Bartlett v. New York State Board of Law Examiners

Full title: MARILYN J. BARTLETT, Plaintiff, v. NEW YORK STATE BOARD OF LAW EXAMINERS…

Court: United States District Court, S.D. New York

Date published: Aug 15, 2001

93 Civ. 4986 (SS) (S.D.N.Y. Aug. 15, 2001)

Opinion

93 Civ. 4986 (SS)

August 15, 2001

Attorney for Plaintiff, JO ANNE SIMON, ESQ., Brooklyn, NY, ROBERTA MUELLER, ESQ., MARIANNE L. ENGELMAN LADO, ESQ., New York Lawyers for the Public Interest, Inc., New York, NY, Attorneys for Plaintiff

Attorneys for Defendants, ELIOT SPITZER, ATTORNEY GENERAL OF THE STATE OF NEW YORK, New York, NY, JUDITH T. KRAMER, ESQ., REBECCA DURDEN, ESQ., Assistant Attorneys General

REMAND FINDINGS OF FACT, CONCLUSIONS OF LAW, AND OPINION AND ORDER

SONIA SOTOMAYOR, U.S.C.J. Sitting by Designation:

INTRODUCTION

Plaintiff Marilyn J. Bartlett claims that she is entitled to reasonable accommodations in taking the New York State Bar Examination because she is an individual with a disability as defined by the Americans with Disabilities Act (the "ADA"), 42 U.S.C. § 12101, et seq., aria Section 504 of the Rehabilitation Act ("Section 504" or the "Rehabilitation Act"), 29 U.S.C. § 700, et seq. Specifically, plaintiff alleges that she has a learning disability¹ that "substantially limits" her "major life activities" of reading, or in the alternative, working, and that defendants New York State Board of Law Examiners and its members (the "Board") discriminated against her under Titles I and III of the ADA and Section 504 by failing to provide her with reasonable accommodations. 42 U.S.C. § 12102 (2)(A) (defining disability as "a physical or mental impairment that substantially limits one or more of the major life activities of [an] individual"); 29 U.S.C. § 705 (9)(B)(same).²

1 Plaintiff and her experts use the terms dyslexia and learning disability interchangeably. See, e.g., Moats Aff. ¶ 11. [l]ikewise use those terms interchangeably in this opinion and order.

2 Title II of the ADA prohibits discrimination by public entities, such as the Board, on the basis of disability. 42 U.S.C. § 12132. Title III of the ADA requires, inter alia, entities offering licensing examinations to provide reasonable accommodations to disabled individuals. 42 U.S.C. § 12189. Section 504 prohibits discrimination against disabled individuals by "any program or activity receiving Federal financial assistance." 29 U.S.C. § 794 (a). Because, in the context of this case, Titles II and III of the ADA and Section 504 "impose largely the same requirements," Bartlett v. New York State Bd. of Law Exam'rs, 226 F.3d 69, 78 n. 2 (2d Cir. 2000), I will treat these claims together.

After holding a 21-day bench trial³ and reviewing thousands of pages of exhibits and briefs, I concluded that plaintiff was entitled to reasonable accommodations on the bar examination because she was a disabled individual within the meaning of the ADA and the Rehabilitation Act. Bartlett v. New York State Bd. of Law Exam'rs, 970 F. Supp. 1094, 1126 (S.D.N.Y. 1997) (Bartlett I). I determined that plaintiff was not substantially limited in the — major life activity of reading because, when I compared her to the "average person in the general population" (as I was required to do by the ADA's implementing regulations, see 28 C.F.R. Pt. 35, App. A § 35.104; 29 C.F.R. § 1630.2 (j)(1)), Bartlett's history of self-accommodation [had] allowed her to achieve . . . roughly average reading skills (on some measures) when compared to the general population. Bartlett I, 970 F. Supp. at 1120. I found, however, that taking the bar: exam "implicates the, major life activity of working" because the bar exam functions like an employment test and that, when I compared plaintiff to persons with "comparable training, skills and abilities" (as I was permitted to do by the ADA's implementing regulations, see 29 C.F.R. § 1630.2 (j)(3)(i)), "her reading skills (which when compared to the general population are barely average) [were] well below normal." Bartlett I, 970 F. Supp. at 1121. I concluded that plaintiff was substantially limited in the major life activity of working because her inability to be accommodated on the bar exam excluded her from a class of jobs — lawyering — for which she was otherwise qualified. Id. at 1123-28. I awarded plaintiff injunctive relief in the form of reasonable accommodations on the bar exam,⁴ id. at 1146, and monetary damages of $2,500 for each of the five bar examinations that she had taken, id. at 1152. Defendants moved for reconsideration, which I denied. Bartlett v. New York State Bd. of Law Exam'rs, 970 F. Supp. 1094 (S.D.N.Y. 1997) (Bartlett II).

3 I use the terms "bench trial" or "original trial" to refer to the trial held before me from October to December 1995. Citations to the bench trial transcript are made with the designation "Trial Tr." followed by the relevant page number(s).

4 I ordered that the Board provide plaintiff with the following accommodations if she chose to retake the bar examination: (1) double the normally allotted time, spaced out over four days; (2) the use of a computer; (3) permission to circle multiple choice answers in the examination booklet; and (4) large print on both the New York State and Multistate Bar Examinations. Id. at 1146-47, 1153.

The Second Circuit affirmed in part, and vacated and remanded in part. Bartlett v. New York State Bd. of Law Exam'rs, 156 F.3d 321 (2d Cir. 1998) (Bartlett III). The Second Circuit agreed that plaintiff was an individual with a disability under the ADA and Section 504, but on different grounds. The Court held that I had erred in considering plaintiff's ability to self-accommodate in determining whether she was disabled and found that absent her self-accommodations, Bartlett was substantially limited in the major life activity of reading when compared to the average person. Id. at 329. Because it found that Bartlett was substantially limited in the major life activity of reading, the Circuit did not consider whether Bartlett was also disabled in the major life activity of working. Id. On those grounds, the Court affirmed my finding that plaintiff was entitled to reasonable accommodations in taking the bar examination. Although the Circuit also affirmed my determination that plaintiff was entitled to compensatory damages, the Court remanded with instructions for me to determine whether any of the Board's denials of Bartlett's requests for accommodations on each of the five bar examinations she took were based on illegal discrimination. Id. at 332.

The United States Supreme Court granted certiorari. New York State Bd. of Law Exam'rs v. Bartlett, 527 U.S. 1031 (1999) (Bartlett IV). The Supreme Court vacated the Second Circuit's judgment and remanded for the Circuit to reconsider this case in light of Sutton v. United Air Lines, Inc., 527 U.S. 471 (1999), Murphy v. United Parcel Serv., Inc., 527 U.S. 516 (1999), and Albertson's, Inc. v. Kirkingburg, 527 U.S. 555 (1999), decisions in which the Supreme Court held that corrective devices and mitigating measures must be considered in determining whether an individual is disabled under the ADA.

On remand, the Second Circuit again affirmed in part, and vacated and remanded in part. Bartlett v. New York State Bd. of Law Exam'rs, 226 F.3d 69 (2d Cir. 2000) (Bartlett V). With respect to whether plaintiff was substantially limited in the major life activity of reading, the Circuit held that although I had properly considered plaintiffs self-accommodations, my finding that plaintiff had roughly average skills on some measures of reading was insufficient to support the conclusion that her reading was not substantially limited. Id. at 81. The Circuit asked me to consider in the first instance whether plaintiffs reading skills are below average on any measure such that plaintiff's ability to read is substantially limited. Id. at 81. Thus, in this opinion and order, I will first address "whether Bartlett is substantially limited in the major life activity of reading by her slow reading rate, or by any other "conditions, manner, or duration" that limits her reading `in comparison to most people.'" Id. at 82.

The Circuit further held that if I find that plaintiff is not substantially limited in reading, I should make further findings with respect to whether plaintiff is substantially limited in the major life activity of working. Specifically, the Court directed me to determine, if necessary, "whether plaintiff has shown that it is her impairment, rather than factors such as her education, experience, or innate ability, that `substantially limits' her ability to work." Id. at 85.

Finally, the Second Circuit stated that it adhered to its original remand with respect to compensatory damages. Id. at 86. It added, however, that if I were to find that plaintiff is disabled, and therefore entitled to compensatory damages, I should limit my damage award to the bar exams, if any, where the Board had before it sufficient information to determine that plaintiff was disabled. Id.

I permitted the parties to submit further evidence on the issues before me on this remand, and I held a four-day remand trial⁵ where both parties had numerous expert and lay witnesses testify. Having witnessed all of this additional testimony and having studied the exhibits and affidavits submitted by the parties on the issues before me on this remand and in the original trial record, I conclude that, when considering both the positive and negative effects of plaintiffs self-accommodations, plaintiff is substantially limited in the major life activity of reading when compared to the average reader by her slow reading rate and by the fatigue caused by her inability to read with automaticity. I also conclude, in the alternative, that plaintiff is substantially limited in the major life activity of working because the Board's failure to accommodate her reading impairment is a substantial factor in her failure to pass the bar. For these reasons, I find that plaintiff is entitled to receive reasonable accommodations in taking the New York State Bar Examination. Accordingly. I award her injunctive and compensatory relief.

5 I use the term "remand trial" to refer to the trial held before me in February 2001. Citations to the remand trial transcript are made with the designation "Fr." followed by the relevant page number(s). Because many of the witnesses repeated themselves at various times, I am citing, as deemed appropriate, to only one place in the record where a witness made a given point.

FINDINGS OF FACT

In my first opinion and order in this action, familiarity with which is assumed, I made extensive findings of fact with respect to the evidence presented during the bench trial. I will not repeat those findings here, but rather will limit my factual findings in this opinion and order to those necessary to resolve the limited issues before me on this remand. Based on the testimony presented and the exhibits admitted during the remand trial, my additional factual findings pursuant to Fed.R.Civ. p. 52 are as follows:

I. Plaintiffs Psychometric Evaluations

Plaintiff has been evaluated by three psychologists: Phillip M. Massad, Ph.D. in December 1989; Richard F. Heath, Ph.D. in May 1993; and Rosa A. Hagin, Ph.D. in September 1994. Each of these psychologists issued a report in connection with his or her respective evaluations. (Pl.'s Ex. 20a, Massad's Psychoeducational Evaluation; Pl.'s Ex. 16, Heath's Psychoeducational Evaluation; Pl.'s Ex. 93, Hagin's Psychological Evaluation.) In my first opinion and order, I made extensive findings of fact regarding these psychologists' evaluations of plaintiff, Bartlett I, 970 F. Supp. at 1104-1110, and I reaffirm those findings here. To the extent that the content of these reports are relevant to my conclusions of law, I discuss them in more detail later in this opinion and order. See infra Conclusions of Law, Section IV.

II. Plaintiffs Bar Examinations

In my first opinion and order, I discussed in detail the circumstances surrounding each of the occasions on which plaintiff has taken the bar examination and, where relevant, the documentation she submitted in support of her requests for accommodations, Bartlett I, 970 F. Supp. at 1102-04, and I also reaffirm those findings here. To the extent this information is relevant to my conclusions of law, I also discuss them in more detail later in this opinion and order. See infra Conclusions of Law, Section IV.

Since my last opinion and order, plaintiff has taken the New York Bar Examination one additional time — in July 1999. (McAlary Aff. ¶ 2⁶; Tr. at 252, 389.) The Board provided plaintiff with the following accommodations on that exam: (1) double the normally allotted time, spaced out over four days; (2) a large print version of the examination; (3) a private testing room; (4) use of a laptop computer for the essay portions of the examination; and (5) permission to mark the multiple choice questions in the examination booklet. (McAlary Aff. ¶ 3; Tr. at 252.) Despite these accommodations, plaintiff failed the July 1999 bar examination with a score of 578 (a passing score was 660). (McAlary Aff. ¶ 4; Tr. at 252.) Plaintiff testified that personal circumstances affected her ability to take the July 1999 bar examination. Specifically, her husband was diagnosed with cancer and, one week before the bar examination, had surgery and began follow-up therapy. (Tr. at 390.)

6 Witnesses gave their direct testimony at the remand trial by way of affidavit. "Aff." refers to the affidavit of direct testimony of the named witness.

III. Plaintiffs Expert Testimony at the Remand Trial

A. Louisa Cook Moats. Ed.D.

Louisa Cook Moats, Ed.D., is an expert in the assessment of learning disabilities in both children and adults. (Moats Aff. ¶ 1.) Dr. Moats's graduate and doctoral studies focused on learning disabilities, and she has developed an expertise in the area of spelling errors experienced by individuals with dyslexia. (Moats Aff. ¶¶ 4-5.) Until 1996, Dr. Moats consulted with the Dartmouth College Academic Skills Center and the Dartmouth Medical School, where she reviewed reports submitted by students who had requested accommodations for their learning disabilities. Dr. Moats also served as an adjunct professor in the department of psychiatry at Dartmouth Medical School, where she trained psychology graduate students in assessing children with learning disabilities. (Moats Aff., Ex. 1, Moats Resume; Tr. at 40.)

Plaintiff contacted Dr. Moats in November 1989 while searching for a clinician to assess her for possible learning disabilities. Dr. Moats referred plaintiff to Dr. Massad. (Moats Aff. ¶ 10; Tr. at 44.) Dr. Moats discussed with Dr. Massad the types of tests he should perform. (Tr. at 44.) With respect to the reading rate problems plaintiff had reported to Dr. Massad, Dr. Moats believed that the measures which existed at the time to document reading rate were limited because the Nelson-Denny Reading Test⁷ was undergoing revision and because those tests could not show how plaintiff read different types of materials for different purposes. (Moats Aff. ¶¶ 12, 17, 18.) Given the limitations of tests of reading rate and the fact that plaintiffs clinical history evidenced that reading rate had been a life-long problem for her, Dr. Massad chose not to administer a formal measure of reading rate, and Dr. Moats concurred. (Moats Aff. ¶¶ 19, 20.)

7 The Nelson-penny Reading Test includes a reading rate test (measured by a words per minute score based on the number of words the individual read in the first minute of the test) and a reading comprehension test. (Pl.'s Ex. 127, M. Kay Runyan, "The Effect of Extra Time on Reading Comprehension Scores for University Students With and Without Learning Disabilities, Journal of Learning Disabilities, vol. 24, no. 2 (Feb. 1991).) This test is normed up to college students. (Defs.' Ex. AAA, Nelson-Denny Reading Test Examiner's Manual.)

After completing his evaluation of plaintiff, Dr. Massad consulted with Dr. Moats regarding the test data and results, including behavioral observations. (Moats Aff. ¶ 13.) Dr. Moats concurred in Dr. Massad's diagnosis of plaintiff as dyslexic. Dr. Moats testified that the most critical factor in assessing whether an adult has a learning disability is his or her history and presenting problems and stated that plaintiffs history and presenting problem, as well as the discrepancies between her comprehension and her word recognition and decoding abilities, were "highly consistent with life-long dyslexia." (Moats Aff. ¶¶ 21, 22; Tr. 48-51.) of further significance to Dr. Moats were the coping strategies that plaintiff employed, including her use of an index card with a cut-out hole for tracking, and the types of errors plaintiff made on the Wide Range Achievement Test-Revised ("WRAT") spelling subtest administered by Dr. Massad.⁸ (Moats Aff. ¶¶ 24, 26; Tr. at 46-47.) Dr. Moats opined at the remand trial that the Board had sufficient information to determine that plaintiff was disabled when presented with Dr. Massad's report. (Moats Aff. ¶ 29.)

8 At the remand trial, Dr. Moats also analyzed the types of spelling errors made by plaintiff on her February 1993 bar examination and concluded that they were consistent with a dyslexia diagnosis. (Tr. at 120-24.)

Dr. Moats testified that plaintiff's ability to read, write, and learn is limited as compared to most people.⁹ (Moats Aff. ¶ 29.) Unlike plaintiff, most people can recognize and process words and letters automatically, can comprehend and retain information they read with little repetition, can easily recognize words they have seen before, do not need to think deliberately about forming letters and consciously remembering what letters look like, and do not read or write significantly better with extended time. According to Dr. Moats, "[m]ost people do not read, write or learn with the level of difficulty experienced by Plaintiff." (Moats Aff. ¶ 30; see also Tr. at 52-53.) Dr. Moats stated that plaintiffs lack of automaticity in reading, which she defined as reading without employing significant conscious attention to the task, causes her to read slowly and laboriously and to seek circuitous routes around reading whenever possible. (Tr. at 53-54.)

9 As I noted in my first opinion and order, "[f]or purposes of this case, plaintiff's claimed disability collapses into an inability to read like the average person on tests like the bar examination, for that is the skill that plaintiff claims constricts her ability to engage in all the other relevant major life activities." Bartlett I, 970 F. Supp. at 1117. Thus, as I did in my first opinion and order, I will focus on the major life activities of reading, and in the alternative, of working.

B. Rosa A. Hagin. Ph.D.

Rosa A. Hagin, Ph.D., examined plaintiff in preparation for and served as plaintiff's lead expert witness in the original trial of this action. Bartlett I, 970 F. Supp. at 1109. Dr. Hagin provided further testimony at the remand trial. As I noted in my first opinion and order, Dr. Hagin is a leading researcher in the field of learning disabilities and has published numerous books and articles on the topic. She is a faculty member at Fordham University and at New York University Medical Center's department of psychiatry and supervises clinics at both schools. Id.

At the remand trial, Dr. Hagin testified about the limitations of using psychometric testing to assess learning disabilities in adults. Like Dr. Moats, Dr. Hagin believes that for adults suspected of having learning disabilities, the most critical element in their assessment is their clinical history, including the type and severity of the difficulties they experienced when learning to read and write and the type of help they have received. (Hagin Supp. Aff. ¶ 7.)

Dr. Hagin noted that "[t]here are no truly appropriate measures of reading assessment for adults and no tests are developed for this purpose, because generally adults are not tested." (Hagin Supp. Aff. ¶ 8.) The few tests that have adult norms merely provide a ceiling by extrapolation and do not establish norms that represent the reading ability of "most people" or an "average person." (Id.) For example, because the skills tested on the Word Attack and Word Identification subtests of the Woodcock Reading Mastery Test-Revised ("Woodcock") are normally mastered by sixth grade, the adult norms for those tests merely establish a ceiling for those skills. (Hagin Supp. Aff. ¶ 14.) Additionally, an individual's scores on the Woodcock do not reflect many decoding problems because the test is untimed, permits an individual numerous tries, and allows for self-corrections. (Tr. at 184.)

Because of these limitations, Dr. Hagin opined that when diagnosing an adult, an evaluator must look at the behaviors the individual uses when taking the tests, not merely the test results. An evaluator should look at the type of word attack skills employed, not merely whether the word is successfully attacked; how automatically an individual decodes words, not whether the word is eventually decoded; how the individual handles extended text; and whether the rate of reading would effectively meet the needs of an adult. (Hagin Supp. Aff. ¶ 9.) For example, the Woodcock score reflects accuracy, not automaticity of functions; instead, automaticity is assessed through clinical observation of how the individual goes about taking the Woodcock. (Hagin Supp. Aff. ¶ 15.) Dr. Hagin stated that an evaluator should also examine an individual's writing competence and ease, which is generally assessed through clinical and behavioral observations such as the manner in which the individual writes and whether the individual uses compensatory techniques or aids when writing. (Hagin Supp. Aff. ¶ 11.)

In evaluating the plaintiff, Dr. Hagin did the following. She noted that with respect to plaintiffs performance on the Weschsler Adult Intelligence Scale-Revised ("WAIS") test given by Dr. Massad, the discrepancy between plaintiffs verbal and performance intelligence quotients ("IQs") was highly significant and merited further investigation. (Pl.'s Ex. 93, at 1-2.) Plaintiff's response on the neuropsychological battery was uneven, and Dr. Hagin observed plaintiff using coping skills while reading, including using pointing cues to assist with keeping her place on visual tests, rehearsing rote sequencing items verbally, and slowing down her rate of response. (Id.) Dr. Hagin further noted that although plaintiff worked slowly and laboriously when providing her with a writing sample, plaintiffs sample contained five spelling errors, including "exciteing," "begining" and "teeching." ( Id. at 2.) Plaintiffs Woodcock Word Attack subtest scores were significantly below expectations (based on the published statistical correlation between that subtest score and IQ), and the discrepancy between her expected and actual performance occurred in less than 5% of the population. ( Id. at 3.)

