Using the Glass Analysis to Improve Decoding, Spelling and Reading Comprehension

1
Using the Glass Analysis to Improve Decoding,
Spelling and Reading Comprehension

• Catherine G. Penney
• Memorial University of Newfoundland

2
Acknowledgements

• Research was supported by
• Janeway Foundation
• Canadian Language and Literacy Research Network
• Memorial University
• Co-investigators Students
• Barbara Hopkins (Glass Analysis, high school
  study)
• James Drover (Tommy Miller)
• Carrie Dyck (phonologist)

3
All About Me

• Cognitive Psychologist
• Did work on memory
• Experimental psychologist - theoretical
• No experience teaching in K to 12 system
• No education courses or internship except two
  short courses in assessment.
• Learned by reading literature working with
  students

4
Overview

• The Glass Analysis
• Description
• Problems with phonics why Glass Analysis is
  better
• Theory
• Evidence
• Repeated-measures design
• Case Study Tommy Miller
• Independent-groups design
• Practical Hints

5
The Glass Analysis

• Developed by Gerald Glass in 1970s
• Child reads spells pronounceable components of
  words
• Words taught in groups organized by components.
• E.g. teach nation, station, inflation,
  confederation

6
The Glass Analysis -2

• To teach excellent
• Components -ex, -ell, -ent
• Rex
• sex
• Tex
• Texas
• extra
• Mexico

7
The Glass Analysis -3

• To teach excellent
• Components -ex, -ell, -ent
• sell
• bell
• tell
• shell
• spell
• spelling
• telling
• yellow
• fellow

8
The Glass Analysis -4

• To teach excellent
• Components -ex, -ell, -ent
• bent
• cent
• sent
• scent - explain difference in meanings
• spent
• penthouse
• basement
• apartment
• fundamental
• excellent

9
The Glass Analysis - 5

• Child learns how letters represent sounds in
  context.
• Glass was ignored. Whole-language approach
  denigrated teaching of phonics.
• Recent return to phonics explicit teaching of
  letter-sound correspondences.

10
Glass Analysis is Better than Phonics

• Young children dyslexics lack phoneme awareness
  (PA)
• PA develops along with literacy, not before
• Glass Analysis does not require phoneme
  awareness need onset-rime awareness.

11
Glass Analysis is Better than Phonics - 2

• English spelling is irregular complex
• daughter, psychology, have, delicious, castle
• Need to look at larger units
• -aught has 2 pronunciations (laughter)
• Ps- is from the Greek pseudo, psalm
• -icious always has same pronunciation
• tio (ratio), tion, tia (inertia, in absentia)
• ture (nature), dure (endure)

12
Glass Analysis is Better than Phonics - 3

• Teaching letter-sound associations is artificial
  abstract, meaningless to many students e.g.
  marking long vowels
• Blending segmenting exercises meaningless to
  many dyslexics
• Understand how letters represent phonemes in
  words
• NOT k makes the kuh sound.

13
Glass Analysis is Better than Phonics - 4

• Spelling pronunciation more regular for rimes
  than individual phonemes (Treiman)
• break (long a) vs. peak (long e)
• sweat (short e) vs. meat (long e)
• -ought is consistent

14
Theory

• Syllable structure onsets and rimes
• Development of literacy and phonological awareness

15
Onsets, Rimes Phonemes - 2

• Phonics assumes all phonemes are equal
• But they are not
• S in sock is different from S in smile or in mess
  or horse or asset
• Syllable have hierarchical structure onsets and
  rimes
• Onset consonants before the vowel
• Rime vowel plus following consonants

16
Onsets, Rimes Phonemes

• Words composed of syllables young children can
  clap syllables

17
Onsets, Rimes Phonemes - 3
18
Syllables, Onsets and Rimes

• Treiman Zukowski
• Preschoolers recognize when words share syllables
  
• return receive have same beginning sound
• expect respect have same ending sound

19
Syllables, Onsets and Rimes - 2

• Treiman Zukowski
• Kindergartners can recognize when words share
  onsets or rimes (i.e. rhyme)
• take toe have same beginning sound
• smoke smile have same beginning
• clock cradle have different beginning X
• rink sink have same ending sound
• tank and park have different ending sound ?
• Cant break up onsets or rimes into phonemes

20
Syllables, Onsets and Rimes - 3

• Treiman Zukowski
• First-graders (who are learning to read) can
  recognize when words begin or end with the same
  phoneme.
• clock cradle have same beginning
• tank and park have same ending sound
• Dyslexics seem to take longer to develop reading
  and phoneme awareness

