BOARD OF EDUCATION CENTER SCHOOL DISTRICT #58

8701 Holmes Road Kansas City, Missouri 64131

There will be a Work Session of the Board of Education of Center School District #58 in the Board Room at Boone Elementary School, 8817 Wornall, on Monday, October 9, 2017 at 7:00pm.

Dr. Sharon K. Nibbelink, Superintendent of Schools_____________________________________

I. CALL TO ORDER AND ROLL CALL

II. APPROVAL OF AGENDA

III. CONSENT ITEMS

a) Control Service Company Contract b) MOU – Universal Screening Project Contract

IV. WORK SESSION TOPICS

a) Early Childhood b) MSBA Conference

V. Closed session Motion to go into Closed Session per Revised Statutes of Missouri Section 610.021 (1) Legal, (3) Personnel, (9) Negotiations and (13) individually identifiable personnel records, performance ratings or records pertaining to employees or applicants for employment.

VI. OPEN SESSION

Motion to go into Open Session

VII. ADJOURNMENT

.

Memorandum of Understanding (MOU) Universal Screening Project

Center School District and University of Missouri

Agency Name: Center School District Address: 8701 Holmes Road, Kansas City, MO 64131 Phone: (816) 349-3300

I. PURPOSE

Researchers: Stephen Kilgus, Ph.D. and Katie Eklund, Ph.D. Titles: Assistant Professors of School Psychology Institution: University of Missouri Address: 16 Hill Hall, Columbia, MO 65211 Phone: 884-4584 and 884-9546

This agreement defines the framework for cooperation between Center School District, herein known as "the district" and Drs. Stephen Kilgus and Katie Eklund, herein known as "the researchers." A universal behavior screener is an important part of behavioral support in schools and service agencies nationwide. To meet that need, the district seeks to employ the use of the Social, Academic, & Emotional Behavior Risk Screener (SAEBRS) created by Dr. Stephen Kilgus (one of the aforementioned researchers) and colleagues.

II. PERIOD OF AGREEMENT The period of agreement shall extend from August I, 2017 to July 31, 2018 with an understanding of automatic renewal of this agreement unless the district violates the agreement or the developers provide 6-months' notice of discontinuation of the agreement. The district may choose to discontinue this agreement at any time. However, if they choose to do so, free access to the SAEBRS will be discontinued.

III. OBJECTIVES I. To provide an efficient, psychometrically valid, universal behavior screener (i.e., SAEBRS)

to be used in each school building for all students enrolled in the district. 2. To ensure that the electronic version of the SAEBRS will be relegated to the secure

University of Missouri (MU) data warehouse, with access to district employees and approved MU researchers only.

3. To ensure that the SAEBRS will remain the property of the researchers and will not be used in a manner inconsistent with the researcher's protocol.

IV. DESCRIPTION OF COMMITMENT I. The researchers will approve access to the SAEBRS for use by the district for all students

enrolled at no cost. 2. The district will not use the SAEBRS protocol for any purpose beyond the private use of staff

and students in the district. 3. The district and researchers will be jointly responsible for adequately training providers

assigned to use and score the SAEBRS. 4. The researchers will be responsible for providing data reports in a timely marmer for each

participating school, while also support school interpretation and use of the data.

5. The researchers will make the district aware of any changes in the SAEBRS protocol during the span of this agreement.

6. The district agrees to share any SAEBRS product created (i.e., Excel spreadsheet for individual classroom use, electronic format designed for districtwide use) to the extent possible with the developers during the span of this agreement.

7. The district will not share the SAEBRS with any outside vender or consultant without the prior written consent of the developers.

V. DATA SHARING The use of the SAEBRS in a large urban district provides an opportunity for research regarding implementation training and further psychometric validation of the SAEBRS, such as via examination of the extent to which SAEBRS scores predict student outcomes (e.g., suspensions, office referrals, attendance, test scores, grades). In return for free access to the SAEBRS, the district agrees to provide the researchers with access to de-identified student data. The following data for the 2017-18 school year will be provided for each screened student:

I. SAEBRS scores (specific to each time the screener was administered) 2. Ethnicity 3. Gender 4. Home language 5. Office discipline referrals (including suspension/expulsion data) 6. Number of absences 7. Curriculum·based measurement (CBM) scores 8. MAP test scores 9. Classroom grades

It is expected that the researchers will use data listed above as part of their research. This research may then be disseminated via a number of outlets, including (but not limited to) presentations at professional conferences and peer-reviewed publications. When the data are presented to any external audience, no information will be included that might permit identification of the district, or any school, teacher, or student.

VI. CONFIDENTIALITY The researchers agree to establish appropriate administrative, technical, and physical safeguards to protect the confidentiality of student, building, and district data. The district agrees to establish similar safeguards to prevent unauthorized use or access to the SAEBRS. All data generated through the use of the SAEBRS will remain the exclusive property of the district.

VII. SIGNATURES In witness whereof, the Agencies' authorized representatives as designated by the district and the researchers attest to and execute this agreement effective with this signing for the period set forth in Article II.

District Representative Date

Researchers Date

Center School BoardWork Session

Early ChildhoodOctober 9, 2017

Our Early Childhood Program

Quick Facts:

• 180 children enrolled (approximately 90 in the morning and 90 in the afternoon):

• 123 Head Start students

• 4 Title I students

• 43 district funded students

• 8 Early Childhood Special Education students

• $8656 – cost per student

Current Funding Sources:

• Early Childhood Special Education $679,080

• Title I $115,170

• District $138,867

• Head Start $625,000

• Total $1,558,117

Facility:

• 8 classrooms with restrooms

• 1 therapy room and 1 therapy office

• Main office – principal and three family advocate offices, conference room, staff workroom

• Multipurpose room

• Access to Success by Six for Early Childhood library

• Access to PAT work room for upstairs quiet area

• 1 social worker office

• 1 downstairs quiet room

• Misc. storage

• Breakfast, lunch and snack are delivered by the custodian from the Boone kitchen

• Playground

Head Start Requirements For 2021

Head Start Grant Application Submitted September 15, 2017

• $682,650.00 for 123 slots beginning November 1, 2018.

• This is an increase from the current amount we get for Head Start - Center must provide an education coordinator who will provide ongoing coaching to the classroom teachers.

• Head Start requires that all programs provide 1020 hours of planned classroom operations over the course of at least eight months per year.

• Center will need to implement at least a six hour a day, five days a week program no later than August of 2021

Strategic Plan

Strategy 1:

We believe that academic achievement is at the center of what we do. Academic achievement is the umbrella under which our goals, action steps, and plans are derived. The future of Center School District depends on our ability to increase academic achievement. Here are ways in which we can do so:

Priority 4:Provide high quality preschool for every child in our community. Increase number of students in the early childhood program and Parents as Teachers program.

Research

• Positive impact on future earnings, health, IQ and crime reduction

http://www.npr.org/sections/ed/2017/04/25/525594764/the-research-argument-for-nycs-preschool-plan-for-3-year-olds http://www.npr.org/sections/ed/2016/11/17/502299963/a-lesson-for-preschools-when-its-done-right-the-benefits-

• Full-day preschoolers had 45 percent fewer chronic absences than half-day preschoolers

• More than 80 percent of full-day preschoolers met national norms for total school readiness compared to 59 percent of half-day preschoolers

http://www.huffingtonpost.com/2014/11/25/benefits-full-day-preschool_n_6221782.html#

• More hours of instruction in the pre-Kindergarten years means less special education along the line.

http://www.nhsa.org/files/resources/june2015researchblast_0.pdf

• “Clearly, the amount of time needs to be accompanied by quality, as defined by good teacher and intentional design, such as curricula and other pedagogical approaches designed to promote learning.”

https://www.newamerica.org/education-policy/policy-papers/making-the-hours-count/

Potential Funding Sources

• The full funding of the formula triggers the beginning of foundation funding for early childhood by the state. (uncertain of impact at this time)

• Missouri Preschool Program may be possible for start-up or expansion only (not ongoing)

• Grant opportunities under consideration

Questions for Consideration

• What is best for our children? Do we want a universal program? A full day program?

• Do we want to continue to be a Head Start Program, given funding and expectations?

• If we convert Head Start and Sped programs to full day, will we do so for all?

• How are Success by 6 and Parents as Teachers impacted? How are local daycares impacted? Are our children receiving high quality learning opportunities in daycare settings?

• What does our community want? What do our ECC parents want?

• If we go to full day, how does that decision impact facility needs?

• What other information do we need to collect to inform these decisions?

24FEB 2017 BLOGPOST

What's the Research Behind Head Start's New Rules?

It's been 40 years since the program performance standards were substantially revised for Head Start, the

national program that has provided education and health support to millions of economically disadvantaged

young children and their parents. Change came in September 2016, when a new set of program rules was

approved and released.

These new rules will likely increase the cost of Head Start by $1 billion or about 12 percent, funds that

Congress has not yet appropriated. To inform the ongoing discussion about this additional investment, here

is some of the research behind the 2016 rule changes:

1.1ncreased Duration and Exposure. The new rules require that Head Start programs expand children's

exposure to the program from a minimum of 3.5 hours a day for 128 days to 6 hours a day for 180 days. For

programs that currently provide the minimum number of hours, this represents an increase of 140 percent

in expected exposure to Head Start education services for the children they serve.

