Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant...
Transcript of Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant...
![Page 1: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/1.jpg)
40th Annual Boston University Conference on Language Development
Franzo Law II, Tristan Mahr, and Jan R. Edwards
Funded by: NIDCD grants RO1 DC02932, R01 DC012513,
and NSF grant BCS 0729140
Exploring effects of expressive vocabulary size and maternal education on lexical processing by preschoolers
using the visual world paradigm
![Page 2: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/2.jpg)
Spoken word recognition
– Distinguishing familiar words from words to be learned.
– Parsing and learning syntactic
structures. – Other aspects of learning.
Cup and saucer I eat cookies because I like them. Lions are bigger than dogs!
• To take advantage of learning opportunities, children need to recognize words efficiently.
BUCLD, 2015 Law, Mahr & Edwards, 1
![Page 3: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/3.jpg)
Studying spoken word recognition in young children
• Two images presented on screen:
• Target words presented: – See the dog! – Find the book!
• Eyetracker records where child looks over time.
Looking-While-Listening (LWL) paradigm
BUCLD, 2015 Law, Mahr & Edwards, 2
![Page 4: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/4.jpg)
Spoken word recognition in young children
• 2-year-olds with larger vocabularies process familiar words more efficiently.. (Fernald et al., 2006)
• Processing speed at age 2 predicts language and working memory scores at age 8. (Marchman
& Fernald, 2008)
• Children who hear more linguistic input process words more efficiently than children who receive less input. (Weisleder & Fernald, 2013)
• 2-year-olds from high-SES families process words more efficiently than children from low-SES families (Fernald et. al, 2013)
BUCLD, 2015 Law, Mahr & Edwards, 3
![Page 5: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/5.jpg)
Socioeconomic status and spoken word recognition
• Why are children from low-SES families slower and less accurate to recognize familiar words than children from high-SES families?
• Non-linguistic consequences of poverty (Noble et al., 2005, 2007) – Poorer attentional skills – Poorer executive function
BUCLD, 2015 Law, Mahr & Edwards, 4
![Page 6: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/6.jpg)
Socioeconomic status and spoken word recognition
• Why are children from low-SES families slower and less accurate to recognize familiar words than children from high-SES families?
• Linguistic consequences of poverty – Decreased linguistic input – Smaller vocabulary size – Non-mainstream dialect
BUCLD, 2015 Law, Mahr & Edwards, 5
![Page 7: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/7.jpg)
Dialect mismatch and academic achievement
• Dialect mismatch: – Home language (NMAE) ≠ School language (MAE) – High levels of non-mainstream dialect at kindergarten
entry è Lower literacy scores in first grade (Terry & Connor, 2012)
BUCLD, 2015 Law, Mahr, & Edwards, 6
![Page 8: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/8.jpg)
Dialect mismatch and spoken word recognition
• Adults – Less effect of semantic predictability
(Clopper, 2012) – Greater effect of noise (Adank et al., 2009)
• Children – 20-month-olds but not 25-month-olds
influenced by dialect differences (van Heugten et al., 2015)
BUCLD, 2015 Law, Mahr, & Edwards, 7
![Page 9: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/9.jpg)
Spoken word recognition in preschool children
• What are the contributions of vocabulary size and maternal education level to spoken word recognition of preschool children? (Law, Mahr, Schneeberg, & Edwards, in revision)
• Differences from previous research: – Children tested in their native dialect. – Individual rather than group differences.
BUCLD, 2015 Law, Mahr & Edwards, 8
![Page 10: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/10.jpg)
Participants
• 60 children, 28-64 months • Half spoke AAE and half spoke MAE • Groups matched by age and sex
BUCLD, 2015 Law, Mahr & Edwards, 9
![Page 11: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/11.jpg)
African American English vs.
Mainstream American English
• Phonological differences • Morpho-syntactic differences
Law, Mahr & Edwards, 10 BUCLD, 2015
![Page 12: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/12.jpg)
Procedure
• Visual world paradigm – Semantic, phonological, and
unrelated foils • Secondary questions
– How do children respond to semantic and phonological competitors?
– Is there an effect of vocabulary size or maternal education level on responses to lexical competitors?
!!
!!!
!
! !!
Semantic foil Target
Phonological foil Unrelated foil
BUCLD, 2015 Law, Mahr & Edwards, 11
![Page 13: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/13.jpg)
Stimuli • Stimuli chosen using age of acquisition norms.
– AOA between 38 and 57 months. • Pictures normed in two preschool classrooms.
– Preschool attended by children from high SES families – Head Start classroom
• Stimuli recorded in both Mainstream American English (MAE) and African American English (AAE).
