patricia MacGregor-Mendoza,
Court Finds Teacher Evaluation System Flawed, LAS CRUCES
SUN NEWS (May 26, 2017, 07:17 PM), https://www.lcsun-
67
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[529]
It has been almost ten years since Race to the Top
brought Mr. Sander’s idea of VAMs from Tennessee to a
national scale, and it seems an appropriate moment to
assess their legal and policy ramifications. Indeed, as we
note, the use of VAMs has triggered a wave of litigation
and policy change that continues today. Many states
continue to use VAMs, while others have reduced their
use under new federal laws.
18
Thus, assessing the legal
and policy landscape forms the basis of this article.
Generally speaking, three lines of legal challenges
have emerged. First, some are grounded in the
substantive Due Process Clause and Equal Protection
Clause of the 14th Amendment, arguing that the laws do
not pass rational basis scrutiny.
19
Second, a line of cases
challenges the authority or jurisdiction of a particular
agency (e.g., state Department of Education) to enact
evaluation regulations or laws that use VAMs. Third,
some cases advance what we refer to as “process”
arguments. These contend that the use of VAMs violates
some agreed-upon or standing procedural terms found in
the Procedural Due Process Clause or collective
bargaining agreements (CBAs). As we note, plaintiffs
have captured the most success (although not always) on
this third line of argument.
That litigants have experienced more success
arguing VAMs offend certain procedural protections
comports with common understanding of procedural due
news.com/story/opinion/2017/05/26/court-finds-teacher-
evaluation-system-flawed/102219102/ [https://perma.cc/ESS8-
SXWX];Valerie Strauss, Judge Calls Evaluation of N.Y.
Teacher “Arbitrary” and “Capricious” in Case Against New U.S.
Secretary of Education, WASH. POST (May 10, 2016),
https://www.washingtonpost.com/news/answer-sheet/wp/
2016/05/10/judge-calls-evaluation-of-n-y-teacher-arbitrary-
and-capricious-in-case-against-new-u-s-secretary-of-
education/ [https://perma.cc/Y645-2T82].
18
See infra Part III.
19
See, e.g., Cook, 792 F.3d at 1298, 1300.
78
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[530]
process. At its core, procedural due process ensures
“fundamental fairness” when the government moves to
take away a protected interest, such as employment.
While courts generally have not overruled a legislature’s
policy choice to use VAMs as violative of the substantive
due process, they (including a federal appeals court case)
have questioned the wisdom of the legislature’s
decision.
20
Where they have overturned the use of VAMs,
they have done so on procedural grounds.
21
This allows
courts to stay within “their lane” and avoid jurisdictional
overreach into the policy area.
The article is organized as follows. Part I
overviews VAMs, their link to teacher evaluation and
employment, and the controversy surrounding their use,
especially as a factor in high-stakes employment
decisions. Part II provides the most current assessment
of cases where the statistical controversy has led to legal
action. Part III discusses the recent policy and legal
developments with respect the use of VAMs in evaluation
that have occurred because changes in federal education
law. In conclusion, we note that VAMs have receded,
somewhat, in terms of their role in evaluation and
employment matters.
II. VAMs: Promise and Controversy
A. A Brief History of VAMs in Educational
Policy
In the simplest of terms, VAMs (e.g., Tennessee’s
TVAAS) are statistical models used to measure the
predicted and the actual “value” a teacher “adds” to (or
detracts from) student achievement from the point at
which students enter a teacher’s classroom to the point
students leave. This is typically done using student
20
See id. at 1301.
21
See id. at 1301–02.
89
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[531]
achievement growth as measured by large-scale
standardized test scores (i.e., the tests mandated by the
No Child Left Behind (NCLB) Act of 2001). The models
attempt to statistically control for outside variables,
including students’ prior test performance, and student-
level background variables (e.g., whether students are
eligible for free-and-reduced lunches).
22
The most widely used VAM is the EVAAS,
developed and used in Tennessee.
23
EVAAS
comes in
different versions for different states (e.g., the EVAAS in
Ohio, North Carolina, and South Carolina, the PVAAS in
Pennsylvania, the TVAAS in Tennessee, and the
TxVAAS in Texas) and different ones based on large and
small school districts (e.g., located within Arkansas,
Georgia, Indiana, Texas, and Virginia). For each
consumer, EVAAS modelers choose one of two
sophisticated statistical models.
24
Using these models, student growth scores are
aggregated at the teacher or classroom level to yield
teacher-level value-added estimates. Depending on where
22
See e.g., Sean Corcoran & Dan Goldhaber, Value Added and
Its Uses: Where You Stand Depends on Where You Sit, 8 EDUC.
FIN. & POL’Y 418, 421 (2013). Other variables include things
such as, English language learners (ELLs), gifted, receiving
special education services, and classroom and school-level
variables (e.g., class sizes, school resources, school leadership).
23
The EVAAS is advertised as “the most comprehensive
reporting package of value-added metrics available in the
educational market” in that the EVAAS offers states, districts,
and schools “precise, reliable and unbiased results that go far
beyond what other simplistic [value-added] models found in the
market today can provide.” SAS® EVAAS ® FOR K-12,
https://www.sas.com/en_us/software/evaas.html [https://perma.cc/
76AY-G47W].
24
For a comprehensive statistical summation of the various
models and options available, see WHITE PAPER: SAS®
EVAAS® FOR K12 STATISTICAL MODELS, https://www.sas.
com/content/dam/SAS/en_us/doc/whitepaper1/sas-evaas-k12-
statistical-models-107411.pdf [https://perma.cc/F5EW-WCB6].
910
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[532]
teachers’ EVAAS
estimates fall, as compared to similar
teachers to whom they are compared (e.g., within
districts) at the same time, teachers’ value-added
determinations are made.
25
Thereafter, EVAAS modelers
make relativistic comparisons and rank teachers
hierarchically along a continuum.
26
Teachers whose
students grow significantly more than the average and/or
surpass projected levels of growth are identified as
“adding value”; teachers whose students grow
significantly less and/or fall short of projected levels are
identified as “detracting value.”
27
Teachers whose
students grow at rates that are not statistically different
from average (i.e., falling within one standard deviation
of the mean) are classified as Not Detectibly Different
(NDD).
28
1. The Rise of VAMs in National Education
Policy: Race to the Top
In 2007, TVAAS/EVAAS entered the national
education policy discussion when developer Dr. William
L. Sanders shared his research with Congress.
Specifically, he testified before the U.S. House of
Representatives Committee on Education and the
Workforce on how TVAAS could improve teacher
25
For a general overview of the use of VAMs and the concepts
noted herein, see WILEY, supra note 5.
26
Id.
27
Id.; Audrey Amrein-Beardsley & Clarin Collins, The SAS
Education Value-Added Assessment System (SAS® EVAAS®)
in the Houston Independent School District (HISD): Intended
and Unintended Consequences, 20 EDUC. POL’Y ANALYSIS
ARCHIVES, no. 12, Apr. 2012, at 1, 7 n.2.
28
WILEY, supra note 5; Amrein-Beardsley & Collins, supra
note 27, at 7 n.2; see, e.g., WILLIAM L. SANDERS, COMPARISONS
AMONG VARIOUS EDUCATIONAL ASSESSMENT VALUE-ADDED
MODELS 18 (2006).
1011
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[533]
accountability and promote educational reform.
29
His
testimony spurred the U.S. Department of Education’s
piloting of VAMs.
30
The use of VAMs nationally grew under the Race
to the Top program. By way of background, RttT was a
competitive federal grant program that amounted to an
injection of $4.35 billion to selected states to support
educational reform efforts.
31
Receipt of the grant was
conditioned on states developing teacher evaluation laws
and policy that used VAMs.
32
States that attached
relatively more serious consequences (e.g., employment
status) to teachers’ VAM-based output were viewed more
favorably than those that did not.
33
High-stakes
consequences included, but were not limited to: teachers’
permanent files being flagged, thus preventing teachers
from changing jobs within states; the revocation of
teacher licenses; teacher tenure; salary increases,
decreases, and merit pay; and teacher probation and
termination.
34
Beyond RttT, the federal government used other
mechanisms to embed VAMs in state evaluation and
employment matters as a matter of law and policy. In
2011, the federal government required that states adopt
the accountability practices discussed above
29
CHRISTOPHER B. SWANSON & JANELLE BARLAGE, INFLUENCE:
A STUDY OF THE FACTORS SHAPING EDUCATION POLICY 41
(2016), https://secure.edweek.org/media/influence_study.pdf
[https://perma.cc/346S-HJSX].
30
Id.
31
U.S. DEP’T OF EDUC., RACE TO THE TOP FACT SHEET (2009),
https://www2.ed.gov/programs/racetothetop/factsheet.pdf
[https://perma.cc/35GG-Y3HM].
32
Id.
33
Arne Duncan, Sec’y, Dep’t of Educ., Remarks at The Race to
the Top Program Announcement: The Race to the Top Begins
(July 4, 2009), https://www.ed.gov/news/speeches/race-top-begins
[https://perma.cc/3RD5-RP7A].
34
See generally PAIGE, supra note 9 (noting that VAMs became
required factors for employment decisions).
1112
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[534]
(notwithstanding if a state applied or received RttT
funds) to secure waivers from the penalties that they
would incur for non-compliance with the No Child Left
Behind Act of 2001.
35
NCLB, passed with bipartisan
support in 2001, required 100 percent of students to
attain proficiency in math and reading state
standardized tests.
36
The utopian goal has been widely
criticized as impractical.