AIR research shows that these increases in Head Start duration and exposure are consistent with state and

local preschool practices across the U.S. For example, Susan Muenchow and Emily Weinberg

</ resou rce/ten-questions-1 ocal-policymakers-should-ask -about -expanding-access-preschool> found that loca I

and state preschool programs they studied are open 17S-190 days per year.

Research from the U.S. Department of Education's Early Childhood Longitudinal Studies </resource/early­

childhood-longitudinal-studies-eels> by AIR's Jill Walston and Kristin Flanagan suggests that expectations

about what children should know and be able to do in kindergarten have drastically changed over a 12-year

span. Their research suggests that investments to get children "kindergarten-ready" are increasingly valued

and expected, especially for economically disadvantaged children such as those eligible for Head Start.

2. Expanded Access. Head Start programs must now give

priority to 3-year olds in the increasing number of communities

that offer full-day public preschool for 4-year-olds.

Recent research <http://nieer.org/wp-

I \ " "~

, li1faa8 Stalf~ Sl~i(i}\tl~ Slaets ' - '

Total Appropriation: $8,285,544,370

Total Children Enrolled: 944,581

Child Age: 44% 4-years; 38% 3-years

Child Race: 43% White; 29% Black; 28% multi-racial, American Indian, Asian or other

Child Ethniclty: Non-Hispanic/Non­Latlno Origin, 62%; 38% Hispanic or Latino

Teachers with a BA degree: 74%

The Revised Rules: As currently written, when the program is fuRy implemented, Head Start will cost an additional $1 biltion a year, up from approximately $8 biUion. That doesn't include increases in teacher salaries (currently an average $18.70 per hour for lead teacher and $11.99 for assistant teachers).

content/uploads/2016/08/1520essential20elements20scan_O.pdf> finds ever increasing rates of public

preschool attendance by the nation's four-year olds (up to 40 percent as of July 2016}. This creates

opportunities for Head Start programs to serve more 3-year olds with high-quality early care and

education. Although rigorous research is limited, well-designed observational studies <https://www.fcd­

us.org/assets/2013/10/evidence20base20on20preschool20education20final.pdf> suggest that two years of

preschool starting at age 31ead to better developmental outcomes than one year, especially for

disadvantaged children.

And, according to my 2015 white paper, What Matters Most for Children: Influencing Inequality at the Start of

Life </resource/what-matters-most-children-influencing-inequality-start-life>, there is a wealth of research

evidence to support education services for economically disadvantaged children even before age 3. Waiting

until preschool to intervene with these children may be too late to close known gaps in their development

and eventual school readiness.

3. Special Supports. Under the new rules, homeless children, foster children and English-language learners

will have more qualified staff and research-based approaches to support their learning. These children

represent a substantial share of Head Start's target population. For example, AIR's National Center on

Family Homelessness </center/national-center-family-homelessness> reports that a historic high 2.5 million

children are homeless each year (one in every 30 children in the United States). Fifty-one percent of these

children are under age 6.

Young children and their parents who are homeless-or children removed from their primary

caregivers-often experience toxic stress </event/toxic-stress-how-economic-inequality-hurts-early­

childhood-development>, which may interfere with their development and their ability to fully benefit from

programs like Head Start, as evidenced in a 2014 study

<http://www.srcd.org/sites/default/files/documents/spr_28_1.pdf> by Mary Dozier at the University of

Delaware. The new rules intend to provide Head Start program staff with the resources to better serve

these children and address their challenges.

4. Quality Teachers. The new rules encourage Head Start programs to replace intermittent teacher

workshops and conferences with regular on-site teacher coaching to support individualized professional

development and quality instruction. An AIR study on coaching< http://www.ad.hhs.gov/opre/resource/the­

descriptive-study-of-heads-start-early-learning-mentor-coach-grant-initiative-final-report-and-appendices> in

Head Start programs was one of several cited in the regulations to support the growing use of coaching for

teachers' professional development. This large, descriptive study found improvement over time in the

quality of Head Start teachers' instructional and behavioral management practices. Teachers also reported

benefiting from ongoing coaching.

Head Start's new rules align with the leading current research on child development and early education.

Ensuring the adequate funding of the implementation of these rules is likely to benefit the children and

families using the program, especially those who are most vulnerable.

Eboni Howard is a Managing Researcher at AIR with 25 years of experience in the areas of early childhood and

family support, as well as program administration and management. She has served as a senior advisor, principal

investigator, or project director for more than 40 projects and is nationally recognized for her expertise in early

childhood development and use of mixed-method evaluation approaches.

FURTHER READING

• What Matters Most for Children: Influencing Inequality at the Start of Life </resource/what-matters­

most-children-influencing-inequality-start-life>

• What to Look for in a High Quality Preschool </resource/what-look-high-quality-preschool>

• AIR Experts to Talk About School Climate, Head Start and Other Topics At Society for Research in Child

Development Biennial Meeting </news/press-release/air-experts-talk-about-school-climate-head-start­

and-other-topics-society>

• Big Changes Coming to Head Start Include More Intensive Coaching </resource/big-changes-coming­

head-start-include-more-intensive-coaching>

• The Descriptive Study of Head Start's Early Learning Mentor Coach Grant Initiative

</resource/descriptive-study-head-starts-early-learning-mentor-coach-grant-initiative>

SHARE print-friendly

CONTACT

<http:/ /www.ai r.org!person/ebon i-howard >

Eboni Howard</person/eboni-howard> Managing Researcher

POLICY CENTER

About the Center </page/about-us-education-policy-center>

Policy Center Blog </page/policy-center-blog>

Products and Publications </page/products-and-publications-policy-center>

lnfographics </page/infographics-policy-center>

Podcasts and Videos </page/podcasts-and-videos-policy-center>

Notice: Undefined variable: file in /srv/bindings/a46164c2dde14b6fa5B765e527e8a525/code/wp~contentlplugins/flx-my-feed~rss-repair/rss­

feed-fixr.php on line 14 Notice: Undefined variable: _SESSION in /srv/bindings/a46164c2dde14b6fa58765e527e8a525/code/wp­

contentlplugins/wp-ticker/wp-ticker.php on line 25 Notice: Undefined index: de install in

/srv/bindings/a46164c2dde14b6fa58765e527e8a525/code/wp-contentlplugins/wp-ticker/wp~ticker.php on tine 267 Notice: Undefined variable:

wpticversion in /srv/bindings/a46164c2dde14b6fa58765e527e8a525/code/wp-contentlplugins/wp-ticker/wp-ticker.php on line 191 Strict

Standards: call_user_func_array() expects parameter 1 to be a valid callback, non-static method GoogleSitemapGeneratorloader::Enable()

should not be catted statically in /srv/bindings/a46164c2dde14b6fa58765e527e8a525/code/wp-includes/plugin.php on line 496 Strict

Standards: call_user_func_array() expects parameter 1 to be a valid callback, non-static method ShareahoticUtilities::add_header_xuaf) should

not be called statically in /srv/bindings/a46164c2dde14b6fa58765e527e8a525/code/wp-includes/plugin.php on line 213 Notice: Undefined

index: custom_ styles in /srv/bindings/a46164c2dde14b6fa58765e527e8a525/code/wp-content/plugins/posts~in-sidebar/posts-in-sidebar.php

on line 1161

What's the Secret Sauce for an Effective Head Start Program?

Ill 4 mins read

November 20, 2014

Shutterstock

H ead Start, the nation's oldest and largest early

education program, has been under increased scrutiny

since the release of the Head Start Impact Study in 2010. The

rigorous evaluation and its follow-up study found that Head

Author Spotlight

Abbie Lieberman (/author/abbielieberman/]

Abbie Lieberman is a policy analyst with New America's Early & Elementary Education Policy team where she provides

Start children showed modest gains research and analysis on

(http:/ /earlyed.newamerica.net/blogposts/20 1 0/thoughts _on_ todays _release_ of_ th~~~{~~~;.\\ff~~f}~tudy-26270) in select school readiness indicators after one year of She recently graduated from

program participation, but the gains over their low-income Georgetown University's

peers had faded by the end of first and third grade. But the

Impact Study data were more nuanced than many reported:

Head Start's impact may be modest on average, but the data

revealed significant disparities within the program.

A recent report (http://www.nber.org/papers/w20639.pdf) by

UC Berkeley economics professor Christopher Walters uncovers

specific factors that might explain many of the disparities

between Head Start programs across the country. In the study,

Walters used Head Start Impact Study data to determine what

exactly is driving these differences in center effectiveness. He

analyzed how various inputs, child characteristics, and center

practices-over many of which center directors maintain

discretion-might explain why some children experience

sustained gains and others do not.

It turns out that program length of day and year and home

visiting are two of the strongest factors impacting center

effectiveness. Offering full-day services

(http ://newamerica.net/publications/policy /making_ the_ hours_ count)

had a particularly significant impact on children's cognitive

skills. Full-day pre-K programs offer more time for high-quality

interactions between adults and children than half-day

programs, and research

(http://www.newamerica.net/bloglearly-ed-

watch/2008/featured-a bstract-its-all-about-interactions-4097)

suggests that these interactions are essential to children's

social-emotional and academic development. In fact, NIEER's

randomized trial examining pre-K dosage

(http://www.nieer.org/resources/research/lsMoreBetter.pdf)

found that "the added hours of preschool education were

substantially effective at closing the achievement gap." Thus,

it's not surprising that children attending full-day Head Start

programs would experience greater cognitive gains than those

attending for only half the time.

