2/7/19

1

Introducing the QUILS: Screening for Early Language Knowledge

Roberta Michnick Golinkoff

Unidel H. Rodney Sharp Professor

University of DelawareKathy Hirsh-Pasek, Mary Wilson

Aquiles Iglesias, Jill deVilliers

Thank you for having me!

• I have escaped

• I have had wonderful,

stimulating

• I got to play with my

grandkids!

Purpose: To acquaint you with QUILS –

in the context of language development

•Unusual step: Starting with a video about the finished project

•We’ll work our way back!

My plan: A talk in 2 parts

•Part I: Why care about language development?

•Part II: What matters for language development?

•Part III: The QUILS

•Part IV: The QUILS and its Baby!

2/7/19

2

The Quick Interactive Language

Screener (QUILS)

6

Part I: Why care about language

development?

•More than 80% of 3rd graders from low-income families

will not be reading at grade 3 in grade 3

• At least half of the school achievement gap between

rich and poor kids starts before kindergarten

The National Governor’s Association

recognizes that strong language skills are

critical if we are to build strong reading

skills!

The National Assessment of Educational Progress reports that ….

Scarborough shows us the many

strands

2/7/19

3

Most of our curricular attempts to remedy the problem focus on word recognition or “code” skills and on vocabulary drill

And they are critical for learning to read

And while important, these interventions alone are not really working

• Systematic review: 31 well-cited interventions revealed that

children generally learn less than 25% of taught words following interventions lasting 2 or more weeks (Wasik et al., 2016)

• Another meta-analysis examining curriculum interventions produced a small and nonsignificant effect size of 0.07 on

children’s vocabulary (Darrow, 2009)

• And, those programs that have targeted specific vocabulary lists generally find that children cannot generalize these words to new contexts (Kaiser et al., 2011 ).

WHAT CAN WE DO?

Maybe we need to spend more time supporting

broader language skillsThe scientific data show both direct and indirect

relationships between language and reading

(NICHD ECCRN, 2002; Dickinson & Tabors, 2001,Lee , 2011, Grissmer, 2011, Munson et al; 2004, 2005; Storkel, 2001, 2003; Whitehurst & Lonigan, 1998, 2001; Silven et al., 2007; Dickinson, Golinkoff & Hirsh-Pasek, 2013)

• Consider phonological awareness, “c-a-t” or “base ball”

• Rhyming, e.g., day

• Asking for definitions – What does this mean?

• Hearing stories to build knowledge of narrative structure

• Exposure to academic language

2/7/19

4

But this is just the tip of the

•Language helps us capture knowledge, acquire concepts. (Grissmer, Grimm, Aiyer, Murrah, & Steele,

2010; Gelman, 1999)

14

• Language helps us engage in self-regulation and executive function (Roben, Cole,

& Armstrong, 2013; Matte & Bernier, 2011; Mendelsohn, Cates et al., 2018)

• Matte & Bernier: Children’s language ability mediates between the ability to profit from parental autonomy support and development of executive function

Back to NAPE…

• NAEP scores – vocab predicts reading comprehension• Lang at ages 3,4 predicts reading ability in grades 3, 4, e.g.,

NICHD 2005• Lang: most important factor affecting reading comprehension in middle grades, e.g., Oakhill & Crane 2012

• Language skill predicts health care outcomes(Adler NE, Boyce WT, Chesney MA, Cohen S, Folkman S, et al., 1994)

15

Our new secondary analyses of the NICHD Child Care data set suggests…

•That language at school entry is the single best predictor of school

outcomes (reading, math, social skills, later language) in grades 1 and 3

• And of gains in outcomes scores from Grades 1 to 3; 3 to 5

Pace, Alper, Burchinal, Hirsh-Pasek & Golinkoff, (2018)

Mountains of data: children from under-resourced environments have poorer general language skills than their

more privileged peers

(see also Hoff, 2002, 2003, 2013; Rowe et al., 2013; 2017;

Golinkoff et al., 2017; Huttenlocher et al., 2010; Pancsofar &

Vernon-Feagans, 2010, but see Sperry et al. 2018; Golinkoff

et al., 2018)

Could this fact be a key

component of the stable

achievement gap we have in our

country between rich and poor?