37
Nevertheless, the federal
government required states to submit waivers to escape
the punitive measures of non-compliance (e.g.,
intervention of state authorities in the operation of local
schools). More specifically, these waivers buttressed the
core policy drivers of RttT by continuing to incorporate
student test scores as a means to hold teachers
accountable for their “value added,” or lack thereof.
38
The cumulative impact of RttT and federal
waivers on the use of VAMs in teacher evaluations was
substantial. By 2014, 40 states and Washington, D.C.,
35
KEVIN CLOSE ET AL., STATE-LEVEL ASSESSMENTS AND
TEACHER EVALUATION SYSTEMS AFTER THE PASSAGE OF THE
EVERY STUDENT SUCCEEDS ACT: SOME STEPS IN THE RIGHT
DIRECTION 5 (Nat’l Educ. Policy Ctr. ed., 2018),
https://nepc.colorado.edu/sites/default/files/publications/PB%20C
lose-Beardsley-Collins_1.pdf [https://perma.cc/RG4N-B8N2].
36
No Child Left Behind Act of 2001, Pub. L. No. 107-110, §
1001, 115 Stat. 1425 (requiring all students obtain proficiency
in specified test areas) (repealed 2015).
37
See, e.g., Bruce Meredith & Mark A. Paige, Opinion,
Rethinking Federal Role in Education Makes Sense. Trump’s
Plan Does Not, ATLANTA J.-CONST.: GET SCHOOLED (Oct. 3,
2018, 11:15 AM) https://www.myajc.com/blog/get-schooled/
opinion-rethinking-fed-education-role-makes-sense-trump-plan-
does-not/T19cWlKAznnDpcoxmvr1nJ/ [https://perma.cc/S3J4-
B4FW] (characterizing the NCLB goal of proficiency as
unrealistic, especially in light of the lack of support from the
federal government to education and other important public
policy areas that impact education success, like housing and
health care).
38
CLOSE ET AL., supra note 35, at 8.
1213
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[535]
(80%) were using or still developing some type of VAM for
increased teacher accountability purposes.
39
While state
department of education leaders recognized and
encouraged the use of VAMs, they did not develop
support mechanisms and resources to help teachers
understand and subsequently use their VAM-based data
to improve their effectiveness.
40
Put differently,
information from VAMs was not actionable. This
disconnect has been the source of serious contention and
concern about the VAM-based teacher and educational
reform enterprise.
B. Statistical and Practical Controversies
Significant statistical and practical concerns
surround VAMs, and these are best understood with
reference to the professional guidelines that govern
education and psychological professions, the Standards
for Educational and Psychological Testing
41
(hereinafter
“Standards”). These issues include, but are not limited to:
(1) reliability, (2) validity, (3) bias, (4) transparency, and
(5) fairness, with emphasis also on (6) whether VAMs are
being used to make consequential decisions using
concrete (e.g., not arbitrary) evidence, and (7) unintended
consequences. These are discussed below.
1. Reliability
Reliability is the degree to which test- or
measurement-based scores “are consistent over repeated
applications of a measurement procedure (e.g., a VAM)
and hence and inferred to be dependable and consistent”
39
Id.
40
Id. at 14.
41
AM. EDUC. RESEARCH ASS’N, AM. PSYCHOLOGICAL ASS’N &
NAT’L COUNCIL ON MEASUREMENT IN EDUC., STANDARDS FOR
EDUCATIONAL AND PSYCHOLOGICAL TESTING (2014)
[hereinafter STANDARDS].
1314
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[536]
for the individuals (e.g., teachers) to whom the scores
pertain.
42
VAMs are reliable when within-group (same
school or district) VAM estimates of teacher effectiveness
are more or less consistent over time, from one year to
the next, regardless of the type of students and subject
areas teachers teach. Consistency over time is typically
captured using particular statistical tools such as
standard errors, reliability coefficients per se, and
generalizability coefficients, among others.
43
These
situate and make explicit VAM estimates and their
(sometimes sizeable) errors and, importantly, help others
understand the errors that come along with VAM
estimates.
Research has documented serious concerns with
respect to VAM reliability (or intemporal stability).
Indeed, teachers classified as “effective” one year might
have a 25–59% chance of being classified as “ineffective”
the next year, or vice versa, with other permutations
possible.
44
If a teacher who is classified as a “strong”
teacher this year is classified as a “weak” teacher next
year, and vice versa, this casts doubt on the reliability of
VAMs for the purpose of identifying and making high-
stakes decisions regarding teachers. Accordingly, across
VAM, reliability is a hindrance, especially when
unreliable measures are to be used for consequential
purposes like decisions to terminate or deny tenure.
42
Id. at 222–23.
43
Id. at 33.
44
For a comprehensive overview of these concepts, see José
Felipe Martínez et al., Approaches for Combining Multiple
Measures of Teacher Performance: Reliability, Validity, and
Implications for Evaluation Policy, 38 EDUC. EVALUATION &
POL’Y ANALYSIS 738-56 (2016); see also Peter Z. Schochet &
Hanley S. Chiang, What are Error Rates for Classifying
Teacher and School Performance Using Value-Added Models?,
38 J. EDUC. & BEHAV. STAT. 142-71 (2013).
1415
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[537]
2. Validity
Validity is “the degree to which evidence and
theory support the interpretations of test scores for [the]
proposed uses of tests.”
45
It is measured by “the degree to
which all the accumulated evidence supports the
intended interpretation of [the test-based] scores for
[their] proposed use[s].”
46
Put another way, validity asks:
Does the model assess what it is supposed to assess?
47
Accordingly, one must be able to support validity
arguments with quantitative or qualitative evidence that
the data derived allows for accurate inferences.
There are various means to assess validity, but of
particular focus for researchers is validity as it concerns
“concurrent-related evidences.”
48
This helps to assess, for
example, whether teachers who post large and small
45
STANDARDS, supra note 41, at 11.
46
Id. at 14.
47
There are sub areas of validity that have been the subject of
considerable research as it relates to VAMs.
These are: (1) content-related evidence of validity; (2)
concurrent-related evidence of validity; (3) predictive-related
evidence of validity; and (4) consequence-related evidence of
validity. See Michael T. Kane, Validating the Interpretations
and Uses of Test Scores, 50 J. EDUC. MEASUREMENT 1, 2, 8
(2013); see generally Samuel Messick, Validity, 3 J. EDUC.
MEASUREMENT 1, 8–103 (1989). However, while all these
evidences of validity help to support construct-related evidence
of validity, in VAM research most researchers rely on
gathering concurrent-related evidence of validity.
48
E.g., Edward Sloat, Audrey Amrein-Beardsley & Jessica
Holloway, Different Teacher-Level Effectiveness Estimates,
Different Results: Inter-Model Concordance Across Six
Generalized Value-Added Models (VAMs), 30 EDUC.
ASSESSMENT EVALUATION & ACCOUNTABILITY 367, 372 (2018);
see also Pam Grossman et al., The Test Matters: The
Relationship Between Classroom Observation Scores and
Teacher Value Added on Multiple Types of Assessment, 43
EDUC. RESEARCHER 293, 293-303 (2014).
1516
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[538]
value-added gains or losses over time are the same
teachers deemed effective or ineffective, respectively,
over the same period using other independent
quantitative and qualitative measures of teacher
effectiveness. Other measures might include supervisors’
observational scores. If all measures line up and
theoretically validate one another, then confidence in
them as independent measures increases.
49
If all
indicators point in different directions, something may be
wrong with either or both indicators (the VAM tool or
observational scores, or both).
50
Researchers have questioned whether measures
of teacher value-added are substantively related to at
least one other criterion of teacher effectiveness (e.g.,
teacher observational or student survey indicators).
51
Moreover, they question whether the concurrent-related
evidence of validity that does exist is strong or
substantive enough to warrant valid inference-making.
3. Bias
Bias pertains to the validity of the inferences that
stakeholders draw from test-based scores.
52
Specific to
49
Kane, supra note 47, at 6–8, 37, 40, 64.
50
Id.
51
E.g., Morgan S. Polikoff & Andrew C. Porter, Instructional
Alignment as a Measure of Teaching Quality, 36 EDUC.
EVALUATION & POL’Y ANALYSIS 399, 399–401 (2014); Tanner
LeBaron Wallace, Benjamin Kelcey & Erik Ruzek, What Can
Student Perception Surveys Tell Us About Teaching?
Empirically Testing the Underlying Structure of the Tripod
Student Perception Survey, 53 AM. EDUC. RES. J. 1834, 1835,
1837–38 (2016).
52
The Standards define bias as follows: as the “construct
underrepresentation of construct-irrelevant components of test
scores that differentially affect the performance of different
groups of test takers and consequently the . . . validity of
interpretations and uses of their test scores.” STANDARDS,
supra note 41, at 216. Biased estimates, also known as
1617
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[539]
VAMs, unpredictable characteristics (variables outside of
the control of a teacher or school) of students can bias
estimates about teachers’ contributions. Student
characteristics include: students’ individual motivation,
capability to learn, and levels of academic achievement.
53
Because schools do not randomly assign teachers, these
variables are not controlled in a way to mitigate bias.
54
Biased results are quite possible, especially when
relatively homogeneous sets of students (e.g., English
Language Learners (ELLs), gifted and special education
students, or free-or-reduced lunch eligible students) are
non-randomly concentrated into schools, purposefully
placed into classrooms, or both.
Statistical models—even the most sophisticated—
cannot control for such bias.