The benefits of home visiting for families with young children

are also well documented

(http://zerotothree.org/zttjournal/new-research-strengthens­

home-visiting.pdf). Home visiting, which is rarely used in

private or state pre-K programs, is a key component of Head

McCourt School with a master's degree in public policy.

Popular Posts

What's the Secret Sauce for an Effective Head Start Program? by Abbie Lieberman

[/author/abbieliebermanl]

New America's Dual Language Learners National Work Group Sets Up Shop (/dllworkgrouplaunch/) by Con or P. Williams

(/author/conorwilliamsl]

Dual Language Learners in Sen. Alexander's No Child Left Behind Bill (falexanderdlls/) by Conor P. Williams

(/author/conorwilliams/]

Start's "whole child" approach. Home visiting fosters

relationships between families and teachers, engages parents

in their child's education, and gives teachers insight into

children's home lives. While all Head Start programs must

conduct home visits, Walters found that centers offering more

than three home visits per year were particularly effective at

improving skills such as making friends easily, enjoying

learning, being able to concentrate, and using self-control.

Unfortunately, only 20 percent of centers in the Head Start

Impact Study offered frequent home visits.

Other factors often thought to increase center effectiveness did

not have an impact, at least in Walters' analysis. Past research

has suggested that higher levels of teacher education, smaller

class sizes, more experienced center directors, and use of the

High/Scope curriculum are all associated with pre-K program

effectiveness. But none of these significantly impacted child

outcomes in Head Start according to Walters' analysis. The

2007 Head Start reauthorization intensified teacher education

requirements in an effort to improve teacher quality, but

Walters' findings suggest that more teacher education is not

the answer to improving the program. His findings also raise

questions about whether the well-respected High/Scope

curriculum, often believed to be "central to the success of the

Perry Preschool Project

(http://www.highscope .org/content.asp?contentid=219}," is as

important as researchers believe.

Walters also found that Head Start had the greatest impact on

children with less educated mothers, suggesting that

disadvantaged children benefit more from program

participation. And even though Walters examines a broad array

of inputs, practices, and characteristics of Head Start programs,

still only about a third of the variation in the programs can be

explained through his analysis. That means there's a lot more

research to be done on the other factors that Head Start centers

can affect-and maybe on the factors they can't change-to

help determine the most essential quality-improvement

activities for programs.

Policymakers have a vested interest in ensuring the federally

funded Head Start program appropriately serves low-income

children and their families. After all, the program's $8 billion

price tag and more than 900,000 children served annually are

not trivial. It's clear that some Head Start centers have figured

out how to provide quality services, and while there likely isn't

one guaranteed method that ensures center effectiveness, there

are certainly lessons to be learned from these high-quality

programs. As researchers continue to dive into the Head Start

Impact Study data to determine "what works," policymakers at

all levels, not to mention program directors, can and should

take these findings into account.

Early Ed (/category/prek·121eartyed/), Featured (/category/featured/), PreK-12

(/category/prek-121)

EdCentral

Early Ed (learlyedl]

PreK-12 [/prek121]

Higher Ed [/higheredl]

Learning Tech (/learningtech/1

Budget Project

[http:l/tebp.newamerica.orgl]

Edcyclopedia [Beta]

[http:l/www.edcyclopedia.org/)

About EdCentral

About Us [/about!]

Contributors (/contributors!)

Authors [/authors!]

Contact [/contact!)

Privacy Policy [/privacy/)

Where WeAre

1899 L Street, N.W., Suite 400 Washington, DC 20036

(/contact!)

@NewAmerlcaEd

Sorry, no Tweets were found.

Authors and Contributors

Aaron Loewen berg (/author/aloewenberg!J

Abbie Lieberman vauthortabbietieberman!J

Abigail Swisher [/author/ablgailswisher/1

Alexander Holt (/author/alexholVJ

Alison Friedman 1/author/alisontriedman/]

Amaya Garcia [/author/amaya-garcia/1

Amy Laitinen [/author/amylaitinen/l

Anna Duncan (/author/annaduncan/)

Anne Hyslop (/author/annehyslopl)

April Ege Vauthor/aprilege!J

Ashley lson (/author/ash\eyison/1

Ashley Simpson Baird

(/author/ashleysimpsonbaird/l

Barbara Ray (/author/barbararay/1

Ben Barrett (/author/benbarretVl

Ben Miller (/author/ben-miller /I

Cary Lou (/author/carylou/l

CJ Libassi (/author/cj·libassil]

Clare McCann (/author/claremccannl)

Conor P. Williams (/author/conorwllliams/)

Danielle Kehl (/author/daniel\ekehVl

Denzel Evans-Bell

(/author/denzelevansbe\VJ

Ebone Williams (/author/ebonewi\Uams/)

Is More Better? The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement

Executive Summary

Numerous studies demonstrate the positive

benefits high-quality preschool programs

can have on children's development.

Confident in this knowledge, policymakers

have expanded the availability of publicly

funded preschool education programs.

What is less well established, however, are

the benefits children derive from programs of various durations and intensity.

Preschool programs vary greatly-from less-than-half-day to full-day-plus

programs. Little rigorous research is available to inform policy decisions

about the relative benefits of programs with shorter and longer hours per

day or days per year. To address this need, NIEER conducted a randomized

trial in which 4-year-olds in a low-income urban district were randomly

assigned to programs of different durations. The programs were otherwise

quite similar: all had teachers with college degrees, a low ratio of children

to teachers, and used the same curriculum.

Our study compared 85 children assigned to an 8-hour program for 45

weeks to 254 children assigned to a 2.5- to 3-hour program for 41 weeks.

Due to the limited number of spaces available in the 8-hour extended year

program, a lottery was used to determine entry, providing the basis for

random assignment. Such studies are the "gold standard" approach to

addressing cause and effect questions in education research. The district's

residents were 50 percent Hispanic, 21 percent African-American, and more

than 20 percent of the families lived in poverty.

THE PEW CHARITABLE TRUSTS Mvancing Quality Pre-Kindecgarten for All

Executive Summary (continued)

NIEER's study was to designed to answer two primary questions:

1. What are the effects of the additional hours of preschool education offered by the extended-day, extended-year public preschool program on children's learning in literacy and mathematics by the end of the school year?

2. If learning gains from the program are evident for children by the end of one school year, are these gains sustained through kindergarten and beyond?

Results of this study indicate that even students who are far behind at entry to preschool can develop vocabulary, math, and literacy skills that approach national norms if provided with extended-duration preschool that maintains reasonable quality standards. By the Spring kindergarten assessment, children in the extended-duration program had improved 11 to 12 standard points on vocabulary and math skills. Children in half-day programs also improved, but to a lesser degree, 6 to 7 standard score points on vocabulary and math). Thus, the added hours of preschool education were substantially effective at closing the achievement gap between these urban children and their more advantaged peers.

Students in the extended program continued to outperform children in the control group in follow-up testing through the spring of first grade. Some of this lasting effect could be due the difference in duration in kindergarten. The treatment group continued to receive 8 hours a day for 45 weeks, and the control group received 6 hours a day for 41 weeks. However, this was a much smaller difference in hours than in preschool. It is noteworthy that by first grade effects of duration are apparent on more complex measures such as reading comprehension and calculation and not just on simple tasks like Jetter and number recognition.

While further research is needed to augment this study of half-day vs. extended-day preschool education, the results clearly indicate that duration matters. Extended-day preschool of good quality had dramatic and lasting effects on children's learning across a broad range of knowledge and skills. As many families need full-day programs for their 4-year-olds to accommodate parent work schedules, the evidence that full-day preschool education can meet child care needs and benefit children's learning should be of high interest to parents and policymakers. Indeed, some children, particularly those in low-income working families, will miss out on high quality preschool education altogether if only a half-day public program is available.

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement [2]

Introduction

Positive effects of high-quality early education have been found for cognitive, linguistic, social, and economic outcomes (Barnett, 200 I). This evidence has led to substantial public investment in preschool programs for economically disadvantaged children by federal, state, and local governments. An important question in the design of public preschool programs is whether learning increases as time in preschool is increased. Time in preschool can be increased by extending daily operating hours, adding days to the school year, or offering multiple years of service. Experimental data are not available to address the effects of any of these approaches. However, results from research conducted with model preschool programs such as the Abecedarian Project (Campbell & Ramey, 1994), the Chicago CPC program (Reynolds, 1993), and the High/Scope Perry Preschool Project (Schweinhart, Barnes, Weikart, Barnett, & Epstein, 1993) and numerous short-term studies provide some support for an "increased intensity and duration" hypothesis that longer lasting interventions are more effective for disadvantaged children (Frede, 1998).