2/7/19

5

Absolutely. And when does the foundation

for the achievement gap start?

Hurt & Betancourt (2017):

Not news: Poverty has a deleterious outcome on development.

They tested:

Low SES: < or = to $23,550 for family of four n = 30

High SES: > than poverty line; both parents H.S. n = 30

Tested kids with Bayley Infant IQ test

and Preschool Language Scales

•What did they find?

“By age 1 year, low socioeconomic status (SES) infants perform less well on cognitive and language evaluations than higher SES infants.

These findings suggest that support for families and children from impoverished circumstances cannot begin too early.”

What happens to children whose language is already behind at 12 months?

• Referred to popularly as the “30-million-word gap,” between rich and poor, it has far reaching consequences:

At age 3, children’s vocabulary size mirrors the language directed to them

Professional - 1,116Working Class - 749

Welfare - 525

Part II: What matters for language

development?

Quantity? Quality? Both?

2/7/19

6

But is it all about number of words that pass children’s ears?

Given the latest interventions, you might think so…

But Hart and Risley (1995) also said that is it not just quantity! Parent talk that is contingent and responsive to child talk is more important than just number of words – quality counts too! (TamisLeMonda et al., 2014; Dunham & Dunham, 1996; Hart & Risley; Mol & Newuman, 2014).

Sometimes that message gets lost in popular translation

We asked two questions:

1. Do low-income children who are successful language learners experience a higher quality of communication than their less able peers?

2. How important is the quantity of language children hear relative to the quality of the communication foundation?

(Hirsh-Pasek, Adamson, Bakeman, Owen, Golinkoff, Pace, Yust, & Summa, (2015). Psychological Science.

2-year olds 3-year olds

High language

scores

Mid language

scores

Struggling

language scores

N = 60 low-income children

Quality =

1) Symbol infused joint engagement (gesture and words)

2) Fluid and connected exchanges (verbal and non-verbal)

3) Playful routines and rituals

Quantity = number of mother’s words per minute

Examining the Quality and Quantity of Communication

during parent-child interaction

3040

10

6070

80

90

100

Language variance explained

QuantityWords per minute

(1%)

Quality + Quantity

(10%)

Quality

(16%)

1. Quantity of input (amount) and Quality of Foundation for Communication are both important for language growth but “communication foundation” matters

more.

2. In our study, it’s not about poverty.

3. Fluid and connected conversations – “Conversational duets” require serve and

return, and return and return and return. …it can’t be a solo performance.

4. It’s “filling the gap” + “building the foundation” – a new metaphor for intervention

Findings and Implications

See Cartmill et al. (2013) for related findings

2/7/19

7

Conversational duets – whether during play or reading storybooks -

-in which what the adult says and does is CONTINGENT on the child’s focus

Are the interactions that fuel language growth

And likely fuel attention and the ability to engage at school

PLAYFUL LEARNING LANDSCAPES

Trying to reduce the word gap with community projects where people

live that encourage talk between caregivers and children

Example 1: The Supermarket StudyRidge, Ilgaz, Weisberg, Hirsh-Pasek & Golinkoff (2015)

• Can the introduction of signs in a supermarket increase caregiver child language interactions?

• Signs up and signs down in middle and low income area supermarkets

• Results show a 33% increase in caregiver - child language when the signs were up in low income neighborhoods.

28

Example 2: Urban ThinkscapeHassinger-Das, Bustamante, Hirsh-Pasek & Golinkoff, (2018) Education Sciences,

Hassinger-Das, Bustamante et al. (in press) Current Perspectives

Funding by:

29

RESULTS (N= 280)

> Parent-child interaction compared to

control playgroundØAdult and child language use

ØIncrease in spatial/numeric language

from 2 to 36%

Transforming a bus stop into a playful learning space

4 activities to increase

activity and conversation

around:• Spatial learning

• Language narrative

• Executive Function

2/7/19

8

Example 3: Parkopolis

• A Human Sized Board Game designed to foster early mathematical skills and scientific reasoning. Compared observations to another STEM, rocket building, activity at the museum

Thanks to Fei Xu, Silvia Bunge and all of our mathematic colleagues!

30

RESULTS (N= 111)

ØParent-child interaction

compared to control ØAdult and child

language use and

question useØTargeted

spatial/numeric

language and fraction language

ØPhysical activity

This project is designed to use our science to

create more conversations through playful learning cities!