55
One influential study
illustrated VAM-based bias when it found that a
systematic error as concerning “[t]he systematic over- or
under-prediction of criterion performance” are observed when
said criterion performance varies for “people belonging to
groups differentiated by characteristics not relevant to the
criterion performance” of measurement. STANDARDS, supra
note 41, at 216, 222.
53
See generally Noelle A. Paufler & Audrey Amrein-Beardsley,
The Random Assignment of Students into Elementary
Classrooms: Implications for Value-Added Analyses and
Interpretations, 51 AM. EDUC. RES. J. 328, 328–62 (2014).
54
See, e.g., Charles T. Clotfelter, Helen F. Ladd, & Jacob L.
Vigdor, Teacher-Student Matching and the Assessment of
Teacher Effectiveness, J. HUM. RESOURCES 778, 779–82 (2006)
(noting the various ways teachers are assigned to schools).
Class assignments in schools are historically a function of a
host of factors, including: pressure from parents for particular
class placement and pressure from teachers for placement of
particular students, especially those who may tend to be
considered “high-achieving.” Id. at 781. Additionally,
placement among schools within a district is similarly subject
to other variables, such as housing patterns. Id.
55
See, e.g., Paufler & Amrein-Beardsley, supra note 53, at
335.
1718
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[540]
student’s 5th grade teacher was a better predictor of a
student’s 4th grade growth than was the student’s 4th
grade teacher.
56
The absurdity of that finding raises
serious questions about the ability of VAMs to control for
bias. Notwithstanding, the primary debate raging across
articles concerns whether statistically controlling for
potential bias by using complex statistical approaches to
account for non-random student assignment makes bias
negligible, or rather “strongly ignorable.”
57
4. Transparency
Transparency is defined as the extent to which
something is accessible and understandable.
58
In terms
of VAMs, this relates to the extent to which VAM-based
estimates may not make sense to those receiving the
information. In education, teachers and principals may
not understand the models being used to evaluate their
performance. Because of this, they are generally unlikely
to use the VAM-generated information for formative
purposes (i.e., as a tool to gather information and change
practice as soon as possible).
59
Practitioners often
56
Jesse Rothstein, Student Sorting and Bias in Value-added
Estimation: Selection and Observables and Unobservables, 4
EDUC. FIN. & POL’Y 537, 546–47 (2009); Jesse Rothstein,
Teacher Quality in Educational Production, Q.J. ECON. 175,
210 (2010).
57
Sean Reardon & Stephen Raudenbush, Assumptions of
Value-Added Models for Estimating School Effects, 4 EDUC.
FIN. & POL’Y 492, 496–97 (2009).
58
STANDARDS, supra note 44.
59
Jonathan M. Eckert & Joan Dabrowski, Should Value-Added
Measures Be Used for Performance Pay?, KAPPAN, May 2010,
at 88, 89–90; Rachel Gabriel & Jessica Nina Lester, Sentinels
Guarding the Grail: Value-Added Measurement and the Quest
for Education Reform, 21 EDUC. POL’Y ANALYSIS ARCHIVES 1,
1–30 (2013); Ellen Goldring et al., Make Room Value Added:
Principals’ Human Capital Decisions and the Emergence of
1819
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[541]
describe value-add data reports as confusing, not
comprehensive in terms of the key concepts and
objectives taught, ambiguous regarding teachers’ efforts
at both the student and composite levels, and often
received months after students leave teachers’
classrooms.
For example, teachers in Houston, Texas,
expressed that they are learning little about what they
did effectively or how they might use their value-added
data to improve their instruction.
60
Teachers in North
Carolina reported that they were “weakly to moderately”
familiar with their value-added data.
61
Tennessee
teachers maintained that there was very limited support
or explanation helping teachers use their value-added
data to improve upon their practice.
62
Quite apart from the statistical concerns noted
above, the “black-box” nature of VAMs raises additional
questions in the field. Indeed, the purported strength of
VAMs is that they will improve instruction by providing
a wealth of positive diagnostic information. The models
are supposed to give practitioners useful, actionable
information. Yet, if practitioners have problems
understanding the models, the value (if you will) of VAMs
is greatly diminished. Unfortunately, statisticians that
have developed the models make “no apologies for the
Teacher Observation Data, 44 EDUC. RESEARCHER 96, 96–97
(2015).
60
Clarin Collins, Houston, We Have a Problem: Teachers Find
No Value in the SAS Education Value-Added Assessment
System, 22 EDUC. POL’Y ANALYSIS ARCHIVES 1, 4, 15, 22 (2014).
61
Kim Kappler Hewitt, Educator Evaluation Policy That
Incorporates EVAAS Value-Added Measures: Undermined
Intentions and Exacerbated Inequities, 23 EDUC. POL’Y
ANALYSIS ARCHIVES 1, 11 (2015).
62
See Eckert & Dabrowski, supra note 59, at 90.
1920
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[542]
fact that [their] methods [are] too complex for most of the
teachers whose jobs depended on them to understand.”
63
5. Fairness
General questions of fairness have been raised
concerning the use of VAMs, especially in the context of
high-stakes employment decisions. Fairness is the
impartiality of “test score interpretations for intended
use(s) for individuals from all relevant subgroups.”
64
But
issues of fairness arise when a test or test use impacts
some more than others in unfair or prejudiced, yet often
consequential ways.
65
Fairness issues are amplified as VAMs are
applied in the field. Indeed, VAMs are generally only
directly applicable to teachers who instruct in areas that
are subjected to standardized tests (typically, math and
reading).
66
States and districts can only produce VAM-
based estimates for approximately 30–40% of all
teachers.
67
The other 60–70%, which sometimes includes
entire campuses of teachers (e.g., early elementary and
high school teachers) or teachers who do not teach the
core subject areas assessed using large-scale
standardized tests (e.g., mathematics and
English/language arts), cannot be evaluated or held
accountable using teacher-level value-added data.
68
Importantly, when districts use this information to make
63
Carey, supra note 1, at 13; see also Gabriel & Lester, supra
note 59, at 20.
64
STANDARDS, supra note 41, at 219 (emphasis added).
65
This concern is consistent with the general argument of this
paper. To wit, courts have sustained objections to the use of
VAMs where they violate procedural due process, the basic
“fundamental fairness.” See Cook v. Bennett, 792 F.3d 1294,
1301 (11th Cir. 2015).
66
E.g., Green et al., supra note 13 (noting that the models only
apply to 30–40% of teachers).
67
Id.; see also Gabriel & Lester, supra note 59, at 7.
68
Green et al., supra note 13, at 15, 27–28.
2021
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[543]
consequential, high-stakes employment decisions the
unfairness can have considerable consequences. Some
teachers in certain grades or subject areas experience the
negative or positive consequences of these VAM-based
data more than their colleagues.
69
6. Consequential Use
Assessing the appropriate use of tests must
consider the social and ethical concerns
70
in addition to
more sterile concerns about statistical methodology.
71
The Standards recommend ongoing evaluation of both
the intended and unintended consequences of any test as
an essential part of any test-based system, including
those based upon VAMs.
72
Typically, ongoing evaluation of social and ethical
consequences rests on the shoulders of the governmental
bodies that mandate such test-based policies.
73
In this
case, local and state education departments would be the
agencies in charge of assessing the social costs and
ethical issues associated with the use of VAMs in high-
stakes contexts. This is because they “provide resources
for a continuing program of research and for
dissemination of research findings concerning both the
69
This has formed the basis of substantive due process claims
against school districts. E.g., Cook, 792 F.3d 1294 (agreeing
that the system of Florida that adopted VAM ratings that apply
to all teachers, including those in non-tested subject areas, was
unwise and unfair but upholding it under rational basis test).
70
E.g., Messick, supra note 47, at 8 noting that “[t]he only form
of validity evidence [typically] bypassed or neglected in these
traditional formulations is that which bears on the social
consequences of test interpretation and use.”
71
See also Kane, supra note 47.
72
STANDARDS, supra note 41.
73
Id.
2122
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[544]
positive and the negative effects of the testing
program.”
74
However, this rarely occurs. The burden typically
rests on the research community who must provide
evidence about the positive and negative effects and
explain these effects to external constituencies, including
policymakers. This group must collectively determine
whether VAM use, given the consequences and issues
identified above, warrant the financial, time, and human
resource investments.
75
Local and state departments of
education typically have not (perhaps for political
reasons) acknowledged or sought to examine the
consequences of their policy actions.
7. Intended Consequences
As noted, the primary intended consequence of
VAM use is to improve teaching and help teachers (and
schools/districts) become better at educating students by
measuring and then holding teachers accountable for
their effects on students. The stronger the consequences,
the stronger the motivation leading to stronger intended
effects. Secondary intended consequences include
74
Position Statement on High-Stakes Testing in Pre-K–12
Education, AM. EDUC. RES. ASS’N (2000), http://www.aera.
net/About-AERA/AERA-Rules-Policies/Association-Policies/
Position-Statement-on-High-Stakes-Testing [https://perma.cc/
969R-8RMR]; see also STANDARDS, supra note 41.
75
Arguably, some “reformers” assume that their ideas are
inviolable and opposition is simply a reflection of a recalcitrant
system, at best, or teachers’ unions at worst. See e.g., Michelle
Rhee, Opting Out of Standardized Tests? Wrong Answer,
WASH. POST (Apr. 4, 2014) https://www.washingtonpost.com/
opinions/michelle-rhee-opting-out-of-standardized-tests-wrong-
answer/2014/04/04/37a6e6a8-b8f9-11e3-96ae-f2c36d2b1245_
story.html [https://perma.cc/JD5L-6APK] (suggesting that an
organization she founded always keeps students’ interests first
and also implying that teachers’ unions do not, especially in
regards to the use of standardized tests).