Clear conclusions regarding the impact of time spent in early education on child development have been elusive, and findings are not entirely uniform. Though long hours in child care (especially during the early years) are sometimes associated with less positive social and emotional development, greater benefits are generally associated with participation in the most intensive, earliest starting and longest lasting programs (Barnett, 1998). Studies specifically designed to measure the effects of preschool duration have tended to focus on the number of years or age at start, rather than daily or yearly time in school (Reynolds, 1994; Sammons et al., 2004; Weikart, 1967) and have found better results for longer duration. One exception to this is the NICHD child care study which examined the effects of hours per week and number of years in non-maternal child care and linked increased time in non-maternal childcare arrangements with non­compliant and aggressive behaviors at 54 months of age and in kindergarten (NICHD Early Child Care Research Network, 200 I, 2003). The NICHD study considered total time spent in non-maternal care through the first 4.5 years of life, so infant and toddler care was included in analyses. The actual implication of these non-experimental findings to the question of preschool intensity and duration as an educational intervention is unclear.

Using a similar approach, the Effective Provision of Preschool Education (EPPE) Project (Sammons et al., 2004) studied a representative sample of nearly 3,000 children from different regions of England who attended targeted types of preschool care. The number of months a child attended was found to be statistically significant for both math and reading achievement at age 6, with longer preschool duration associated with better cognitive outcomes. Effects were particularly strong for the highest quality programs. The EPPE findings may underestimate the effects of longer duration in that nearly 25% of the sample spent time in a non-target preschool setting during the course of the study, and tinoe in these settings was not included in the data used for analyses.

Findings from the EPPE project also provide evidence that changes in preschool quantity and changes in preschool quality may offer different types of benefits (Sammons et al., 2003). While duration was significantly related to children's cognitive development, data did not reveal similar effects of duration on social-behavioral gains during the preschool year. Neither number of months in attendance nor age-at -start of preschool was significantly related to any of four social outcome measures but in contrast to the NICHD findings more time in preschool was also not associated with lower social outcomes. In fact, this study found quality to matter, with higher scores on the Social Interaction and Language and Reasoning subscales of the Early Childhood Environment Rating Scale-Revised (ECERS-R) (Harms, Clifford, & Cryer, 1998) related to positive gains on social-behavioral measures.

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement [3]

More research exists comparing the effects of full- and half-day kindergarten programs though much of it was completed over two decades ago and few studies used randomized trials. In addition, kindergarten outcomes may not generalize to younger ages. Findings indicate that participation in full-day kindergarten has positive effects on academic and social outcomes (Clark & Kirk, 2000; Gullo, 2000).

In reviews of the earlier literature, authors selected studies for review based on different criteria such as population studied or research design but all concluded that students who attended full-day kindergarten showed significantly greater achievement than half-day attendees (Fusaro, 1997; Karweit, 1987; Olsen &

Zigler, 1989; Puleo, 1988; and Stinard, 1982). Children in full-day programs scored higher on standard­ized tests, had fewer grade retentions, demonstrated more regular school attendance, and had fewer Chapter I placements than those who attended half or alternating days. Positive effects of increased kindergarten duration were generally sustained through the early elementary years.

Lee and colleagues (2006) found in a recent review that, although rare, studies that used random assignment and matched control group designs regularly found advantages for full-day kindergarten across varying groups of children.

Studies with the largest sample use data from the Early Childhood Longitudinal Study- Kindergarten cohort (ECLS-K), a nationally representative sample of over 17,000 first-time kindergarteners. Approximately 58% of the children who participated in the ECLS-K attended full-day programs. Among public school children, gains in reading and math achievement over the kindergarten year favor children who attended full-day classes with the effectiveness of full-day kindergarten roughly equal for children of different social backgrounds (Denton, West, & Walston, 2003; Lee, Burkam, Honigman, and Meisels, 2001; Lee, Burkam, Ready, Honigman, and Meisels, 2006).

Other reports based on ECLS-K data offer qualifications to these findings. One study including both public and private schools in the analyses failed to find that kindergarten achievement in reading and math differed by length of day (West, Denton, & Reaney, 2001). A more recent report shows no signifi­cant differences by length of program day in math, science or reading achievement at the end of third grade for public and private school children collectively (Rathbun, West, & Germino-Hausken, 2004), though small positive initial effects did exist. However, high attrition, including only children tested at grade level (i.e. those not retained in grade or placed in special education), and the exclusion of children who were ever tested in a language other than English raise serious concerns about drawing conclusions from these analyses.

Concerns have been raised about child stress, fatigue and irritability caused by increased kindergarten duration (Good, 1996), but research does not support the existence of these effects (da Costa & Bell, 2000; Elicker & Mathur, 1997). Hough and Bryde (1996) found no difference between fatigue levels of full- and half-day kindergarten attendees. Elicker and Mathur ( 1997) suggest that a full day in class may be less stressful for young children because it allows time for more developmentally appropriate instruction. Teachers spend more time in one-to-one instruction and less time engaged in large group activities, chil­dren take part in more self-directed learning, and parents report greater satisfaction. Research has found a deleterious effect of highly structured, skills-based practices in kindergarten which could be exacerbated if these same practices were just increased in longer day (Burts, Hart, Charlesworth & Kirk, 1990).

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement [4]

There is less research on length of the kindergarten year. Using a matched-pairs quasi-experimental design, Frazier & Morrison {1998) compared an extended-year (210 days) kindergarten program to a traditional program (180 days). No differences existed between groups at kindergarten entry. Test results at entrance to grade I favored children from the extended-year program in math, reading, and general knowledge. Benefits for extended-year students were not associated with program quality or teacher characteristics.

The study described here is a randomized trial in which outcome comparisons are made between children who attended an extended-day, extended-year public preschool program and peers who attended half-day public school programs. Although it is obvious that full-day, year round schedules are more accommodating of parental work schedules, it is less clear how much they might contribute to children's learning and development. Nevertheless, preschool policy makers across the nation face costly decisions about operating hours within the limitations of available resources. In 2002, 10 of 44 state preschool initiatives (23%) operated on a full-day schedule (Barnett, Hustedt, Robin, & Schulman, 2004). Many other states allowed local providers to determine the length of the operating day. Yearly schedule, or number of days offered per year, was also usually decided locally. Most state preschool providers followed the academic year and offered about 180 days of service, but in a few states such as Hawaii, Massachusetts, and Connecticut, preschool programs operated year-round.

Between 1997 and 2003, the percent of funded slots in federal Head Start reported to be full-day, 5 days per week nearly doubled, from 24% to 46% (U.S. Department of Health and Human Services, 1998, 2004a). The addition of nearly 250,000 full-time slots over 6 years exceeded the increase in total funded slots, and contributed to a substantial increase in cost per child. Adjusting for inflation, cost per child in Head Start increased by nearly 20% between fiscal years 1998 and 2003 (U.S. Department of Health and Human Services, 1999, 2004b). Evidence regarding the impacts of duration on the educational benefits of preschool would provide information about the value of Head Start's movement toward longer hours of service.

Three principle research questions are addressed by this study:

I. What are the effects of attending an extended-day, extended-year public school on children's learning in literacy and mathematics by the end of the preschool year?

2. If learning gains are evident for children in the extended duration school, are these gains sustained through kindergarten and beyond?

3. How is children's learning affected by family background characteristics?

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement [51

Research Design & Methods

Setting

This research was conducted in a low-income, urban school district in the northeastern United States with a total population of about 120,000 (U.S. Census 2000). The district's residents were 50o/o Hispanic or Latino, 21 o/o African-American, and over 20o/o of families with children less than 5 years of age lived in poverty. The educational attainment of adults age 25 and over was relatively low, with 38o/o not completing high school and only 12o/o holding a bachelor's degree. The median household income in 1999 was $35,175, compared to a national average of $41,994 and a state average of $55,146. The percentage of pupils eligible for free or reduced lunch was approximately 75o/o.

The district was one of then 28 urban, low-income districts involved in a long-standing school funding court case in New jersey, Abbott v. Burke. As part of the New jersey Supreme Court effort to ensure a "thorough and efficient" education as required by the state constitution, the Court ordered at least half­day preschool for 3 and 4 year olds in the 28 districts. The district studied here began a magnet program that offered an 8 hour, 10 month program for 4 year olds while providing only half-day, 9 month programs for other children.

Sample

Study participants were selected from the pool of preschoolers registered in a lottery to attend the extended­day, extended-year program for either the 1999-2000 or 2000-2001 school years. Preschool eligibility criteria required that a child reach the age of 4 on or before October 1st in order to enroll. Lottery registration was open to all parents or guardians of such children in the school district.

Those who were not offered enrollment through the lottery were offered enrollment in half-day classrooms in the district. A few parents elected either a private full-day child care program or some other non-center­based arrangement. All of these programs were supported by state dollars, and curriculum training and other oversight was provided by the school district.

All of the 1999 applicants to the lottery participated in this study. Of the 180 total children, 40 were admitted to the extended-duration program. The 1999 applicants are referred to in this study as Wave 1. In 2000, 318 children applied to the lottery and 45 were admitted. Due to a limited research budget, the first half of the lottery applicants were recruited for this study. Thus, Wave 2 consisted of 45 experi­mental and 114 control group children. The total sample invited to participate, across waves, was 339 (85 experimental+ 254 control).

