Guide to Identification of
1998 New York State
John O. Willis,
Ron Dumont, Ed.D.
Fairleigh Dickinson University
1. ß SLD means a disorder in one or more of the basic
psychological processes involved in understanding or in using
language . . . You cannot identify a SLD without specifying a
disorder and showing how it impairs school work.
2. ß Severity of SLD is not measured by the severity of the
weakness in basic process(es). It is measured by the severity of the
impact on academic achievement. Real-life academic achievement is
often more important than achievement test scores. Careful,
diagnostic assessment of achievement is often the core of the
3. ß The "exclusions" are not as important as most people
think, as long as the disorder in basic process(es) and impact on
academic achievement have been documented thoroughly and
4. ß Students may have SLD as a disability secondary to
another disability, even intellectual handicap.
5. ß Global intelligence measured by total IQ scores (GCA,
MPC, BCA, etc.) is usually not a helpful construct for understanding
the cognitive functioning of students with specific learning
disabilities. Part 200 Regulations do not define "intellectual
ability" even as "intelligence," much less as an "intelligence test
score." Cognitive ability factors are usually the most helpful level
6. ß Beware of assessing cognitive factors with fewer than
three "clean," independent measures.
7. ß The factor structure of a test for persons with
certain, specific disabilities may be very different from the factor
structure for the norming sample.
8. ß It is almost always better to adopt an appropriate test
than to adapt an inappropriate one for a student with a severe
9. ß Parents, teachers, and the students themselves make
important contributions to the evaluation, and they must be included
in the process. Examiners should elicit genuine referral questions ,
not just "issues" to be answered by the evaluation. Interviews and
questionnaires are essential parts of a complete evaluation.
10. ß To ensure the valuable contributions of parents,
teachers, and students, evaluation results must be as clear as
possible. Jargon and statistics must be defined very clearly.
11. ß The total evaluation must be integrated, which is not
achieved with a staple.
12. ß SLD identification is a professional judgment by a
team, not an exercise in arithmetic.
13. ß However, any arithmetic involved should be
14. ß Tests and scores are not interchangeable. A student's
age- and grade-equivalent scores (which are horrible statistics) will
not come out in the same rank order as the student's standard scores.
Age-based and grade-based norms differ, and often both must be
reported. Discontinuities between fall, winter, and spring norms can
be dramatic. The same performance yields very different scores on
different tests. The same grade equivalent yields different standard
scores on different tests.
15. ß Reading tests use an extraordinary number of means of
assessing reading. They are not interchangeable. Often, several
reading tests must be used for a complete picture. Reading fluency
and study skills are important.
16. ß Math tests almost always require limits-testing.
Correct test scores, including "math-fact" errors and misreading of
operations signs, must not be confused with the reality of the
student's math skills.
17. ß Writing must be assessed carefully. The best formal
written expression tests have many flaws. Writing samples may be
needed in addition to tests.
18. ß Achievement testing should include detailed
descriptions of actual skills, gaps, and weaknesses. It is useless
simply to print a table of test scores and describe the scores in
19. ß Tests that combine separate skills in single scores
(e.g., reading decoding and reading comprehension or math calculation
and math applications) are as useful as a second handle on a snow
20. ß Discrepancy formulae are statistical nightmares. The
Part 200 Regulations do not state that a "50% discrepancy" (whatever
that may be) is the only way to identify a SLD. Legal precedent
indicates that the absence of a "50% discrepancy" cannot be used to
exclude children from SLD classification.
21. ß Norms are much more important than most people think.
Norms are worse than most people think. Validity and reliability
matter. Validity for specific purposes and reliability over realistic
spans of time are rarely documented.
22. ß Diagnoses are not political or economic decisions.
23. ß Relative and transient weaknesses must be taken
24. ß Examiners must use the best instruments available.
Inadequate tests should be used only when they are absolutely
necessary and the best existing for the purpose.
25. ß All disabilities, including SLD, can be seen as
mismatches between learning style and instruction. Changes in
circumstances can alter the need for special education.
26. ß Evaluation processes that always or never recommend
highly restrictive placements are equally suspect.
27. ß Fads in diagnosis and treatment must be avoided.
28. ß Evaluations must be a careful, thoughtful, thorough
process, whether initial or re-evaluations
29. ß Concrete recommendations, individually planned for the
student, are an important goal of the evaluation. Stock, boilerplate
recommendations are not much help. A useful evaluation with
recommendations does not cost much more than a useless one without
30. ß Computers don't write reports. People do.
31. ß Evaluations should be individual and humanistic,
consider multiple intelligences, reflect reality beyond test scores,
and accept the possibility of improvement in areas of weakness as
well as potential stability of individual patterns of strengths and
32. ß There is often a huge gap between the science of SLD
identification and the social policy involved with the identification
33. ß Best practice and educational law are often in
34. ß There is often a distinct difference between an
evaluation for classification and an evaluation for diagnosis of