During her evaluation of plaintiff, Dr. Hagin administered the Diagnostic Reading Test ("DRT"), a test of reading rate and comprehension, under both timed and untimed conditions, to determine: (1) if plaintiff's performance improved significantly with extra time; (2) if plaintiffs rate of reading decreased significantly when she had more time; and (3) if, regardless of timing, plaintiff used physical tracking techniques to maintain her place in text. (Hagin Supp. Aff. ¶ 16; Tr. at 172-73.) The DRT is written at an upper elementary school level in terms of its construction and the complexity of the language it uses. (Hagin Supp. Aff. ¶ 17.) Dr. Hagin avoided administering the Nelson-Denny, another measure of reading rate, because she was concerned plaintiff might be familiar with the test from having been an assistant school superintendent. (Id.)

Plaintiff read at 195 words per minute ("wpm") under timed conditions and scored at the 16th percentile for comprehension¹⁰ when compared to the oldest norm group, college freshmen. (Hagin Supp. Aff. ¶ 21.) Under timed conditions, plaintiff read at 156 wpm and scored at the 98 percentile for comprehension when compared to college freshmen. (Hagin Supp. Aff. ¶ 21; Tr. at 173-76.) Plaintiffs reading rate of 195 wpm under timed conditions translated to the 4th percentile when compared to college freshmen. (Hagin Supp. Aff. ¶ 22.) Using the norms for timed conditions (because there are no separate norms for untimed conditions (Tr. at 229.)), plaintiff's rate of 156 wpm translated to the 1st percentile when compared to college freshmen. (Hagin Supp. Aff. ¶ 22; Tr. at 173-76.) Dr. Hagin calculated plaintiffs "reading ease level," which is defined by the DRT manual as the grade level at which the individual's score is roughly comparable to the raw score at the 50th percentile, to be below the fourth grad. level. (Hagin Supp. Aff. ¶¶ 30, 31; Tr. at 233.)

10 Dr. Hagin's evaluation report erroneously stated that plaintiff scored at the 50th percentile for comprehension. Dr. Hagin testified at the remand trial that she had inadvertently transferred the percent correct to the percentile comprehension score. (Hagin Supp. Aff. ¶ 21; Tr. at 142.)

Dr. Hagin disagreed with the opinion of the defense experts that the highest norm group available — college freshmen — could be a proxy for "most people." (Tr. at 230-31.) College freshmen do not have the experience and exposure to language and reading that most adults over forty have, and therefore, plaintiffs true comprehension abilities on the DRT should be interpreted as even lower than her scores indicate. (Hagin Supp. Aff. ¶ 29.) To demonstrate. the severity of plaintiffs reading rate impairment, Dr. Hagin compared plaintiff to public high school students. Using plaintiffs timed rate of 195 wpm, plaintiff scored at the 10th percentile when compared to high school seniors, the 13th percentile when compared to high school juniors, and the 22nd percentile when compared to high school freshmen and sophomores. (Hagin Supp. Aff. ¶ 33.) Using plaintiffs untimed rate of 156 wpm (and using the timed DRT norms), plaintiff scored at the 2nd percentile when compared to high school seniors, the 3rd percentile when compared to high school juniors, the 5th percentile when compared to high school sophomores, and the 8th percentile when compared to high school freshmen. (Hagin Supp. Aff. ¶ 34.) Dr. Hagin concluded that plaintiffs scores on the DRT, when interpreted in conjunction with plaintiffs history and the observations of the clinicians who tested her, show that plaintiff's reading is "greatly compromised when compared with most people." (Hagin Supp. Aff. ¶ 36.)

In the context of explaining why the DRT does not directly measure automaticity, Dr. Hagin described the difference between automaticity and reading rate as follows: There are two systems involved in reading: a decoding system and a comprehension system. (Tr. at 176.) An individual like plaintiff who has problems decoding uses the majority of his or her mental energy figuring out what the words say, and therefore has difficulty understanding the ideas. (Tr. at 176-77.) She explained that reading rate is merely a superficial observation of reading and that the rate tells an evaluator nothing about whether the person understood. (Tr. at 177, 181.) Dr. Hagin distinguished plaintiffs reading problems from those of a "garden variety" poor reader, stating that a poor reader would not have the word attack problems that plaintiff consistently demonstrated. (Tr. at 175.) She stated that someone who is merely a poor reader — who may be lacking in the intelligence, background, or education necessary for reading — would be able to — decode fairly well but would not understand the ideas. (Id.) She also testified that a "garden variety" poor reader would not have the significant intra-ability discrepancies that plaintiff exhibited. (Tr. at 184.)

Dr. Hagin opined that plaintiffs ability to read, write, and learn is limited as compared to most people. (Hagin Supp. Aff. ¶ 51; Tr. 210-11.) Unlike plaintiff, most people can visually recognize and process words and letters automatically, can comprehend what they visually perceive with little or no repetition, can remember words they have seen before, can write easily and do not have to think deliberately about the process of writing itself, do not read or write significantly better with extended time, do not visibly struggle not to write in mirror image, automatically return to the left side of the page at the end of each line, and are not exhausted by the process of reading and writing. (Hagin Supp. Aff. ¶ 52.) Dr. Hagin opined that "[m]ost people do not read, write or learn with the level of difficulty experienced by Plaintiff." (Id.)

Dr. Hagin also testified that the Board had sufficient evidence before it of plaintiff's disability to reach this same conclusion. She stated that Drs. Massad's and Heath's evaluation reports, which were before the Board, were consistent with accepted practice at the time and clearly showed plaintiffs lack of automaticity. (Hagin Supp. Aff. ¶¶ 37-46.) For example, their reports explained plaintiff's history of reading problems, described behaviors evidencing lack of automaticity, and, in the case of Dr. Heath's report, explicitly stated that plaintiff reads "slowly and without automaticity." ( Id.; Tr. at 235-36.) The pattern of scores contained in those reports further supported Drs. Massad's and Heath's dyslexia diagnosis. (Hagin Supp. Aff. ¶ 38.)

Dr. Hagin also compared plaintiffs file with the files of other applicants who had requested accommodations from the Board due to their learning disabilities. She testified that she was unable to establish a clear method by which the Board's outside expert, Frank Vellutino, Ph.D., determined whether applicants were to receive accommodations and that she believed the decisions to be quite arbitrary. (Hagin Supp. Aff. ¶ 47.) Dr. Hagin stated that, on the whole, Dr. Vellutino focused on an individual's scores on the Woodcock Word Identification and Word Attack subtests but that occasionally he granted accommodations to applicants with higher Woodcock scores than plaintiff or to applicants whose files did not contain Woodcock scores at all. (Id.) For example, Dr. Hagin compared plaintiff's file to that of Applicant 57. She observed that plaintiff's report was more complete and detailed than that of Applicant 57 and that the difference between the applicants' Word Identification and Word Attack scores were "fairly narrow." (Hagin Second Supp. Aff. ¶ 11-17; Tr. at 203-05.) She also compared plaintiff's file to that of Applicant 6 and stated that the report in Applicant 6's file contained far less information than Dr. Massad's report and that the only difference between plaintiff's and Applicant 6's test scores was that the latter scored below the 30th percentile on the Letter Identification subtest. (Tr. at 205-07.)

C. Noel Gregg, Ph.D.

Noel Gregg, Ph.D., is an expert in the field of learning disabilities and has extensive clinical and research experience working with dyslexic adults. (Gregg Aff. ¶ 3; Tr. at 259.) She is the director of the University of Georgia Learning Disabilities Center and of the Regents Center for Learning Disorders, in which capacity she evaluates approximately 250 requests for accommodations each year (of which she rejects approximately 51 percent as insufficient). (Gregg Aff. ¶ 2; Tr. at 256.) She also is a full professor in special education at the University of Georgia. (Gregg Aff. ¶ 2.) During her tenure at the University of Georgia, Dr. Gregg has participated in or helped staff evaluations for approximately 2,000 clients. (Gregg Aff. ¶ 6.) Dr. Gregg also reviews requests for accommodations on the Scholastic Aptitude test ("SAT") by students with learning disabilities for the College Boards and the Educational Testing Service. (Tr. at 302.) Dr. Gregg also testified as an expert in learning disabilities in Root v. Georgia State Bd. of Veterinary Medicine, 114 F. Supp.2d 1324 (N.D. Ga. 2000).

Dr. Gregg testified in detail regarding each of the primary variables that impact reading — word recognition, listening comprehension, metacognition,¹¹ working memory, verbal ability, and background knowledge — and addressed each of these variables in relation to plaintiff (Gregg Aff. ¶ 9.1-9.6f.) Dr. Gregg concluded that the primary cause of plaintiff's reading problems is her deficits in the area of word recognition.

11 Dr. Gregg explained that readers use their metacognitive skills by "construct[ing] a mental model of context (inferencing) and monitor[ing] the accuracy and sense of that model as they read (comprehension monitoring)." (Gregg Aff. ¶ 9.2.)

Dr. Gregg stated that plaintiff has a high verbal ability (as evidenced by her superior verbal IQ score) and strong metacognitive skills. With respect to her metacognitive skills, although plaintiff can use her strong inferencing skills to compensate for her decoding problems in some situations, this strategy makes it difficult for her to grasp the details of the test, particularly in timed testing situations, because this strategy only allows her to get the "gist" of the text. Dr. Gregg opined that this problem could be greatly reduced by plaintiff being provided with alternative means to reading silently, such as a reader or assistive technology. (Gregg. Aff. ¶ 9.2.) Plaintiffs background knowledge, including her familiarity with legal literature, facilitates her ability to use her inferencing skills on legal exams. This background knowledge, however, does not help her with decoding specific words or reading quickly. (Gregg Aff. ¶ 9.3.) With respect to listening comprehension, Dr. Gregg theorized that if plaintiff had been administered a listening comprehension measure, her score on that test would be a better indicator of her reading ability than her verbal ability scores on intelligence tests because, although there is generally a high correlation between verbal intelligence and reading comprehension, that correlation does not hold for individuals with learning disabilities. (Greg Aff. ¶ 9.4.) Finally, Dr. Gregg theorized that, based on plaintiff's discussion of how she writes, plaintiff may have some problems with working memory that affect her reading. Specifically, Dr. Gregg hypothesized that plaintiff writes using short descriptions in lieu of longer, more detailed descriptions, because it is difficult for her to hold information in her working memory long enough to transcribe her ideas onto paper. (Gregg Aff. ¶ 9.5.)

Dr. Gregg testified, however, that the predominant cause of plaintiff's reading problems are her difficulties with word recognition. As a general matter, recent studies show that while the impact of word recognition on reading competence decreases as an individual gets older, the opposite pattern is true for adults with dyslexia. (Gregg Aff. ¶ 9.6.) Adults with dyslexia continue to demonstrate significant phonological¹² and orthographic¹³ processing deficits. Dr. Gregg referred to a study that compared adults with dyslexia to controls with a sixth grade reading level and found that the dyslexic adults read more slowly and relied more on sound-symbol information and context when reading words, although the adults demonstrated adequate reading comprehension as tested by a multiple-choice type of reading comprehension format. (Id.)

12 Phonology is the study of patterns of sound in speech. See Webster s II New Riverside University Dictionary (1994).

13 Orthography is the study of letter and spelling patterns in speech. See id.

Dr. Gregg explained that plaintiff lacks automaticity, which causes her to have difficulty reading and comprehending text because it takes her so long to decode individual words. (Gregg Aff. ¶ 9.6a, 9.6d.) Most people, even those with only an elementary school education, have automatized the processes of reading, spelling, and writing. (Gregg Aff. ¶ 9.6d.) Research has confirmed that the key to efficient reading is automaticity in processing words quickly and without conscious attention. Dr. Gregg compared the way that plaintiff reads — by focusing on the words themselves, word by word — to that of a young child learning to read. (Gregg Aff. ¶ 9.6e.) Additionally, while most readers vary their approach to reading depending on the level of difficulty of the material to be read and the time constraints within which that individual must complete the task, plaintiff cannot make these types of adjustments. (Gregg Aff. ¶ 9.6f.)

Dr. Gregg testified that plaintiff's history was consistent with a diagnosis of adult dyslexia. of particular importance is plaintiffs use of study groups and readers. Also, most adults do not routinely have others edit their work for basic grammar and spelling. (Gregg Aff. ¶ 9.6f.) Dr. Gregg observed that plaintiffs academic success is not inconsistent with her having dyslexia because she has achieved these results through non-traditional methods of self-accommodation. (Tr. at 294-95.) She also stated that plaintiffs scores on the Law School Aptitude Test ("LSAT"), SAT, and Graduate Record Examination ("GRE") were not inconsistent with her having dyslexia. (Tr. at 295-305.)

Dr. Gregg discussed how to assess an individual's reading ability. Reading involves two elements — a comprehension factor and a decoding factor; an accurate assessment requires measures aimed at both of these elements. (Gregg Aff. ¶ 11.) Based on statistical research that Dr. Gregg has recently completed, adequate performance on one measure of reading comprehension on portions of the Woodcock or Nelson-Denny will not predict the same behavior on other measures of comprehension. For example, a fill-in-the-blank task is processed at the sentence level and uses different cognitive and linguistic abilities than tasks processed at the word level. The result of this research confirms Dr. Gregg's opinion that a single test cannot be used as the sole predictor of ability and that professionals should not generalize reading competence based on a single measure. (Gregg Aff. ¶ 10.) Rather, patterns of scores are the most reliable means to identify performance levels. Dr. Gregg opined that average adults will not show the significant intra-ability discrepancies that plaintiff exhibits. (Gregg Aff. ¶ 13.) Additionally, particularly when diagnosing adults, background information and clinical observations are very important to an assessment. (Tr. at 269.) For example, because the Woodcock is untimed, the scores do not reflect plaintiff's fluency and rate problems; rather the clinicians' observation of time is vital to diagnosing those problems. (Tr. at 272.)

Dr. Gregg testified that normal readers rely more on the basis of direct visual orthographic information rather on their knowledge of spelling-sound information. Individuals with dyslexia, however, have problems developing reliable orthographic representations of words on a direct visual route — what Dr. Moats calls sight vocabulary and one of defendants' experts . . . calls raudamatized words — because of their weak spelling-sound correspondence skills. (Gregg Aff. ¶¶ 19, 20.) As a result, adults with dyslexia have inaccurate and slow word-recognition skills. (Gregg Aff. ¶ 21.) There are no measures that directly test orthographic processing in adults. (Tr. at 277.) Rather, orthography and phonology are evaluated through the analysis of the types of errors an individual makes on assessment tests. (Tr. at 280.)

Dr. Gregg testified that plaintiffs ability to read, write, and learn is limited as compared to most people. (Gregg Aff. ¶ 24.) Unlike plaintiff, most people can read and write quickly and automatically, can recognize words and letters automatically, remember what a word look like and quickly recognize that word the next time they see it, can write spontaneously without thinking about letter orientation, do not labor through a reading passage, do not need extended time for writing and reading, and are not exhausted by the process of reading and writing. Dr. Gregg opined that "[m]ost people do not read. write or learn with the level of difficulty experienced by Plaintiff." (Gregg Aff. ¶ 25.)

D. Nancy Mather, Ph.D.

Nancy Mather, Ph.D., is an expert in assessing learning disabilities and has personally assessed both children and adults with suspected learning disabilities. (Mather Aff. ¶ 3.) For the last fifteen years, she has taught courses on assessment at the University of Arizona. She serves on the College Board's panel which reviews requests for accommodations by high school students on the SAT. (Tr. at 413.) She consulted with Dr. Richard Woodcock in the development of the Woodcock Reading Mastery Test-Revised ("Woodcock") and co-authored a chapter in the manual for that test. Dr. Mather also wrote the manuals for the WoodcockJohnson Psychoeducational Battery ("Woodcock-Johnson") and co-authored and wrote manuals for the major revision of that test, the Woodcock-Johnson III. (Mather Aff. ¶ 4.)

With respect to using the Woodcock on adults, Dr. Mather stated that although there are adult norms, most of the basic reading skills on the test (including Word Attack and Word Identification subtests) are normally mastered by most people by the fifth or sixth grade. (Mather Aff. ¶ 15; Tr. at 419.) The decoding scores on those subtests alone are not a valid indicator of functional reading because they do not reflect the manner by which one decodes, the speed at which one reads, or the ability to comprehend what one reads. (Mather Aff. ¶ 15.) The Word Attack and Word Identification subtests only test the accuracy of word identification and are not designed to measure speed or automaticity. Rather, according to Dr. Mather, automaticity can only be assessed through clinical observations, such as whether the individual makes self-corrections. (Mather Aff. ¶ 16; Tr. at 428-29.)

Dr. Mather opined that one cannot rely solely on Woodcock scores for diagnosing learning disabilities and stated that the conclusions that Dr. Vellutino draws from an individual's Woodcock scores are incorrect and "based on fundamental misperceptions of the purposes and properties of the tests." (Mather Aff. ¶ 7.) The manual Dr. Mather wrote for the Woodcock-Johnson III recommends that the evaluator look at four levels in interpreting the scores: (1) qualitative data, including behavioral observations and error analysis; (2) age and grade equivalent scores; (3) the tasks the individual mastered; and (4) standard score percentile rank. In order to perform an appropriate evaluation, a clinician must look at all four types of information. (Tr. at 417-18.)

Dr. Mather also testified that most people learn and perform roughly equal to their intellectual capacities, and that the Woodcock-Johnson tests are constructed around this basic — principle. Thus, part of the task of identifying individuals with learning disabilities using the Woodcock-Johnson is documenting intra-ability discrepancies. (Mather Aff. ¶ 12.)

Dr. Mather' s review of plaintiff's assessment scores, history, and clinical observations lead her to believe that plaintiff has significant problems with orthography, which is reflected in her spelling errors and slow word perception. Plaintiff's orthographic problems are evidenced in the presenting problems and background sections of Dr. Massad's report, which discusses her difficulties with perceiving letters and words accurately, writing backwards and making reversals and inversions, and recalling word endings. (Mather Aff. ¶ 9.) Dr. Mather explained that plaintiffs orthographic problems prevent her from developing automaticity with sight vocabulary, which forces plaintiff to "reinvent" the spelling of a word each time she uses it based upon the word's sound. (Tr. at 426-28.) Further evidence of plaintiff's lack of automaticity and orthographic problems is the fact that, although for most people there is a high correlation between their listening and reading comprehension, plaintiff understands much better when she hears than when she reads. (Tr. at 436.) In addition to having problems with developing sight vocabulary, individuals with severe orthographic problems, like plaintiff, often lose their place or miss lines of text, confuse words that are similar in appearance, misread small words, and need to reread a word several times in order to decode it. (Tr. at 429.) Dr. Mather reiterated that scores alone do not capture these problems; rather, behavioral observations are also required. (Tr. at 430.)

Dr. Mather described plaintiffs reading problems as "severe" (Tr. at 429.) and compared her reading to that of a child:

Typically, by third grade, . . . most children . . . are reading to learn, they're not learning to read anymore. Dr. Bartlett still has components of the learning to read in terms of her behaviors. So, usually by the fourth, fifth grade most children are automatic, their attention is directed to meaning and comprehension, not [to] perceiving words.

(Tr. at 434.)

Significant to Dr. Mather is plaintiff's history of having others read to her, which is not a behavior of an average adult. (Mather Aff. ¶ 11.) Plaintiff's use of a cut-out index car for tracking evidences her automaticity problems, and the use of such a coping mechanism causes fatigue. (Tr. at 434.) As did plaintiffs other experts, Dr. Mather opined that the key to assessing learning disabilities in adults is a review of their clinical history, including the type and severity of difficulties in reading and writing they have experienced and the compensatory mechanisms they have developed. (Mather Aff. ¶ 13.) She also stated that the testing of plaintiff was consistent with practice at that time and that Drs. Massad's and Heath's reports contained ample evidence of learning disability. (Mather Aff. ¶ 18-25.)

Dr. Mather concluded that plaintiffs ability to read, write, and learn is limited when compared to most people. (Mather Aff. ¶ 26.) Unlike plaintiff, most people recognize and process words and letters automatically, comprehend what they visually perceive with little or no repetition, do not labor through a passage, do not need extended time for reading and writing, do not struggle to recognize words to which they have had extensive exposure, and are not exhausted by the processes of reading and writing. Dr. Mather opined that "[m]ost people do not read, write, or learn with the level of difficulty experienced by Plaintiff." (Mather Aff. ¶ 27.)