Of the 217 control group children who completed baseline assessments, 186 (86o/o) enrolled in half-day programs in public schools. The remaining 31 received care at home ( 13), or at one of 16 available private child care programs (18). The private centers offered full-day, year-round programs.

The Effects of Fuii·Day vs. Half-Day Preschool on Early School Achievement [6]

Attrition

Baseline child assessments were completed in Fall 1999 and Fall 2000 with 297 of the 339 children invited to participate. Of the 42 who were not assessed, 7 were in the treatment group and 35 in the control group. Reasons for non-assessment included: parental refusal (25), child did not speak English or Spanish (9), assessors were unable to arrange a meeting with the parent and child (6), and child refusal to cooperate with the testing conditions (2). Three of the children tested at baseline were excluded from analyses after a review of their test protocols indicated that valid administration had not been possible. All three of these children were referred for evaluation for special education services. Analyses reported are thus based on an initial sample of 294 children, 77 of whom attended the treatment program. Parent interviews were completed at baseline for all but 12 of the children who were assessed. In 7 of these cases, parents were unreachable after repeated phone calls and at least two home visits. Five parents declined to be interviewed.

The majority of post baseline attrition was due to geographic movement, with children either moving out of state or "missing" from their reported address and school. Overall rate of attrition at Spring kindergarten testing was the same for both groups at 14%. By the Spring of first grade, 26 of the initial sample of 158 Wave I students had left the study (16%). No statistically significant characteristic differentiated children who left the study from those who remained involved.

Description of Treatment

In 1998, the district involved in this study began an extended-day, extended-year program for children in prekindergarten through grade 8 as a district-wide magnet school. This program was designed to be more attractive to working parents due to daily operating hours from 7:30am to 3:30pm, and an academic calendar beginning in late August and running through the end of June. The district's other preschool classrooms operated for 2.5 to 3 hours and followed a standard academic-year schedule, operating from early September through late June (4 weeks fewer than the extended program).

Children who were admitted to the treatment program were assigned to one of three classrooms. In 1999-2000, two classrooms were comprised of 16 4-year-olds, and one classroom offered a mixed-age program with eight 4-year-olds and eight 5-year-olds. Each room was staffed by a certified teacher and a teaching assistant. Class sizes in the public half-day programs varied, with an average of 18.5 during Year I (1999-2000) and 13.2 in Year 2. The reduction was due to the state requirement that preschool programs in Abbott districts maintain class sizes of 15 or fewer. Although class size for children in the control group was generally smaller for children in Wave 2, regression analyses did not reveal differential treatment impact by Wave. A certified teacher and a teaching assistant staffed all rooms in half-day programs. All preschool classrooms in the district, regardless of duration, used the High/Scope curriculum and all other services and supports were identical.

Classrooms were rated twice each year using the Early Childhood Environment Rating Scale - Revised (ECERS-R; Harms, Clifford and Cryer, 1998). The ECERS-R is a widely used measure of program practices in early childhood care and education. Classrooms are rated on seven subscale content areas by a trained observer, and each item is scored from I to 7, with I being inadequate and 7 being optimal. Analyses examined total ECERS scores and subscale scores. The mean total ECERS-R score for extended­day classrooms was 4.80 (N=6), while half-day classrooms averaged a nearly identical4.79 (N=62). ECERS-R scores were not obtained for private full-day programs, so data are unavailable for those classrooms.

The Effects of Fuii~Day vs. Half~Day Preschool on Early School Achievement [71

Data Collection

A letter explaining the nature of the research with an informed consent form was sent in early September to all families recruited for the study. Consenting parents completed a Parent Interview of Family Characteristics and Home Activities (Center for Early Education Research, 1999) by telephone each Fall. The interview gathers information regarding family background characteristics such as ethnicity, home language, maternal employment and income. Children were tested near the beginning and end of each preschool and kindergarten year, and at the end of first grade for Wave 1 only. Program quality in public school preschool classrooms was assessed through systematic observations in both the Fall and Spring.

Language of child assessment was either English or Spanish, based on the judgments of parents and teachers regarding which language was most often and most comfortably used by each child. Of the 297 baseline assessments, 31 o/o were conducted in Spanish. Over the course of the study, as English proficiency increased for many children, and they became accustomed to classroom activities being in English, language of testing was switched to maintain the goal of optimal performance. By the beginning of their kindergarten year, only 6% of children were assessed in Spanish.

Measures

Woodcock-Jo/mson Psycho-Educational Battery- Revised (WJ-R). Five subtests of the W)-R (Woodcock & Johnson, 1989) were used to assess cognitive ability and achievement: Picture Vocabulary, Letter-Word Identification, Passage Comprehension, Calculation, and Applied Problems. The W)-R provides a standardized measure of children's individual progress over time in basic skill areas widely endorsed by practitioners as critical for the young child's success in school. These include early literacy, language, and math skills and knowledge. The same subtests from the Woodcock-Mufioz Revised, the Spanish­language version of the W)-R, were used for Spanish-language assessments.

Peabody Picture Vocabulary Test, Third Edition (PPVT Ill). The PPVT-III (Dunn and Dunn, 1997) is a norm-referenced measure of listening comprehension and a quick test of verbal ability from age 2'/,. Reliability coefficients for the PPVT-III range from .85 (test-retest) to .97 (internal consistency), and concurrent validity has been established through correlations with general intelligence tests of between .4 and .8. Children assessed in Spanish were tested using the Test de Vocabulario en Imagenes Peabody (TVIP; Dunn, Padilla, Lugo, & Dunn, 1986), a Spanish version of the PPVT.

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement [8]

Data Analyses

Multiple methods were used to estimate the effects of duration on child outcomes. Simple !-tests and regression analyses were used to estimate effects at the end of preschool, the beginning and end of kinder­garten, and the end of grade I. Independent variables in the regression analyses include treatment, maternal employment and education, income, ethnicity, primary home language, and baseline test score. Growth curve modeling was used to estimate the pattern of effects over time by creating a continuous trajectory of change for both the treatment and comparison groups. The growth curve model allows for more precise estimates of impact (Love et a!., 2002), and provides for an exploration of both linear and quadratic effects. The Passage Comprehension and Calculation subtests proved difficult for most study participants in preschool and kindergarten. Prior to first grade, the vast majority of raw scores were either 0 or I, and valid standard scores could not be computed. Growth curve analyses were conducted on these two outcome measures using raw scores. In addition, all analyses were conducted with experimental vs. full control (all children from the waiting list) and experimental vs. half-day control (just children who attended the half-day program) to determine if the choice of a full-day, full-year child care program by a small subset of the parents biased the results.

Results

Baseline Comparison

Table I provides baseline data for child outcome measures. The simple comparison of pastiest means using at-test assumes that initial baseline test differences are due to differences existing prior to baseline assessment. The regression analyses control for differences in baseline test scores and selected demographic characteristics. These data were collected primarily during October and November of students' preschool year. Children in the extended-day program scored significantly higher than controls on the PPVT, as well as the Picture Vocabulary and Applied Problems subtests of the W)-R at baseline. The treatment and control groups also differed significantly on the average number of hours that mothers worked per week. Mothers with children in extended-day preschool spent an average of 39.9 hours per week at work, compared to 35.4 hours for mothers whose children attended half-day programs. Controlling for ethnicity and maternal education, treatment was a significant (p < .05) predictor of weekly maternal employment hours. The control group also contained a somewhat higher percentage of children identified by parents as Latino (78% compared to 66% for the experimental group), though this difference was not statistically significant. Group differences on the PPVT and Picture Vocabulary tests were significant whether comparing the treatment group to all control children (including those in private daycare), or to half-day students only. The baseline difference on Applied Problems was slightly reduced by restricting the comparison group to half-day attendees, but remained significant at p < .I 0.

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement [9]

Table 1 Sample Characteristics at Baseline

Characteristic

Children: (N = 294)

Average age

%female

Average PPVT standard score*

Average WJ-R standard scores

Picture-Vocabulary*

Letter-Word ldent.

Applied Problems*

(N = 273)

% Spanish-speaking

%Latino

% African American

% White non-Latino

% Other or Missing Race

Families· 1

Average # people in household (N = 272)

% married (N = 271)

% mothers working (N = 273)

Average # work hours/week* (N = 158)

% high school graduate (N = 250)

% income below $25,000 (N = 247)

* Groups differ significantly at p < .05

(N = 77)

Treatment

52 months

49%

85.4

83.0

93.1

89.4

33%

66%

20%

9%

5%

4.3

62%

49%

39.9

79%

52%

1 Sample size reflects # of respondents to specific survey questions

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement

(N = 217)

Control

52 months

51%

81.1

79.1

91.5

84.6

42%

78%

12%

5%

5%

4.3

64%

61%

35.4

73%

53%

[10]

Regression analyses were conducted to examine the degree to which differences in test scores at baseline could be explained by incidental differences between groups in background characteristics such as ethnicity, language spoken at home, mother's employment status, and mother's education. Baseline test score differences were reduced somewhat by controlling for these family background characteristics. Scores no longer differed significantly on any test at p < .05, though the treatment group remained favored by about 3 raw score points on the PPVT, and I point on Applied Problems. Without considering background characteristics, these differences were about 5 and I.4 points, respectively.

