DESCRIPTION OF TESTS
Reading tests do not "rule out" problems.
Instead, reading tests are controlled observations of
the processes that are needed to develop good reading skills (including accuracy,
fluency, and comprehension). A variety of tests is used
to identify strengths and weaknesses in the processes (such as attention, phonemic
awareness, word retrieval) and knowledge (such as word identification, spelling).
The ultimate purpose of testing reading should be to
direct intervention to where it will produce the greatest good. While there is a
central core of tests that are generally useful, no student takes every available test.
1. Attention
2. Non-verbal processing
3. Language
4. Working
memory
5. Manipulating
sounds in syllables
6. Retrieval
from memory: Rapid Automatic Naming
7. Word level reading skills
8. Text level reading skills
9. Spelling
- Attention
: A "core
deficit" has not yet been identified in Attention Deficit Disorder (ADD), which
includes various forms of attention difficulty -- focusing attention, sustaining
attention, saliency determination, marshaling resources, and impulsivity. Diagnosis
requires sufficient characteristic behaviors to be present for more than six months and in
more than one setting (for example, school and home).
We generally ask parents to complete the Parent checklist
(Connors 48), a behavioral checklist. A computerized Continuous Performance Test
(Connors CPT) provides objective measures of some aspects of attention, distractibility,
and impulsivity that are sensitive to medication effect. The CPT is useful in monitoring
the response to medication.
- Non-verbal processing: The Test
of Non Verbal Intelligence (TONI) is a test of puzzle solving. The student chooses a
geometric design that best fits a sequence. A non-verbal test is helpful in
differentiating a language disorder from overall learning problems.
- Language:
Vocabulary is a major factor
in comprehension. Listening vocabulary (Peabody Picture Vocabulary Test, PPVT)
requires the student to point to a picture that represents a spoken word. ("Point to
the picture of the pagoda.") Limited vocabulary will impair comprehension even if the
words are read accurately. Speaking vocabulary (Expressive Vocabulary Test, EVT)
requires that the student look at a picture and produce a synonym for a spoken word.
("What is another word for ‘blazing’?") Low speaking vocabulary often
shows as sparse language and limited choice of vocabulary in describing and writing.
Oral Comprehension (Token Test) is a test of understanding of
oral directions. It uses very limited vocabulary (5 colors, big and little, square and
circle), but puts a greater emphasis on the understanding of syntax. ("Put the yellow
square behind the red circle.")
A general language-screening test may be used to determine
whether administering a battery of language tests is likely to uncover a primary language
disorder. The Clinical Evaluation of Language Fundamentals (CELF) is a battery of
language tests that are used to document significant difficulties in receptive, expressive
and total language.
- Working memory is the label used for
a temporary holding place in memory. A student with low working memory may correctly
"sound out" each sound in a word, but leave out a sound when trying to blend
them together. (/b/ /l/ /a/ /k/ --> "back") Another student may read the
words correctly, but flatly, without expression. Because he doesn’t recognize how
words group together into logical units, comprehension and retention may be impaired.
("Early in the morning" "before dad left for work" "the street
cleaners" "were busy sweeping the confetti" "that had rained
down" "during the parade.") Normally a sentence of many words breaks into a
much smaller group of ideas (or thematic units), each of which provokes related memories.
A group of thematic units that provokes widespread activation may be analyzed deeper and
retained better. It is not surprising then that short-term memory measures are dependent
on the meaningfulness of what is to be remembered.
The CELF-R Recalling Sentences subtest asks the student to
repeat sentences that increase gradually in length and complexity. Memory for Digits
(repeating number sequences: 9-4-7 3-5-1-8 etc) and Non-Word Repetition (repeating
nonsense words of increasing length) are subtests of the Comprehensive Test of
Phonological Processing. They are used to calculate the Phonological Memory Composite.
5. Manipulating sounds in syllables: Phonemic awareness is
defined as the awareness of the number, type, and order of sounds within a syllable.
("Green" consists of 4 sounds -- /g/ /r/ /ee/ /n/.) A level of phonemic
awareness is necessary for the student to be explicitly aware of the sounds in a word.
With that awareness, he can map the sounds to the available letters and extend his
knowledge of sound-symbol and symbol-sound relationships. (The student’s thought
process might go like this: "Clean" makes sense here and it has four sounds: /k/
/l/ /ee/ /n/. I have seen the /k/ sound written as the letter c, and I know what l and n
stand for. There is only one sound left, the /ee/ sound, so the letters ea might be
another way of writing the /ee/ sound. Cool!") The student gradually masters the
alphabetic principle, is able to read new words that he has never seen before, and
discovers new sound-symbol associations.
To have sufficient phonemic awareness to provide a stable foundation
for reading, the student needs to be able to divide words into syllables orally and to
manipulate five-sound syllables accurately and easily. 90% of students can do the first
with little or no instruction; it is the latter (awareness of sounds within syllables)
that is the initiating problem for most poor readers.