IV. Defendants' Expert Testimony at the Remand Trial

A. Ronald P. Carver, Ph.D.

Ronald P. Carver, Ph.D., is an expert in the area of reading rate and has spent over thirty years researching and writing on this topic. (Carver Aff. ¶ 3.) He is a professor of education at the University of Missouri-Kansas City. (Carver Aff. ¶ 2.) Dr. Carver's expert testimony addressed the question of "whether plaintiff's test results on the Diagnostic Reading Test (DRT) show a lack of automaticity, or a below average reading rate, or a disability in rate." (Carver Aff. ¶ 1.)

Carver has developed a reading theory called "rauding theory." (Pl.'s Ex. 188, R.P. Carver, "Progress in Understanding Reading From the Viewpoints of Stanovich and Rauding Theory," Journal of Literary Research, vol. 33, no. 2 (2001).) Rauding theory addresses how individuals normally read relatively easy text. ( Id. at 1.) Rauding theory provides that "poor readers use exactly the same normal reading processes as good readers" and that "[t]here is no malfunction or deficit in the normal reading process used by poor readers." ( Id. at 6.) In contrast to the theories of the majority of experts in the learning disability field (including all of plaintiff's experts and defendants' other expert), Dr. Carver believes that poor readers and individuals with what other experts call learning disabilities use the same processes for reading. According to Dr. Carver, a poor reader's lack of fluency is caused by words that have not been raudamatized by the reader, that is, become part of what plaintiff's experts call an individual's sight vocabulary.¹⁴ ( Id. at 6; Tr. at 577.) Rauding theory suggests that poor readers, including those individuals that others label as learning disabled, can increase their reading level by raudamatizing new printed words. ( Id. at 11.)

14 Raudamatized words are "words that an individual (a) knows when spoken, (b)can be pronounced accurately when seen in print, and (c) can be recognized relatively quickly because they have been over learned through practice to a speed limit." (Pl.'s Ex. 188 at 2.)

Dr. Carver has conducted research regarding the average reading rate of students from first grade to graduate school, using a representative sample of about thirty to fifty students in each grade and testing their reading rate based on a one-minute test. (Tr. at 514, 516.) According to Dr. Carver, individuals use five "gears" when reading text. Individuals read in "Gear 3," their "normal reading rate," when they are reading text that is not too difficult for them to understand, i.e., text for which they know the meaning of the words and easily understand the ideas in the text. Based on his research, the average college student's normal reading rate is 300 wpm. (Carver Aff. ¶ 4; Tr. at 522.) If an individual knows that he or she is going to be tested on what he or she reads, the individual will shift down to "Gear 2," the "learning gear," in which college students read about 200 wpm. (Carver Aff. ¶ 4.) If an individual is to be asked to recall facts about what he or she has read, such as being asked to write down what he or she remembers from the text, he or she will shift down to "Gear 1," the "recalling gear." College students read approximately 156 wpm in Gear 1. ( Id.) If an individual merely wants to get the gist of the text, he or she will read in Gear 4, the "skimming" gear, in which college students read around 450 wpm. (Carver Aff. ¶ 4; Tr. at 523.) Finally, if an individual only wants to locate certain target words in text, he or she will read in Gear 5, the "scanning" gear. College students scan text at approximately 600 wpm. ( Id.) Based on this research, Dr. Carver has developed grade equivalents for normal reading rates. (Carver Aff. ¶ 9.) He has not, however, conducted empirical research to determine whether his gear-shifting theories hold true for individuals who have been identified as reading disabled (under Dr. Carver's system) or identified as learning disabled under more traditional diagnostic systems. (Tr. at 573-74.)

Based on research regarding reading efficiency, which Dr. Carver defined as the number of sentences or thoughts an individual is comprehending during a certain amount of time, he has determined that the optimal rate of reading for a college student is approximately 300 wpm¹⁵ with accurate comprehension,¹⁶ regardless of the level of difficulty of the text, so as long as the text is below the individual's reading level. (Tr. at 516-17, 519-21.) Dr. Carver concluded that, at least with text below their reading level, readers do not vary their reading rate with the difficulty of the material. (Tr. at 518, 526.)

15 Dr. Carver noted that these two studies showed that the normal reading rate for a college student and the most efficient rate for a college student were the same — 300 wpm. He testified that this research confirmed that people tend to read at a rate that is most efficient for them, which is their normal reading rate. (Tr. at 520-21.)

16 Dr. Carver defined accurate comprehension as comprehending approximately 75 percent of the material read. (Tr. at 517.)

Dr. Carver also discussed his research regarding a fully functional literate adult's reading level, which is the grade level difficulty of text that an average adult can sufficiently comprehend. (Carver Aff. ¶ 9.) Dr. Carver sampled newspapers from all fifty states in the United States and determined that the average difficulty of the newspaper articles is at the eighth grade level. He, therefore, concluded that the average adult reads at the eighth grade level. (Carver Aff. ¶ 9; Tr. at 540-41, 578) Dr. Carver conceded that his conclusion was based on logic rather than empirical data. (Tr. at 578.)

Dr. Carver has developed a standardized system to diagnose reading disabilities. His purpose in developing this system was to create an "objective" diagnostic system that "does not rely on the subjective judgments of clinicians." (Carver Aff. ¶ 10; Tr. at 543, 549.) He — believes that clinicians have a vested interest in finding that an individual is disabled and that clinical observations can be unreliable. (Tr. at 546.) Thus, Dr. Carver uses a computer to administer and to interpret the scores of the tests and to determine if a student has a disability in reading level or reading rate.¹⁷ (Carver Aff. ¶ 10.) Dr. Carver testified that his diagnostic system can test for what he considers to be the important reading disabilities — reading rate, reading level, word identification, listening, and naming speed disabilities. (Tr. at 561.) Adult readers are considered disabled under Dr. Carver's diagnostic system "if their scores [are] below the sixth grade level, that is two [grade equivalents] below the eighth grade level, which is the traditional level expected for a functional literate adult." (Carver Aff. ¶ 10.) Dr. Carver has not examined whether there is a correlation between individuals identified as reading disabled under his system and individuals identified as learning disabled under more traditional systems. (Tr. at 560-61.) As far as Dr. Carver is aware, only one charter school¹⁸ is currently using his diagnostic system, and no state or school system has adopted it. (Tr. at 543, 552.)

17 The words are administered by difficulty level (easy to difficult) using adaptive testing, which avoids giving an individual all of the items on the test. (Tr. at 557, 586.) Under this system, if an individual gets a harder word correct, the computer scores as if the individual got all of the easier words correct. Readers hit their highest level only once they have answered ten words incorrectly in a row. (Tr. at 557-58.)

18 A charter school is a public school established and operated under a charter from a local board of school directors and independently of a local board of education. See generally, e.g., N.J. Stat. Ann. § 18A:36A-3 (defining charter schools); 24 Pa. Cons. Stat. Ann. § 17-703-A (same).

Dr. Carver also testified about his research on the effect of extra time on readers. He has found that people's comprehension is very influenced by a lack of time if they do not have time to finish the text they are reading. He determined that once individuals reach a point where they have time to finish the text, their comprehension will increase gradually with time — until a length of time that is equal to approximately twice the time it took them to read it once. (Tr. at 545.) After that point, there is no further increase in comprehension when an individual is given extra time. (Tr. at 545-46.)

With respect to whether he believes plaintiff reads without automaticity, Dr. Carver testified that there is no standard method for measuring automaticity in reading. (Carver Aff. ¶ 6.) He stated that while some evaluators might consider hearing an individual who reads aloud "in a halting manner while mispronouncing several words" to be an indicator of lack of automaticity, he opined that the most credible indicator of the degree of automaticity is silent reading rate in words per minute while reading normally, that is, when the text is below the individual's grade level reading. (Carver Aff. ¶ 6; Tr. at 538.) Dr. Carver cautioned, however, that there is not an "objective way to measure the normal reading rate of a person who now claims to read silently very slowly" because "[a] person can always perform at levels lower than their [sic] capacity or ability." (Carver Aff. ¶ 11; see also Tr. at 545.)

He also discussed whether plaintiff's DRT scores evidence a reading rate disability. He testified that plaintiff's reading rate on the DRT under timed conditions — 195 wpm — is a "reasonably valid" measure of plaintiff's normal reading rate because the text of the DRT is below plaintiff's reading level.¹⁹ (Carver Aff. ¶ 7; Tr. at 529.) However, he considers the 195 wpm rate to be a low estimate of her normal rate because she may have shifted down to "Gear 2," the learning gear in anticipation of being asked questions on what she read. (Carver Aff. ¶ 8.) Under Dr. Carver's system of grade equivalents for normal reading rate, plaintiffs rate of 195 wpm is equivalent to the eighth grade level²⁰ (Carver Aff ¶ 9; Tr. at 539.), which Dr. Carver considers to be the average rate expected of fully functionally literate adults based on his newspaper research. (Carver Aff. ¶ 12; Tr. at 539.) Dr. Carver concluded that "[t]his reading rate at the eighth grade level disqualifies [plaintiff] as being reading disabled because the only published standard for reading rate disability i.e. Dr. Carver's system] requires that an adult have a normal reading rate at the fifth grade level or below." (Carver Aff. ¶ 12; see also Tr. at 542.)) He also opined that a rate of 195 wpm on the DRT, with 75 percent comprehension,²¹ "would not be showing an automaticity problem at all." (Tr. at 535.) Specifically, he stated that plaintiff's reading rate is only a tenth of a millisecond per word slower than the average college student and is inconsistent with plaintiff's description of how she reads. (Tr. at 536.)

19 Dr. Carver noted that the DRT is given to seventh graders (Carver Aff. ¶ 7.), and stated that he believed it to be written around the seventh to ninth grade level. (Tr. at 533.)

20 This rate means that individuals in the eighth grade, reading eighth grade reading level material, will read at a rate of 195 wpm. (Tr. at 542.)

21 Dr. Carver noted that plaintiff got 15 of 20 questions correct, or 75 percent, which he considered to be adequate comprehension. (Tr. at 530.) The DRT manual provides that an individual must get at least 75 percent correct to be considered comprehending the text. (Tr. at 576. 594.) Dr. Carver was dismissive of the fact that 75 percent correct translated into the 16th percentile (Tr. at 530) because he believed the norms established by the DRT are skewed by the fact that schools giving the DRT choose whether to report their scores to the company. (Tr. at 527-28.)

B. Dawn Flanagan, Ph.D.

Dawn Flanagan, Ph.D., is an expert in the intellectual assessment, evaluation, and testing of individuals suspected of having learning disabilities. (Flanagan Aff. ¶ 2.) She has developed the diagnostic theory of cross-battery assessment, in which practitioners select — diagnostic tests to augment intelligence tests in an effort to measure a broader range of abilities. (Tr. at 602-03.) She is an associate professor of psychology at St. John's University and teaches a number of courses on assessment of learning disabilities. (Flanagan Aff. ¶¶ 1, 2.) Dr. Flanagan consults with the following organizations, which she assists in determining whether applicants are entitled to accommodations: the National Board of Medical Examiners, the Association for Investment Management and Research, and the Board of Law Examiners in Pennsylvania, Minnesota, and Maryland. She also consults with a number of major test publishing companies regarding the development of intelligence tests. (Flanagan Aff. ¶ 3.) Dr. Flanagan has testified as an expert in two cases involving requests for accommodations by individuals alleging that they have learning disabilities — Gonzales v. Nat'l Bd. of Med. Exam'rs, 60 F. Supp.2d 703, aff'd, 225 F.3d 620 (6th Cir. 2000), cert. denied, 121 S.Ct. 1999 (2001), and Price v. Nat'l Bd. of Med. Exam'rs, 966 F. Supp. 419 (D.W.Va. 1997). (Tr. at 605.)

The Board asked Dr. Flanagan to assess whether plaintiff's "documentation substantiates a diagnosis of a learning disability (i.e., dyslexia or reading disability)." (Flanagan Aff. ¶ 5.) In answering this question in the negative, Dr. Flanagan relied on the following information related to plaintiffs reading performance: (1) the results of the psychometric tests given by Drs. Massad, Heath. and Hagin; (2) plaintiffs performance on standardized, timed tests she took without accommodations; and (3) plaintiff's record of academic success without accommodations. (Flanagan Aff. ¶ 7.) Dr. Flanagan stated that reading performance involves four major skills — reading comprehension, reading decoding, reading fluency, and general fund of knowledge and information — and that an individual cannot be considered significantly limited in reading relative to most people unless he or she is deficient in at least one of these skills. — (Flanagan Aff. ¶ 8.)

Dr. Flanagan testified regarding the results of the psychometric tests administered by Drs. Massad, Heath, and Hagin. The general population, which Dr. Flanagan equated with "most people," falls within plus or minus one standard deviation of the mean, which corresponds to a percentile range of the 16th to 84th percentiles²² (Flanagan Aff. ¶ 9 (a)(i); Tr. at 616.) With respect to each of the psychometric tests plaintiff took, except the DRT, Dr. Flanagan noted that plaintiff scored within the average range. (Flanagan Aff. ¶ 10.) Dr. Flanagan correlated each of the psychometric tests with the four skills that make up reading performance and determined that the test results show that plaintiff is not deficient in any of these skills. (Flanagan Aff., Exs. C-G; Tr. at 618-19.) According to Dr. Flanagan, plaintiffs scores on the comprehension portions of the Woodcock, which were untimed, suggest that her skills of decoding, speed, and vocabulary knowledge are sufficient for an average adult. (Flanagan Aff. ¶ 10(a).) Plaintiffs decoding scores, such as on the Woodcock's Word Attack and Identification subtests, also suggest that plaintiff's requisite skills of phonological awareness and processing are sufficiently developed. (Flanagan Aff. ¶ 10(b).) Finally, plaintiffs superior verbal IQ on the WAIS shows that she has a superior fund of general knowledge and information. (Flanagan Aff. ¶ 10(d).)

22 Dr. Flanagan noted that some other experts use the range between the 25th and 75th percentiles, where fifty percent of the population falls, as the range of average. (Tr. at 617.)

Despite plaintiff's low scores on the DRT, Dr. Flanagan testified that plaintiff's reading fluency, which she equated with reading rate,²³ is average when compared to most people. Dr. Flanagan stated that plaintiff's percentile ranks on both the reading rate and comprehension tests on the DRT are skewed because she was compared to college freshmen, a . . . category that is not synonymous with "most people." (Flanagan Aff. ¶ 10(c) at 13-14; Tr. at 620, 640.) Specifically, fluency is impacted by speed of processing, which declines naturally with age. Therefore, according to Dr. Flanagan, by comparing plaintiff to individuals more than 25 years her junior, plaintiffs percentile scores on the DRT significantly underestimate her abilities relative to most people.²⁴ (Flanagan Aff. ¶ 10(c) at 14; Tr. at 640-43.) In any event, plaintiffs reading rate on the. DRT under timed conditions, falling in the 16th percentile, was within the average (albeit borderline) range. (Tr. at 616, 618-19, 643.) Further, Dr. Flanagan assumed Dr. Massad found plaintiffs fluency to be average based on his comment at original trial that he did not report plaintiffs performance on the Oral Grey Reading test because he did not see anything remarkable to report. (Flanagan Aff. ¶ 10(c) at 15.) Therefore, Dr. Flanagan concluded that the psychometric tests administered to plaintiff show that plaintiff performs within normal limits on all indicators of reading performance. (Flanagan Aff. ¶ 10(e).)

23 Dr. Flanagan used the terms fluency and reading rate interchangeably. (Tr. at 668.) She stated that automaticity, defined as the ability to immediately recognize a word, precedes fluency. (Flanagan Aff. ¶ 10(c).) She testified that plaintiffs decoding test scores indicate that her automaticity is similar to that of "most people." ( Id.)

24 For demonstrative purposes, Dr. Flanagan showed that an individual who scores at the 50th percentile on the Woodcock-Johnson III Reading Fluency Test when compared to 46 year-olds (plaintiffs age when she took the DRT) falls between the 13th and 21st percentiles when compared to college students. ( Id.)

Dr. Flanagan opined that plaintiff's experts' conclusion that plaintiff has a learning disability is unsubstantiated. (Flanagan Aff. ¶ 15(b).) She stated that plaintiffs experts' reliance on plaintiff's discrepant but average test scores on psychological measures was misplaced because "[p]erformance that falls within the [a]verage range of functioning cannot be misconstrued as [d]eficient." (Flanagan Aff. ¶ 15(b) at 27.) A significant discrepancy between an individual's WAIS verbal and performance IQs is not automatically indicative of a learning disability. (Flanagan Aff. ¶ 15(c) at 28.) She criticized Dr. Hagin for equating the abnormality of the discrepancy with impairment. (Flanagan Aff. ¶ 15(c) at 29.) Dr. Flanagan also noted that it is typical to find at least two significant discrepancies in the profiles of normal people. (Flanagan Aff. ¶ 15(c) at 30; Tr. at 648-50.) Further, Dr. Flanagan believes that the clinicians' personal observations of plaintiff have little value because the data unequivocally indicates that plaintiff is not impaired. (Flanagan Aff. ¶ 16.) She opined that actuarial methods are superior to clinical methods because examiners become predisposed to seeing only those patterns in the data that support their assumptions and to minimizing the data that are counter to those assumptions. (Flanagan Aff. ¶ 16.) With respect to the clinical observations of plaintiff, she found the evaluations to be inconsistent in many respects and therefore placed little weight on them. (Tr. at 609-13, 644-46.)

Dr. Flanagan testified that the real-world outcomes of plaintiff's reading performance also show that she reads as well or better than most people. Plaintiff has taken numerous timed, standardized tests and has scored within average range. (Flanagan Aff. ¶ 11(a).) In particular, Dr. Flanagan stated that the LSAT is a "sophisticated reading comprehension test" that compares the test-taker to a "highly competitive normative group." (Dr. Flanagan Aff. ¶ 11(a); 624-45.) She opined that plaintiff's LSAT score at the 51st percentile²⁵ "is representative of a person of average ability in comparison to other students competing for law school, not the score of an impaired individual who is substantially limited in her reading abilities as compared to most people." (Flanagan Aff. ¶ 11(a).) Dr. Flanagan further stated that plaintiff's SAT and GRE scores were generally within average range, although a few were more than one standard deviation below the mean. (Tr. at 627-28, 712; Pl.'s Ex. 187.) Dr. Flanagan also testified that plaintiff's demonstrated knowledge of German, such that she could pass an informal examination to meet her Ph.D. foreign language requirement, is inconsistent with plaintiff's purported phonological difficulties, especially given that German is a more — phonetically complex language than English. (Flanagan Aff. ¶ 11(b) at 18-19; Tr. at 630.) Finally, Dr. Flanagan stated that plaintiffs average performance in college (2.10 grade point average ("GPA")) and graduate school (3.8 GPA) without accommodations also evidences that plaintiffs reading ability is not significantly restricted. (Flanagan Aff. ¶ 11(c); Tr. at 631.) Thus. Dr. Flanagan concluded that plaintiffs reading performance in real-world situations demonstrates that her ability to read is not significantly restricted as compared to most people. (Flanagan Aff. ¶ 11(d); Tr. at 631.)

25 The 51st percentile rank is based on scores from June 1988-February 1991. (Flanagan Aff., Ex. I.) Defendants presented evidence that plaintiffs LSAT score was actually at the 58th percentile in the year she took it. (Defs.' Ex. NNNN.)

Dr. Flanagan noted that the only real-world manifestation involving reading performance on which plaintiff has been unsuccessful is the bar examination. (Flanagan Aff. ¶ 12.) Dr. Flanagan hypothesized that plaintiff's repeated failures on the bar exam is caused by her lack of domain-specific knowledge and "[t]hat her manner, condition. [and] duration in terms of how she reads is not the primary cause for failing the bar exam." (Tr. at 655; see also Flanagan Aff. ¶ 13; Tr. at 656.) According to Dr. Flanagan, the indices of plaintiff's knowledge of the law suggest that plaintiff does not have a sufficient mastery of the law necessary to pass the bar exam. Plaintiff had a 2.09 GPA after her first year of law school without accommodations, and her performance did not improve appreciably once she was given accommodations. Plaintiff graduated with a 2.32 GPA, which placed her in the bottom ten percent of the class. (Flanagan Aff. ¶ 13.) As further evidence that a lack of domain-specific knowledge, and not reading problems, caused plaintiff to fail the bar exams, Dr. Flanagan noted that plaintiff did not pass the bar exam on three separate occasions in which she was provided with accommodations. (Flanagan Aff. ¶ 13; Tr. at 655.)