Some children in both the treatment and control groups had attended a preschool program for over 2 months before baseline testing was conducted. A reasonable explanation for baseline differences is that initial exposure to treatment improved the test performance of children in the experimental group. However, regression analyses failed to find that the number of days between program inception and base­line testing had a statistically significant effect on baseline scores. There is some evidence that children who were least proficient on the measures may have benefited most from the first few weeks of extended­day preschool duration. When raw scores of less than 2 were removed from analyses ofWJ-R baseline scores, initial group differences were reduced. A larger percentage of the control group scored 0 or I on all three valid W)-R subtests at baseline compared to children in the treatment group. Analyses were conducted omitting children with these very low scores in order to evaluate the extent to which this difference in baseline scores affected the results. As discussed below, the treatment effects over time remained essentially unchanged when the lowest W)-R scores were omitted from the analyses, indicating that estimated effects are not explained by the larger percentage of control children who tested poorly at baseline.

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement [11 I

Table 2 Mean Raw Child Test Scores by Group and Time

Control

PPVT Treatment

Difference

Picture Control

Vocabulary Treatment

Difference

Control Letter-Word Identification

Treatment

Difference

Control Passage

Treatment Comprehension

Difference

Calculation Control

Treatment .03

Difference

Control Applied

Treatment Problems

Difference

p < .05

PK-Fall (N=216, 77)

27.85

32.70

4.85*

16.53

17.49

.96*

4.84

5.22

.38

.71

.97

.26

.05

.33

-.02

5.99

7.36

1.37*

PK-Spring K-Fall (N=203, 73) (N=197, 71)

40.77 49.50

49.51 58.52

8.74* 9.02*

18.50 19.93

19.58 20.94

1.08* 1.01 *

7.23 8.32

8.89 10.31

1.66* 1.99*

1.02 1.46

1.27 1.70

.25 .24

.20 .50

.62 2.65

.13 .12

10.09 12.38

11.48 14.03

1.39* 1.65*

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement

K-Spring G1-Spring (N=186, 66) (N=100, 32)

62.79 79.89

74.00 93.13

11.21 * 13.24*

22.03 24.28

23.20 26.22

1.17* 1.94*

12.80 23.94

13.45 25.56

.65 1.62

1.66 8.85

2.17 11.84

.51 2.99*

1.90 8.02

9.72

.75* 1.70*

16.06 22.41

18.33 26.28

2.27* 3.87*

[12]

Comparison of Means

Table 2 presents mean raw test scores by group and time. Because there was no significant difference in age between groups, comparisons of mean test performance can be made using raw scores, with age entered as a factor. The treatment group scored better on all measures at all post-tests, with statistically significant differences on all measures at one or more post-tests. With the exception of Letter-Word Identification, differences between groups tended to grow over time. By the end of grade !, the Wave I treatment group scored more than 13 points higher than controls on the PPVT, compared to a baseline difference for Wave I of less than 5 points. The divergence between groups over time was similar on Applied Problems, with a grade I difference nearly 3 times the size found at baseline. The comparison of means shown in Table 2 was conducted using independent-sample ttests, without controlling for other variables. Adjusting for incidental group differences in family background characteristics diminishes baseline discrepancies, but does not eliminate them completely.

Growth Curve Analyses

Growth curve analyses were conducted to determine whether students in the experimental group experienced a significantly different trajectory of achievement over time on the PPVT and five WJ-R subtests. The data were centered by converting age to a deviation score around the total group mean, and then a regression equation was created to examine growth as it occurred by month. Both linear

and non-linear trends were examined.

As shown in Table 3, students who attended the treatment program showed a significantly higher rate of growth on four of six child outcome measures, and were favored on a fifth. Standard scores on the PPVT improved for all children at an average rate of .40 points per month. A significant time by treatment interaction (t*x) reveals that scores for students who received extended-day preschool education rose an additional .21 standard score points per month (p < .01). This indicates that the "value added" by the extended-day program was to increase growth rates by 53%. Scores on the Picture Vocabulary subtest measuring expressive verbal ability also favored the experimental group, with added growth of .16 standard points per month. A similar pattern was found for measures of mathematic abilities. Children in the treatment group demonstrated twice the growth rate of controls on the Applied Problems subtest of the W)-R (.35 additional standard points per month, (p < .01). No significant differences in trends were found for scores on the Letter/Word Identification W)-R sub test.

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement 113]

Table 3 Regression Estimates of Treatment Effect on Growth in Child Standard Scores

Test (Standard) OF t X

PPVT-111 638 .40 3.77

Picture Vocab. 637 .30 2.44

UW Identification 637 .24 1.62

Applied Problems 637 .35 5.64

Test (Raw)

Passage Comp. 637 .12 .00

Calculation 637 .16 .24

Dependent Variable: PPVT-111 and WJ-R subtest standard scores

t represents within subject standard score growth per month across treatment groups

x represents an estimate of baseline differences between groups

t*x

.21*

.15

.05

.35*

.06*

.05*

t*x represents a treatment effect in terms of additional growth per month for the experimental group

t*t*x represents an estimate of quadratic treatment effects on growth in standard scores per month

* p < .05

t*t*x

.00

.01

.00

.02*

.01*

.00

Analyses conducted on the Passage Comprehension and Calculation subtests are limited to Wave 1 because standard scores for younger children on these tests were not considered appropriate for analysis due to unreliable age-norms. Even the most basic skills required to provide accurate responses on these tests typically develop in kindergarten and beyond. A raw score of 0 corresponds to an age equivalent of 4 years 11 months on Calculation, and 5 years 6 months on Passage Comprehension. For the study sample, a restricted range of performance was observed for children prior to first grade, with over 90% of raw scores at kindergarten posttest falling between 0 and 4 for Passage Comprehension and between 0 and 5 for Calculation. By contrast, in Spring of grade 1, children had experienced considerable gains, with 95% of scores between 0 and 17 for Passage Comprehension, and 0 and 11 for Calculation. Growth analyses were conducted on raw scores to compare improvement between groups by the end of first grade. The growth curve model was appropriate for analyses of these scores due to the sensitivity of this approach to group differences regardless of the timing or nature of the difference. The extended-day program was found to significantly improve performance on both measures (p < .05), though the estimated effect was less than one added raw score point per year. In Spring of grade 1, students assigned to the control group achieved mean raw scores of 8.85 on Passage Comprehension and 8.02 on Calculation, as compared to 11.84 and 9.72, respectively, for the experimental group.

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement [14]

Regression Analyses Compariltg Treatment and Control Groups

Ordinary least squares regression was used to estimate the effect of the extended-duration program on test scores at the end of the preschool, kindergarten, and first grade years. In addition to the treatment variable, regression equations included child age and gender, mother's education, mother's employment status, a dummy variable for Hispanic race, and a dummy variable for African-American race. To control for baseline differences in test performance between groups, all regression equations included baseline scores on the dependent measure. For reasons discussed earlier, analyses were conducted for passage comprehension and calculation at the end of first grade only.

Table 4 Regression Results at Spring Preschool Testing

Total df

Adj. R'

Standardized regression coefficients

Baseline test score

Mother's education

Mother employed

Child male

Child Latino

Child African-American

Treatment/Control

* p < .05

- p < .10

PPVT

255

.647

.696*

.112*

-.005

.036

-.094•

-.009

.123*

Picture Vocab.

255

.387

.481*

.117*

.067

-.157*

-.128•

.039

.037

uw identification

255

.322

.482*

.103•

.033

-.055

-.112

.003

.159*

Applied Problems

254

.489

.654*

.098*

.094*

.017

-.013

.016

.078•

Results for regressions performed on Spring preschool test scores are presented in Table 4. The regression model is significant (p <.OJ) for all four outcome measures, explaining between one-third (for Letter-Word Identification) and two-thirds (for PPVT) of the variance of Spring preschool scores. Children who had attended extended-day preschool significantly outperformed children in the control group on two of the four measures analyzed (p < .05 ). Group differences were found for the PPVT and Letter-Word Identification.

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement [15)

Table 5 Regression Results at Spring Kindergarten Testing

Total df

Adj. R'

Standardized regression coefficients

Baseline test score

Mother's education

Mother employed

Child male

Child Latino

Child African-American

Treatment/Control

* p < .05 • p < .10

PPVT

232

.496

.532*

.054

.045

-.008

-.220*

-.128•

.235*

Picture Vocab.

229

.344

.434*

.119*

.008

-.237*

-.139•

.004

.087

uw Identification

229

.123

.320*

.094

-.004

.019

-.162•

-.119

.045

Applied Problems

228

.417

.564*

.045

.151*

.024

-.081

.013

.196*

Scores obtained at the beginning of the kindergarten year were almost identical to the results from Spring preschool assessments. Summer break did not seem to affect the advantages demonstrated by children

who had attended extended-day preschool. Regression results for child test scores at the end of the kindergarten year are shown in Table 5. The model continued to predict test scores well, with adjusted R' values above .33 for all analyses except Letter-Word Identification scores. Statistically significant effects

(p < .01) were found for PPVT and Applied Problems scores. Compared to the end of preschool results, regression coefficients rose from .16 to .24 on the PPVT, and from .09 to .20 on Applied Problems. No

effect on Letter-Word Identification scores was evident at Spring of the kindergarten year. However, a trend favored students in the extended-duration program on WJ-R Picture Vocabulary scores (p < .10).