2223
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[545]
replacing the nation’s antiquated teacher evaluation
systems which have been criticized by all corners of the
education research.
76
Yet, in practice, research evidence supporting
whether VAM use has led to these intended consequences
is suspect. Indeed, numerous studies have noted that
there is a lack of evidence linking VAMs to improved
teacher quality. First, VAM estimates have not produced
useable information for teachers about how teachers,
schools, and states might improve upon their instruction,
or how all involved might collectively improve student
learning and achievement over time.
77
Likewise, recent
evidence suggests the use of VAMs has not led to
improvements in teacher evaluation systems.
78
In sum,
strong evidence suggest that VAMs have not promoted
the intended benefits of providing actionable information
for teachers to improve instruction or teacher evaluation
systems.
8. Unintended Consequences
Simultaneously, ethical and research standards
require that the use of testing data must also recognize
VAMs’ unintended consequences.
79
Policymakers must
present evidence on whether VAMs cause unintended
effects and if those effects outweigh their intended
impact. This means that the educative goals at issue (e.g.,
increased student learning and achievement) should be
76
See, e.g., DANIEL WEISBERG ET AL., THE WIDGET EFFECT
(2009) (criticizing the evaluation models that treat teachers as
“widgets” and fail to recognize their differences and value).
77
Henry Braun, The Value in Value-Added Depends on the
Ecology, 44 EDUC. RES. 2 (2015); Corcoran, supra note 12.
78
Matthew A. Kraft & Allison Gilmour, Revisiting the Widget
Effect: Teacher Evaluation Reforms and the Distribution of
Teacher Effectiveness, 46 EDUC. RES. 234–49 (2017).
79
See AM. EDUC. RES. ASS’N, supra note 74; STANDARDS, supra
note 41.
2324
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[546]
examined alongside the positive and negative
implications for both the science and ethics of using
VAMs in practice.
80
Researchers have produced an exhaustive list of
these unintended consequences.
81
First, the use of VAMs
leads to teacher isolation whereby teachers “literally or
figuratively ‘close their classroom door’ and revert to
working alone.”
82
Sadly, teacher isolation is at cross-
purposes with collaboration among colleagues,
something that is an essential part to improving
schools.
83
Second, the use of high-stakes testing causes
teachers to leave the profession and avoid high-needs
schools that most need the best teachers.
84
Because of the
very nature of VAM-based teacher evaluation which
rewards testing achievement, teachers avoid teaching
high-needs students. This is rational: if they perceive
themselves to be at greater risk of teaching students who
may be more likely to hinder their value-added
85
they
“seek safer [grade level, subject area, classroom, or
school] assignments, where they can avoid the risk of low
VAMS scores.”
86
Of course, the flip side of this, teachers
avoid challenging assignments or leave the profession all
together.
87
Third, and most troubling perhaps, is the
dehumanization that high-stakes testing causes. Indeed,
under such regimes, teachers view and react to students
as “potential score increasers or score compressors,” not
children.
88
80
Messick, supra note 47.
81
See, e.g., Susan Moore Johnson, Will VAMS Reinforce the
Walls of the Egg-Crate School?, 44 EDUC. RES. 117–26 (2015).
82
Id. at 120.
83
Id.
84
Id.
85
Id.
86
Id.
87
Id.
88
Hewitt, supra note 61, at 32.
2425
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[547]
III. The Cases
This section discusses cases where the central
issue was the role VAMs played in adverse employment
actions. It first traces those cases related to arguments
grounded in the substantive Due Process and Equal
Protection clauses of the U.S. Constitution. It then
highlights the series of cases where plaintiffs challenged
the use of VAMs on jurisdictional grounds (i.e., that a
particular government agency superseded its authority
or other statutes in requiring the use of VAMs). The final
subsection assesses the cases where process arguments
have been advanced by the plaintiffs.
A. Federal Substantive Due Process Rights &
Equal Protection Arguments: VAMs May Be
Unwise But Still Constitutional
1. Cook v. Bennett
In 2015, a group of teachers challenged Florida’s
use of student test scores to evaluate their job
performance.
89
As part of that state’s application for Race
to the Top funds, the state legislature enacted a new
teacher performance evaluation regimen in their law of
teacher evaluation.
90
Specifically, the legislature
required that at least 50% of a teacher’s performance
evaluation be based on student growth on state
standardized tests in math and English (the Florida
Comprehensive Assessment Test, or FCAT).
91
The
remaining portion of the teacher’s evaluation was
89
Cook v. Bennett, 792 F.3d 1294 (11th Cir. 2015).
90
FLA. STAT. ANN. § 1012.34 (West 2011).
91
Id. A teacher’s final evaluation was based on the student test
growth (the VAM rating) on the FCAT (50%) and a VAM rating
based on the school’s contribution to a student’s growth. Cook,
792 F.3d at 1297.
2526
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[548]
calculated based on a school-wide VAM rating.
92
Not all
students take the math and English tests. In fact,
students took the English FCAT exam in grades 3
through 10 and the mathematics FCAT exam in grades 3
through 8.
Under the evaluation law, Florida teachers fell
under one of three types of categories.
93
“Type A” teachers
were those that taught the tested subjects (math and
English) in the years that the FCAT was administered
for those subjects. In effect, as the Eleventh Circuit Court
of Appeals noted, the model adopted by the state
education commissioner only worked as designed in
evaluating teachers of English in grades 4 through 10
and math in grades 4 through 8.
94
The rest of Florida’s
public school teachers fell into two groups. “Type B”
teachers taught students in grades 4 through 10, but in
subjects other than English or math.
95
“Type C” teachers
taught students in grades below 4 or above 10 or their
students did not take standardized tests (e.g., art).
96
The thrust of the legal problem, according to the
teachers challenging the evaluation scheme, related to
the evaluation of Type B and C teachers. As a practical
matter, school districts evaluated Type B teachers using
student FCAT scores for math and English,
notwithstanding the fact that those teachers did not
instruct the students in those subjects.
97
Type C teachers’
VAM scores were calculated based on school-wide FCAT
scores derived from student scores in subjects they did
not teach.
98
Under this scenario, for example, a second
92
Id.
93
The district court designated the classification set forth in
this discussion and, for ease of reference, the appeals court
adopted it in its analysis.
94
Cook, 792 F.3d at 1297.
95
Id.
96
Id.
97
Id.
98
Id. at 1298.
2627
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[549]
grade art teacher’s VAM rating could be calculated based
on a 3rd grade student’s math and English test growth.
The plaintiff-teachers argued that the evaluation
laws violated the Substantive Due Process and Equal
Protection clauses of the Fourteenth Amendment.
99
Because no fundamental right was at issue, the court
applied the rational basis test to determine whether the
government’s actions had a legitimate purpose and
whether the chosen methods were rationally related to
that purpose.
100
Ultimately, the court sided with the
government, finding that there was a legitimate interest
which was to “increas[e] student academic performance
by improving the quality of instructional, administrative,
and supervisory services in the public schools of the
state.”
101
The court also concluded that there was a rational
relationship between this purpose and the use of the
FCAT VAMs.
102
The court concluded—and the plaintiffs
conceded at oral argument—that the government “could
have reasonably believed that (1) a teacher can improve
student performance through his or her presence in a
99
U.S. CONST. amend. XIV provides, in relevant part, that: “No
state shall . . . deprive any person of life, liberty, or property,
without due process of law; nor deny to any person within its
jurisdiction the equal protection of the laws.”
100
Cook, 792 F.3d at 1300 (citing Fresenius Med. Care
Holdings, Inc. v. Tucker, 704 F.3d 935, 945 (11th Cir. 2013);
FCC v. Beach Comm’ns, Inc., 508 U.S. 307, 314 n.6 (1993)).
101
Id. at 1301 (citing FLA. STAT. § 1012.34(1)(a) (2013)); see also
Houston Fed’n of Teachers, Local 2415 v. Houston Indep. Sch.
Dist., 251 F. Supp. 3d 1168, 1182 (S.D. Tex. 2017) (concluding
that plaintiff’s substantive due process claims failed because
“[e]ven accepting plaintiffs’ criticisms at face value, the loose
constitutional standard of rationality allows governments to
use blunt tools which may produce only marginal results.”).
The Houston court, however, ruled that the plaintiff’s
allegations of procedural due process violations survived
summary judgment dismissal. Id. at 1183.
102
Cook, 792 F.3d at 1301.
2728
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[550]
school and (2) the FCAT VAM can measure those school-
wide performance improvements, even if the model was
not designed to do so.”
103
To be sure, both the appellate
and district courts criticized the chosen model.
104
The court similarly applied the rational basis
review to dismiss the equal protection claims.
105
Under
this claim, the teachers argued that the evaluation law
created a separate class of teachers: “those whose
evaluations are based on student growth data for
students assigned to the teacher in the subjects taught
by the teacher, and those whose evaluations are based on
student growth data for students and/or subjects they do
not teach.”
106
However, because this classification did not
implicate a suspect class (e.g., race, gender) rational basis
applied and, under the same line of reasoning of the
substantive due process claim, the equal protection claim
was dismissed.
107
103
Id.
104
Id. at 1301 (noting that “[w]hile the FCAT VAM may not be
the best method—or may even be a poor one—for achieving this
goal, it is still rational to think that the challenged evaluation
procedures would advance the government's stated purpose.”).
