Radboud University Nijmegen

Report
Reading skills and
challenged phoneme perception
Cecile Kuijpers, Louis ten Bosch, Renske Schilte
Radboud Universiteit Nijmegen
Pedagogische Wetenschappen en Onderwijskunde
CLST/Dept Linguistics
1
Introduction
• Reading disability, or dyslexia, is the most common learning
disability
• Adult dyslexics
- read more slowly than non-dyslexics
- problems for nonsense word reading (non-lexical, phonological decoding)
- problems at spelling
• Dyslexia and IQ are not related
2
Introduction
• Dyslexia (broad term): a learning disability that impairs a person's
fluency in being able to read
- can manifest itself as a difficulty with phonological awareness, phonological
decoding, orthographic coding, auditory short-term memory, or rapid
naming.
• Dyslexia is separate and distinct from reading difficulties resulting
from other causes:
- non-neurological deficiency with vision or hearing
- poor or inadequate reading instruction
• Estimated 5 to 10 percent of the population
3
Aim of this study
• To investigate
whether difficulty to identify phonemes in speech is related to
difficulty to acquire reading skills
whether phonological representations are deficient
4
METHOD
• 114 students
• Grade 4 / 5 / 6 (9- to 12-y-old)
• 55 male, 59 female
• 100% letter knowledge
Test
EMT
Klepel
EMT
Klepel
EMT
Klepel
EMT
Klepel
EMT
Klepel
Mean SS
3,63
4,81
5,28
7,71
7,71
5,28
10,92
11,00
ss < 7
ss < 7
ss <7
ss ≥ 7
ss ≥ 7
ss <7
ss ≥ 7
ss ≥ 7
ss ≥12
ss ≥12
16
22
problem
60
normal
16
good
• School results and timed reading tests
• One-Minute-test (words), de Klepel (pseudowords)
• Severe reading problem, reading problem, normal readers,
good readers
5
severe
problem
Task
• Phoneme identification: two-alternatives forced choice
- combined with grapheme presentation
• 4 vowels (a,o,e,u), 16 consonants (p,t,k,b,d,f,s,v,z,w,r,l,j,w,m,n)
• 64 auditory stimuli VCV (e.g. /aba/, /utu/, /efe/, /ono/)
• Speech-shaped noise (Stuart Rosen): 3 conditions
a. no noise (>80 dB SNR)
b. noise (6 dB SNR)
c. noise (3 dB SNR)
• Trained speaker
• Close pair (e.g. b-p) or distant pair (e.g. b-r)
• Close: 1.04 artic. features vs. distant: 3.26 artic. features
6
experiment
Graphemes
(target - close/distant alternative)
empty screen
+ beep
+ silence
+ auditory stimulus
response
+ silence
total
Design (each subject)
• 6 blocks of 30 random stimuli (180 trials)
• equal number of left–right position target letter
• equal number of close and distant alternatives
• equal number of stimuli with noise level 0,1,2
• two training sessions (feedback correct/incorrect)
7
ms
1000
150
400
1000
2000
500
5050
Keyboard [c] [m]
f
8
s
f
9
s
f
10
s
k
11
j
k
12
j
k
13
j
NOISE
RT
RT
main effect p<.001
main effect p<.001
0.97-0.91-0.87
686-744-764
N. S
RT (ms)
accuracy (%)
Accuracy
ACCURACY
Alternative (distant-close)
Accuracy
ACCURACY
RT
main effect p<.001
7 2 5 - 739
N. S
RT (ms)
accuracy (%)
main effect p<.001
0. 9 4 - 0.89
RT
Interaction Noise * Alternative
ACCURACY
RT
RT (ms)
Accuracy (%)
Accuracy
Conclusions
• Noise hampers correct identification
- in noise slower and less accurate response, for all groups
- poor readers suffer most (accuracy)
• More errors in case of close alternatives as compared to distant
- increases with noise
- no difference between groups
• Children with (severe) reading problems are slower in their response
than normal and good (independent of noise level)
• Some weaknesses:
- 6-graders were primarily normal
17
ACCURACY
READING GROUP
RT
Accuracy
RT
n.s.
0.94-0.92-0.91-0.90
p<.01
822-814-689-692
N. S

similar documents