The strength of the association between maternal education and child test scores appeared to weaken over

time, remaining significant at Spring kindergarten only for Picture Vocabulary scores. As found at the end of the preschool year, Spring kindergarten scores showed significantly greater achievement on Applied Problems for children with working mothers compared to students whose mothers did not work. Latino children scored significantly lower than peers on the PPVT, and may have performed slightly more poorly

on both Picture Vocabulary and Letter-Word Identification.

Two considerations distinguish analyses of treatment effect at Spring grade 1 from previous analyses.

These scores only include children in Wave 1 (N = 132), and analyses could be conducted on two additional measures, the WJ-R Passage Comprehension and Calculation subtests. Regression results for Spring grade 1 scores are presented in Table 6. Children attending extended-day preschool scored significantly

higher than the control group on five of the six measures analyzed (p < .05). Performance did not differ between groups on Letter-Word Identification.

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement [16]

With one exception, family background characteristics did not significantly impact Spring grade I test scores. Achievement of Latino children no longer differed significantly on any measure compared to peers. In addition, no significant results were found related to maternal education or employment. However, consistent with all previous assessments, girls performed significantly better than boys on Picture Vocabulary (p < .05).

Table 6 Regression Results at Spring Grade 1 Testing (Wave 1 Only)

PPVT

Total df 131

Adj. R' .446

Standardized regression coefficients

Baseline test score

Mother's education

Mother employed

Child male

Child Latino

Child African-American

Treatment/Control

* p < .05 ' p < .10

.553*

.103

-.060

-.070

-.089

-.054

.270*

Picture Vocab.

130

.261

.386*

.146"

.023

-.179*

-.108

-.009

.179*

uw I dent.

130

.080

.346*

.040

-.029

-.109

-.026

-.070

.041

Passage Comp.

130

.059

.093

.123

-.103

-.075

-.012

.091

.187*

Regression Analyses Comparing Treatment and Half-day Programs Only

Applied Calculation Problems

130 129

.086 .320

.218* .415*

.137 .052

-.132 -.130"

-.107 -.139"

.100 .049

.141 .052

.210* .343*

Because 31 children assigned to the control group either enrolled in full-day private childcare or stayed home at baseline assessment, regressions were conducted comparing the treatment group to children in half-day public programs only. These analyses were intended to focus group differences on duration alone, and to test the hypothesis that the treatment effect could in part be explained by poor test performance of children in private care. The estimated impact of treatment was generally the same as found in analyses that included the entire control group. The treatment group advantage at Spring grade I on Picture Vocabulary was no longer statistically significant, though scores for children in the treatment program were still higher (p < .10), and the effect size was essentially unchanged (.17 compared to .18). Scores on the Calculation subtest no longer significantly differed between groups at Spring grade I. These results seem likely to be due to reduced power with the smaller Wave I sample.

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement [17]

Discussion

Children who attended an extended-day, extended-year preschool program experienced greater improvement in test scores compared to peers who attended half-day programs. The difference in performance gains over time was evident for measures of both verbal and mathematic abilities. These results indicate that duration is an important consideration for the effectiveness of preschool education. Common sense and other research suggests that increased time in the classroom yields better results because it provides greater opportunity for teachers to work individually with students and allows for a more relaxing atmos­phere with less time proportionately spent on routines such as meals, tying shoelaces and hand washing.

The average classroom quality score on the ECERS-R was just under 5 which is higher than scores found for the rest of the Abbott preschool program during that time period (Barnett, Tarr, Esposito-Lamy, &

Frede, 2001) and higher that typical results from other programs (Espinosa, 2002), but roughly equal to scores of Head Start classrooms reported in the 2000 Family and Child Experiences Survey (U.S. Department of Health and Human Services, 2003 ). The magnitude of treatment effect found in this study may be dependent on program quality. Effects of increased duration could be even larger with higher quality programs. Conversely, the effect of increased duration could be reduced, eliminated, or even reversed if program quality is low. If disparities in child growth rates occurred independent of classroom quality; then benefits experienced by children in extended-day preschool were not caused by exposure to something generally better, but rather by increased exposure to an equally advantageous environment. However, different assessment of the classroom practices would be required to fully understand the extent to which the proportion of time spent in various activities may have differed with duration of day.

Results of this study indicate that students in a high-poverty district who are far behind at entry to preschool can develop vocabulary, math, and literacy skills that approach national norms if provided with extended-duration preschool that maintains reasonable quality standards. At Spring kindergarten assessment, children in the extended-duration program had improved more than 10 standard points on two measures (from 85 to 96 on the PPVT, and from 89 to !OJ on Applied Problems), though they remained somewhat behind national norms on all measures except Applied Problems. Scores of children in control programs also improved, but not to the degree demonstrated by their treatment group peers. From baseline to Spring kindergarten, the average standard score for the control group rose from 81 to 87 on the PPVT, and from 85 to 92 on Applied Problems. Due to their accelerated rate of improvement, children in the treatment group were within .3 standard deviations of the national mean on two of four valid outcome measures and had achieved the national mean on a third. The control group remained approximately a full standard deviation behind the national mean on two measures and was not within .3 standard deviations on any measure.

The impact of family and child background characteristics on learning was weaker at the end of first grade than at earlier assessments. The apparent fade-out of the influence of such background characteristics as ethnicity and maternal education over the course of the study may be evidence that quality early education compensates somewhat for disadvantaged circumstances. If so, then preschool education is an effective tool for enhancing equality of opportunity as well as increasing achievement generally.

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement (18(

One interpretation of baseline discrepancies in test scores between groups is that pre-testing was not conducted early enough in the academic year to avoid early impacts of treatment. Scores for the experi­mental group may have been boosted during the first two months of attendance at the extended duration preschool. The existence of such an immediate impact could only be fully examined using a different methodology than was employed for this research. Another possibility is that the treatment group began with some advantages despite random assignment, and started the program with accelerated scores. There is no evidence that the randomization procedure was somehow faulty, but baseline differences could have occurred by chance. No difference at baseline was statistically significant at (p <.OJ) and given that these factors tend to fade in their power to predict child progress, confidence can be placed in the statistical procedures controlling for the initial differences.

The difference in duration of program continued into kindergarten. The treatment group continued to

receive 8 hours of service daily in kindergarten, while the control group moved to a 6-hour daily kinder­garten schedule. Given the research design of this study, it is not possible to analyze how much later gains by the treatment group were due to the 5-hour advantage in preschool duration or the 2-hour advantage in kindergarten day. A proportional allocation of the benefits would attribute the majority of the gains to the preschool program, and the magnitude of difference in gains between groups was larger at the end of preschool and beginning of kindergarten than at the end of kindergarten. However, this is another question that calls for further empirical research.

Although further research is needed to augment this single study of half-day vs. extended-day preschool, the results clearly indicate that duration and intensity matter. Extended-day preschool seems to have dramatic and lasting effects when it is high quality. All teachers in the study classrooms were certified, public school employees paid on union scale. A comprehensive curriculum was implemented with strong supervisory support offered to classroom staff. Classrooms were also well supplied, and both children and families received support services. Given the evident need of many families for full-day care for their 4 year olds and the evidence presented here that full-day preschool has important benefits for child learning, policy makers should strongly consider implementation of full-day preschool.

About the Authors

Kenneth B. Robin, Ph.D., is an Assistant Research Professor at the National Institute for Early Education Research. Ellen C. Frede, Ph.D., is a developmental psychologist specializing in early childhood education and co-director of the National Institute for Early Education Research. W. Steven Barnett, Ph.D., is a professor of education economics and director of the National Institute for Early Education Research.

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement [19]

References

Barnett, W. S. (1998). Long term effects on cognitive development and school success. In W. S. Barnett & S. S. Boocock (Eds.), Early care and education for children in poverty: Promises, programs, and long term results (pp. 11-44). Albany, NY: SUNY Press.

Barnett, W.S. (2001). Preschool education for economically disadvantaged children: Effects on reading achievement and related outcomes. InS. Neuman and D. Dickinson (Eds.), Handbook of Early Literacy Research (pp. 421-443). New York, NY: Guilford Press.

Barnett, W. S., Tarr, J.E., Esposito-Lamy, C., & Frede, E. C. (2001). Fragile lives, shattered dreams: A report on implementation of preschool education in New Jersey's Abbott districts. CEER report. New Brunswick, NJ: Rutgers University.

Barnett, W. $., Hustedt, J. T., Robin, K. B., & Schulman, K. L. (2004). The state of preschool: 2004 state preschool yearbook. New Brunswick, NJ: National Institute for Early Education Research.

Burts, D.C., Hart, C.H., Charlesworth, R., & Kirk, L. (1990). A comparison of frequencies of stress behaviors observed in kinder­garten children in classrooms with developmentally appropriate versus developmentally inappropriate instructional practices. Early Childhood Research Quarterly, 5, 407~423.

Campbell, F. A., & Ramey, C. T. ( 1994). Effects of early intervention on intellectual and academic achievement. Child Development, 65, 684~698.

Center for Early Education Research. (1999). Parent interview of family characteristics and home activities. New Brunswick, NJ: Rutgers University.