The district court in finding for the government concluded, in
dicta, that “[t]he unfairness of the evaluation system as
implemented is not lost on this Court” and that “this Court
would be hard-pressed to find anyone who would find this
evaluation system fair to non-FCAT teachers, let alone be
willing to submit to a similar evaluation system.” Cook v.
Stewart, 28 F. Supp. 3d 1207, 1215–16 (N.D. Fla. 2014), aff’d
sub nom. Cook v. Bennett, 792 F.3d 1294 (11th Cir. 2015).
105
Cook, 792 F.3d at 1301.
106
Stewart, 28 F. Supp. 3d at 1213.
107
Cook, 792 F.3d at 1301 (citing City of Cleburne v. Cleburne
Living Ctr., 473 U.S. 432, 440 (1985) (internal citations
omitted)).
2829
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[551]
2. Trout v. Knox County Board of Education
Plaintiff teachers in Trout v. Knox County Board
of Education brought substantive and procedural due
process claims based on their evaluations that used
VAMs for purposes of teacher evaluations.
108
In Trout,
the teachers challenged the use of Tennessee’s VAM
rating (the EVAAS). Specifically, two teachers (one a
math teacher and the other a science teacher) were
denied bonuses based on their VAM rating.
109
Both teachers involved (Trout and Taylor,
respectively) argued that the use of the VAMs was
arbitrary and capricious and, therefore, could not be
sustained under the rational basis test. Echoing
criticisms of the reliability and validity of VAMs,
110
the
plaintiffs argued that the VAMs were too imprecise to be
used to assess their effectiveness
111
and therefore
violated substantive due process rights.
The federal district court ruled in favor of the
government. It began its analysis by noting that the
plaintiffs failed to state a substantive due process
claim.
112
By way of background, a substantive due
process claim requires that there be some property
interest at stake. Here, under an analysis of property
interest rights in the Sixth Circuit Court of Appeals, the
court concluded that the plaintiffs did not have an
interest in bonuses.
113
For sake of argument, however, the court went on
to apply the rational test and found that the
government’s use of the VAMs in this case satisfied that
108
Trout v. Knox Cty. Bd. of Educ., 163 F. Supp. 3d 492, 494
(E.D. Tenn. 2016).
109
Id.
110
See supra Part I.
111
Trout, 163 F.Supp. 3d at 500.
112
Id.
113
Id at 501.
2930
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[552]
test.
114
The use of VAMs to identify and support
instruction to lead to increased student achievement was
not in dispute as a legitimate government interest.
115
The
plaintiffs, similar to Cook v. Bennett,
116
argued that
various statistical infirmities made reliance on VAMs
irrational, however.
117
In rejecting these arguments, the
district court noted, among other things, that there was
no legal authority requiring the court to apply a standard
with respect to the confidence level of a test.
118
To be sure, the Trout court was sympathetic to the
plaintiffs’ complaints regarding the statistical
inadequacy of the VAMs.
119
Yet, at bottom, there was no
legal authority that required the court to apply a certain
level of statistical confidence with respect to the
government’s chosen method for purposes of measuring
teacher effectiveness.
120
114
Id.
115
Id. at 503.
116
Cook, 792 F.3d at 1297.
117
For example, the plaintiffs took issue with the confidence
level of the statistical test (68%). Trout, 163 F.Supp. 3d at 503.
118
Id.
119
Id. at 504 (writing that the Court notes that Plaintiffs'
criticisms of the statistical methods of TVAAS are not
unfounded.)
120
Id. at 504–05. The court wrote that while “[p]laintiffs
bemoan the statistical imprecision of TVAAS,” no legal
authority “support[s] the proposition that the United States
Constitution requires legislative decision making regarding
the use of statistics to require ‘statistically significant’ results.
Absent controlling authority to the contrary, this Court refuses
to extend the rational basis test this far—where no suspect
class or fundamental right is at issue, the Constitution
requires a rational basis, not a statistically significant basis,
for the law in question.” Id.
3031
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[553]
3. Wagner v. Haslam
Another set of teachers in Tennessee challenged
the use of VAMs in Wagner v. Haslam.
121
Pursuant to
state and district evaluation policies, teachers of non-
tested subjects were evaluated based on school wide data
of student performance on test subjects.
122
Similar to
Cook v. Bennett, the teachers claimed that this practice
violated the substantive Due Process and Equal
Protection clauses of the U.S. Constitution.
123
The federal court, however, echoing the decisions
of other federal courts assessing similar claims, rejected
the teachers’ arguments. With respect to the substantive
due process claim, the court enumerated several reasons
why the policies at issue passed constitutional muster. It
noted that “the State Board could rationally believe that
a school-wide score provides some measure (albeit a crude
one) of evaluating an individual teacher’s
performance.”
124
The court also added that the legislature
had continued to amend its teacher evaluation laws to
address some of the concerns raised by the plaintiffs.
125
While the Wagner court concluded that the use of
VAMs was constitutional, it expressed concerns over
fairness similar to those found in Cook and Trout. Indeed,
the Wagner court wrote that although the current
evaluation processes may produce “unfair results” for
certain teachers, it did not rise to the level of being
irrational.
126
At the same time, the court was explicit
about its use of judicial restraint, especially with respect
to education policy questions. Indeed, subject to limited
121
112 F. Supp. 3d 673 (M.D. Tenn. 2015).
122
Id.
123
See Cook, 792 F.3d at 1297.
124
Wagner, 112 F. Supp. 3d at 694 (emphasis added).
125
Id.
126
Id. at 695.
3132
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[554]
exceptions,
127
the states have “unfettered”
128
discretion to
regulate education, and state legislators can make both
“excellent decisions and terrible decisions,” so long as
there is some “modicum of rationality.”
129
Put another
way, a court may disagree with the policy choice of a
governing body, but it is not the role of the courts to
second-guess policy judgments of elected officials.
130
4. Matter of Lederman v. King
The one extant case that succeeded in
demonstrating the government’s use of VAMs rose to the
high bar of arbitrary and capricious is found in Matter of
Lederman v. King.
131
In this case, a well-regarded
veteran teacher who had previously had positive
evaluations received an “ineffective” review under New
York’s new evaluation system.
132
This new system
required the use of VAMs. The teacher, Sheryl
Lederman, submitted “overwhelming” and ample
evidence from experts in the field that the court
concluded satisfied her burden in the record before the
court.
133
In contrast, the court noted that state defendants
left numerous statistical issues unaddressed, including
the potential VAM biases against teachers with high-
127
Some exceptions, of course, would include the use of race to
segregate schools. See generally Brown v. Bd. of Educ., 373 U.S.
483 (1954).
128
Id. at 692.
129
Id. at 693.
130
But see PAIGE, supra note 9 (arguing that for scholars of
educational policy the appropriate question is determining
which institutions—courts, legislatures, or markets—have the
capacity to best address a particular policy need in education,
like teacher evaluation).
131
Lederman v. King, 54 Misc. 3d 886 (N.Y. Sup. Ct. 2016).
132
Id. at 888.
133
Id. at 897–98.
3233
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[555]
performing students.
134
Critically, how Mrs. Lederman’s
scores swung so wildly from the second-highest level of
effective all the way to the lowest level of ineffective in a
single year with statistically similar scoring students,
among others.
135
In sum, the court was constrained to the
record before it and, on that evidence, found Ms.
Lederman satisfied her burden.
136
B. Legislative State Agency Authority
Questioned
Litigants have also challenged the use of VAMs in
teacher evaluation on jurisdictional grounds. In these
cases, organizations (typically unions) have argued that
a legislative or executive agency exceeded their
respective authority in requiring VAMs for purposes of
evaluation or high-stakes employment decisions. These
cases are discussed below.
1. Leff v. Clark County School District
At issue in Leff v. Clark County School District
was the constitutionality of changes made to state laws
governing teacher evaluation and post-probationary (or
continuing contract) status.
137
By way of background, up
until 2011, a teacher who completed a probationary
period of employment (three years) and was subsequently
rehired by a school district received post-probationary
status.
138
Post-probationary status conferred to a teacher
certain procedural protections should they face
termination and required that termination be “for
134
Id.
135
Id.
136
Id. at 898.
137
Leff v. Clark Cnty. Sch. Dist., 210 F. Supp. 3d 1242, 1244–
45 (D. Nev. 2016).
138
Id. at 1245.
3334
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[556]
cause.”
139
In contrast, probationary teachers could be
non-renewed without cause and did not have similar
procedural protections.
In 2011, the Nevada legislature changed its
teacher evaluation and post-probationary statutes. In
particular, it required that VAMs be used as part of
teacher evaluations. The legislature also required that if
a post-probationary teacher achieved two negative
evaluations, they would revert back to probationary
teacher status.
140
Put another way, a teacher could lose
the protections (e.g., a teacher’s termination could only
be for “cause”) because of the changes to the state
statutes.
Teachers contested the changes based on the
federal Constitution’s Contracts Clause.
141
That clause,
in relevant part, reads as follows: “No State shall . . .
pass any . . . Law impairing the Obligation of
Contracts[.]”
142
In essence, the post-probationary
teachers claimed that they had a binding contract with
the state once they achieved post-probationary status. In
exchange for meeting the demands of satisfactory
performance, the state had agreed to give them
procedural protections and the only grounds for
termination were cause. By passing the 2011 amendment
that tied teacher contract status to teacher evaluations
(that incorporated VAMs), the state breached the
contract, something not permitted under the U.S.
Constitution.
The federal court declined to adopt the teachers’
position and held that the statute prior to 2011 did not
create a contractual obligation between the state and
teachers. In its analysis, the court determined that there
is a strong presumption in law against the idea that a
139
Id.