BUCLD, 2015 Law, Mahr & Edwards, 12
![Page 14: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/14.jpg)
Stimulus dialect
• All children tested in their home dialect • Home dialect determined by a number of factors.
BUCLD, 2015 Law, Mahr, & Edwards, 13
![Page 15: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/15.jpg)
BUCLD, 2015 Law, Mahr & Edwards, 14
Child-level variables • Age • Vocabulary size (EVT-2) • SES: Maternal education level
– 24: high – 14: middle – 22: low
![Page 16: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/16.jpg)
Results
BUCLD, 2015 Law, Mahr & Edwards, 15
Time from target word onset (ms)
Look
s to
targ
et (p
erce
nt)
Target Semantic foil Phonological foil Unrelated foil
![Page 17: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/17.jpg)
Analytic strategy • Growth curve analysis (Barr, 2008, Mirman et
al., 2008, Mirman, 2014)
– Restrict analysis to a meaningful time window.
– Model how fixations to a target area of interest (AOI) change as a function of time.
– Include random effects for participant. – Transform to empirical log-odds so models
work. – Subject-level variables: age, vocabulary size, ,
maternal education level, – Condition: stimulus dialect
BUCLD, 2015 Law, Mahr & Edwards, 16
![Page 18: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/18.jpg)
Results: Stimulus dialect
• No main effect of dialect and dialect did not interact with any of the other predictors.
• Combined data across the two stimulus-dialect groups.
• Methodologically feasible to test children in their native dialect.
BUCLD, 2015 Law, Mahr, & Edwards, 17
![Page 19: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/19.jpg)
Results: Expressive vocabulary and maternal education level
• Expressive vocabulary size is significant predictor of both accuracy and speed.
• No significant main effect of maternal education level.
• Interaction between maternal education level and vocabulary size.
BUCLD, 2015 Law, Mahr & Edwards, 18
![Page 20: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/20.jpg)
Looks to semantic and unrelated foils
BUCLD, 2015 Law, Mahr, & Edwards, 19
• Compare looks to target for trials were children were looking at target onset to: • semantic foil • unrelated foil
• Children more distracted by semantic foil
• Significant effect of vocabulary size for both trial-types
• No significant effect of maternal education level
![Page 21: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/21.jpg)
Looks to phonological and unrelated foils
BUCLD, 2015 Law, Mahr, & Edwards, 20
• Compare looks to target for trials were children were looking at target onset to: • phonological foil • unrelated foil
• Children (slightly) more distracted by phonological foil
• Significant effect of vocabulary size for both trial-types
• No significant effect of maternal education level
![Page 22: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/22.jpg)
Discussion
• Methodologically feasible to test children in native dialect.
• .What about dialect mismatch? – Not directly addressed in this study
because all children received stimuli in their native dialect.
– Ongoing study with both MAE and AAE speakers.
BUCLD, 2015 Law, Mahr, & Edwards, 21
![Page 23: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/23.jpg)
Discussion
• Spoken word recognition in children – Preschool children, like adults, were
sensitive to phonological and semantic competitors
– Vocabulary size did not interact with inhibition of semantic/phonological competitors
– Results argue for a continuity between children and adults in spoken word recognition (Mayor & Plunkett, 2014)
BUCLD, 2015 Law, Mahr, & Edwards, 22
![Page 24: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/24.jpg)
Discussion
• Why do children from high-SES families process familiar words more quickly and accurately than children from low-SES families?
• Linguistic versus non-linguistic explanations. – Vocabulary size explained much of this effect
• Insulating effect of high maternal education level.
BUCLD, 2015 Law, Mahr & Edwards, 23
![Page 25: Learning to Talk | Learning to Talk - Exploring effects of … · 2017. 4. 17. · and NSF grant BCS 0729140 Exploring effects of expressive vocabulary size and maternal education](https://reader036.fdocuments.net/reader036/viewer/2022071523/613cd4924c23507cb635a1c7/html5/thumbnails/25.jpg)
Acknowledgments
• Multiple PI team: Mary E. Beckman and Ben Munson • Research team (at UW): Ruby Braxton, Nicole Breunig,
Michelle Erskine, Megan Flood, Allie Johnson, Kayla Kristensen, Amy Muczynoski, Michelle Minter, Alissa Schneeberg, Janet Schwartz, Tatiana Thonsevanh, and Nancy Wermuth
• Funding sources: NIH and NSF
• Participation of the children and cooperation from their parents For all of which, a heartfelt:
謝謝 thank you ευχαριστώ πολύ ありがとう
BUCLD, 2015 Law, Mahr & Edwards, 24