Dr. Flanagan further testified regarding her operational definition of learning disabilities and its application to plaintiff She described a five-level analysis, where the criteria of each lower level must be met before proceeding to the next level of the model and the criteria for all levels must be met for a learning disability diagnosis. (Tr, at 631, 634; Defs.' Ex. MMMM.) The first level is measurement of specific academic skills and measured knowledge. Based on the testimony of plaintiff and her experts, Dr. Flanagan believes that plaintiff established a deficit in basic reading. (Tr. at 633.) Level II involves exclusionary factors for explaining the deficit, such as insufficient education, language differences, and cultural differences, and Dr. Flanagan stated that plaintiff's history did not present any of these factors. (Tr. at 633.) At Level III, any deficit noted at Level I must be explained in terms of an underlying cognitive processing problem. Dr. Flanagan stated that at Level III, there are no data to substantiate that there is any underlying cognitive processing problem, and that she would not move on to Level IV of the model but rather would go back to Level II and re-examine possible exclusionary factors. (Tr. at 634-35.) In her analysis of plaintiff, Dr. Flanagan did not, in fact, go back and re-examine Level II. Rather, Dr. Flanagan concluded that plaintiff does not have a learning disability based on the lack of test scores below the average range at Level III.

Dr. Flanagan also testified that, even if I find that plaintiff has a reading disability, the Board had insufficient information before it to make that determination. (Flanagan Supp. Aff. ¶ 1.) She noted that only the reports from Drs. Massad and Heath, and not from Dr. Hagin, were before the Board. (Flanagan Supp. Aff. ¶ 2.) Neither Drs. Massad nor Heath reported scores below the average range or administered a reading rate test, and Dr. Heath's observation that plaintiff reads without automaticity was unsupported by test scores. (Flanagan Supp. Aff. ¶¶ 3, 4.) Dr. Flanagan concluded that, based on the scores reported by Drs. Massad and Heath, neither the Board nor its expert had sufficient evidence before them to conclude that plaintiff was disabled. (Flanagan Supp. Aff. ¶¶ 7.)

Dr. Flanagan also compared plaintiffs file to that of Applicant 57. The difference between the two files, and the reason, in Dr. Flanagan's opinion, that Applicant 57 was accommodated when plaintiff was not, was that Applicant 57 performed below the average range (of the 25th to 75th perventiles) and below Dr. Vellutino's cut-off (of the 30th percentile) on the Woodcock Word Identification and Word Attack subtests. (Flanagan Supp. Aff. ¶¶ 12, 15.) Also significant was the fact that Applicant 57 provided the Board with additional information regarding her early diagnostic history, which had been documented since she began the first grade. (Flanagan Supp. Aff. ¶ 16.) As to Applicant 6, Dr. Flanagan conceded that the majority of Applicant 6's Woodcock scores were similar to those of plaintiff but stated that the difference between the two applicants was Applicant 6's "very low score" of the 12th percentile on the Letter Identification. (Flanagan Supp. Aff. ¶ 18.) Dr. Flanagan opined, however, that a low score on the Letter Identification test is clinically meaningful but not indicative of an impairment that — warrants an accommodation. (Flanagan Supp. Aff. ¶ 21.)

V. Lay Testimony at the Remand Trial

A. Marilyn Bartlett

Plaintiff described the methods and coping techniques she uses for reading under both timed and untimed conditions. When reading under timed conditions, plaintiff tries to get the general meaning of the text by reading in "chunks" by searching for "important" words like nouns and skipping over less important words such as "who," "where," "when," etc. (Tr. at 383.) She tries to use context to assist her. She employs a totally different strategy in untimed situations. There, she gives herself the "luxury" of reading each of the words by first deciphering the letters, then sounding out the words, then ascertaining the context of the words, and then determining the meaning of the entire sentence. (Tr. at 384.) She uses different devices, depending on what is available and the sophistication of the text, to assist her with reading, including using her fingers to assist with tracking or a card with a hole cut out to help her focus when reading difficult text. (Tr. at 384.)

Plaintiff testified about how her reading impairment affects her everyday life. She reads all types of material, whether simple e-mails or professional level materials, slowly and haltingly, sounding out each word. (Bartlett Supp. Aff. ¶¶ 5-6.) She rereads material two or three times to make sure she understands it, and highlights negatives ( e.g., not) to make sure she sees these words when rereading the text. (Bartlett Supp. Aff. ¶ 8.) She often asks her colleagues, secretary, and husband to read material to her. (Bartlett Supp. Aff. ¶ 8.) She avoids reading for pleasure because it is too laborious and exhausting. (Bartlett Supp. Aff. ¶ 9-10.) Because of her spelling difficulties and her inability to skim, she is unable to use research tools such as electronic databases or computerized card catalogues. (Bartlett Supp. Aff. ¶ 13.) For example, she described the process she uses if she has to read an e-mail. She prints it out, brings it to the photocopier and enlarges it, and then starts reading letter by letter to decipher the word and understand its meaning in context. (Tr. at 386.)

Plaintiff also discussed the difficulties she experiences with respect to writing. She must consciously think about what each letter looks like and must, for example, remind herself of the difference between a "b" and a "d" or a "g" and a "q." (Bartlett Aff. ¶ 23.) Further, the sophistication of her writing is affected by her spelling problems because she chooses words she can spell. (Tr. at 389.) She stated that she writes very slowly and laboriously. (Tr. at 387.)

Plaintiff described the strategies that she has used to succeed in various academic settings. She made it through high school with the assistance of her mother, who would help her by reading her homework assignments to her. (Tr. at 364.) In college, while she continued to have others read to her, she also started participating in study groups, where she could learn by listening to others. (Tr. at 364-65.) She was successful in graduate school because there was less of an emphasis on exams and more of an emphasis on term papers, on which plaintiff could receive assistance from others in typing and editing. (Tr. at 365.) As one example, she described how she completed her dissertation. She used a reader to go through the library's card catalog and to read the sources to her which she would select, she dictated the substance of the dissertation, and then she hired a typist/editor to help her write it. (Tr. at 387, 398-99.) In law school, she used note-takers, readers, and participated in study groups. (Tr. at 365-66.) Plaintiff stated, however, that although she was given time-and-a-half to complete her law school exams starting in her second year, she generally never had enough time to finish her exams. (Tr. at 365.)

Plaintiff also described the strategies that she used in taking standardized timed tests. For example, on the LSAT, she took a course that taught her a strategy called "hit and move on," where she would read the question at the end of the paragraph first and, if possible, make a quick guess. She did not have time to read the entire text on the LSAT. (Tr. at 366.) She did not remember employing any particular strategy when taking the SAT and GRE, but she stated that her scores were so low that she had difficulty getting into college and graduate school. (Tr. at 368-70.)

Plaintiff described how she was able to learn and pass a German proficiency examination. Plaintiff learned German while living in Germany for four years. She is fluent in speaking, but not reading, German. The single instance she took a formal course in German, which required her to read and write German, she received a D as "a gift." (Tr. at 373-78.) Plaintiff explained that she passed out of her doctoral program's foreign language requirement by speaking to the head of the German department in German, reading silently a paragraph from a first year German textbook in an untimed manner, and answering questions the professor asked about the paragraph. (Tr. at 379-80.)

B. Maryann Marino

Maryann Marino is plaintiff's secretary at the New York Institute of Technology School of Education and Professional Services. (Marino Aff. ¶ 1.) Marino testified regarding the type of assistance with reading she provides to plaintiff. For example, Marino reads all of — plaintiff's e-mails to her, dials outgoing calls for plaintiff, and accompanies plaintiff to meetings to take substantive notes for her. (Marino Aff. ¶ 7-11.) Plaintiff often asks Marino to read correspondence to her. (Tr. at 135-36.) Marino does not provide a similar level of assistance to the other professor with whom she works. (Marino Aff. ¶ 13.) Marino has observed plaintiff read and noted that plaintiff reads slowly and rereads material several times. (Marino Aff. ¶ 12.)

C. John J. McAlary

John J. McAlary, the Deputy Executive Director of the New York State Board of Law Examiners, testified that plaintiff sat for the New York State Bar Examination in July 1999. (McAlary Aff. ¶¶ 1-2.) He stated that although she was provided with accommodations, she did not pass the exam. (McAlary Aff. ¶ 3-4.)

VI. Evidence of Disability

A. Psychometric Testing

Defendants claim that plaintiff cannot be "substantially limited" in reading when compared to "most people" because she scored within the average range, defined as between the 16th and 84th percentiles, on all of the psychometric measures administered to her. Defendants and their experts argue that I should give little or no weight to clinical observations of plaintiff's reading problems because they are "subjective" and biased; they claim that clinicians are predisposed to seeing patterns in data that support a finding of disability and to minimizing data that are counter to those findings. (Flanagan Aff. ¶ 16; Carver Aff. ¶ 10; Tr. at 543, 546, 549.) Plaintiff's experts have convinced me, however, that learning disabilities cannot be captured by psychometric measures alone and that clinical observations are essential to a diagnosis of — learning disabilities.

While I discuss in more detail below the limitations of the specific psychometric measures administered to plaintiff, the following are observations about the limitations of using psychometric measures to diagnose learning disabilities generally, particularly with adults. First, all of the experts, including defendants' experts, agree that no psychometric test directly measures automaticity. (Hagin Aff. ¶ 15; Mather Aff. ¶ 16; Carver Aff. ¶ 6; Tr. 237, 241.) For example, the Woodcock tests decoding accuracy, not decoding fluency or automaticity. (Mather Aff. ¶ 16; Tr. at 429.) Thus, no matter how low plaintiff scored on any psychometric measure, that score could not reflect what plaintiff's experts diagnose as a significant aspect of her disability — lack of automaticity.

Second, the psychometric measures are similarly limited with respect to diagnosing reading rate problems, another significant aspect of plaintiffs disability according to her experts. Reading rate tests like the DRT and Nelson-Denny merely give a snapshot of where in a paragraph of text an individual is after a short period of time (three minutes for the DRT and one minute for the Nelson-Denny). (Pl.'s Ex. 186; Defs.' Ex. AAA.) The reading rate score does not take into account, for example, whether the individual read each word or skipped words or whether he or she reread that text before later answering the comprehension questions. There is no direct correlation between reading rate and comprehension on these tests. While there was some testimony at trial regarding the new Woodcock-Johnson III Reading Fluency test, which plaintiff has never taken, the co-author of that test noted that it too has limitations — it measures an individual's ability to read very easy sentences and does not measure how fast an individual can read material with heavy comprehension demands. (Tr. at 450-52.)

Third, most psychometric measures are extremely limited in their application to adults. Many psychometric tests, such as the DRT and Nelson-Denny, lack adult norms. (Pl.'s Ex. 186; Defs.' Ex. AAA.) Even when a psychometric measure does have adult norms, like the Woodcock, those norms merely set a ceiling for low-level skills like decoding and letter identification. (Hagin Supp. Aff. ¶¶ 8, 14.) While defendants' attempt to extrapolate test results from psychometric measures with no adult norms using adult norms from other psychometric measures has superficial appeal (Flanagan Aff. ¶ 10(c) at 14.), I agree with plaintiffs experts that such an analysis relies on untested assumptions and is not statistically defensible.²⁶ (Tr. at 194-95; 452-53.)

26 I note that in my first opinion and order, I rejected Dr. Vellutino's attempt to do what Dr. Flanagan does here — extrapolate plaintiff's DRT scores using the norms from other psychometric measures. Bartlett I, 970 F. Supp. at 1114 ("Finally, I do not credit Dr. Vellutino's attempt to equate Bartlett's low DRT reading rate score with an average rate by extrapolation to other tests. This approach is seriously infirm in that it attempts to compare scores on different tests with different subject populations.").

Finally, although these psychometric measures are relevant diagnostic tools, they cannot alone answer the question of whether an individual is substantially limited in the major life activity of reading as compared to "most people." Because the purpose of these measures is to diagnose learning disabilities, not to determine whether an individual is disabled under the ADA or Rehabilitation Act, these test are not normed to "most people."²⁷ While psychometric measures available to diagnose adults have improved since this case was instituted almost a decade ago, no measure has yet been developed that can completely test the complexity of the reading process and whose scores alone can show the way that a flaw in one or more component of reading can affect an individual. As I discuss in more detail below, even if tests results appear average, those average scores tell nothing about the processes an individual uses and the difficulties that individual experiences in achieving those results. In sum, I cannot determine whether plaintiff is "substantially limited" in the "manner, condition or duration" in which she reads compared to "most people" solely based on her scores from the psychometric measures that have been administered to her. Rather, clinical judgment remains an important part of my determination of whether plaintiff is disabled.

27 The Second Circuit's most recent opinion suggests its dissatisfaction with the norm groups on the psychometric measures administered to plaintiff, particularly the measure of her reading rate. Bartlett V, 226 F.3d at 81-82 ("Bartlett's DRT test results, which show a reading rate in the fourth percentile or below as compared to college freshmen, are of limited value, because the proper reference group is 'most people,' not college freshmen.").

1. Woodcock Scores

Defendants' current argument is but a slight variation of the one they made, and I rejected, in the original trial of this action. There, defendants argued that I should rely on the results of psychometric testing alone, particularly the Woodcock, in determining whether plaintiff is disabled and that I should reject plaintiffs evaluators' clinical judgments as subjective. I disagreed:

In short, I do not accept [defendants' experts'] conclusions that reading disabled individuals are incapable of having the test scores reflected by plaintiff. Plaintiff's experts have convinced me that a reading disability is not quantifiable merely in test scores. A learning disability is not measurable in the same way a blood disease can be measured in a serum test. By its very nature, diagnosing a learning disability requires clinical judgment. . . . Moreover, I accept the opinion of plaintiff's experts . . . that tests like the Woodcock are "poor discriminators" for adults. Thus, as much as the Board would like to find an easy test discriminator for a reading disability in its applicants, such a test does not exist. Finally, I also do not accept the position of defendants' experts that clinical judgments of a lack of automaticity must be rejected as subjective. Clearly, plaintiff's low, albeit within the average range, test scores on the Woodcock, combined with clinical observations of her manner of reading amply support a conclusion that she has an automaticity and a reading rate problem.

Bartlett I, 970 F. Supp. at 1114 (internal citations omitted). Specifically, I credited plaintiff's experts' opinion that the Woodcock is unable to measure plaintiffs disability because: (1) it is an untimed measure of decoding ability and does not score for reaction time; (2) it does not measure for lack of automaticity; (3) the test is designed principally for children and does not have enough difficult items to test an adult adequately; and (4) research shows that over one-third of adult dyslexics score above the 30th percentile (and thus, a fortiori, an even higher number score within the average range of the 16th to 84th percentiles) on this test.²⁸ Id.

28 In my first opinion and order, I noted that while Dr. Vellutino stated that he relied principally on the studies of adult dyslexics conducted by Dr. Maggie Bruck in setting his 30th percentile cut-off on the Woodcock Word Attack and Word Identification subtests for determining the existence of a learning disability, Dr. Bruck's studies actually found that the Woodcock subtests are poor discriminators for learning disabilities and that the 3rd percentile cut-off was under-inclusive because one-third of dyslexics score above this cut-off. Bartlett I, 970 F. Supp. at 1113.

While defendants now try to couch their argument in terms of the entire gestalt of tests administered to plaintiff, this is a distinction without a difference. Dr. Flanagan concluded that plaintiff's performance on "multiple, individually administered" psychometric measures "indicates that she performs within normal limits or better on all indicators of reading performance." (Flanagan Aff. ¶ 10(e).) of the 35 scores from psychometric measures on which Dr. Flanagan relied, however, thirty are from the Woodcock.²⁹ For the reasons I discussed in my first opinion and order, I am convinced that the Woodcock cannot be used as the principal instrument to diagnose a reading disability.

29 Plaintiff was administered the Woodcock on three separate occasions. Dr. Massad administered Form H of the Woodcock in 1989. Dr. Heath administered Forms H and G of the Woodcock in 1993. Dr. Hagin administered the Word Attack subtests from Forms G and H in 1994. The other five scores upon which Dr. Flanagan based her conclusion are plaintiffs score on the WRAT Oral Reading Test and plaintiffs four scores from Dr. Hagin's administration of the DRT to her. (Flanagan Aff. ¶ 10, Tables 1-3.) I discuss the DRT scores in detail below.

Evidence from Dr. Mather, who was involved in the development of the — Woodcock and co-authored portions of the manual for this test, reinforces this view. Dr. Mather emphasized that one cannot rely solely on Woodcock scores for diagnosing learning disabilities. (Mather Aff. ¶ 17.) Scores from the Woodcock, which is an untimed test that allows for self corrections, do not directly reflect functional reading ability because those scores do not show the manner by which one decodes, the speed at which one reads, or the ability to comprehend what one reads. (Mather Aff. ¶ 15.) For example, the Word Attack and Word Identification subtests (upon which Dr. Vellutino focused almost exclusively and which constitute almost one third of the scores upon which Dr. Flanagan now relies) test only accuracy of word identification and were not designed to measure speed or automaticity. Rather, automaticity can only be assessed on the Woodcock through error analysis and clinical observations, such as whether the individual makes self-corrections. (Mather Aff. ¶ 16; Tr. at 429.) Dr. Mather observed that the manual from the related Woodcock-Johnson III specifically directs an evaluator to examine, inter alia, qualitative data, including behavioral observations and error analysis, and she criticized defendants' attempt to rely solely on percentile rankings. (Tr. at 417-18.)

Contrary to Dr. Flanagan's contention, plaintiff's experts testified that plaintiff's Woodcock test scores are relevant to their diagnostic analysis. Plaintiff's experts stated that discrepancies between some of her test scores on the psychometric measures provide evidence of her learning disabilities. For example, Dr. Massad noted a significant discrepancy between plaintiffs Verbal IQ and her scores on Word Identification and Word Attack subtests (Pl.'s Ex. 20a, at 4.); Dr. Heath observed a significant discrepancy between plaintiffs basic reading skills and her "superior level of intellectual functioning" (Pl.'s Ex. 16, at 1.); and Dr. Hagin discussed a significant discrepancy between plaintiffs verbal and performance IQs (Pl.'s Ex. 93, at 1.). Dr. Moats stated that plaintiffs configuration of test results conformed with those of high-functioning adults with dyslexia, noting that plaintiffs spelling score was lower than her word recognition score, which was lower than her comprehension score. (Tr. at 84-85) Dr. Hagin testified that most individuals without learning disabilities generally have a flat profile and that the highs and lows in plaintiffs test profile were significant. (Tr. at 182-83.)

Dr. Flanagan criticized plaintiffs experts' reliance on plaintiffs intra-ability discrepancies as evidence of her learning disabilities. (Flanagan Aff. ¶ 15(b) at 27.) However, even Dr. Flanagan's operational definition of learning disabilities requires the existence of intraability discrepancies for a learning disability diagnosis. (Defs.' Ex. MMMM.) I also note that the Diagnostic and Statistical Manual of Mental Disorders ("DSM") requires the existence of intra-ability discrepancies as a diagnostic criteria for dyslexia. (Pl.'s Ex. 99, DSM-III-R, at 44; Defs.' Ex. AA, DSM-IV, at 50.)

Because I discussed the controversy surrounding the discrepancy theory in my first opinion and order, I need not revisit that issue here. Bartlett I, 970 F. Supp. at 1115 n. 11. I have already determined "that deviations or discrepancies in test scores should only be used as an indication that a learning disability exists. They do not, standing alone, identify' a learning disabled person," id. at 1116, but they do help to confirm such a diagnosis when other reasons for the discrepancy are eliminated. Drs. Flanagan and Vellutino erroneously assume that because the existence of intra-ability discrepancies is not a discriminator for learning disabilities, it cannot be an indicator of learning disabilities. To take one example, while discrepant spelling scores are not a discriminator for dyslexia because all poor spellers do not have dyslexia, they can be an indicator of dyslexia because most dyslexics have poor spelling. Plaintiff's experts rely on discrepancy analysis solely as a further indicator of her learning disabilities.

In sum, I reaffirm my findings that I cannot rely solely on plaintiffs Woodcock scores alone in determining whether she has a reading disability and that plaintiffs intra-ability discrepancies evidenced by her Woodcock scores lend support to plaintiffs other evidence of her disability.