140
Id.
141
Id. at 1244.
142
U.S. CONST. art. I, § 10.
3435
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[557]
legislative action creates a private contract.
143
Absent
any expression of the legislature that they were creating
a contract, it is generally assumed that typical legislative
activity simply reflects a policy determination that can be
changed.
144
Accordingly, the teachers’ claim that the
state legislature exceeded its authority with the
statutory amendments failed.
2. Stapleton v. Skandera
In Stapleton v. Skandera, teachers challenged the
use of VAMs in teacher evaluation on several
jurisdictional grounds related to statutory and agency
authority.
145
By way of brief background, the New Mexico
legislature attempted—but failed—to make several
amendments to its existing teacher evaluation laws in
2012. Notwithstanding this, the New Mexico Department
of Education Secretary (through the Department)
promulgated new regulations relative to the evaluation
of teachers.
146
The teachers sought judicial relief in that
the court would suspend the use of the regulations.
147
The teachers argued that the Secretary exceeded
her authority—that, in effect, she acted in a legislative
capacity. They raised particular objection to the
incorporation of VAMs in teacher evaluation, arguing
that such a move could only be done by way of legislative
action because it represented a shift in public policy
under exclusive legislative purview.
148
However, the New
Mexico Court of Appeals sided with the Department on
143
Leff, 210 F. Supp. 3d at 1246–47 (citing Nat’l R.R. Passenger
Corp. v. Atchinson Topeka and Santa Fe Ry. Co., 470 U.S. 451,
465–66.)
144
Id.
145
Stapleton v. Skandera, 346 P.3d 1191, 1194 (N.M. App.
2015).
146
Id. at 1193 (citing N.M. CODE R. § 6.69.8).
147
Id.
148
Id. at 1194.
3536
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[558]
this issue. It noted that the enabling statute required
only that the Department enact evaluation regulations
that were “uniform statewide” and “highly objective.”
149
Accordingly, the legislature left the Secretary “broad
authority” to enact regulations reflecting these
requirements and, in the view of the court, including
VAMs in teacher evaluation protocol did not exceed her
authority.
150
The teachers in Stapleton raised other claims
related to agency authority. In particular, they raised two
additional objections. They contended the new
departmental regulations permitted “assistant
principals” to observe teachers which violated the state
evaluation law that only gave such authority to
“principals.”
151
Similarly, they argued that the provisions
in the regulations that exempted charter schools from
coverage of the evaluations violated the state law
requirement that the Department enact a system of
“uniform” evaluation.
152
The court of appeals rejected both of the
arguments. With respect to the first claim (that only
principals could observe teachers), the court read the
state statute as allowing others to observe teachers,
including assistant principals. The court wrote, “We
agree with the district court that the regulation does not
necessarily conflict with the statute because the statute
‘mandates the participation of school principals [but]
does not limit the persons who may [also] observe
[teachers].’”
153
Regarding the claim that the regulations
inappropriately exempted charter schools, the state court
of appeals noted that the state Charter School Act
specifically allowed the Department to waive certain
149
Id. at 1195 (citing N.M. STAT. ANN. § 22-10A-19(A) (1978)).
150
Id.
151
Id. at 1196.
152
Id.
153
Id. (alterations in original).
3637
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[559]
regulations normally applicable to public schools.
154
Because the teachers could not cite to any other legal
authority that suggested the waiver was not permitted
under the Charter School Act, this theory was also
rejected.
155
3. Louisiana Federation of Teachers v. State
In Louisiana Federation of Teachers v. State, a
teacher’s union challenged Louisiana’s enactment,
amendment, and repeal of multiple state laws related to
public education, including those related to teacher
evaluation requirements.
156
During the 2012 legislative
sessions, the state legislature amended and re-enacted
nine different statutes, enacted two new distinct
statutes, and repealed twenty-eight statutes all related
to education.
157
The plaintiffs alleged that these actions, which all
occurred through one legislative act, violated the state
constitution’s “single object” requirement.
158
That
requirement stipulates that the legislature enacts bills
that have “one object” and that various pieces of a bill
must have a relationship to one another.
159
The teachers
argued that the bill contained unrelated subjects, such as
the changes to teacher evaluation, reduction in force
issues, rules governing contracts with superintendents,
among others.
160
Louisiana’s supreme court rejected the plaintiffs’
arguments.
161
The court began its assessment by noting
154
Id.
155
Id. at 1196–97.
156
La. Fed’n of Teachers v. State, 171 So. 3d 835, 841 (La.
2014).
157
Id.
158
Id. at 838.
159
Id. at 841.
160
Id. at 842.
161
Id. at 851.
3738
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[560]
that there is a general presumption that a legislature’s
acts satisfy the “one object” rule.
162
It also noted that the
purpose of the rule was to prevent “logrolling,” or the
practice of packaging many measures into one bill
because any of those measures, alone, would not pass the
legislature.
163
The court noted that under such a “grave
and palpable” scenario, the legislature would violate the
single object rule.
164
Yet, in this case, the court concluded
that the object of the act at issue “is improving
elementary and secondary education through tenure
reform and performance standards based on
effectiveness.”
165
The court concluded that various
components of that bill could be broadly related to this
objective.
166
4. Robinson v. Stewart
Another Florida case, Robinson v. Stewart,
167
also
involved a challenge to the authority of the state Board
of Education to implement teacher evaluation
regulations using VAMs.
168
In Robinson, the plaintiffs
sought to declare the 2011 Student Success Act
unconstitutional on the grounds that it impermissibly
delegated legislative control over public education to the
executive branch.
169
The act revised teacher evaluation
procedures and required the use of “student learning
growth measures” (or VAMs) to evaluate teachers and
make significant employment decisions, such as
tenure.
170
The act left it to the Department of Education
162
Id. at 845.
163
Id. at 845–46.
164
Id. at 851.
165
Id. at 850.
166
Id.
167
161 So. 3d 589 (Fla. Dist. Ct. App. 2015).
168
Id.at 590–91.
169
Id.
170
Id. at 591.
3839
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[561]
Commissioner (the executive branch) to develop the
formula to achieve these goals
171
and required the use of
standardized test scores.
172
The Florida District Court of Appeals rejected the
plaintiffs’ argument that the legislature, in requiring the
Commissioner to develop the formula, violated the non-
delagability doctrine of the state constitution that
ensures a separation of powers.
173
Its analysis noted that
the plaintiffs carried a high burden of proof: that the
legislature’s action violated the doctrine “beyond a
reasonable doubt,” the highest standard of proof under
the law.
174
The court further interpreted the act as simply
requiring the Commissioner to provide technical
implementation support, as opposed to allowing the
executive to make policy determinations.
175
5. Filed but not Adjudicated
Another case that deserves some attention as it
also related to a claim that a state agency exceeded its
authority by incorporating VAMs in evaluating teachers.
In Texas Teachers Association v. Texas Education Agency,
the Texas Department of Education adopted teacher
171
Id.
172
Id. at 592.
173
Id. at 590–91.
174
Id. at 591.
175
Id. at 592. But see id. at 597 (Benton, J., dissenting) (noting
that the legislature “has conferred on the State Board of
Education power to designate some of them—perhaps nearly
all of them—professionally ‘unsatisfactory,’ and therefore,
among other things, subject to being laid off, for reasons that
are so unclear and indefinite that the Legislature has
abandoned its responsibility to set public policy in this
important area, and delegated legislative authority it should
have exercised itself to the State Board of Education, an
executive branch agency.”)
3940
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[562]
evaluation regulations requiring the use of VAMs.
176
Numerous plaintiffs, including teachers’ unions, sought
to enjoin the use of VAMs on the grounds that the
regulations exceeded the power vested in the state
Department of Education.
177
The case settled and the
state ultimately agreed to eliminate the required use of
VAMs in teacher evaluation regulations.
178
In New Mexico ex rel Stewart v. New Mexico
Public Education Department, a group of plaintiffs
consisting of legislators, unions, and teachers filed a
complaint on the grounds that the state Department of
Education improperly infringed other state laws when it
promulgated its teacher evaluation regulations.
179
Plaintiffs argued that the School Personnel Act provides
for the processes associated with teacher evaluation and
termination.
180
Similarly, plaintiffs allege that the Department’s
regulation conflicts with New Mexico’s Public
176
Sean Collins Walsh, Union Sues to Block Texas Teacher
Evaluation Change, AUSTIN AM.-STATESMAN, (Aug. 13. 2016),
https://www.statesman.com/news/20160813/union-sues-to-
block-texas-teacher-evaluation-change [https://perma.cc/MQ
C2-FATW].
177
Id.
178
Melissa B. Taboada, Lawsuit Settled: Texas Teacher
Appraisals Won’t Be Tied to STAAR Scores, AUSTIN AM.-
STATESMAN (last updated Sept. 25, 2018),
https://www.statesman.com/news/20170504/lawsuit-settled-
texas-teacher-appraisals-wont-be-tied-to-staar-scores
[https://perma.cc/XP3C-H2WB].
179
Complaint, State ex rel Stewart v. N.M. Pub. Educ. Dep’t,
No. D-101-CV-2015-00409 (N.M. 1st Jud. Dist. Feb. 13, 2015),
https://www.aft.org/sites/default/files/nm-complaint-
teacherevals_1114.pdf [https://perma.cc/7T99-FG89]. The
plaintiffs also claim substantive and procedural due process
violations.
180
See e.g., N.M. STAT. ANN. § 22-10A-19(D) (2010) (providing
that evaluations should be determined in part by how well
professional development was carried out).