21
Syllables, Onsets and Rimes - 4

• Treiman Zukowski
• Phonological Development
• Syllable awareness first (preschool K)
• Awareness of onsets rimes (K Grade 1)
• Awareness of phonemes within onsets and rimes
• Grade 1-2 (normal children)
• Later as literacy develops (dyslexics)

22
Onsets and Letters

• First step in learning to read and spell is
  learning the association between letters and
  sounds for simple onsets. (Tommy Miller)
• Need to know letter names recognize when two
  words start with same sound
• B is for book, ball, Billy, Bob,

23
Onset Deletion Say make without the /m/ sound
24
Deletion from Onset ClusterSay smile without the
/s/ sound.
25
Evidence for Effectiveness of Glass Analysis

• Repeated-measures design
• Case Study Tommy Miller
• Independent-groups design

26
Evidence Experiment 1

• Participants 7 females and 9 males
• Grades 2.8 to 11.5 at first assessment, M 6.76
• mean reading grade equivalent of 2.99, range
  0.9 to 4.8.
• mean discrepancy between actual school grade and
  reading grade equivalent was 3.76 years, range
  1.2 to 7.7 years.

27
Experiment 1 contd

• Control condition first and then experimental
  condition.
• Procedure
• First assessment of reading spelling
• Control period approximately 8 months
• Second assessment
• Experimental period 30 hours of tutoring,
  approximately ten months on average
• Third assessment

28
Experiment 1 contd

• Prediction
• Greater improvement in word identification, word
  attack spelling between 2nd 3rd assessments
  (experimental period) than between 1st and 2nd
  assessments (control period)

29
Experiment 1 Results

• Mean Reading Raw Scores
• Assessment
• 1 2 3
• Woodcock Reading Mastery Test
• Word Identification 53.8 58.9 66.1
• Word Attack 15.0 16.9 23.5
• Passage Comprehension 31.8 33.9 40.3
• Decoding Skills Grade Equivalent
• (WI WA) 2.99 3.36 4.39
• A3 A2 gt A2 A1, p lt .05

30
Experiment 1 Results

• Mean Spelling Raw Scores
• Assessment
• 1 2 3
• Test of Written Spelling
• Predictable Words 13.6 15.2 20.6
• Unpredictable Words 10.1 12.0 13.9
• Wide Range Achievement Test
• 23.2 23.3 27.1
• A3 A2 gt A2 A1, p lt .05

31
Experiment 1 Conclusion

• Glass Analysis improved
• Reading comprehension
• Spelling
• Decoding
• Why did reading comprehension improve?
• Ss had higher Listening Comprehension than
  Reading Comprehension
• Poor decoding impaired reading comprehension
• Improve decoding ? improve reading comprehension

32
Evidence contdCase Study Tommy Miller

• At the end of first grade, Tommy
• Did not know all letter names
• Could not read any words
• Distractible, Inattentive
• Impulsive, Energetic
• Behaviour problems

33
Evidence Case Study Tommy Miller

• Phonological Problems
• Could not tap syllables
• Could not delete onsets
• Slow letter naming word-finding difficulties
• Could not discriminate many phonemes in nonsense
  words
• e.g. chisser vs. chizzer, himmer vs. hinger
• Correct on only 33 of vowel discriminations

34
Evidence Case Study Tommy Miller - 2

• Woodcock-Johnson Battery
• Low on 2 expressive vocabulary measures, average
  on Peabody Vocabulary (receptive)
• Low on Visual-auditory learning (symbol to sound
  associations)
• Average on Memory for Names (sound to picture
  associations)
• Low on Sound Blending
• Many measures within average range

35
Evidence Case Study Tommy Miller - 3

• Reading Development
• After 10 months of tutoring (21 lessons), could
  read I, a, the.
• Could select a written word to match a spoken
  word on basis of first letter sound
• Show dog, book, car ask which is book.

36
Evidence Case Study Tommy Miller - 4

• Reading Development
• After 19 months of tutoring (37 lessons), could
  select a written word to match a spoken word on
  basis of last letter sound (90 or more
  correct)
• Show book, ball, big, ask which is book.
• Month 20 could read 7 words

37
Evidence Case Study Tommy Miller - 5

• Reading Development
• Glass Analysis drills were done in all lessons
• After 37 lessons (19 months), TM was beginning to
  show some transfer from Glass drills to reading
  trained words in sentences
• correct ranged from 18.2 92.3 during months
  17 to 30

38
Evidence Case Study Tommy Miller - 6
39
Evidence Case Study Tommy Miller - 6

• Grade Equivalent Reading and Spelling Achievement
  after 124 lessons
• (Late 5th grade)
• Woodcock Reading Mastery Test
• Word Identification 3.5
• Word Attack 1.6
• Passage Comprehension 3.5
• Gray Oral Reading Test
• Rate 3.4
• Accuracy 5.9
• Passage Comprehension 4.4
• Test of Written Spelling 2.2