2. DRT Scores

While a significant portion of the original trial focused on the Woodcock (particularly on the Word Attack and Word Identification subtests), the remand trial has focused in large part on the DRT. The DRT measures both reading rate and reading comprehension on narrative and textbook text. (Pl.'s Ex. 186, DRT Manual, Directions for Administering, at 1.) Dr. Hagin administered Form C under timed conditions and Form A under untimed conditions. Dr. Hagin directed plaintiff under timed conditions to "[r]ead as rapidly as you can and still understand what you read." (Pl.'s Ex. 186, DRT Manual, General Directions, at 3.) Dr. Hagin determined plaintiff's reading rate by having her mark the place to where she had read in the narrative section after three minutes and by using the manual to compute a wpm rate.³⁰ (Pl's Ex. 186, DRT Manual, Directions for Administering at Table III.) After the three minutes, plaintiff continued to read the passage and answered comprehension questions at the end. (Pl.'s Ex. 186, DRT Manual, General Directions, at 3 n. 3.) When using the oldest grade norms, those for — college freshmen, plaintiff's results on the DRT are as follows:

30 Thus, the wpm reading rate is simply a snapshot of how fast the reader read for the first three minutes of the test. It is not indicative of whether the reader read at that speed for the entire test, nor does it indicate whether the reader went back and reread the text before answering the comprehension questions.

FORM C (timed) FORM A (untimed) Reading Rate 195 wpm 156 wpm 4th percentile 1st percentile Comprehension Score 16th percentile 98th percentile — Narrative Section 75% correct 95% correct — Textbook Section 50% correct 95% correct

(Hagin Supp. Aff. ¶¶ 21, 22.) Plaintiff relies on the DRT as evidence of her slow reading rate and of the fact that her performance improves significantly with extra time. Defendants attempt to undermine the importance of plaintiffs low percentile reading rate score by: (1) using plaintiff's reading rate from the DRT in Dr. Carver's diagnostic system; and (2) criticizing the norms Dr. Hagin used.

Dr. Carver testified that plaintiffs results on the DRT do not evidence a lack of automaticity, below-average reading rate, or disability in rate. Specifically, based on his research on reading rates at different grade levels, plaintiff's rate of 195 wpm, which Dr. Carver believed was a valid measure of plaintiff's normal reading rate, equated to an eighth grade level. (Carver Aff. ¶¶ 7, 12.) In applying his diagnostic system to adult readers, Dr. Carver assumes that an average, literate adult's scores should equate with the eighth grade level because, based on his newspaper research, the average literate adult reads at the eighth grade level. (Carver Aff. ¶¶ 9 10; Tr. at 540-41.) Premised on his translation of plaintiff's wpm score on the DRT to a grade equivalent under his diagnostic system, Dr. Carver concluded that "[t]his reading rate at the eighth grade level disqualifies [plaintiff] as being reading disabled because the only published standard for reading disability [ i.e., Dr. Carver's system] requires that an adult have a normal reading rate at the fifth grade level or below." (Carver Aff. ¶ 12; Tr. at 542.) He also opined that plaintiff's score of 195 wpm, which he stated was only one tenth of a millisecond per word slower than the average student, did not evidence any automaticity problems and was inconsistent with the way plaintiff describes how she reads. (Tr. at 535-36.)

Dr. Carver's testimony does not convince me that plaintiff does not have a significant reading rate problem. With all due respect to Dr. Carver, I do not credit the rationale underlying his conclusion that plaintiff does not have a reading rate disability. First, Dr. Carver sets the standard for an average adult based on his reasoning that an average adult reads at the eighth grade level because newspapers around the country are written at that level. There is no empirical data to support this standard. Employing this same type of non-empirical reasoning, I find it hard to believe that newspapers are written at the level of an average adult. It is more reasonable to assume that newspapers are written at a level that is comprehensible to the vast majority of readers, including those who read at a level below an average adult. Furthermore, I agree with Dr. Hagin's observation that the manner in which people read newspapers is different than the manner in which they read other types of text. Because I credit Dr. Hagin's testimony that most individuals skim the first few lines of a newspaper article to get the gist of the story (Tr. at 167.), logic would dictate that a newspaper is likely to be written at a level below that of an average adult to facilitate this type of reading.

More importantly, Dr. Carver's diagnostic system is extremely conservative and is not used in the profession for diagnosing reading disabilities. For example, when Dr. Carver applied his diagnostic system to community college students who scored so low (that is, more than one standard deviation below the mean) on a battery of reading tests that they were required to take a remedial reading course, Dr. Carver found a significant proportion of these students to be "advanced readers" such that they did not need "special attention or remediation for functional literacy." (Tr. at 565-69; Pl.'s Ex. 192, Carver Clark, Investigating Reading Disabilities Using the Rauding Diagnostic System, Journal of Learning Disabilities 30(5) (1999).) Under Dr. Carver's system, plaintiff is not disabled because her reading rate under timed conditions of 195 wpm puts her above fifth grade level. Although I have serious doubts that an adult must read below the fifth grade level in order to be considered substantially limited in reading, even if I applied this standard using the DRT norms, plaintiffs "reading ease level" — which is defined by the DRT manual as the grade level at which the individual's score is roughly comparable to the raw score at the 50th percentile — is below the fourth grade level. (Hagin Supp. Aff. ¶¶ 30, 31.) Furthermore, even assuming arguendo that the average adult reads at the eighth grade level, plaintiff's reading rate is far below that of the vast majority of eighth graders according to the DRT norms. (Pl.'s Ex. 186, Public School 8th Grade Norms) (plaintiffs timed rate of 195 wpm and untimed rate of 156 wpm equates to the 22nd percentile and the 5th percentile, respectively, in comparison to 8th graders); see also Hagin Supp. Aff. ¶¶ 33, 34 (stating that plaintiff's timed rate of 195 wpm and untimed rate of 156 wpm equates to the 22nd percentile and the 8th percentile, respectively, in comparison to high school freshmen).

Finally, while Dr. Carver's contention that the difference in plaintiff's reading rate and that of an average college student is not significant because plaintiff only reads a fraction of a millisecond slower per word has superficial appeal, assuming arguendo that plaintiff can continue at a pace of 195 wpm for extended periods of time, those fractions of a millisecond can add up into seconds, minutes, and even hours when reading longer texts. Additionally, plaintiff's 195 wpm reading rate is more than one-third slower than the 300 wpm normal reading rate of average college students under Dr. Carver's system.³¹

31 I also note that plaintiffs experts have raised some valid concerns with respect to the use of adaptive testing in Dr. Carver's system. Specifically, they argue that adaptive testing may produce skewed results for learning disabled individuals because, while the theory that an individual will get easier items correct if they get harder items correct may hold true with most people, it may not hold true with learning disabled individuals who have an unstable sight vocabulary. (Tr. at 109, 557-58.) Dr. Mather also noted that Dr. Carver's theory that an individual with a reading disability can become a better reader by raudamatizing more words is applicable to an individual with the type of reading problems that plaintiff has exhibited. (Tr. at 44 1-42.) I note that when the very disability at issue is the inability to read words with automaticity, then the very process of raudamatizing is affected.

Dr. Flanagan's major criticism of plaintiff's DRT scores is that Dr. Hagin did not use appropriate norms. Dr. Carver's testimony suggests that Dr. Hagin should have compared plaintiff to the grade equivalent norm of most adults — eighth grade.³² However, Dr. Flanagan claimed the exact opposite. Dr. Flanagan stated that because the speed of processing naturally declines with age, by comparing plaintiff to individuals more than 25 years younger, plaintiff's percentile scores on the DRT significantly underestimate her abilities relative to most people. (Flanagan Aff. ¶ 10(c) at 14; Tr. at 640-34.) Plaintiff's experts disagreed that an individual's reading rate would necessarily decrease with age because, while certain types of processing — speed may decrease, other factors may offset that decline, including the possibility that the individual may have developed a larger sight vocabulary due to additional exposure to language. (Tr. at 287-88; 452.) Dr. Mather, a co-author of the Woodcock-Johnson III, criticized Dr. Flanagan's comparison of the DRT norms to those of the Woodcock-Johnson III Reading Fluency test because those two tests measure different skills. (Tr. at 450-52.) In any event, Dr. Flanagan's criticism of the norms is merely hypothetical because the oldest norm group on the DRT is college freshmen and is reflective of what all of plaintiff's experts have repeatedly emphasized (and which was not disputed by defendants' experts) — that no good reading rate measure for adults existed when plaintiff was evaluated (and remains true, to a large extent, even now).

32 I read the Second Circuit's most recent opinion likewise as implying that the norm group of college freshmen is too high, that is, that college freshmen may read better than "most people." Bartlett V, 226 F.3d at 81-82 ("Bartlett's DRT results, which show a reading rate in the fourth percentile or below as compared to college freshmen, are of limited value, because the proper reference group is "most people," not college freshmen.").

This debate over which norm group equates to "most people" highlights why I cannot rely on test scores alone in determining whether plaintiff is disabled. Tests like the DRT and Woodcock are developed to assist with the diagnosis of learning disabilities, particularly in children having problems in school. They are diagnostic tools. They are not ADA and Rehabilitation Act tests. These tests are not normed to "most people." Whatever "most people" means in the context of the DRT, it cannot mean that a highly educated, intelligent adult who reads slower than 78 percent of fourteen year-olds on a reading test written at an upper elementary school level has no reading rate disability

With all of its limitations, however, the DRT does provide some evidence of plaintiff's reading rate problems. While plaintiff's low percentile rank (no matter which norm is used) lends support to plaintiff's other evidence of her slow reading rate, Dr. Hagin has convinced me that, regardless of which norms are most appropriate, plaintiff's DRT scores provide an important piece of information regarding her disability — that her performance improves significantly with additional time. When Dr. Hagin gave plaintiff approximately 100 percent more time, plaintiff's comprehension increased from the 16th percentile when compared to college freshmen to the 98th percentile.³³ In terms of percentage correct, the number of items plaintiff answered correctly in the Textbook Section³⁴ increased from 50 to 95 percent, and from 75 to 95 percent in the Narrative Section.³⁵ Research shows that on the Nelson-Denny, a test similar to the DRT, there is a significant difference between the scores of normally achieving students and students with learning disabilities under timed conditions that does not exist when learning disabled students are provided with extra time. (Pl.'s Ex. 127, M. Kay Runyan, The Effect of Extra Time on Reading Comprehension Scores for University Students With and Without Learning Disabilities, Journal of Learning Disabilities, vol. 24, no. 2 (Feb. 1991).) This research shows that while students with learning disabilities perform significantly better with extra time, extra time does not have a significant impact on the performance of normally — achieving students. ( Id.) Plaintiff's experts agreed. (Moats Aff. ¶ 30; Tr. at 258-59; 320-21.) While Dr. Carver testified that extra time significantly affects individuals who do not have time to finish the text they are reading (Tr. at 545-46), the DRT, which is written at an upper elementary school level and which has "generous" time limits (because they are set for children, who do not read as fast as adults), does not contain the type of text that "most people" would have trouble completing in the set time limits.³⁶ (Hagin Aff. ¶¶ 17, 19.)

33 Dr. Carver pointed out that Dr. Hagin erroneously marked one question right, such that plaintiffs comprehension rate was at the 94th percentile, not the 98th percentile. (Tr. at 590-91.)

34 The Textbook Section (also known as the Comprehension subtest) contains reading material similar to that found in social studies and science textbooks (written at a level appropriate for seventh graders). (Pl.'s Ex. 186, DRT Manual, Directions for Administering, at 1.)

35 The Narrative Section (also known as the General Reading subtest) contains "interesting story-type material with a generally simple vocabulary load" (again written at a level appropriate for seventh graders). (Pl.'s Ex. 186. DRT Manual, Directions for Administering, at 1.)

36 Throughout the remand trial, defendants criticized plaintiff for failing to document her reading rate problems with scores from psychometric measures. I observe, however, that Dr. Massad, in consultation with Dr. Moats, decided not to administer to plaintiff a psychometric measure of reading rate because of her documented history of reading problems, because of those tests' limitations, and in light of the fact that the Nelson-Denny was undergoing revision. (Moats Aff. ¶¶ 12, 17, 18; Tr. at 44.) Dr. Hagin decided to administer to plaintiff the DRT, rather than the Nelson-Denny, because she was concerned that plaintiff might be familiar with the Nelson-Denny from her work as an assistant school superintendent. (Hagin Supp. Aff. ¶ 17.) Finally, while there has been some testimony regarding the Woodcock Johnson III Reading Fluency test, this test did not exist at the time plaintiff was evaluated.

In sum, I reject defendants' argument that plaintiff's scores on the psychometric measures administered to her are necessarily inconsistent with a finding of disability. I agree with plaintiff's experts that plaintiff's DRT scores and the discrepancies between some of her Woodcock scores lend support to the other evidence of plaintiff's reading disability.

B. Clinical Observations

After considering the additional evidence presented to me in the remand trial, I conclude — as I did in the original trial — that a reading disability cannot be diagnosed by test scores alone. Rather, diagnosing a learning disability requires clinical judgment. Bartlett I, 970 F. Supp. at 1114. As Dr. Mather stated, "[i]f you just look at scores you miss the boat here." — (Tr. at 484.) I do not accept the position of defendants' experts that clinical judgments regarding plaintiff's lack of automaticity and reading rate must be disregarded (as suggested by Dr. Carver) or given little weight (as suggested by Dr. Flanagan) because they are "subjective"; plaintiffs experts have convinced me that clinical judgments are objective (albeit qualitative) and are a critical part of any proper assessment. (Tr. 269, 494.) In this case, I find that given the inability of tests like the Woodcock to measure automaticity and orthographic problems, the clinical observations of plaintiff's manner of reading are the most probative evidence of her reading disability. The clinicians who evaluated plaintiff consistently reported that plaintiff reads and decodes slowly, employs coping techniques to assist with reading, self-corrects, commits atypical spelling errors (which are typical for individuals with dyslexia), reads without automaticity, and rereads text. See infra Conclusions of Law, Section W. These clinical observations, in combination with plaintiffs low (albeit average) scores on the decoding portions of the Woodcock and with plaintiffs very low scores on the DRT, lead me to conclude that plaintiff has a lack of automaticity and significant reading rate problems.³⁷

37 If Dr. Flanagan had considered the clinical observations when applying her learning disability model to plaintiff, I believe she would have concluded that plaintiff has a learning disability. Dr. Flanagan stated that plaintiffs experts would probably find, based on clinical observations, that plaintiff had deficiencies in her language abilities on Level III. (Tr. at 634.) The evidence shows that plaintiff has intra-ability discrepancies (a requirement of Level IV) and that her learning problems significantly interfere with learning and daily life activities (a requirement of Level V). (Defs.' Ex. MMMM.)

CONCLUSIONS OF LAW

I adopt herein any Finding of Fact previously set forth which might more properly be deemed a Conclusion of Law, Similarly, I adopt in my Findings of Fact any Conclusion of Law which might more properly be deemed a Finding of Fact.

I. Substantial Limitation Under the Law for the Major Life Activity of Reading

In order to find that plaintiff is entitled to accommodations under the ADA or Section 504, I must find that she has a disability, defined as "a physical or mental impairment that substantially limits one or more of the major life activities" of an individual. 42 U.S.C. § 12102(2)(A); 29 U.S.C. § 705 (9)(B).³⁸ I must employ the three-step process set forth by the supreme Court for determining whether plaintiff is disabled and decide: (1) whether plaintiff's reading problems constitute a mental impairment; (2) whether the life activities upon which plaintiff relies (here, reading and, in the alternative, working) are major life activities under the ADA and Section 504; and (3) whether plaintiffs impairment substantially limits these major life activities. See Bragdon v. Abbott, 524 U.S. 624, 631 (1998). In my first opinion and order, I found that plaintiff had a mental impairment and that reading and, in the altemative, working, were major life activities, Bartlett I, 970 F. Supp. at 1098, 1116, and the Second Circuit affirmed these findings. Bartlett V, 226 F.3d at 79-80. Thus, the sole remaining issue I must resolve in determining whether plaintiff has a reading disability is whether plaintiff's reading impairment "substantially limits" her major life activity of reading.

38 As I mentioned earlier, because in the context of this case, Titles II and III of the ADA and Section 504 "impose largely the same requirements," I will treat these claims together. Bartlett V, 226 F.3d at 78 n. 2.

Neither the text of the ADA, Section 504 nor the Department of Justice's ("DOJ") implementing regulations define the term "substantially limits," see Bartlett V, 226 F.3d at 80,. but the preamble to the regulations provides: "A person is considered an individual with a disability . . . when the individual's important life activities are restricted as to the conditions, manner, or duration under which they can be performed in comparison to most people." 28 C.F.R. Pt. 35, App. A § 35.104. This interpretation of the phrase "substantially limits" is nearly identical to that of the Equal Employment Opportunity Commission ("EEOC") in connection with Title I of the ADA. 29 C.F.R § 1630.2 (j)(1)(ii) (defining substantially limited as "[s]ignificantly restricted as to the condition, manner or duration under which an individual can perform a particular major life activity as compared to the condition, manner, or duration under which the average person in the general population can perform that same major life activity").

In interpreting the phrase "substantially limits," the Supreme Court has rejected the argument that an individual is not substantially limited in a major life activity unless it is impossible for him or her to perform that activity, stating that the ADA "addresses substantial limitations on major life activities, not utter inabilities." Bragdon, 524 U.S. at 641. Therefore, "[w]hen significant limitations result from the impairment, the definition is met even if the difficulties are not insurmountable." Id. at 641.³⁹ This definition, however, is not met by a plaintiff showing a mere difference in the "conditions, manner or duration" in which he or she performs a major life activity; a plaintiff must show a "significant restriction." Albertson's, 527 U.S. at 565. Thus, I must determine whether plaintiff has proven that her reading impairment causes significant limitations to the "conditions, manner or duration" in which she reads when compared to "most people" or whether she has proven only a "mere difference." In its remand,. the Second Circuit noted that plaintiff's burden in showing that she is "substantially limited" in reading "is not necessarily an onerous one" and that "certain limitations will `ordinarily' qualify as disabilities." Bartlett V, 226 F.3d at 80 (citing Albertson's, 527 U.S. at 567). A court must, however, consider the existence of a disability on a case-by-case basis. Albertson's, 527 U.S. at 566. According to the EEOC's interpretation, which I find instructive, my determination of whether plaintiff is disabled should focus "on the effect of that impairment on [plaintiff's life]." 29 C.F.R. Pt. 1630, App. A § 1630.2(j). As I discuss more fully below, I must consider both the positive and negative effects of the mitigating measures which plaintiff employs in determining whether she is disabled. Sutton, 527 U.S. at 482.

39 Applying this standard, the Supreme Court in Bragdon found that an asymptomatic HIV-positive plaintiff was substantially limited in the major life activity of reproducing because of the risks the plaintiff faced when trying to conceive — there was a twenty percent risk that she would transmit the disease to her partner and a 25 percent risk that she would infect her child. Id. at 639-40.

A. Corrective Devices or Mitigating Measures

After my opinion and order and the Second Circuit's first opinion issued in this case, the Supreme Court decided a series of cases in which it held that the corrective devices or mitigating measures employed by an individual are relevant to the determination of whether that individual is "substantially limited" in a major life activity. In Sutton, the lead case on this issue, the Supreme Court stated that "if a person is taking measures to correct for, or mitigate, a physical or mental impairment, the effects of those measures — both positive and negative — must be taken into account when judging whether that person is `substantially limited' in a major life activity." Sutton, 527 U.S. at 482; accord Murphy, 527 U.S. at 520; Albertson's, 527 U.S. at 565-66. The Supreme Court rejected the suggestion, however, that an individual's use of a — mitigating measure or corrective device is determinative of whether that individual is disabled:

The use of a corrective device does not, by itself, relieve one's disability. Rather, one has a disability . . . if, notwithstanding the use of a corrective device, that individual is substantially limited in a major life activity. For example, individuals who use prosthetic limbs or wheelchairs may be mobile and capable of functioning in society but still be disabled because of a substantial limitation on their ability to walk or run. The same may be true of individuals who take medicine to lessen the symptoms of an impairment so that they can function but nevertheless remain substantially limited.

Sutton, 527 U.S. at 488. Thus, the Supreme Court concluded that the test for determining whether an individual is disabled is not whether he or she uses a corrective device or mitigating measure, but "whether the limitations an individual with an impairment actually faces are in fact substantially limiting." Id. at 488 (emphasis in original).