4041
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[563]
Employment Bargaining Law (the state’s enabling
collective bargaining statute) that governs “the terms
and conditions of employment.”
181
More specifically, that
law provides that local school districts must negotiate
terms and conditions of employment with the
representative union.
182
The case is pending with various
motions before the court.
183
C. Process & “Fundamental Fairness” Cases
1. Houston Federation of Teachers
A group of Houston teachers sought declaratory
and injunctive relief in the case of Houston Federation of
Teachers v. Houston Independent School District.
184
At
issue for the court was the constitutional protections
afforded teachers in the instance where the Houston
public school districts used VAMs to rate and make
employment decisions for its teachers.
185
The Houston
Independent School District (HISD) had contracted with
a third-party vendor who had created certain algorithms
to classify and rate teachers based on their students’ test
performance.
186
This third party vendor, citing trade
secrecy, refused to reveal the algorithms when they were
requested for review by the teachers.
187
Therefore,
teachers who faced adverse employment consequences
181
Complaint at 31, Stewart, No. D-101-CV-2015-00409.
182
See generally N.M. STAT. ANN. § 10-7E-17 (New Mexico’s
Public Employment Labor Relations Statute).
183
See Motion for Summary Judgment Filed in New Mexico
Teacher Evaluation Lawsuit, (Feb. 13, 2018), http://www.
krwg.org/post/motion-summary-judgment-filed-new-mexico-
teacher-evaluation-lawsuit [https://perma.cc/R8CU-DYHN].
184
Houston Fed’n of Teachers, Local 2415 v. Houston Indep.
Sch. Dist., 251 F. Supp. 3d 1168, 1174 (S.D. Tex. 2017).
185
Id. at 1171.
186
Id.
187
Id. at 1172.
4142
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[564]
could not review the underlying formulas that
contributed to these decisions.
188
The teachers claimed that the use of the value
added models constituted violation of the substantive and
procedural due process clauses of the Constitution.
189
Repeating a line of reasoning in Cook v. Bennett, and
other cases, the federal district court ruled that the
district’s use of VAMs did not amount to a substantive
due process violation.
190
The court concluded the
following: “Even accepting plaintiffs’ criticisms at face
value, the loose constitutional standard of rationality
allows governments to use blunt tools which may produce
only marginal results. HISD’s motion for summary
judgment on this substantive due process claim is
granted.”
191
Yet the court found in favor of the plaintiffs’
procedural due process claims.
192
The court’s analysis is
instructive because it relied heavily on procedural due
process as ensuring fundamental fairness.
193
The court
wrote:
“[The] purpose of procedural due process is to
convey to the individual a feeling that the
government has dealt with him fairly, as well as
to minimize the risk of mistaken deprivations of
protected interests.” [] In short, due process is
designed to foster government decision-making
that is both fair and accurate.
194
188
Id. at 1172–73.
189
Id.
190
Id. at 1181–82.
191
Id. at 1182.
192
Id. at 1180.
193
Id.
194
Id. at 1176 (alteration in original) (quoting Carey v. Piphus,
435 U.S. 247, 262 (1978)).
4243
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[565]
The court then listed the factors required for procedural
due process to be satisfied in the case of a teacher
termination in Texas.
195
Of particular note was that a
teacher facing termination must “be advised of the cause
for his termination in sufficient detail so as to enable him
to show any error that may exist.”
196
Teachers contended—and the court agreed—that
they were not being afforded due process protections
because the school district violated the requirement that
afforded a teacher “sufficient detail” to show that there
may be an error in the government’s decision.
197
Because
the district’s third party vendor would not release the
underlying formulas, teachers could not possibly assess
the accuracy of the district’s value-added rating.
198
The court listed numerous potential errors that
could be revealed if inspection of the formulas was
permitted.
199
As the court stated: “The [] score “might be
erroneously calculated for any number of reasons,
ranging from data-entry mistakes to glitches in the
computer code itself. . . . HISD has acknowledged that
mistakes can occur in calculating a teacher’s EVAAS
score . . . .”
200
The court was troubled by the district’s
stipulation that it could not correct a single teacher’s
score, even if an error was found, because correcting one
score would alter the results of all other teachers.
201
195
Id.
196
Id. The court also noted that a teacher facing termination
must be afforded: “the names and testimony of the witnesses
against him; [] a meaningful opportunity to be heard in his own
defense within a reasonable time; [] and a hearing before a
tribunal that possesses some academic expertise and an
apparent impartiality toward the charges.” Id. (citing Ferguson
v. Thomas, 430 F.2d 852, 856 (5th Cir. 1970).
197
Id. at 1176–77 (citing Levitt v. Univ. of Tex. at El Paso, 759
F.2d 1224, 1228 (5th Cir. 1985)).
198
Id.
199
Id. at 1177.
200
Id.
201
Id. at 1178.
4344
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[566]
Indeed, it is worth recalling that value added
scores are comparative in nature, assessing one teacher
against others.
202
This means that, if one teacher’s score
is adjusted for an error, it alters all others.
203
The court
characterized the underlying foundation of the VAM
ratings as built upon a “house of cards.”
204
Accordingly, it
denied the school district’s summary judgment claim
with respect to procedural due process.
205
2. Washington Teachers’ Union v. D.C. Public
Schools
The collective bargaining forum has also been
another forum wherein teachers have successfully
appealed the use of VAMs in teacher evaluations. By way
of background, collective bargaining agreements (CBAs)
provide for a process (grievance arbitration), to redress
violations of the contract. This arbitration process can be
important, especially when a contract calls for certain
specifications concerning how teacher evaluations can be
conducted. Indeed, districts’ decisions to non-renew or
terminate a teacher for performance have been called
into question because a district fails to follow
contractually mandated processes.
206
With some limited
202
Id. at 1172.
203
Id. at 1177.
204
Id. at 1178.
205
Id. at 1180. To be sure, procedural due process claims made
in Wagner v. Haslam, see supra notes 121129 and
accompanying discussion, did not survive. However, at issue in
that case was whether the teachers' bonuses could be linked to
their VAM scores. Wagner v. Haslam, 112 F. Supp. 3d 673, 688
(M.D. Tenn. 2015). In that context, the court concluded that
bonuses were not a property interest sufficient to trigger due
process protections. Id. at 698.
206
See, e.g., Dennis Yarmouth Teachers v. Dennis Yarmouth
Reg’l Sch. Dist, 360 N.E.3d 883, 884–885 (1977) (reversing a
school district’s decision to non-renew a probationary teacher
4445
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[567]
exceptions, scholarship has omitted consideration of the
value and importance of collective bargaining
agreements in relation to legal challenges to the use of
VAMs in teacher evaluations.
207
Cases emerging from Washington, D.C., illustrate
this theme. In Washington, a teacher’s union grieved the
public district’s performance ratings based on VAMs of
hundreds of teachers. As an initial matter, the school
district challenged whether the issue could, in fact, be
subject to the grievance arbitration procedures in the
contract. Indeed, as a general matter, disputes are
subject to the grievance process only if both parties
agreed to arbitrate the dispute under the CBA.
208
In Washington Teachers’ Union, a lower court had
concluded that the district’s final evaluation decisions
made under the evaluation systems were not arbitrable
but the district’s use of evaluation procedures under the
collective bargaining was, in fact, arbitrable.
209
Put
another way, the parties did not, under the CBA, agree
to arbitrate disputes over the judgment of the teachers’
final performance, but they did agree to arbitrate
whether or not the evaluation procedures outlined were
because school district violated terms of the collective
bargaining agreement that specified evaluation processes).
207
But see PAIGE, supra note 9, at 63–73 (arguing the use of
VAMs is susceptible to the grievance arbitration process and
the failures of VAMs to accurately assess teacher effectiveness
could be remedied through the collective bargaining process.);
see also Mark A. Paige, Applying the Paradox Theory: A Law
and Policy Analysis of Collective Bargaining Rights and
Teacher Evaluation Reform From Selected States, 2013 BYU
EDUC. & L.J. 21, 41–42 (highlighting the benefits of a more
collaborative collective bargaining process understood as
“interest-based” bargaining particularly with respect to
teacher evaluation).
208
Wash. Teachers’ Union v. D.C. Pub. Schs., 77 A.3d 441 (D.C.
Cir. 2013)
209
Id. at 444.
4546
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[568]
followed.
210
On appeal, the District of Columbia Court of
Appeals upheld the decision that the district’s final
judgments were not arbitrable. However, the school
district did not challenge the lower court’s determination
that the issue of whether the district followed evaluation
procedures was subject to evaluation.
211
In at least one other well-publicized case, the
Washington Teachers’ Union succeeded in frustrating
the D.C. Public Schools use of the IMPACT evaluation
system.
212
In this case, the union alleged that the school
district violated various evaluation procedures when they
terminated a seventeen year veteran teacher, Thomas
O’Rourke, under the district’s evaluation procedures.
213
As noted above, the controlling courts in the District of
Columbia have concluded that “process arguments”
under the collective bargaining agreement are arbitrable,
although the school district’s final judgment with respect
to evaluation categorization (e.g., ineffective,
satisfactory, etc.) is not.
In the District of Columbia Public Schools matter,
the arbitrator found that the district violated evaluation
procedures governing the length of observation visits,
which, according to the contract, should be “at least 30
minutes.”
214
In this case, the administrators evaluating
the teacher exceeded that length by substantial amounts
(e.g., observations lasted 80 minutes), which, in the eyes
of the arbitrator, amounted to a procedural violation of
evaluation processes.