• With pilots now in Philadelphia, Seattle, Chicago, Tulsa and Johannesburg, South Africa

• We are testing a new kind of dissemination that can be used in public spaces and in ”trapped spaces” like waiting rooms, supermarkets, laundromats, etc. Places where people wait and where we might increase the contingent conversations in ways that reduce the achievement gap

• All through playful learning that speaks to how families use the 80% of children’s waking time

when they are not in school or care.

BUT WAIT! The existence of the word gap has been challenged! Sperry, Sperry, and Miller (2018) Child Development

Argument: Based on what counts

as input to the child

32

“When more expansive definitions of the verbal environment

[our itals] were employed …the evidence pointed in a

different direction. Not only did the Word Gap disappear,

but also some poor and working-class communities showed

an advantage in the number of words children heard, compared with middle-class communities. “

What is missing in the definition of input?

Overheard speech.

Yet the Sperry et al. paper was questioned in two commentaries, both to appear in Child Development:

Golinkoff, Hoff, Rowe, Tamis-LeMonda, & Hirsh-Pasek ; Purpura

• SSM claim that they replicated HR à Flawed argument.• No professional class or upper SES group included

• Overheard speech is not effective for learning language• Lab experiments by Fisher et al. under tightly controlled conditions (also

Akhtar); Weisleder & Fernald, Goldin-Meadow- overheard speech not effective

• Children learn most when language is addressed to them re: their interests (e.g., Dunham & Dunham, 1992)

• Much variability within SES classes

• True! But doesn’t deny on average, main effects between classes

33

2/7/19

9

Neurological evidence for effects of the

word gap is increasing

Romeo et al. (2018): Conversational turns with a parent, rather than

the sheer amount of language, predicted

greater activation in Broca’s area and children’s

language scores in diverse sample of 4-6 yr olds

Components of “dialogic reading” (Whitehurst et

al., 1988), e.g., open-ended questions in back

and forth conversations , were related to

differences in brain activation in Broca’s area

in age 4 low SES females.

34

Hutton et al. (2017):

Perry et al. (2018): Conversational turns in preschool

classroom, low SES 2-3 yr olds• Positively related to vocabulary size

• “It is not enough for adults to simply talk around children [overheard speech];

instead adults need to talk with children.”

35

Gilkerson et al. (2018): Conversational turns (LENA)

• 18 to 24 months predict 14-27% of variance in IQ, verbal

comprehension, rec & exp vocabulary 10 years later

• Adult word counts also correlated but much more weakly

especially after controlling for SES

And studies on the role of conversation in language

learning:

Cartmill et al. (2013): Quality of parental speech• Measured “referential transparency” in conversation 14-18 mos.

• Predicted child language 3 years later! And SES did not predict quality;

only quantity of parent language

Part III: The QUILS Why do we need good language screeners?

We’ve all heard stories of children

sitting in classrooms, not identified

until age of 4 or 5 as having language

issues!

Professional assessment takes time

and money

We needed a better way to screen

for potential language issues

2/7/19

10

Enter The QUILS: Quick Interactive

Language Screener™ (Brookes Publishing)

• For children 3 through 5

• Assesses language comprehension

• 15-20 minutes to give

• Touch-screen tablet or computer

gives test and records responses

• Culture and dialect neutral

• 45-item monolingual English version

• Spanish-English bilingual version (in

press)

Roberta Golinkoff, Jill de Villiers, Kathy Hirsh-Pasek, Aquiles Iglesias, Mary Sweig Wilson

The literature on language development

emphasizes that:

•Assessing vocabulary is not enough to assess language; syntax is critical too.

•We need to pay attention to whether the child can learn new words and structures,

not just assess what they have alreadylearned.

1 VOCABULARY:

2

3

What words do children already know?

SYNTAX:What do children know about how words go together in

sentences?

PROCESS:How good are children at learning new language items?

QUILS ™ examines three critical language areas

Not just what language children know

but how children learn language!

Existing tests…

• Require a skilled examiner – like a speech-language pathologist [read that as expensive].

• Often take a long time to administer.

• Are not based on the latest evidence of what children

know about language.

• Do not capitalize on technology.