Plaintiff argues that although I must consider the mitigating measures she uses in determining whether she is disabled, I should not take into account those measures that do not affect her ability to perform the major life activity of reading,⁴⁰ such as having other people read to her or participating in study groups. (Pl.'s Post-Hearing Br. at 46-47.) While Sutton, Murphy, and Albertson'S make clear that I must consider both the positive and negative effects of the mitigating devices she employs, they are not instructive on whether I should make the distinction that plaintiff suggests because all of the corrective devices employed by the plaintiffs in those cases affected their ability to perform the major life activities they had alleged were substantially limited by their impairment. See Sutton, 527 U.S. at 475 (use of corrective lenses improved plaintiffs' ability to see); Murphy, 527 U.S. at 519 (use of blood pressure medication improved plaintiff's ability to walk, lift, climb, etc.); Albertson's, 527 U.S. at 565 (plaintiff's brain's subconscious adjustments improved plaintiff's ability to see). A number of courts have considered this issue post- Sutton and have concluded that a court should only take into account mitigating measures or corrective devices that affect the individual's ability to perform the major life activity the plaintiff alleges is substantially limited by his or her impairment. I agree.

40 Throughout this litigation, the term reading has been used to mean being able to see and quickly comprehend visual images on a page. See, e.g., Bartlett I, 970 F. Supp. at 1130.

For example, in Finical v. Collections Unlimited, Inc., 65 F. Supp.2d 1032 (D. Ariz. 1999), the district court examined whether an individual with an impairment that he alleged affected his major life activity of hearing was disabled within the meaning of the ADA. The district court held that "[b]ecause the analysis of disability focuses on the major life activity alleged to be substantially limited, the relevant mitigating measures are those that affect the claimant's ability to perform that major life activity." Finical, 65 F. Supp. 2d at 1039-40. The district court discussed this distinction in relation to the facts of Albertson's, which involved an individual who alleged his monocular vision substantially limited his major life activity of seeing:

For example, in [ Albertson's], the Court indicated that corrective devices, including coping mechanisms developed by the individual's own brain, are to be considered to the extent that they affect the individual's ability to perform the major life activity of seeing by, for example, adjusting the manner in which the individual senses depth and perceives peripheral objects. In contrast, the brain may compensate for lack of vision in other ways, such as by making the affected individual's hearing more acute, but this compensatory measure does not improve the individual's ability to see and should not be considered in determining whether the individual is substantially limited in the major life activity of seeing.

Id. at 1040 (internal citations omitted).

The Finical court found that the corrective devices employed by the plaintiff that actually improved her ability to hear — including plaintiff's use of an amplification device on her phone, her sitting in the front of her classes close to the teacher, and her requesting co-workers to face her while speaking — were relevant to its determination of whether plaintiff was substantially limited in the major life activity of hearing. Id. at 1041-42. In contrast, the court found that those measures that merely helped the plaintiff communicate, but did not improve her hearing, should not be considered:

Measures such as lip-reading and telephone lights improve Plaintiff's ability to communicate despite a hearing impairment and, as a result, her ability to engage in normal activities. However, these measures do not necessarily improve Plaintiff's ability to hear. Lip-reading is unlike the use of corrective lenses at issue in Sutton, because the lenses actually improved the pilot applicants' ability to see. For the same reasons, lip-reading is even unlike corrective measures taken by the body to improve the ability of a person with monocular vision to see both range and depth, at issue in [ Albertson's]. . . . In contrast to a corrective device that would actually improve an individual's ability to hear, lip-reading is a means by which an individual whose hearing is impaired may obtain information visually rather than orally. Thus, lip-reading is more akin to use of a wheelchair or braille — all are devices that assist an impaired individual to remain functional, even highly so, but do not alter the status of the impairment.

Id. at 1041. Based on the evidence that plaintiff had a 37 percent hearing loss and had significant difficulties in using hearing in her everyday life, the district court found that plaintiff had raised a genuine issue of material fact as to whether she was substantially limited in hearing and denied defendant's motion for summary judgment. Id. at 1040-1042.

The district court employed a similar analysis in Root v. Ga. State Bd. of Veterinary Med., 114 F. Supp.2d 1324 (N.D. Ga. 2000), rev'd in part, vacated in part on other grounds 252 F.3d 443 (11th Cir. 2001) (unpublished disposition).⁴¹ The plaintiff in Root alleged that her learning disabilities substantially limited her major life activity of learning. Like the district court in Finical, the district court in Root distinguished between measures that improved plaintiff's ability to learn and other measures:

41 The Eleventh Circuit's decision addressed only the defendant's motion to dismiss on Eleventh Amendment grounds and did not call into question the district court's disability analysis. (Pl.'s Post-Hearing Br. at 41 n. 37.)

Plaintiff has testified that she has had a lifelong struggle with learning, and has tried to compensate for her difficulties by having others read to her, assist her with spelling and in writing papers, and by choosing courses and degree programs that did not require much reading and writing and in which she could take her exams orally. Plaintiff's testimony is corroborated by the testimony of several witnesses, including her mother, sister, classmates, and doctors. Plaintiff's attempts to compensate for her learning difficulties do not correct her disability in the same way that glasses or the body's internal systems can correct or compensate for impaired vision. In this regard, Dr. Gregg⁴² has opined that while "[m]ost remedial efforts can produce some benefits, [t]hey do not eliminate a learning disability, nor can they `correct it.'"

Id. at 1333 (internal citations omitted). Based on the evidence of plaintiff's learning difficulties, and applying the above analysis, the court found that plaintiff had raised a genuine issue of material fact as to whether she was "substantially limited" in learning and denied defendant's motion for summary judgment. Id.

42 This is the same Dr. Noel Gregg that testified as one of plaintiff's experts in the remand trial.

Finally, the dissent in Gonzales v. Nat'l Bd. of Med. Exam'rs, 225 F.3d 620 (6th Cir. 2000), cert. denied, 121 S.Ct. 1999 (2001), employed a similar analysis with respect to the mitigating measures used by a student with learning disabilities. Like Dr. Bartlett, the plaintiff in Gonzales alleged that his learning disabilities substantially limited his major life activities of reading, writing, and working. Id. at 627-31. The majority found that plaintiff was not substantially limited in any of these major life activities. Id. The dissent noted that plaintiff had employed a number of strategies to assist him in his academic endeavors that "allowed him to get by while doing the bare minimum of reading," including tape-recording lectures and having his friends read their lecture notes to him. Gonzales, 225 F.3d at 633. The dissent criticized the majority's reliance on these types of self-accommodations as mitigating plaintiff's substantial limitation in the major life activity of reading:

The majority's analysis in the present case relies on [ Albertson's] reasoning that Gonzales has learned to self-accommodate in a similar fashion. I respectfully disagree. In plain although admittedly unscientific terms, Gonzales's claim is that the part of his ;brain responsible for decoding written language is not wired the same as, and functions substantially worse than, that of the average person, even though in other respects his mental faculties are significantly better than average. If, despite this faulty "wiring," Gonzales had been able to adapt so that his ability to read was substantially no worse than that of the average person, then Gonzales would not be considered to have a reading disability under [ Albertson's] rationale. But that is not Gonzales's claim. Rather, Gonzales has asserted that despite his best efforts to work around his problem, he is still not able to read nearly as well as the average person. I do not believe that working around a reading impairment by pursuing strategies in school that minimize the necessity for reading is the type of self-accommodation the Supreme Court had in mind in Murphy, Sutton, or [ Albertson's].

Id. at 633-4. Based on plaintiff's evidence of the significant limitations with respect to reading he experienced in his everyday life, the dissent stated that it would have reversed the district court's finding that plaintiff was not disabled because it was not supported by the evidence.⁴³ Id. at 637.

43 The district court in Price v. Nat'l Bd. of Med. Exam'rs, 966 F. Supp. 419 (S.D.W.Va. 1997) applied an analysis similar to that of the majority in Gonzales, finding that the plaintiffs were not disabled because, inter alia, they had not exhibited a pattern of substantial academic difficulties. Price, 966 F. Supp. at 423-24, 426-28. The Price Court, however, did not have the benefit of the Supreme Court's decisions in Sutton, Murphy's, and Albertson's or the subsequent cases I discussed above.

I agree with these courts that I should consider only the mitigating measures that affect plaintiff's ability to read in determining whether she has a reading disability. I believe this result is correct in light of the language of the ADA and Section 504, Supreme Court precedent, and the broad remedial purpose of these acts. The ADA and Section 504 define disability in terms of a "substantial limitation" with respect to a particular major life activity. 42 U.S.C. § 12102; 29 U.S.C. § 705 (9)(B). As I discussed earlier, the Supreme Court has interpreted this language as setting forth a three-step analysis for determining whether an individual is disabled. See Bragdon, 524 U.S. at 631. It makes no sense for a court to first examine whether the life activity a plaintiff alleges to be substantially limited constitutes a major life activity if later, in determining substantial limitation, corrective measures that do not affect the individual's ability to perform that particular life activity are to be considered. Additionally, an analysis that distinguishes between measures that affect the life activity alleged to be limited and those that do not is consistent with the rationale of Sutton, Murphy, and Albertson's, which focused on whether an impairment in fact significantly restricts an individual's ability to perform a major life activity. It would be inconsistent with this rationale for a court to find that compensatory measures that permit an individual to avoid performing a particular major life activity mitigate the individual's substantial limitation in performing that life activity. Along these lines, for example, the Supreme Court in Sutton noted that an individual who uses a wheelchair to be mobile, that is, uses a wheelchair to avoid performing the major life activities of walking and running, could still be limited in those major life activities. Sutton, 527 U.S. at 488. To conclude that an individual is not disabled because he or she is able to function by "getting around" using a major life activity would undermine the remedial purpose of the ADA, see Bartlett V, 226 F.3d at 84 n. 3., and would "[exclude] from coverage all individuals who, though possessing an impairment substantially limiting a major life activity, remain highly-functional overall." Finical, 65 F. Supp. 2d at 1038.

This legal landscape with respect to how a court should analyze mitigating measures and corrective devices did not exist four years ago when I issued my first opinion and order. My consideration of what I called plaintiff's "self-accommodations" at that time was correct, as a basic principle. In light of this new caselaw, however, I realize that my analysis was incomplete. I did not distinguish, as I now believe I should, between the mitigating measures plaintiff uses that affect her ability to read and those that merely assist her in functioning in her daily life. I also did not examine the negative effects of the mitigating measures plaintiff employs, as Sutton, Murphy, and Albertson's now make clear that I must. In sum, my analysis of the effect of plaintiff's use of "self-accommodations" must be more specific and precise than that of my first opinion and order. When applying this refined analysis, considering both the positive and negative effects of the mitigating measures that affect plaintiff's ability to read, I conclude that plaintiff is substantially limited in the major life activity of reading when compared to most people.⁴⁴

44 I note that I would reach this same conclusion even if I considered those measures that plaintiff employs that do not affect her ability to read, such as participating in study groups or having other read to her.

Plaintiff uses certain coping strategies that have helped her be successful in her academic and work endeavors but that do not assist her in reading. These methods include having other individuals read to her, studying in groups, dictating letters for others to write for her, and taking classes that grade based on written papers rather than timed exams.⁴⁵ Because these methods merely help plaintiff function, but do not affect her ability to read, these measures are not relevant to my determination of whether plaintiff has a reading disability.

45 Defendants agree that under the rationale of Sutton, Murphy, and Albertson's, plaintiff's use of an amanuensis or a reader in reading and writing are appropriately viewed as accommodations and not "mitigating measures." (Defs.' Resp. to Pl.'s Post-Hearing Br. at 21.) Defendants argue, however, that the other mitigating measures employed by plaintiff enable her to read at least as well as the average person. ( Id. at 22.) I also observe that a number of these self-accommodations, such as having other people read to plaintiff, should not be counted as mitigating measures for another reason — because their use in not solely within plaintiffs control. Cf. Finical, 65 F. Supp.2d at 1038 (finding that telephone set provided by and within the control of the employer could not be considered a mitigating measure).

In contrast, plaintiff uses other coping strategies that assist her with reading. These include using her fingers or a card to move from line to line, using an index card with a hole cut out when reading more difficult text, re-reading text multiple times, subvocalizing ( i.e., sounding out words), and highlighting important words in the text.⁴⁶ While many of these measures increase plaintiff's decoding accuracy, they do so at the cost of speed and cognitive energy.

46 While I do not separately examine the major life activity of writing, see supra note 9, I note that when writing, plaintiff employs the following measures: she draws a line along the left edge of the paper to remind her to return to the left side at the end of a line, writes short sentences, and only uses words that she knows how to spell.

All of plaintiff's experts discussed the negative effects of using such coping mechanisms. For example, Dr. Mather testified:

Using a cut-out index card to keep track of a line of text, and rereading numerous times also make[s] it greatly fatiguing to do what most people do effortlessly. All of the mechanisms that Plaintiff has used to compensate, such as the cut-out card, reading and re-reading, subvocalizing, etc., improve her accuracy at the cost of speed and efficiency.

(Mather Aff. ¶ 25.) Dr. Mather explained that the use of these devices also can negatively affect reading comprehension:

[I]f you are needing to block out print with a card, the strain on the perceptual system — there is a strain just in terms of even your eyes. It is physically exhausting because you are exerting so much energy to decoding that it's hard to comprehend. So often what happens is you have to reread a phrase or reread a word or stop to understand what you are reading. . . . [Y]our comprehension is always being disrupted by your lack of automaticity with words.

(Tr. at 434.) Dr. Moats also discussed the effect of using coping mechanisms like the cut-out card:

Using an index card the way Plaintiff does is incongruent with fluency or automaticity. It is too restrictive to yield minimally efficient reading because it slows down the process and precludes automaticity. Yet it is clear from Plaintiff's record, that the cut-out card helps her; that without that assistance, her reading is even less efficient and accurate.

(Moats Aff. ¶ 24.) Dr. Hagin agreed that using these coping mechanisms consumes a considerable amount of plaintiff's cognitive energy and causes her fatigue. (Tr. at 55, 161.) The use of many of these strategies, including rereading text multiple times and reading word by word through a hole in an index card, increases the amount of time it take plaintiff to read text. Plaintiff's testimony confirms the increased time and fatigue the use of these techniques causes her. Plaintiff stated that reading is exhausting to her, that reading material on her own "takes up a huge amount of . . . time and energy," and that employing these coping strategies "comes at the cost of an incredible amount of time and energy." (Bartlett Supp. Aff. ¶¶ 17-19.)⁴⁷ Thus, the evidence has convinced me that, while the coping measures that plaintiff employs have the positive effect of increasing her decoding accuracy, they also have the negative effects of causing plaintiff fatigue, which prevents her from reading for sustained periods of time and affects her comprehension, and of decreasing her already slow reading rate.

47 While I do not separately consider her major life activity of writing, see supra note 9, I note that plaintiff employs coping mechanisms when writing, such as choosing words that she can spell and writing in short sentences, that negatively affect the quality and maturity of her written work.

When I view the evidence as a whole, including both the positive and negative side effects of the mitigating measures plaintiff uses, I find that plaintiff is substantially limited in the major life activity of reading when compared to most people. All of plaintiff's experts agree that most people do not read with the level of difficulty experienced by plaintiff. (Gregg Aff. ¶ 25; Hagin Aff. ¶ 52; Moats Aff. ¶ 30; Mather Aff. ¶ 27.) The testimony of plaintiff and her experts have convinced me that "most people" can do the following things that are extremely difficult or impossible for plaintiff: read and write quickly and automatically, recognize words and letters automatically, develop a sight vocabulary, and form letters without consciously thinking what they look like. "Most people" do not need other people to read to them, write for them, or edit their basic grammar and spelling. ( Id.) "Most people" are able to skim or scan text. (Tr. at 440.) "Most people" can read for longer than thirty minutes without experiencing fatigue. (Tr. at 163.)

The effect of plaintiff's reading impairment on her life, even with all of her self-accommodations, is profound Cf. 29 C.F.R. Pt. 1630, App. A § 1630.2(j) ("The determination of whether an individual has a disability is . . . based on . . . the effect of that impairment on the life of the individual."). Plaintiff has difficulty with tasks that most people perform effortlessly, including reading short e-mails, using a telephone directory or electronic database, writing a shopping list, or following a recipe. (Bartlett Aff. ¶¶ 11, 12, 13, 22.) Plaintiff generally avoids reading any unnecessary material and does not read for pleasure. (Bartlett Aff. ¶¶ 9, 10, 14.) As plaintiff and her experts stated, plaintiff consistently tries to find alternative routes around reading. Dr. Hagin testified that based on her experience, plaintiff's "reading was more limited than the average person I might see, even the average person with a learning disability." (Tr. at 163.)

Based on all of the evidence before me, I conclude that plaintiff has proven that she is substantially limited in the major life activity of reading in comparison to "most people."

B. Measure of Substantial Limitation By Outcomes Alone

Defendants contend that plaintiff cannot have a reading disability because she has been at least moderately successful in tasks involving reading. Defendants argue that no matter how differently plaintiff reads, she is not substantially limited in reading because she has performed within the average range on tasks involving reading. Specifically, in addition to stating that plaintiffs scores on the psychometric measures administered to her are inconsistent with her having a reading disability, Dr. Flanagan testified that the following real-world outcomes also belie plaintiffs claim that she has a reading disability: (1) plaintiff learned German and passed a proficiency test in connection with her doctoral program; (2) plaintiff scored within the average range on the LSAT, GRE, and SAT without accommodations; and (3) plaintiff completed college and graduate school without accommodations.

The problem with Dr. Flanagan's analysis is that she relies on quantitative outcomes alone.⁴⁸ The ADA and Section 504, however, require that I examine the manner in which plaintiff achieved each of these outcomes in determining whether plaintiff's dyslexia "substantially limits" her reading. The DOJ's definition of "substantially limits" does not state that a court should merely examine the end result of tasks that involve a particular major life — activity. Rather, the DOJ Interpretive Guidance states that the qualitative aspects of how an individual performed those tasks, including the manner in which the individual uses the life activity and the amount of time for which the individual can perform that life activity, are relevant to a disability determination. 28 C.F.R. Pt. 35, App. A § 35.104 ("A person is considered an individual with a disability . . . when the individual's important life activities are restricted as to the conditions, manner, or duration under which they can be performed in comparison to most people.") (emphasis added). Similarly, the EEOC's guidance, which I find instructive, states that the determination of whether an individual has a disability should be based "on the effect of that impairment on the life of the individual," again implying that the manner in which an individual performs tasks, not merely an end result, is relevant. 29 C.F.R. Pt. 1630, App. A § 1630.2(j). Further, the EEOC also defines "major life activities" using qualitative terms; "`major life activities' are those basic activities that the average person in the general population can perform with little or no difficulty." 29 C.F.R. Pt. 1630, App. A § 1630.2(i).

48 I observe that the dissent in Gonzales also criticized Dr. Flanagan's exclusive reliance on test scores and rejected her hypothesis that the test scores of the plaintiff there were inconsistent with his having a reading disability. Gonzales, 225 F.3d at 635. The dissent also noted that, as was brought out in testimony in the remand trial in this case, Dr. Flanagan is an academician and not a clinical psychologist qualified to diagnose learning disabilities in specific individuals (Tr. 209-10; 600-01.). Id. at 634. Finally, the dissent questioned whether Dr. Flanagan's testimony could satisfy an analysis under Kuhmo Tire Co. v. Carmichael, 526 U.S. 137 (1999), and Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993). Id.

A definition of disability based on outcomes alone, particularly in the context of learning disabilities, would prevent a court from finding a disability in the case of any individual like Dr. Bartlett who is extremely bright and hardworking, and who uses alternative routes to achieve academic success. While the majority of the panel in Gonzales found that the plaintiffs academic success belied his claim of disability, the dissent recognized that how the plaintiff achieved that academic success was more relevant to the determination of disability than the fact of his success:

[A] person's ability to get good grades is not the bottom line. If it were, then a student's ability to achieve reasonably high marks (by whatever means) without formal accommodations would, as a practical matter, foreclose a finding that he has a reading disability — a conclusion that I believe is incorrect. No one, for example, would argue that a blind student who is able to get good grades in college with the help of friends and an appropriately configured personal computer — but without any formal accommodation from the school — would not be considered "disabled" for purposes of the ADA. Similarly, if a student with severe reading difficulties can get reasonably high marks in school even though it takes him three times as long as the average person to read the required course materials, it would make little sense to say that he does not have a disability in reading. One might say that he is overcoming his disability as far as getting good grades is concerned, but his method or manner of reading would still be substantially limited as compared to the average person.

Gonzales, 225 F.3d at 633.

I agree with the Gonzalez dissent that in determining whether plaintiff's real world achievements are inconsistent with her claim that she is substantially limited in reading, I must look at how she achieved these results. As Dr. Moats stated, in examining plaintiff's successes, the "struggle is a very meaningful part" of the picture. (Tr. at 69.) When I examine the manner in which plaintiff has achieved all of these successes, I conclude not only does this evidence not belie a finding of disability but it confirms my conclusion that plaintiff has a reading disability.

