Predicting Literacy: Correlating Working Memory, Phonological Awareness,

1
Predicting Literacy Correlating Working
Memory, Phonological Awareness, Early Decoding.
Preliminary Data.

• Cynthia W. Core, PhD, CCC-SLP
• Florida Atlantic University
• Elena Zaretsky, PhD, CCC-SLP
• University of Massachusetts Amherst

2
Agenda

• Overview of specific skills required for literacy
• Current study
• Premises
• Research Questions
• Methodology
• Results
• Discussion
• Future direction

3
Specific Skills Required for Literacy
Phonological Awareness as a Predictor for Reading
Acquisition

• The relationship between phonological awareness
  (PA) and subsequent reading acquisition has been
  a focus of research since 1980.
• PA, or sound segmentation skills, refers to the
  ability of consciously thinking about spoken word
  in terms of its components syllables, onsets and
  rimes, and individual phonemes, as well as
  manipulating these elements in a range of tasks.
• Empirical evidence suggests that
• PA is a significant predictor of success in early
  reading (Blackman, 1991),
• there is a relationship between acquisition of
  alphabetical literacy and the development of PA
  (Catts, 1989).
• preschoolers who show good sound segmentation
  skills tend to become good readers,
• children who show poor skills in this area are
  likely to become less proficient readers (Goswami
  Bryant, 1990 Wimmer, Landerl, Linortner
  Hummer, 1992).

4
Learning to Read in Alphabetical System

• Learning to read and write in the alphabetic
  orthography requires young learners to understand
  how the components of spoken language are
  represented by the orthographic system in the
  language of exposure.
• PA by itself may not be enough to acquire
  spelling skills
• Requires alphabetic principle, i.e.,
  phoneme/grapheme conversion

5
Bidirectional Relationship Between PA and
Learning to Read

• Evidence from studies in the area of reading
  acquisition also suggests that the relationship
  between the sound segmentation skill and reading
  abilities in alphabetic orthography is
  bi-directional
• PA will be enhanced by the exposure to print
  while influencing reading proficiency.
• If formal reading instruction may be the
  triggering events in developing PA (Wesseling
  Reitsma, 1998), it is of interest to see the
  readiness level of children in the classroom
  before they start formal reading instruction.

6
Factors that Predict Reading Difficulties (Known
and Unknown)

• Already known factors
• PA (Share Stanovich, 1995)
• naming skills (Wolf, 1991 Swan Goswami,
  1997b).
• vocabulary (Gathercole, Willis, Emslie
  Baddeley, 1992).
• The question still remains as to what underlying
  processes allow some children to acquire such an
  important skill as PA with ease, and why some
  other children have difficulties even at the
  earliest stages of reading acquisition.

7
Possible Underlying Processes

• One possible explanation - good phonological
  representation of lexicon (vocabulary) children
  already have (Elbro, Nielsen, Peterson, 1994,
  Metsala, 1999).
• Children who have difficulties performing PA
  tasks may have less refined, imprecise system of
  phonological representations (Elbro, 1996
  Wesseling Retsma, 2001).
• Children who are poor readers may have somewhat
  weaker receptive vocabularies (Gathercole,
  Willis, Emslie Baddeley, 1992).

8
Can we explain the underlying processes?

• Correlation between vocabulary and phonological
  representations suggests a possible mechanism
  making this connection possible
• infants may process the initial phonological
  information encoded in the words as a whole
• segmentation of lexical items for accurate
  representation of words presented auditorily is a
  later developmental process (Fowler,1991)
• The quality of phonological representations may
  be directly related to how much phonological
  information one may hold to represent variety of
  lexical items.

9
Verbal Working Memory (VWM) as Possible
Explanation for Deficits

• One way to assess the relationship between early
  whole words representations and the more
  refined ability to segment words that appears
  later may be through the application of the
  Baddeleys model VWM.
• This model proposes that the phonological loop,
  also known as phonological memory (PM), is
  essential for establishing phonological
  representations that consequently will promote
  vocabulary acquisition.
• PM is assessed through the nonsense word
  repetition task (NWR).
• Baddeley (2005) argues for the validity of
  nonsense word repetition as it closely resembles
  the natural situation facing any language leaner.