Auditory conceptualization (LAC, Lindamood Auditory
Conceptualization test, Form X) assesses the student’s ability to represent with
colored blocks the number, order, and type of speech sounds made by the examiner. The
nonsense word /stip/ would be represented by four different colored blocks. Changing
/stip/ to /stis/ would require replacing the fourth block with a block matching the first
block’s color. It is an abstract test (colored blocks, no letters, nonsense words)
that can be quite difficult for a kindergarten student.
The Rosner Test of Auditory Analysis is a test of Elision.
("What is hat without the /h/? -- /at/"). Elision requires working memory to
hold the original word, segmenting to cut off a sound, and blending to build the resulting
new word. It uses real words. Some older students with poor phonemic awareness can score
relatively well if they have memorized some of the words on the test. Nonsense word
spelling can help detect these students.
Elision and Blending subtests, from the Comprehensive
Test of Phonological Processing, combine to form a Phonological Awareness Composite.
For planning instruction, it is useful to know the actual level of
syllables (2-5 sounds) that the student can handle. This is determined with informal
measures of blending (/c/ /a/ /t/ à cat), segmentation (cat -> /c/ /a/
/t/), and elision (cat without /c/ à at). For example, if it is determined that a
student can handle 3 sound words but needs help with 4-sound words, fluency instruction
would start at the 3-sound level and decoding instruction at the 4-sound level.
- Retrieval from memory: Rapid
Automatic Naming (RAN) is measured by timing the student while he "reads" a
series of numbers, letters, or colored dots (no words involved) as quickly as possible. It
is considered a measure of the speed at which the sound codes for words can be retrieved
from memory.
At a minimum, slow word retrieval will slow reading rate. If
retrieval is even slower, the student will attempt to decode words that are already in
memory (but he doesn’t realize it because they have not yet been retrieved). Rapid
word retrieval allows the student to read some new words by comparing to words already
known. For example, a student knows the word "light" and comes across the (to
him) new word "slight." The word "light" is retrieved and placed in
working memory. The phonological processor adds the sound /s/ to the sound code for
"light" and the student recognizes "slight." This is a much faster
process than decoding the word from left to right. Rapid word retrieval is an important
part of the mechanism by which the student generalizes word attack skills – develops
and learns to apply an array of over 600 sound-symbol relationships that are needed to
decode new words.
In later grades, slow word retrieval can impair comprehension. In
science, history, or other content course books, as the length of sentences and the
percentage of new words increase, slow retrieval produces a logjam in working memory.
Words cannot be recognized quickly enough to group together into logical units in working
memory, and text becomes a string of words, rather than a string of ideas. The student may
"see the trees" (read the words accurately, albeit slowly) but "miss the
forest" (miss the idea). Slow reading impairs comprehension, just as looking at a
famous painting though a pinhole in a card may not allow one to identify what you are
seeing, or hearing 2 second snippets of a song may not allow one to identify it.
Finally, spelling is affected. The "that doesn’t look
right" phenomenon that guides good spellers to try an alternative spelling or check a
word in a dictionary, does not happen if words don’t fly out of memory while one is
proof reading.
Rapid Digit Naming and Rapid Letter Naming, which form
the Rapid Naming Composite, are new standardized measures from the Comprehensive Test of
Phonological Processing.
Slow retrieval speed is often (but not invariably) associated with inaccurate
retrieval. The Test of Word Finding (TWF) provides a measure of the accuracy of
retrieval.
- Word level reading skills:
At
the most fundamental level, Names for letters asks the student to identify by name
the letters of the alphabet. The Sounds for symbols test asks the student to
"read" 50 basic symbol-sound associations: What sound would you "read"
if you saw "p" on the paper? "oy"? "ai"? "eigh"?
etc. In federal law terminology, word level reading skills
can be referred to as "basic reading" (which is a separate category from reading
comprehension). Basic reading consists of word attack and word identification skills, from
which a basic reading index is computed. The test of basic reading that is most commonly
used in research is the Woodcock Reading Mastery (WRM).
Word attack is a measure of the ability to read unknown
words. It is always measured using nonsense words ("bem," "nin"). Also
called phonemic or alphabetic reading, word attack is the single most reliable predictor
of future growth in basic reading skills.
Word identification is an estimate of the store of words that
the student can read relatively quickly. It consists of words that the student can decode
by word attack skills (/b/ /a/ /t/ -> "bat"), by analogy to a known word
(reading "slight" by adding /s/ to "light"), and by immediate
recognition ("sight words" in the sense of words read automatically and without
conscious thought). (If the student takes a full 5 seconds to read each word, he receives
full credit. If all words were read that slowly, reading rate would be 12 words per
minute, extremely slow.)
The Basic Reading Cluster is computed from word attack
and word identification. Good word attack and word identification are necessary for
reading accuracy, the first step toward reading proficiency. (Grade-level reading
proficiency requires that accuracy, rate, and comprehension be at grade level.)
Word attack is a useful skill for discussing an appropriate use of Grade Equivalent (GE). Let us suppose that George, a
fourth-grade student, has word attack skills at a 2.0- grade equivalent level.