215
Importantly, the arbitrator noted
210
Id.
211
Id.
212
D.C. Pub. Sch. v. Wash. Teachers Union, Local 6, AAA No.
16-20-1300-0499 AVH (Feigenbaum, Arb.); see also Perry
Stein, Teachers Union Touts Victory in Evaluation Fight WASH.
POST (Apr. 5, 2016), https://www.washingtonpost.com/news/
education/wp/2016/04/05/teachers-union-touts-victory-in-
evaluation-fight/ [https://perma.cc/P7RU-PSP7].
213
D.C. Pub. Schs., AAA No. 16-20-1300-0499 AVH.
214
Id. at 26–28.
215
Id. at 18.
4647
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[569]
two other significant factual findings to his decision. He
concluded that the administrator evaluating the teacher
had a reputation of using the observation system to
penalize teachers “he did not like.”
216
A school district
administrator, as well, testified that an observation that
exceeded or did not meet the thirty minute threshold
would amount to a process violation.
217
In sum, and under
these circumstances, therefore, procedural violations
could be seen as simply pretext for terminating a
teacher.
218
In arbitration cases, the remedy for a bargaining
violation can be a contested issue. In Washington, D.C.,
an arbitrator cannot issue a remedy in the form of
recategorizing a teacher’s evaluation from ineffective to
effective.
219
Reinstatement and back pay, however, are
typical arbitration remedies,
220
and these were, in fact,
used in the case.
IV. Current Policy Landscape in Wake of ESSA
This section discusses the current policy
landscape following the reauthorization of the
Elementary and Secondary Education Act of 1965 by
Congressional passage of the Every Student Succeeds Act
(ESSA) of 2015. It illustrates that the ESSA
reauthorization allowed for more state-level flexibility
with regards to VAM use. It then highlights how the new
policies have essentially shifted the emphasis from VAMs
216
Id. at 19.
217
Id. at 7.
218
Id. at 19.
219
Wash. Teachers’ Union v. D.C. Pub. Sch., 77 A.3d 441, 458
(D.C. Cir. 2013).
220
See e.g., DISCIPLINE AND DISCHARGE IN ARBITRATION ch.
13.I.A. (Norman Brand & Melissa Birens, eds., 3d ed.) (noting
that back-pay and reinstatement are two essential remedies for
making an employee whole).
4748
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[570]
in high stakes decision making to, perhaps, other ways of
measurement.
A. ESSA Reauthorization
In 2015, Congress passed a reauthorization of the
Elementary and Secondary Education Act under a new
name, the Every Student Succeeds Act.
221
In general,
ESSA reduced some federal mandates and incentives tied
to accountability system effectively limiting some of the
federal control promoted by RttT and other waiver
requirements.
222
Specifically, ESSA allowed state
departments of education two main changes: (a) ESSA
gave state departments leniency to interpret key terms
like, “including, as a significant factor, data on student
growth for all students,” and (b) ESSA gave state
departments more control to determine state goals and
measures for success with a federal framework.
223
Put
simply, ESSA allowed more flexibility.
To break down the policy changes further, the first
main change, allowing states to interpret “data on
student growth” differently, allowed state departments of
education to step back from the statistically-based
measures of student growth such as VAMs. ESSA
allowed states to use some measures which could include
qualitative measures as data showing student growth,
such as student learning objectives (SLOs), which are
objectives for the growth of students developed at the
beginning of the year by teachers (sometimes in
conjunction with others).
224
SLOs still rely on evidence which can still include
VAM scores, but the evidence can also include course
221
Every Students Succeeds Act of 2015, Pub. L. No. 114-95 §
114 Stat. 1177 (2015).
222
Race to the Top Act of 2011, S. Res. 844, 112th Cong. (2011).
223
ESEA Flexibility, U.S. DEP’T EDUC. (2012),
https://www.ed.gov/esea/flexibility [https://perma.cc/95A7-FLFA].
224
CLOSE ET AL., supra note 35, at 18.
4849
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[571]
exams, performance demonstrations, and other types of
evidence. In short, ESSA allowed states to incorporate
more nuanced and qualitative measures of student
growth without removing the requirement that states
must use evidence of student growth. The distinction is
small but significant. It signals a redefinition of “data” to
include information beyond large standardized testing
(although, importantly, it can still include these test
scores).
The second main change, allowing states to set
their own goals and measures for success, marks a
backing away from the strict adequate yearly progress
(AYP) goals established by NCLB. Although states still
must meet AYP for certain subgroups of students, the
consequences and the interventions that must be
imposed can be decided by the states themselves.
Essentially, ESSA removes the punitive bite
demonstrated previously by NCLB, the bite that
encouraged many states to apply for waivers and adopt
VAMs in the first place, and replaces it with flexibility.
States choose their own bite now. The standards remain,
but the consequence, the type of intervention required for
a failure to meet AYP, is decided by state departments of
education.
These two changes, though small, rolled back
some of the features that encouraged, or forced, states to
use large standardized statewide systems that leaned on
VAM results to measure teacher achievement.
225
The
new policy meant states did not need to create large-scale
comparable data about teacher achievement. States no
longer needed to structure their systems top-down and
could allow for more bottom-up control, essentially
handing more control to local educational authorities
such as school districts. ESSA marked a shift of power.
The federal government loosened reigns on state
225
Cindy Long, Six Ways ESSA Will Improve Assessments,
NEATODAY (2016), http://neatoday.org/2016/03/10/essa-
assessments/ [https://perma.cc/92AW-UC6A].
4950
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[572]
departments of education, who, in turn, had the freedom
to deviate from establishing one-size-fits-all teacher
evaluation systems across their state, handing more of
the power to make decisions to local educational
authorities, such as districts.
B. State Plans
Though ESSA allowed for many of the changes
stated above, it did not require or guarantee these
changes. The work of exercising the flexibility was for the
states, not the federal government. Hence, this section on
state plans reveals how state teacher evaluation plans
changed as a whole after the passage of ESSA through
state legislative and regulatory action. The changes, as
expected, trend toward less use of VAMs in high-stakes
decision making, though the trend is somewhat muted.
In general, less states are currently using growth
models or VAMs for teacher evaluation. The percentage
dropped from 42% in 2014 to 30% in 2018.
226
However,
that percentage drop fails to highlight the magnitude of
change. The study showing that the percentage
decreased measured whether some states currently use
or, importantly, endorse statewide use of VAMs. Some of
these states endorse VAMs but allow for local educational
authorities to avoid VAMs completely. For example,
Maine, encourages the use of VAMs, but offers two
models from which local education authorities can
choose, one of which measures student growth with
SLOs, not VAMs.
227
In this case, VAMs play a role in the
state’s teacher evaluation process, but, ultimately, the
choice is made locally. This represents a major departure
from the trend of heavy-handed state teacher evaluation
systems before the passage of ESSA.
Additionally, some states have maintained their
VAMs but use them in novel ways. North Carolina still
226
CLOSE ET AL., supra note 35, at 12.
227
Id. at 13.
5051
TENNESSEE’S NATIONAL IMPACT
13 TENN. J.L. & POL’Y 523 (2019)
[573]
uses a VAM, called EVAAS, which featured heavily in
many of the lawsuits.
228
However, the state does not use
the results to make high-stakes decisions. Rather, North
Carolina uses and reports the scores to foster
professional development.
229
In other words, the state
does not shy from using VAM data as a part of their
system, but they do shy from using VAMs for
consequential decisions such as tenure decisions and
others.
Additionally, and of note, recent state plans
demonstrate increased focus on formative feedback
practices compared to state plans collected in 2012, with
31 of 51 education plans stating that their evaluation
systems use formative data.
230
This shift indicates a
significant change in the stated values present in this
new set of state documents.
V. Conclusions
Quite apart from what education scholars and
policymakers believe with respect to the merits of added
models, all would likely agree that their introduction has
had significant consequences. Of course, there is
widespread disagreement with respect to how these
statistical models should be used. Teachers and unions
seeking to block the use of VAMs in high-stakes
employment decisions have sought judicial relief with
mixed success. That said, while courts may uphold the
use of VAMs under a rational basis test, they are suspect
about the wisdom of using VAMs to make significant
decisions with respect to teacher employment status.
But that does not mean that VAMs should be
relegated to the dustbin of educational policy history.
They may have important contributions to improving
teacher quality. They may be important “flags” for
228
See Hewitt, supra note 61, at 32.
229
CLOSE ET AL., supra note 35, at 14.
230
Id.
5152
TENNESSEE JOURNAL OF LAW AND POLICY
VOLUME 13 | WINTER 2019 | ISSUE 2
[574]
teachers, alerting them to investigate their practice a bit
further. VAMs may, someday, play an important role in
helping teachers.
Importantly, however, the use of VAMs must be
judicious, especially in light of their severe limitations.
VAMs cannot tell a teacher what causes a particular
result (the type of robust and actionable feedback a
teacher would want) and they are highly sensitive to
demographics and variables outside of a teacher’s control.
Yet, because VAMs were incorporated in high-stakes
decisions with such haste, especially with the impetus of
the Race to the Top, they were brought to scale, warts
and all.
Thankfully, states have a rare opportunity in
educational policy to take a bit more control over their
destiny under the Every Student Succeeds Act. They
can—and are—placing VAMs as a piece of a puzzle to
solve teacher quality issues. Many are beginning to adopt
laws and policies that minimize or eliminate their use in
high-stakes employment. That is a step in the right
direction, one that recognizes a relative value to VAMs in
the larger quest to improve public education.
5253