10
VWM and Vocabulary Acquisition

• Research has demonstrated, that young children
  who have better ability to repeat nonsense words
  turn out to be better learners of new vocabulary
  and vice versa, thus establishing the connection
  between phonological representation and good
  vocabulary skills (Gathercole Baddeley, 1990b).
  
• The ability to repeat nonsense words turned out
  to be a distinguishing factor between the typical
  language developing population and subgroups of
  children with specific language impairment (SLI)
  (Mongomery, 1995 Gathercole Baddeley, 1990a,
  Conti-Ramsden, 2003).

11
Other Views on the Interaction Between
Vocabulary, PA and Ability to Repeat Nonwords

• Chiat (2006) argues that children come to the
  nonsense word repetition with already established
  underlying phonological representations.
• This interconnectivity has been established by
  previous research (Snowling, Chait, Hulme,
  1991).
• Metsala (1999) found that children with larger
  vocabularies were more likely to perform better
  on nonsense word repetition tasks.
• There is no clear consensus to what degree
  nonsense word repetition influences PA skills.
• Baddeley (2005) suggested that nonword repetition
  and PA skills are predictors for future reading
  performance
• Strongly argued in favor of PM rather than
  phonological processing abilities (Snowling et
  al.,1991), as a better specified hypothesis for
  the acquisition of phonological skills.

12
Other Factors that May Influence Establishment of
Phonological Representations WM Capacity

• Baddeleys model of VWM suggests that complex
  language processing (reading is considered to be
  such a process) not only relies on PM, but
  requires the support of capacity based central
  executive component (WM capacity).
• The central executive component allows one to
  store some information while integrating new
  information to achieve full comprehension of
  complex linguistic material
• larger capacity allows for better comprehension.
  
• WM capacity is assessed by a complex memory span
  (CMS) paradigm storing some information while
  being engaged in concurrent processing activity
  Competing Language Processing Task (CLPT).
• CLPT requires a subject to make a semantic
  judgment on a sequence of individual sentence and
  then recall the last word of each sentence within
  the sequence.

13
How WM Capacity May Relate to NWR, Language and
Reading?

• Gathercole (2006) suggested that NWR, as a
  measure of PM, may not be sufficient enough to
  account for future deficits in language abilities
  and may depend on WM capacity.
• There is evidence, that poor performance by
  children with SLI on NWR is related to
  significantly diminished WM capacity (Zaretsky,
  2003).
• Current research is actively studying
• the inter-relationship between the PM, WM
  capacity, and the ensuing performance on language
  tasks by different groups of children (Swanson,
  1993 Marton Schwartz, 2003).
• Relationship between NWR and measure of WM
  capacity in typically developing children as a
  possible predictor for PA and subsequent reading
  acquisition is an important area of research.

14
The Significance for Identifying the Interaction
Between WM capacity, PM, PA and Vocabulary for
Reading

• Based on the reviewed previous research, there is
  a need to investigate the linguistic and
  cognitive mechanisms that relate to the
  acquisition of the PA, (Gathercole Baddeley,
  1989), as well as the role of PM and WM capacity
  in the process of reading acquisition.
• This research was designed to answer specific
  questions that may predict success in reading
  acquisition based on the performance of typically
  developing kindergarteners on the tasks that
  measure PA, PM and WM capacity, vocabulary and
  early literacy skills.

15
Research Questions

• Our specific questions are as follows
• a) Is there a correlation between PM, WM capacity
  and vocabulary?
• b) Is there a correlation between PM, WM capacity
  and PA?
• c) Is there a correlation between PM, WM capacity
  and early literacy skills?
• d) To what degree do PM and/or WM capacity
  predict PA skills?
• e) To what degree do PM and/or WM capacity
  predict early reading (decoding) and spelling
  skills prior to the start of formal reading
  instructions?