40
Experiment 2 High School Students

• Participants All students enrolled in Reading
  course, 3 schools were recruited.
• 46 permission forms signed ended with 33
  students
• 21 students tutored (9 males, 12 females) 12
  not tutored (2 females, 9 males)
• Tests Woodcock Word Identification, Word Attack,
  Passage Comprehension administered Dec/Jan and
  again April/May

41
High School Students - 2

• 15 18 sessions (school period)
• Tutors were teachers or teaching assistants hired
  by schools part-time or as substitutes
• Content taken from course material designed to
  interest students
• Included newspaper articles, leveled readers,
  provincial driver-training handbook

42
High School Students - 3

• Students read aloud from prescribed reading
  material
• Tutors noted reading errors and used Glass
  Analysis on these words
• Student reread passage.
• Outside tutoring sessions, students did
  assignments as set by course teacher

43
High School Students - 4

• Tutored Control
• Age 208 (15.5) 206 (13.8)
• PPVT stand. 74.7 (12.5) 83.9 (12.9)
• Ravens raw 35.3 (9.9) 38.8 (8.4)
• (mean 42)

44
High School Students - 5

• Raw Scores on Reading Measures
• Tutored Control
• Pre Post Pre Post
• Word Id. 65.3 75.9 69.9 74.0
• Word At. 22.3 28.0 26.2 27.3
• Pass Comp 35.7 43.3 41.3
  43.6
• Total Reading Grade Equivalent
• 4.5 7.2 5.6
  6.9
• P lt .05
• plt .001

45
High School Students - 6

• Summary
• Train decoding using Glass Analysis
• Get improvements in decoding but also reading
  comprehension
• All Ss in course designed to improve reading
  comprehension, but Ss given decoding training
  improved more on reading comprehension than did
  control Ss.
• No point giving comprehension training if problem
  lies with decoding.

46
Some Practical Hints

• Teach spelling reading together
• Practice printing or cursive writing as part of
  spelling drills. Insist on correct formation of
  letters
• Insist on correct pronunciation of words.
• Do not permit elision of phonemes, especially at
  end of word. Inflections must be pronounced
• Tinned milk, bringing
• Teach standard Canadian English pronunciations
• TM th pronounced as d or t
• Pronunciation guides spelling apricot ? ah pry
  cot (long I)

47
Some Practical Hints - 2

• Consonant clusters in onsets
• Problematic for dyslexics
• Insert r or l sounds when reading aloud, or omit
  consonant sounds
• Built clusters when doing Glass drills
• Sit, pit, spit
• Sit, lit, slit
• Pit, lit, plit
• Split
• Practice troublesome clusters with other rimes
• Include nonsense words (marked) if necessary

48
Some Practical Hints - 3

• Teach vocabulary as needed (e.g. homophones)
• Integrate reading, spelling (oral written),
  penmanship, vocabulary pronunciation when doing
  word drills.
• Can teach other language arts points when doing
  word drills.
• Dont interrupt oral reading. Supply unknown
  words.
• Second reading, focus on fluent reading, good
  expression.

49
Some Practical Hints - 4

• Dont interrupt oral reading. Supply unknown
  words.
• Second reading, focus on fluent reading, good
  expression.
• On 2nd or 3rd reading, focus switches to
  comprehension
• Ask comprehension questions after second or third
  reading. Get decoding first.
• Get decoding first, then focus on comprehension

50
Some Practical Hints - 5

• When oral reading fluency reading comprehension
  are good, focus more on spelling writing.
• Assign writing exercises (answer questions
  summarize chapter discuss themes, characters
  etc.)
• Correct sentence structure (syntax), punctuation,
  spelling, word usage etc.
• Have student rewrite

51
Some Practical Hints - 6

• Build consonant clusters
• Sit, pit, lit, spit, plit, split
• Always have a pronounceable word or nonword
• Insist on correct pronunciation
• Do not break up onsets or rimes into individual
  phonemes until the student talks about sounds of
  individual letters.

52
Some Practical Hints - 7

• Word study should be an essential part of
  language arts programs
• Pronunciation, spelling, part of speech,
  meanings, usage
• University students dont understand parts of
  speech, transitive intransitive verbs, count
  vs. mass nouns etc.
• Misuse words poor vocabulary knowledge
• Use words in inappropriate ways
• Dyslexics have a lot of language problems, not
  just reading and spelling

53

• Questions ?
• Comments ?