I have already addressed and rejected defendants' contention that I should focus solely on plaintiffs scores on psychometric measures like the Woodcock. Scores from such tests cannot measure significant aspects of plaintiff's reading disability, including her lack of — automaticity and orthographic problems. Thus, plaintiff's scores, without the accompanying clinical observations of how plaintiff went about achieving those scores, tell nothing about how plaintiff achieved those results, including whether she reads with automaticity, whether she reads slowly, and whether reading causes her fatigue. Rather, the clinicians' observations about how long it took plaintiff to complete particular tasks, the types of errors she made, whether she reread or self-corrected, and how she used coping techniques provide with probative information about whether she is substantially limited in reading. See supra Findings of Fact, Section IV; Bartlett I, 970 F. Supp. at 1113-14. My conclusion that test scores alone are insufficient for diagnosing learning disabilities is confirmed by Dr. Bruck's research showing that a significant percentage of individuals with dyslexia score within average range on tests like the Woodcock. Id. at 1113; see also supra note 28.

With respect to plaintiff's other accomplishments, Dr. Flanagan concluded that "Dr. Bartlett's mastery of a second language, in the absence of any reported accommodations, demonstrates that she does not have an impairment in phonological processing as compared to most people." (Flanagan Aff. ¶ 11(b) at 20.) Dr. Flanagan testified that it would be "highly unusual" for an individual with significant phonological problems to demonstrate adequate proficiency of German, which is a more phonetically complex language than English. ( Id. at 19.)

However, underlying Dr. Flanagan's conclusion are two assumptions that plaintiff has proven are not true: (1) that plaintiff is fluent in reading German; and (2) that the proficiency exam required plaintiff to demonstrate a mastery of reading German. Plaintiff learned German in an informal manner when living in Germany for more than four years. While in Germany, she generally did not have to read or write German. (Tr. at 373-75.) Plaintiff testified that she is fluent in — speaking, but not in reading, German and explained that she experiences the same types of problems and uses the same methods for decoding when reading German that she does when reading English. (Tr. at 373, 380-81.) To place out of her foreign language requirement for her doctoral program, she did not have to read text aloud, write, answer written questions, or perform a literal translation. Rather, she merely engaged in a conversation in German with the head of the German department, silently read a passage from a first-year German textbook in an untimed manner, and orally summarized the gist of the paragraph. (Tr. at 379-80.) In fact, the single time that plaintiff took a formal German course that required her to read, write, and take exams, she stated that she received a D "as a gift." (Tr. at 378.) Therefore, I disagree with Dr. Flanagan that plaintiff's ability to pass this informal German proficiency test is inconsistent with her having a reading disability.

Similarly, the method plaintiff used to score in the average range on the LSAT is also consistent with the other evidence of plaintiff's disability. Plaintiff testified that she took a preparatory course that taught her to avoid reading the majority of the text on the test by using the "hit and move on" strategy, where she would read the questions at the end of the paragraph first and try to guess the correct answer. (Tr. at 366.) Dr. Gregg testified that the test strategy plaintiff used — minimizing the amount of reading and relying on her strong inferencing skills — could help a dyslexic applicant like plaintiff be successful on this type of test. (Tr. at 296-97.) She noted, however, that such a strategy would not work on a test like the bar exam, which involves analyzing detailed fact patterns, answering multiple choice questions with fine distinctions between the correct and incorrect answers, and composing written answers. (Tr. 305-08; Gregg Aff. ¶ 9.2.) When Dr. Mather was asked whether plaintiff's average performance on the LSAT using this "hit and move on" strategy was inconsistent with her having a learning disability, Dr. Mather answered in the negative because "the way she approached the task . . . wasn't reading." (Tr. at 468.)⁴⁹

49 While defendants point out that one LSAT tutoring service recommends against using the type of strategy that plaintiff employed (Defs.' Ex. 0000), this is irrelevant to the issue of whether such a strategy could help someone with the type of reading problems that plaintiff experiences achieve a very low average score.

As for plaintiff's scores on the SAT and GRE, although she did not specifically remember using a particular method in taking these tests, many of her scores were more than one standard deviation below the mean. Compare Tr. at 626-29 (Dr. Flanagan testifying that the average range of plus or minus one standard deviation from the mean on the SAT and GRE is 400-600 on each subtest) with Pl.'s Ex. 187, Pl.'s Report of Pl.'s SAT and GRE scores (reflecting half of SAT and GRE subtest scores slightly below average range).⁵⁰ Dr. Gregg testified that for the general population, there is a close correlation between IQ (particularly verbal IQ) and SAT scores, and based on that, she would have expected plaintiff's SAT scores to be significantly higher. She noted, however, that the correlation does not hold for learning disabled individuals. (Tr. at 302-04.) Further, plaintiff testified that her SAT scores were so low that she was rejected from every college to which she applied and was eventually admitted to a college only because she personally met with the president and convinced him to accept her. (Tr. at 368-69.) Likewise, she was only able to enter graduate school by first taking some classes as a non-matriculating student and convincing the school that her grades were more reflective of her abilities than were her GRE scores. (Tr. at 370.)

50 Plaintiff took the SAT three times and received the following scores: (1) 372 (verbal) and 378 (math); (2) 492 (verbal) and 393 (math); and (3) 436 (verbal) and 470 (math). Plaintiff took the GRE twice and received the following scores: (1) 470 (verbal) and 350 (quantitative); and (2) 420 (verbal) and 390 (quantitative). (Pl.'s Ex. 187.)

Finally, plaintiff's successful performance in college and graduate school is not inconsistent with her claim that she is substantially limited in reading. While plaintiff did not receive formal accommodations in college or during her Master's degree program, she testified about the informal accommodations she used there, including having people read to her, participating in study groups, and choosing classes that graded based on term papers (for which she had others assist her with typing and editing) rather than timed exams. (Tr. at 364-65.) She stated that she succeeded by "avoiding reading at all costs." (Tr. at 365.) As for plaintiff's Ph.D. program, she was granted accommodations⁵¹ in all of the courses taught in her department, and she was able to complete her dissertation using the same types of strategies she had used in college and graduate school — using a reader to assist her with collecting and reading research and having others type and edit her dissertation. (Tr. at 386-87, 397-98.)⁵²

Plaintiff testified about what she believes to be the key to her success:

I have been successful because I have been willing to work the extra hours my disability requires. I have been creative in finding ways around tasks which are difficult and impossible for me and have been lucky to have people around me to assist in those tasks I cannot do at an effective pace.

51 The accommodations New York University granted plaintiff included unlimited time to take the written comprehensive exams, use of an electronic typewriter with correction capability to take exams, and the use of a department secretary as an amanuensis. Bartlett I, 970 F. Supp. at 1101.

52 Defendants also point to the fact that plaintiff received her highest grade in law school on a first-year torts exam, on which she received no accommodations, as being inconsistent with her having a reading disability that requires her to be accommodated on the bar exam. (Tr. at 394.) At the original trial, however, plaintiff explained that she did well on the torts exam because it was primarily multiple choice and true-false, with short, single-issue questions, and with a generous amount of time allotted to complete the exam. (Trial Tr. at 626-35.) For comparative purposes, plaintiff noted that she had far more time to answer the multiple choice questions on the torts exam (where plaintiff had thirty minutes to answer 8 multiple-choice questions) than on the New York State multiple choice portion of the bar exam (where plaintiff had to answer approximately 17 questions in thirty minutes) or on the Multistate portion of the bar exam (where plaintiff had to answer approximately 25 questions in thirty minutes). (Trial Tr. at 628-29) Plaintiff also testified that true-false is an even easier format for her than multiple-choice. (Trial Tr. at 632.) Therefore, I conclude that the fact that plaintiff received her highest law school grade on a primarily multiple-choice and true-false torts exam with generous time allotted does not belie her claim that she is entitled to accommodations on the bar exam.

(Bartlett Aff. ¶ 20.) The above evidence certainly supports this statement.

In sum, while Dr. Flanagan's claim that plaintiff's accomplishments are inconsistent with her having a reading disability has superficial appeal, when I examine the manner in which plaintiff achieved each of these results, as I believe I should under the ADA and Section 504, I conclude that her successes are a result of her creativity in finding methods around reading and her extremely hard work and are further evidence that she is substantially limited in reading when compared to "most people."

C. The Comparison Group of "Most People"

In order to find that plaintiff has a reading disability, I must find that she is substantially limited in reading in comparison to "most people." 28 C.F.R. Pt. 35, App. A § 35.104. Dr. Flanagan contends that "most people" should be defined in reference to a bell curve on standardized tests. Specifically, she testified that individuals can only be substantially limited when compared to "most people" if their scores fall outside the average range of the 16th to 84th percentiles. (Flanagan Aff. ¶ 9(a)(i); Tr. at 616.) While I have already rejected the proposition that plaintiff's disability can be diagnosed solely on her scores outcome on tests without reference to the manner in which she achieved those outcomes, see supra Conclusions of Law, Section I.B., I also disagree with Dr. Flanagan that, to the extent that such test scores are relevant, I should equate "most people" with the average range on the bell curve. While the DOJ has not specifically defined the term "most people," I find the EEOC's discussion of this term⁵³ to be instructive:

53 I note that the DOJ Interpretive Guidance uses the phrase "most people," while the EEOC regulations use the phrase "average person in the general population." Cf. 28 C.F.R. Pt. 35, App. A § 35.104; 29 C.F.R. § 1630.2 (j)(l).

[An] impairment is substantially limiting if it significantly restricts the duration, manner or condition under which an individual can perform a particular major life activity as compared to the average person in the general population's ability to perform that same major life activity. . . . It should be noted that the term "average person" is not intended to imply a precise mathematical "average."

29 C.F.R. Pt. 1630, App. A § 1630.2(j) (emphasis added). In practical terms, Dr. Flanagan is attempting to do the same thing that I found "seriously infirm" in the first trial — setting a cut-off for the existence of disability. Bartlett I, 970 F. Supp. at 1113. In fact, Dr. Flanagan's suggested cut-off — below the 16th percentile — is even more conservative than Dr. Vellutino's cut-off — below the 30th percentile. To the extent that I found a cut-off of the 30th percentile on tests like the Woodcock and the WRAT to be under-inclusive based on research showing that one-third of dyslexics score above the 30th percentile on those tests, I find a 16th percentile cut-off to be even more problematic.

This Court is not unsympathetic to the Board's desire to have a clear standard by which to determine whether an applicant claiming to be learning disabled is entitled to accommodations on the bar exam under the ADA and Section 504. It is clear to me, however, that such a standard must take into consideration an applicant's evaluation report in toto, rather than focusing exclusively on psychometric test scores. Reading is a complex process composed of numerous cognitive functions. A deficit in one or more of these underlying processes can seriously affect an individual's ability to read. The Board (like many others in the public) wants the comfort of a test score to measure this complex process. While research about learning disabilities continues to advance and diagnostic tools are being improved, no test exists today whose scores alone can diagnose learning disabilities. Today, reliance on clinical judgments is necessary to diagnose learning disabilities (as it is with other types of disabilities, such as mental illness).

The Board's preoccupation with test scores and its distrust of clinical judgments, however, seems to be driven, at least in part, by misperceptions and stereotypes about learning disabilities. I noted in my first opinion and order that the Board appears to view applicants who claim to be learning disabled with suspicion. Bartlett I, 970 F. Supp. at 1136. of particular concern to me were alleged comments by James T. Fuller, the Executive Secretary of the Board, which if made, evidenced "a certain cynicism as to the existence of learning disabilities to begin with," his doubts as to the legitimacy of testing reports, and his view that "anyone who has the money can pay for a report [concerning a learning disability]." Id. This same attitude was evidenced at the remand trial when defendants and their experts implied on numerous occasions that plaintiff might be "faking" her reading problems or contriving her errors.⁵⁴ See, e.g., Tr. 109, 545; Carver Aff. ¶ 11. Defendants and their experts also repeatedly questioned the integrity of the clinicians who diagnosed plaintiff and of clinicians who diagnose learning disabilities generally. See, e.g., Tr. at 546, 609-13, 644-46; Flanagan Aff. ¶ 16.)

54 While defendants' preoccupation with plaintiff faking her symptoms is clear from the weeks of testimony before me, I find that Dr. Bartlett is a person of high integrity and that her testimony was credible and sincere.

While the Board's concern with protecting the integrity of the bar is laudable, the Board cannot turn this legitimate concern into a bias against learning disabled applicants. This type of bias is evidenced by alleged comments by Fuller, suggesting that people with learning disabilities are incapable of being competent lawyers. See Bartlett I, 970 F. Supp. at 1136 (noting that Fuller allegedly stated that "the law is a learned profession and I am not sure that a person with learning disabilities should aspire to such a goal," that "it was his job to protect the public from incompetent and incapable lawyers," and that the public would be "unaware that they would be purchasing a defective product in the case of learning disabilities"). While this attitude certainly is invidious, Fuller, if he in fact made these comments, would not be alone in holding it. For example, a few years ago, Boston University's provost delivered a number of speeches suggesting that "the learning disability movement is a great mortuary for the ethics of hard work, individual responsibility, and pursuit of excellence," that "the learning disabilities movement cripples allegedly disabled students who could overcome their academic difficulties with concentrated effort, demoralizes non-disabled students who recognize hoaxes performed by their peers and wreaks educational havoc," and that "hundreds of thousands of children are being improperly diagnosed . . . by self-proclaimed experts who fail to accept that behavioral and — performance difficulties exist." Guckenberger v. Boston University, 974 F. Supp. 106, 118-19 (D. Mass. 1997) (internal quotation marks and alterations omitted).

As I noted in my first opinion and order, much of this bias appears to arise from the assumption that giving extra time to applicants with learning disabilities gives them an unfair advantage over other applicants. Bartlett I, 970 F. Supp. at 1136. However, as I noted earlier, this assumption is belied by research showing that extra time does not have a significant impact on the performance of individuals who do not have learning disabilities. See supra Findings of Fact, Section IV.A.2. Further, as defendants concede, the bar is not a reading rate test. Plaintiff's experts have convinced me that the extra time provided to learning disabled applicants merely levels the playing field and allows these individuals to be tested on their knowledge; it does not provide them with an unfair advantage.

In practice, the Board's fear of compromising the integrity of the bar exam creates a presumption against accommodating a disability whose very essence is difficult to test. Many other institutions manage to protect their integrity, however, while at the same time reviewing requests for accommodations by learning disabled students in other ways than relying solely on test results from limited psychometric measures. For example, the University of California Hastings College of the Law and the University of Houston Law Center have set the following four criteria which must be met in order for an individual to receive accommodations for learning disabilities: (1) average or above-average intelligence as measured by a standard intelligence test which includes assessment of verbal and non-verbal abilities; (2) the presence of a cognitive-achievement discrepancy or an intra-cognitive discrepancy indicated by a score on a standardized test of achievement which is 1.5 standard deviations or more below the level corresponding to a student's sub-scale or full-scale IQ; (3) the presence of disorders in cognitive or sensory processing such as those related to memory, language or attention; and (4) an absence of other primary causal factors leading to achievement below expectations, such as visual or auditory disabilities, emotional or behavioral disorders, a lack of opportunity to learn due to cultural of socio-economic circumstances, or deficiencies in intellectual ability. These law schools required a report to be prepared by "a professional qualified to diagnose a learning disability, including but not limited to a licensed physician, learning disability specialist, or psychologist." Bartlett I, 970 F. Supp. at 1144 (internal quotation marks omitted). Similarly, Dr. Hagin testified that a proper evaluation report should be prepared by a person trained or licensed to diagnose learning disabilities, such as clinical or school psychologists, and should include information concerning the applicant's history, cognitive development, educational ability, and reading sub-skills. Id. Many states use experts to review applicants' documentation for accommodations on the bar exam, and at least one state uses a panel of four individuals, including a psychiatrist, a learning disabilities specialist, and a judge, to evaluate learning disabled applicants' documentation. Id. While it is unquestionably difficult to determine whether an individual has learning disabilities that require accommodations, other institutions have managed to create review standards without using cut-offs based on scores alone.

On the totality of evidence before me, including plaintiffs psychometric test scores, I find that plaintiff has proven that she is an individual with a disability under the ADA because she is substantially limited in the major life activity of reading when compared to most people.

II. Substantial Limitation Under the Law for the Major Life Activity of Working

The Second Circuit remanded for me "to determine, if necessary, whether plaintiff has shown that it is her impairment, rather than factors such as her education, experience or innate ability, that `substantially limits' her ability to work." Bartlett V, 226 F.3d at 85. I interpret the Circuit's use of the phrase "if necessary" to refer to the fact that (at least in the context of Title I of the ADA), a court should only examine whether an individual is substantially limited in the major life activity of working if it finds that the individual is not substantially limited in any another major life activity. See Sutton, 527 U.S. at 492 (stating that the EEOC "has suggested that working be viewed as a residual life activity, considered as a last resort"); 29 C.F.R. Pt 1630, App. § 1630.2(j) ("If an individual is not substantially limited with respect to any other major life activity, the individual's ability to perform the major life activity of working should be considered. If an individual is substantially limited in any other major life activity, no determination should be made as to whether the individual is substantially limited in working."). While ordinarily I would not reach the issue of whether plaintiff has a working disability because I find that she has a reading disability, given the long procedural history of this case, I will exercise an abundance of caution and discuss this issue in the event that the Circuit disagrees with my conclusion that plaintiff is substantially limited in the major life activity of reading.

The issue before me with respect to whether plaintiff has a working disability is extremely narrow. The Circuit has already affirmed my application of the EEOC's regulation defining the term "substantially limits" with respect to working and my determination that jobs utilizing legal training constitute a "class of jobs" for which plaintiff is otherwise qualified. — Bartlett V, 226 F.3d at 82-84. To the extent that the Second Circuit has not explicitly affirmed my conclusion that when plaintiff is compared to persons of "comparable training, skills, and abilities," in accordance with those EEOC regulations, plaintiffs reading skills are well below normal, I reaffirm that conclusion.⁵⁵

55 Obviously, because I now find that plaintiff's reading skills are below normal when compared to the general population, I also find that plaintiff's reading skills are below normal when compared to other law students, who represent a highly educated subgroup of the general population.

In my first opinion and order, I found that the bar examination functions like an employment examination that prevents plaintiff from working in her chosen profession:

If plaintiff's disability prevents her from competing on a level playing field with other bar examination applicants, then her disability has implicated the major life activity of working because if she is not given a chance to compete fairly on what is essentially an employment test, she is necessarily precluded from potential employment in that field. In this sense, the bar examination clearly implicates the major life activity of working.

Bartlett I, 970 F. Supp. at 1121. While I might have chosen a different word than "implicates" to describe how the bar exam functions like an employment test had I known the confusion its use would cause, I was merely addressing defendants' argument that the Title I definition with regard to working was not proper in a Title II case because the bar exam is not an employment test. I in no way intended to imply that plaintiff had shown only that her reading impairment "implicates" the major life activity of working rather than substantially limits it.

With that said, I still must answer the limited question posed to me by the Circuit on remand — whether "the denial of accommodations was a substantial factor preventing her from passing the [bar] exam." Bartlett V, 226 F.3d at 85. The Second Circuit emphasized that — plaintiff "need not prove that she would have passed the bar examination `but for' the denial of accommodations, because even under the best of circumstances a well qualified candidate may not pass on any given sitting." Id.

Plaintiff has shown a sufficient causal connection between her reading impairment and her failure to pass the bar. The bar examination involves extensive reading and writing. As! discussed in detail above, plaintiff has proven that her reading impairment seriously decreases the rate at which she reads and causes her significant fatigue. See supra Conclusions of Law, Section 1.A. Plaintiff also presented substantial evidence, which I credit, that her impairment also prevents her from writing quickly and easily, both physically (because she must consciously think of how to form letters) and compositionally (because she consciously uses only words she can spell). (Bartlett Aff. ¶ 23; Tr. at 387-89.) Moreover, plaintiff has convinced me that she needs extra time on tests in order for her true abilities and knowledge to be assessed. While the DRT is strong evidence of how her comprehension is significantly improved with extra time generally, plaintiff also testified at the original trial that when taking practice bar exam questions under timed conditions, her scores drastically improved when given double time as compared to her scores under the prescribed time. (Trial Tr. at 643-44.)