George’s fourth-grade teacher sends George to the second-grade classroom to find an
average second-grade reader – one who is just average for second grade. George brings
this average second-grade student back to the fourth-grade class. The teacher hands
the second-grader the fourth-grade science book and tells the student to start reading
aloud to the class. Because both George and the second-grader have word attack skills at
the 2.0 grade level, each has the same statistical probability of correctly reading the
new words of the textbook. (This would be child abuse.)
The Woodcock Reading Mastery provides a another interesting measure
that helps to understand why George no longer likes school and says he’s dumb, etc.
It is called the Relative Performance Indicator (RPI). Let's say
that George has an RPI for word identification of 40/90. This means that George
reads correctly only 40% of the words that his age-peers are reading with 90% accuracy.
Such a gap is a set up for stress, failure, and loss of self-esteem. Yet it is silently
tolerated or even defended by adults who at other times proclaim the virtues of
"early detection and intervention" and of "preparing the child for the
future."
The reading part of the Wide Range Achievement Test (WRAT) is
similar to the Woodcock Reading Mastery word identification measure. However, the
student has a 10 full seconds to read each word! A good score of this test indicates
accuracy, but could cover up a significant lack of fluency. (If a student took 10 seconds
to read each word, his reading rate would be only 6 words a minute!)
Sight Word Efficiency and Phonemic Decoding Efficiency
are timed measures of how many real words (sight word efficiency) and nonsense words
(phonemic decoding) the student can read in 45 seconds. According to the authors, any
score below 30%ile on these efficiency measures suggests a high risk for reading problems
that will interfere with the student’s ability to make adequate progress in school.
- Text level read skills.
The Gray Oral Reading Test (GORT) is commonly used in research and clinical
testing because it provides standardized measures of rate, accuracy, and comprehension.
While comprehension is the goal of reading, "adequate"
comprehension scores do not mean that the student has compensated and that basic reading
issues can be ignored. In the first place, reading comprehension is a separate category
from basic reading under federal law. Secondly, there are difficulties in measuring
comprehension that limit its value as a overall measure of the student’s reading
skill or as a predictor of future reading progress. Reading comprehension measures are not
pure measures of reading comprehension, but are usually strongly contaminated by
oral language comprehension. In the case of the GORT, the comprehension questions are read
aloud to the student. The format is multiple choice, requiring recognition (not
production) of a correct answer. The test does not control for the student’s prior
knowledge (background knowledge) and therefore can’t measure what information
actually came from reading. For example, most people can answer this question (only
minimally altered from the GORT), based on their prior knowledge and without reading the
required text: "How does a farmer feel when blackbirds wipe out his corn crop? a)
indifferent, b) defeated, c) enthused, d) inferior." Finally, comprehension questions
based on narrative (story) text is not measuring the same degree of skill as is needed
with expository (e.g. science, history) text. "Grade-level" comprehension of
stories does not imply grade-level comprehension of the more difficult expository text
that predominates content area courses.
The Scholastic Reading Inventory (SRI) is a new
methodology (the Lexile Framework) for assessing student comprehension and the readability
of text (text difficulty). Its measurement ultimately depends on the vocabulary and
sentence structure of the passage. It appears particularly useful for selecting reading
material that matches a student’s reading needs. (One level is chosen for the
objective of challenging the student to build vocabulary and comprehension skills. A quite
different level is chosen for the objective of improving reading fluency.) It provides a
standardized, "unit interval" score of comprehension, but does not provide a
grade equivalent. Before using the Lexile Framework to guide reading selections, however,
the student should have adequate decoding skills. Students with decoding problems should
work with "decodable texts" to improve accuracy and fluency, before switching to
the Lexile Framework to guide reading. Decodable text is defined as texts in which 95% of
the words can be decoded using the decoding skills that have been taught.
For high school and college students, the Nelson-Denny Reading
Test provides measures of vocabulary, reading comprehension, and reading rate. The
reading passages on this test are from actual high school and college textbooks
(expository text). The Nelson-Denny provides a useful foretaste of the reading
requirements of community college, technical school, and university.
- Spelling:
Beginning spelling
requires awareness of each sound that is in a word (phonemic awareness) and awareness of a
basic way to represent each sound. The core symbols test checks the student’s
knowledge of basic ways to write the sounds of English. The student is asked, "What
would you write for the sound /p/?" The pseudowords test requires the student
to integrate phonemic awareness and core-symbol knowledge to write a readable
representation of nonsense words. ("How would you spell the silly word
‘blim’? They bought a new blim at K-Mart.") Transitional spelling requires
(in addition) understanding of meaning units (morphemes). (The final /t/ sound of bat
and walked are different.) Formal spelling requires memory of specific words
(orthographic knowledge). (See A Reading Crisis? for a brief description of the
phonologic and orthographic challenges in learning to read and write English.) The Test
of Written Spelling is a nationally standardized measure of formal spelling. The
student is asked, "Spell ‘hospital.’ They took the injured player to the
hospital.") Spelling is the single best measure of the depth of a student’s
knowledge of English orthography, the complicated sound-symbol system of our language. A
well-developed orthographic system is both the result of reading and what makes proficient
reading possible.
.
|
|