16
Current Study

• Participants
• 21 kindergarten students between 51 and 61
  years of age (M55, SD.32)
• Typically developing language abilities
• All children passed school language screening
• Monolingual speakers of English

17
Methods Tests administered

• PPVT, receptive vocabulary
• The Phonological Awareness Test, phonological
  awareness
• Segmenting, blending, rhyming, substitution (with
  manipulatives)
• Competing Language Processing Test, working
  memory capacity
• 50 item Nonword Repetition Task, Phonological
  Memory
• Early Reading Screening Inventory
• Alphabet Knowledge (upper case and lower case
  recognition, production)
• Invented Spelling
• Basal Word Reading
• Decoding, CVC Words

18
Correlation of Phonological Memory, Working
Memory Capacity, and PPVT
r.695, plt.001
r.415, p.06
19
Correlation of PM and WM Capacity
r.598, p.004
20
Results Correlations of PM and PA

• Correlations of PAT Subtest Scores
  and NWR
• PAT Total Rhyme
• r.609, p.003
• PAT Rhyme Discrimination
• r.617, p.004
• PAT Phoneme Deletion
• r.471, p.03
• PAT Segmentation r.604, p.004
• PAT Blending Phonemes
• r.449, p.04
• PAT Substitution w/ Manipulatives
• r.456, p.03

r.52, p.01
21
Correlation of WM Capacity and PA

• Correlations of PAT Subtests with WM Capacity
• PAT Total Rhyme
• r.730, plt.001
• Rhyme Discrimination r.465, p.039
• Rhyme Production r.567, p.005
• Phoneme Segmentation r.668, p.001

r.52, p.01
22
Correlation of PM, WM Capacity and Early Literacy
Skills

• PM and WM Capacity did not correlate with early
  literacy skills
• PM did not correlate with any of the reading and
  spelling skills
• WM capacity did not correlate with any of the
  reading and spelling skills

23
Do PM and WM predict phonological awareness?

• Linear regression was conducted with PA Composite
  as dependent variable
• PPVT, WM Capacity and PM were independent
  variables
• R2.414, df 3, F4.005, p.025
• NWR accounted for 32.6 of the variances
• PPVT accounted for 38.1 of the variances
• WM Capacity accounted for 6.1 of the variances

24
Do PM and WM Capacity predict early literacy
skills?

• Our results showed no correlation between PM or
  WM and early literacy skills
• Alphabet recognition (Upper and lower case)
• Alphabet production (writing)
• Invented spelling
• Basal word reading
• Decoding

25
Discussion

• Only WM Capacity correlated with vocabulary
• PM (as measured by NWR) - did not
• Both PM and WM Capacity strongly correlated with
  Composite Phonological Awareness Scores
• Specific PAT subtests correlated with PM and WM
  Capacity
• rhyming, deletion, blending, segmentation,
  substitution with manipulatives

26
PM and WM Capacity relationship to early literacy
skills

• Neither PM nor WM Capacity correlated with early
  literacy skills
• Literacy involves skill of phoneme-grapheme
  conversion
• Phoneme-grapheme correspondences strengthen with
  exposure to reading and spelling instruction

27
Memory and Vocabulary Contributions to PA and
Early Literacy Skills

• PM, WM Capacity and PPVT accounted for variance
  in Composite PA Scores
• PM and PPVT contributed the most to the variance
  in Composite PA Scores
• 32.6 and 38.1 respectively
• WM Capacity added little to the variance in PA
  scores, 6
• WM Capacity and PM did not contribute to measures
  of early literacy skills

28
Conclusions

• PA skills strongly rely on PM and WM Capacity
• The PA subtests that most correlated with memory
  tasks were rhyme, phoneme deletion, phoneme
  blending, and phoneme substitution
• PA also strongly relies on vocabulary skills
• Although this was not our original research
  question, we found that PPVT scores correlated
  with PAT rhyming (r.634, p.002) and blending
  (r.470, p.03)
• The role of vocabulary in PA should be further
  investigated
• PM and WM Capacity were strongly correlated
• This supports the view that NWR requires more
  cognitive resources than previously thought
• Memory skills did not play a role in early
  literacy skills
• These children were not readers and lacked
  critical skill to convert phonemes to graphemes

29
Future Directions

• Follow children longitudinally to the end of
  kindergarten or beginning of first grade to see
  if Memory and PA abilities predicted success in
  acquisition of early literacy skills
• Investigate the role of vocabulary in
  phonological awareness skills
• Investigate memory and PA skills in early
  literacy skills of bilingual children and
  children with language impairment

30
Acknowledgements

• AD Henderson University School, Florida Atlantic
  University
• Pelham Elementary School, Pelham, MA
• Department of Communication Sciences and
  Disorders at FAU
• Department of Communication Disorders, UMass
  Amherst
• Research Assistants at FAU and UMass Amherst