Plaintiff's causal evidence is not undermined by the fact that plaintiff has failed the New York Bar on two separate occasions with accommodations — once in July 1993 and once in July 1999. As the Second Circuit noted, "[t]he July 1993 examination is an especially poor indicator, given that Bartlett was not granted accommodations until two days before the exam and apparently had no opportunity to practice with her amanuensis, an accommodation she had never previously used." Bartlett V, 226 F.3d at 85 (internal quotation marks omitted). I agree. . . . Similarly, I find that plaintiff's performance on the July 1999 examination also is not determinative because, as she testified, her performance was substantially affected by the fact that her husband was diagnosed with cancer and had surgery only one week before the bar exam. (Tr. at 390.)⁵⁶

56 In July 1993, plaintiff also took and failed the Pennsylvania Bar Exam with partial accommodations. Plaintiff, however, was only permitted to use an amanuensis (the primary accommodation she was granted) for the essay portion of the bar and was not permitted that accommodation on the Multistate Bar Examination (MBE). (Tr. at 693.) Therefore, I do not believe that this exam is indicative of her what performance might be with full reasonable accommodations.

Defendants contend that I must determine whether plaintiff has sufficient domainspecific knowledge to pass the bar exam. I seriously doubt that the Second Circuit intended for me to answer this question because, as plaintiff correctly points out, there is no real way for me to make such an assessment. I can say, however, that plaintiff is bright, articulate, and hardworking. Despite her significant problems with reading and writing, plaintiff successfully completed law school, albeit with only mediocre grades.⁵⁷ She was nevertheless given excellent reviews when she worked as an associate at a law firm. Based on this, and to the extent that I can make a determination of her subjective knowledge, I believe she has sufficient knowledge of the law to pass the bar.

57 She testified at trial that although the accommodations she received in law school were insufficient, she believes that she would have failed her courses without them. (Trial Tr. at 623.)

When I view the evidence as a whole, I find that plaintiff has shown that the Board's denial of accommodations was a significant factor in her failure to pass the bar exam. While on any single exam there may have been other factors that affected plaintiff's ability to pass the bar (as I have found to be the case with the July 1993 and 1999 exams), I believe that the Board's failure to provide her with accommodations on those exams deprived her of a fair opportunity to be tested on her knowledge.

Therefore, I find that plaintiff has proven that her impairment of dyslexia substantially limits her major life activity of working and, therefore, that she is an individual with a disability under the ADA and Section 504.

III. Iniunctive and Declaratory Relief

Having demonstrated that plaintiff is disabled under the ADA and Section 504, plaintiff is entitled to injunctive relief in the form of reasonable accommodations on the bar examination. I reaffirm, to the extent that the Second Circuit's remand required that I revisit these issues from my first opinion and order, my conclusion that the accommodations plaintiff seeks are reasonable and order the Board to provide plaintiff with the following accommodations if she should decide to retake the New York Bar Examination in the future: (1) double the normally allotted time, spread over four days; (2) the use of a computer (3) permission to circle multiple choice answers in the examination booklet; and (4) large print on both the New York State and Multistate Bar Examinations. Bartlett I, 970 F. Supp. at 1146-47, 1153. For the reasons I stated in my earlier opinion and order, I find that declaratory relief is not appropriate in this case. Bartlett I, 970 F. Supp. at 1147.

IV. Compensatory Damages

In my earlier opinion and order, I found that plaintiff was entitled to compensatory damages. I determined that, to the extent that intent is required for recovery of compensatory damages under the ADA or Section 504, plaintiff had met her burden of proof on this issue. Bartlett I, 970 F. Supp. at 1151. In previously considering damages, I found that plaintiff's purported losses of salary and benefits were unduly speculative and declined to award her compensatory damages for those losses. However, I determined that plaintiff's taking of the bar examination without the accommodations to which she was entitled under the law "was a waste of her time and money." Bartlett I, 970 F. Supp. at 1152. Therefore, I awarded plaintiff $2,500 for each of the five bar examinations that she took, totaling $12,500. Id.⁵⁸

58 With respect to plaintiff's request for punitive damages, I did not find the requisite level of "malice" or "reckless indifference" to federally protected rights that would justify such an award, and therefore, declined to award such damages. Id. at 1152-53.

In Bartlett III, the Second Circuit affirmed my conclusion that plaintiff was entitled to compensatory damages. Bartlett III, 156 F.3d at 331. With respect to my calculation of damages, however, the Second Circuit stated that I had erred by failing to "examine whether, for each bar examination, there was a denial of accommodations due to illegal discrimination." Id. at 332. The Second Circuit held that because plaintiff did not seek accommodations for the February 1992 bar examination, the Board could not be held liable for damages for that particular exam. In contrast, the Circuit stated that because the Board illegally denied plaintiff's timely request for accommodations in taking the February 1993 bar exam, the Board would be held liable for plaintiff's damages arising from that examination. With respect to the remaining three bar examinations, the Second Circuit remanded for me to make additional findings of fact and to recalculate the compensatory damages due to plaintiff. Id. at 332.

In Bartlett V, the Second Circuit stated that it adhered to its original reasoning in Bartlett III regarding compensatory damages. Bartlett V, 226 F.3d at 86. The Circuit, however, added an additional factor for me to consider in my recalculation of damages:

In addition, we note that the district court should consider, if it finds a disability, whether the Board had enough information to determine that Bartlett was disabled. In other words, the district court may conclude that Bartlett is substantially limited by her slow rate of reading, but it may not be the case that the requests. for accommodations and the accompanying reports provided to the Board reflected that limitation. If Bartlett's requests for accommodations did not provide enough information for the Board to determine that she was disabled, the Board would not be liable.

Id. (internal citations omitted). The Second Circuit in Bartlett V did not specify whether I should consider this additional factor only with respect to the three exams for which it had remanded for additional findings in Bartlett III (the July 1991, July 1992, and July 1993 bar exams) or whether I should also consider this factor with respect to the February 1993 exam, for which the Circuit in Bartlett III had affirmed my compensatory damages award. Barlett III, 156 F.3d at 332. In an abundance of caution, I will address this additional factor with respect to each of the exams plaintiff has taken.

Before doing so, however, I note that the Circuit's instruction to consider this additional factor appears to be in some tension with the findings in my first opinion and order regarding the Board's practice of using cut-offs to determine whether a learning disabled applicant was entitled to accommodations. Given my previous finding that the Board denied plaintiff's request for accommodations solely because of her test scores, Bartlett I, 970 F. Supp. at 1111-12, it seems unnecessary for me to consider a post-hoc justification for the Board's denial of accommodations — that the clinical observations (which the Board consistently has argued should be disregarded as subjective) in the evaluators' reports did not provide the Board with enough information of plaintiff's disability. Dr. Vellutino testified at the original trial that "a diagnosis of dyslexia can only be based exclusively on measures of reading ability, in particular measures of Word Identification and phonetic decoding or word analysis skills (ability to `sound out' a word), deficiencies in which are characteristic of individuals with severe reading disability.'" Bartlett I, 970 F. Supp. at 1111-12. Based on this view, Dr. Vellutino recommended to the Board that it automatically grant accommodations to applicants claiming to have reading disabilities if their scores on the Woodcock Word Attack and Word Identification subtests were below the 30th percentile (when age and grade normed). Id.; see also Hagin Second Supp. Aff. Ex. 1, Pl.'s Ex. 123-A57, File of Applicant 57, at A57-6c-6d. Because the evidence showed that the Board would have denied plaintiff accommodations no matter what clinicians observations were contained in her report because her test scores were above its cut-off, I would award plaintiff compensatory damages for each of the times she applied for, and was denied, accommodations for taking the bar examination.

The evidence at the remand trial lends further support to this analysis. For example, in reviewing Applicant 57's file, Dr. Hagin noted that, although the report in support of Applicant 57's request for accommodations clearly related details of clinical observations and functional reading and writing behaviors showing that this applicant lacked automaticity and read and wrote slowly and with great effort, Dr. Vellutino was dismissive of the report, and only granted that applicant accommodations after she scored below the 30th percentile on the Woodcock Word Attack and Word Identification subtests. (Hagin Second Supp. Aff. ¶¶ 3, 4, 7, 8, 9.) Even Dr. Flanagan conceded that "[g]iven Vellutino's cut-off score of 30th percentile it is not surprising to me that the Board granted A57 accommodations because 3 of her 4 [Word Attack and Word Identification subtest] scores were below the 30th percentile" and that "because 7 of Plaintiff's 8 [Word Attack and Word Identification subtest] scores were above the 30th percentile . . . it is not surprising to me that the Board did not grant Plaintiff accommodations." (Flanagan Supp. Aff. ¶ 15.)

Despite the preceding misgivings about the Circuit's instruction to consider whether plaintiff presented the Board with sufficient information, I now consider this question with respect to each of the five occasions on which plaintiff took the bar exam.⁵⁹ With respect to plaintiff's request for accommodations on the July 1991 bar exam, plaintiff submitted Dr. Massad's 1989 evaluation of her. Bartlett I, 970 F. Supp. at 1102. The Board denied her request for accommodations because: (I) her request had been submitted fewer than 45 days before the examination, as required by then-existing policies; (2) Dr. Massad's report was not current enough; and (3) plaintiff's test scores on the Woodcock did not qualify her for accommodations. (Pl.'s Ex. 56, July 1, 1991 Letter from James T. Fuller.) I have already rejected the argument that plaintiff should not recover for the July 1991 bar exam because her application for accommodations was not timely, finding that "[t]he defendants have consistently through the years considered untimely applications" and that the letter from the Board to plaintiff denying her accommodations on this test "made clear that the Board considered the lack of merit in plaintiffs application as the primary reason for denying an accommodation." Bartlett II, 2 F. — Supp. 2d at 396 n. 7. While the Board also stated that it did not consider Dr. Massad's 1989 evaluation of plaintiff as current enough to support her July 1991 request for accommodations, other evidence in the record shows that the Board only considered evaluations more than three years old to be "not current." (Hagin Second Supp. Aft Ex. 1, Pl.'s Ex. 123-A57, File of Applicant 57, at A57-5b.) Therefore, the sole question with respect to the July 1991 bar exam is whether Dr. Massad's report provided sufficient evidence to allow the Board to conclude that plaintiff is disabled. I find that it did.

59 I note that, in the end, it does not matter whether I apply the standard which I believe to be correct under the facts of this case or whether I apply the standard as articulated by the Circuit in its most recent remand, because I reach the same result under both tests.

Dr. Massad's evaluation clearly set forth plaintiff's extensive history of reading problems. It stated that plaintiff complained that "it takes her a long time to read material and that her writing lacks cohesion and appropriate use of transitional devices." (Pl.'s Ex. 20a, at 1.) He reported that plaintiff used the following techniques to assist with her reading problems: having others read to her, participating in study groups, using a card to help her move from line to line, and putting a hole in a card to help her read word by word when reading particularly difficult material. ( Id. at 1-3.) He also stated that plaintiff reported that she had a tendency to write backwards and reverse or invert numbers and letters. ( Id. at 1.) In interpreting the scores from the tests he had administered to plaintiff, Dr. Massad reported that the scores reflecting her decoding skills, including the Woodcock Word Identification and Word Attack subtests, were very discrepant (albeit still in the average range) relative to plaintiff's Verbal IQ. ( Id. at 4.) This discrepancy, when combined with plaintiff's history and presenting problems, in Dr. Massad's opinion, suggested that plaintiff had a "phonologically based decoding difficulty (i.e., dyslexia)." ( Id.) He also stated that her below average spelling (albeit still within the average range) also suggested difficulty with phonetic decoding of words. ( Id.)⁶⁰

60 In concluding that Dr. Massad's report did not provide sufficient information to the Board to determine plaintiff's disability, Dr. Flanagan relied on Dr. Massad's comment that he did not report plaintiff's score on the Oral Gray Reading test because there was nothing remarkable to report. (Flanagan Supp. Aff. ¶ 7.) However, because this comment is not in the report, but rather was made at the bench trial, it is irrelevant to my determination of whether Dr. Massad's report provided the Board with sufficient information. Compare Pl.'s Ex. 20a with Trial Tr. at 224.

In discussing Dr. Massad's report, plaintiff's experts agreed that it was quite extensive and sophisticated for the time, especially given that few clinicians were experienced in assessing adults then. (Tr. at 496-97.) The Board's main criticism of Dr. Massad's report is that it contained no objective test scores to support his diagnosis. As I have already discussed, I do not believe that test scores alone can diagnose plaintiff's disability. In fact, each of plaintiff's experts testified that in diagnosing an adult with suspected dyslexia or learning disabilities, the individual's history and presenting problem are most important. (Moats Aff. ¶ 21; Mather Aff. ¶ 13; Hagin Supp. Aff. ¶ 7; Tr. at 269.) I find that those sections of Dr. Massad's report sufficiently evidenced plaintiff's disability.

While defendants would like me to view Dr. Massad's 1989 evaluation through a 2001 lens, with the benefit of over ten years of research in the field of learning disabilities and dyslexia, I decline to do so. Plaintiff's experts have convinced me that their present belief that plaintiff's dyslexia is both phonologically and orthographically based is not inconsistent with Dr. Massad diagnosis of plaintiff with phonologically-based dyslexia but is merely a function of additional research regarding the underlying role of orthography in dyslexia. The fact that Dr. Massad did not use the term "automaticity" is likewise not determinative because, as all of plaintiff's experts explained, the behavioral observations he reported sufficiently evidence — plaintiff's lack of automaticity. (Tr. 235-36, 339, 501-02; see also Moats Aff. ¶ 22; Mather Aff. ¶ 22.)

Because I find that the Board had sufficient information before it to determine that plaintiff was disabled at the time she applied for accommodations on the July 1991 bar exam, I award her $2,500 in compensatory damages for taking that exam without accommodations.

Plaintiff did not request accommodations for the February 1992 examination, and, as a result, the Second Circuit determined in Bartlett III that plaintiff was not entitled to compensatory damages for that exam. Bartlett III, 156 F.3d at 332. As for the July 1992 bar exam, plaintiff presented no evidence to support her claim that she applied for accommodations for that exam, Bartlett I, 970 F. Supp. at 1102, and, in any event, the briefs plaintiff submitted after the remand trial focus exclusively on her request for compensatory damages for the July 1991, February 1993, and July 1993 bar exams. (Pl.'s Post-Hearing Br. at 61-62; Pl.'s Post-Hearing Reply Br. at 22-27.) Therefore, I find that plaintiff is not entitled to compensatory damages for either the February 1992 or July 1992 bar exams.

In support of her application for accommodations on the February 1993 bar examination, plaintiff submitted Dr. Massad's 1989 report and a letter from him dated November 1992 reasserting the opinion he had set forth in his 1989 evaluation. Bartlett I, 970 F. Supp. at 1102-03. The Board denied plaintiff's request for accommodations, stating that "the documentation you submitted is insufficient to establish a basis for granting you the accommodations you request." (Pl.'s Ex. 38, Jan. 26, 1993 Letter from James T. Fuller.) I find that Dr. Massad's 1989 report contained sufficient information for the Board to determine that plaintiff was disabled for the reasons I discussed in connection with the July 1991 bar exam. Therefore, I find that plaintiff is entitled to $2,500 in compensatory damages for taking the July 1993 bar exam without accommodations.

Finally, plaintiff submitted the evaluations from both Drs. Massad and Heath in support of her request for accommodations for the July 1993 bar exam. While I believe that Dr. Massad's report alone gave the Board sufficient information to determine that plaintiff was disabled, the Board undoubtedly had more than enough information to make that determination with Dr. Heath's additional report. As I noted earlier, the purpose of Dr. Heath's evaluation was to confirm Dr. Massad's diagnosis and to supply plaintiff with the Woodcock scores requested by the Board. With respect to plaintiff's history, Dr. Heath stated that she "reports a long history of reading problems, primarily with reading rate and decoding." (Pl.'s Ex. 16, at 1.) In discussing plaintiff's scores on the Woodcock forms he had administered, Dr. Heath observed that some of her scores were below expectations based on her superior level of intellectual functioning, particularly her score of 18th percentile on the Letter Identification subtest. In two separate places in the report, he reported his observation that plaintiff made letter reversals when taking the test — one b/q reversal on the Letter Identification subtest and two b/d reversals on the Word Attack and Word Identification subtests. ( Id. at 1-2.) He also observed that when taking the Woodcock, plaintiff self-corrected, read the passages slowly, and reread the more complex passages two or three times. ( Id.) He specifically stated that plaintiff "decoded words slowly and without automaticity." ( Id. at 1.) Based on plaintiff's history, test scores, and behaviors exhibited when taking the Woodcock, Dr. Heath confirmed Dr. Massad's diagnosis of plaintiff. ( Id. at 2.)⁶¹

61 In concluding that Dr. Heath's report did not provide the Board with sufficient information, Dr. Flanagan relied on Dr. Heath's diagnosis of plaintiff as having mild to moderate dyslexia. (Flanagan Supp. Aff. ¶ 7.) As with Dr. Massad's comment regarding the Oral Gray, Dr. Heath's comment was made at the bench trial and was not contained in his report. Compare Pl.'s Ex. 16 with Trial Tr. at 507. Thus, this comment is not relevant to the question of whether Dr. Heath's report provided the Board with sufficient information.

Again, defendants' argue that the test scores do not support Dr. Heath's diagnosis. I have repeatedly found that the Board cannot rely on test scores alone, and that the Board chose to ignore Dr. Heath's clinical observations that explicitly reported plaintiff's lack of automaticity and her slow reading rate. Defendants also claim that Dr. Health should have administered to plaintiff a reading rate test. That argument is, to say the least, disingenuous, given that the primary purpose of Dr. Heath's evaluation was to administer the Woodcock tests which the Board had specificaily requested. Additionally, as I noted previously, the reading rate measures available for adults were (and still are to a large extent) very limited. See supra at note 36. In any event, I agree that Dr. Heath's clinical observations more than sufficiently evidence plaintiff's lack of automaticity and reading rate problems. Because the Board had sufficient information before it in July 1993 to determine that plaintiff was disabled, I find that plaintiff is entitled to $2,500 in compensatory damages for taking that exam.⁶²

62 I have already held that plaintiff's entitlement to compensatory damages for the July 1993 exam is not undermined by the fact that the Board granted plaintiff some limited accommodations on that exam (under a stipulation entered into on the eve of the bar exam) because "the accommodations granted were neither those that the plaintiff requested nor those to which the Court has deemed plaintiff was entitled." Bartlett II, 2 F. Supp. 2d at 396 n. 7.

In sum, I find that plaintiff is entitled to compensatory damages of $2,500 per exam, totaling $7,500, for her taking of the New York Bar Examination in July 1991, February 1993, and July 1993.⁶³

63 After the Second Circuit issued its most recent decision and after I had held the remand trial, the Supreme Court issued its decision in Board of Trustees of the University of Alabama v. Garrett, 121 S.Ct. 955 (2001), where the Supreme Court held that Congress exceeded its powers under the Fourteenth Amendment when it abrogated States' Eleventh Amendment immunity under Title I of the ADA. Because the decision addressed exclusively Title I of the ADA, it is an open question whether the same rationale applies to the other titles of the ADA or to Section 504. The parties have notified me that the Second Circuit in Garcia v. State University of New York, No. 00-9223, is in the process of answering the Eleventh Amendment questions with respect to the other titles of the ADA and Section 504 Therefore, rather than delving into this thorny issue myself when the Second Circuit will likely decide it before any appeal of this opinion and order, I do not address this issue and limit myself to the issues remanded to me by the Second Circuit in Bartlett V.

CONCLUSION

For the reasons discussed, I find that plaintiff is an individual with a disability as defined by the ADA and Section 504. With respect to the specific questions before me on remand, I find that when considering both the positive and negative effects of the mitigating measures which plaintiff employs, plaintiff is substantially limited in the major life activity of reading when compared to most people by her slow reading rate and by the fatigue caused by her lack of automaticity. In the alternative, I find that plaintiff is substantially limited in the major life activity of working because she has shown that her reading impairment is a substantial factor in her failing the bar examination. Therefore, I find that the Board is liable under the ADA and Section 504 for failing to provide reasonable accommodations to plaintiff on the New York State Bar Examination.

I award plaintiff her requested injunctive relief and order that plaintiff shall receive the following reasonable accommodations should she decide to retake the New York Bar Examination in the future: (I) double the normally allotted time, spread over four days; (2) the . . . use of a computer; (3) permission to circle multiple choice answer in the examination booklet; and (4) large print on both the New York State and Multistate Bar Examinations.

I also find that plaintiff provided the Board with sufficient information to determine her disability in her requests for accommodations on the July 1991, February 1993, and July 1993 bar examinations. I award plaintiff compensatory damages of $2,500 for each of these exams, totaling.$7,500.