Clark, P., & Kirk, E. (2000). Review of recent research on ali-day kindergarten. Childhood Education, 76(4), 228-231.

daCosta, J. L., & Bell, S. (2000). Full day kindergarten at an inner city elementary school: perceived and actual effects. Paper present­ed at the annual meeting of the American Educational Research Association, New Orleans, LA.

Denton, K., West,}., & Walston, J, (2003). Reading- Young children's achievement and classroom experiences: Findings from the coll­dition of education 2003 (NCES 2003-070). Washington, DC: National Center for Education Statistics.

Dunn, L., & Dunn, L. (1997). Peabody Picture VOcabulary Test, Third Edition. Circle Pines, MN: American Guidance Service.

Dunn, L M., Padilla, E. R, Lugo, D. E., & Dunn, L. M. (1986). Test de Vocabulario en Imagenes Peabody. Circle Pines, MN: American Guidance Service.

Elicker,}., & Mathur, S. (1997). What do they do all day? Comprehensive evaluation of a full-day kindergarten. Early Childhood Research Quarterly, 12, 459~480.

Espinosa, L. (2002). High quality preschool: Why we need it and what it looks like. NIEER Policy Brief (Issue l, November 2002), from http:Unieer.org.

Frazier, J, A., & Morrison, F. J. {1998}. The influence of extended-year schooling on growth of achievement and perceived compe­tence in early elementary school. Child Development, 69(2), 495-517.

Frede, E. C. (1998). A sociocultural analysis of the long-term benefits of preschool for children in poverty. In W. S. Barnett & S. S. Boocock (Eds.), Early care and education for children in poverty: Promises, programs, and long term results (pp. 11-44). Albany, NY: SUNY Press.

Fusaro, J, A. { 1997). The effect of full-day kindergarten on student achievement: A meta-analysis. Child Study Journal, 27(4), 269-277.

Good, L. (1996). Teachers' perceptions of the ali-day, alternating day kindergarten schedule. Brookings, SD: South Dakota State University. (ERIC Document Reproduction Service No. ED396853)

Gullo, D. F. (2000). The long term educational effects of half-day vs. full-day kindergarten. Early Child Development and Care, 160, 17-24.

Harms, T., Clifford, R. M., & Cryer, D. (1998). Early Childhood Environmetlt Rating Scale, Revised Edition. New York, NY: Teachers College Press.

Hough, D., & Bryde, S. (1996, April). The effects of full-day kindergarten on student achievement and affect. Paper presented at the annual conference of the American Educational Research Association, New York.

Karweit, N. (1987). The kindergarten experience. Educational Leadership, 49(6), 82-86.

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement [20]

Lee, V. E., Burkam, D. T., Honigman, J., & Meisels, S. J. (2001 ). Full-day vs. halFday kindergarten: Which children leam more in which program? Paper presented at the annual meeting of the American Sociological Association, Anaheim, CA.

Lee, V. E., Burkam, D. T., Honigman, J., Ready, D. D., & Meisels, S. J. (2006). Full-day vs. half-day kindergarten: In which program do children learn more? American Journal of Education, 112, 163-208.

Love, J. M., Kisker, E. E., Ross, C. M., Schochet, P. Z., Brooks-Gunn, J., Paulsell, D., Boller, K., Constantine, J., Vogel, C., Fuligni, A. $., & Brady-Smith, C. (2002). Making a difference in tlte lives of infants and toddlers and their families: The impacts of Early Head Start. Executive Summary. Washington, DC: U.S. Department of Health and Human Services.

NICHD Early Child Care Research Network. (2001). Child care and children's peer interaction at 24 and 36 months: The NICHD Study of Early Child Care. Child Development, 72, 1478-1500.

NICHD Early Child Care Research Network. (2003). Does amount of time spent in child care predict socioemotional adjustment during the transition to kindergarten? Child Development, 74, 976-1005.

Olson, D., & Zigler, E. (1989). An assessment of the ali-day kindergarten movement. Early Childhood Research Quarterly, 4, 167-186.

Puleo, V. T. ( 1988). A review and critique of research on full-day kindergarten. Elemeutary School Journal, 88( 4), 427-439.

Rathbun, A. H., West, J., & Germino-Hausken, E. (2004). From kindergarten through third grade: Children's beginning school expe­riettces (NCES 2004-007). Washington, DC: National Center for Education Statistics.

Reynolds, A. (1993, March). Effects of a preschool plus follow-on intervention program for children at risk. Paper presented at the 1993 biennial meeting of the Society for Research in Child Development, New Orleans, LA.

Reynolds, A. ( 1994). Effects of a preschool plus follow-on intervention for children at risk. Developmetltal Psychology, 30(6), 787-804.

Sammons, P., Sylva, K., Melhuish, E., Siraj-Blatchford, 1., Taggart, B., & Elliot, K. (2003). The Effective Provision of Pre-School Education (EPPE) Project: A longitudinal study funded by the Department for Education and Skills. Technical paper 8b: Measuring the impact of pre-school on children's social/behavioral development over the pre-school period. London, England: Institute of Education, University of London.

Sammons, P., Sylva, K., Melhuish, E., Siraj-Blatchford, I., Taggart, B., Elliot, K., & Marsh, A. (2004). The Effective Provision of Pre­School Education (EPPE) Project: A longitudinal study funded by the Department for Education and Skills. Technical paper 9: Report on age 6 assessment. London, England: Institute of Education, University of London.

Schweinhart, L. }., Barnes, H. V., Weikart, D.P., Barnett, W. S., & Epstein, A. S. (1993). Significant benefits: The High/Scope Perry Preschool study through age 27. Monographs of the High/Scope Educational Research Foundation, 10.

Stinard, T.A. (1982). Synopsis of research on kindergarten scheduling: Half-day, everyday; Full-day, alternate day; and full-day, everyday. Cedar Rapids, IA: Grant Wood Area Education Agency.

U.S. Census 2000. [Data ftleJ. Washington, DC: U.S. Census Bureau.

U.S. Department of Health and Human Services. (1998). Head Start Program Information Report for 1996-1997 program year. Washington, DC: U.S. Department of Health and Human Services, Head Start Bureau.

U.S. Department of Health and Human Services. (1999). 1999 Head Start Fact Sheet. Retrieved December 8, 2004, from http://www.acfhhs.gov/programs/hsb/research/factsheets/99 hsfs.htm

U.S. Department of Health and Human Services. (2003). Head Start Faces 2000: A whole child perspective on program performance, Washington, DC: U.S. Department of Health and Human Services, Head Start Bureau.

U.S. Department of Health and Human Services. (2004b). Head Start Program Fact Sheet. Retrieved December 8, 2004, from http://www.acf.bhs gov/programs/hsb/research/2004.htm

U.S. Department of Health and Human Services. (2004a). Head Start Program Information Report for 2002-2003 program year. Washington, DC: U.S. Department of Health and Human Services, Head Start Bureau.

Weikart, D. P. ( 1967). Preschool intervention: A preliminary report of the Perry Preschool Project. Ann Arbor, MI: Campus.

West,]., Denton, K., & Reaney, L. M. (2001 ). The kindergarten year: Findings front the Early Childhood Longitudinal Study, Kindergarten Class of 1998-99 (NCES 2001-023). Washington, DC: National Center for Education Statistics.

Woodcock, R. W., & Johnson, M. B. (1989). Woodcock-Johnson Psycfw.Educational Battery. Revised. Itasca, IL: Riverside.

The Effects of Full-Day vs. Half-Day Preschool on Early School Achievement [21]

2126

Research

Original Investigation

Association of a Full-Day vs Part-Day Preschool Intervention With School Readiness. Attendance. and Parent Involvement Arthur J. Reynolds, PhD; Brandt A. Richardson, BA; Momoko Hayakawa, PhD; Erin M. Lease, MA; Mallory Warner­Richter, MPP; Michelle M. Englund, PhD; Suh-Ruu Ou, PhD; Molly Sullivan. MPP

IMPORTANCE Early childhood interventions have demonstrated positive effects on well-being. Whether full-day vs part-day attendance improves outcomes is unknown.

OBJECTIVE To evaluate the association between a full· vs part-day early childhood program and school readiness. attendance, and parent involvement.

DESIGN, SETTING, AND PARTICIPANTS End-of-preschool follow-up Of a nonrandomized, matched-group cohort of predominantly low-income. ethnic minority children enrolled in the Child-Parent Centers (CPC) for the full day \1 hours; n = 409) or part day (3 hours on average; n = 573) in the 2012·2013 school year in 11 schools in Chicago, Illinois.

INTERVENTION The Midwest CPC Education Program provides comprehensive instruction, family-support, and health services from preschool to third grade.

MAIN OUTCOMES AND MEASURES School readiness skills at the end of preschool, attendance and chronic absences, and parental involvement. The readiness domains in the Teaching Strategies GOLD Assessment System include a total of 49 items with a score range of 105-418. The specific domains are socioemotional with 9 items (score range, 20·81), language with 6 items (score range, 15·54), literacy with 12 items (score range, 9·104), math with 7 items (score, 8·60), physical health with 5 kerns (score range, 14·45), and cognitive development with 10 items (score range, 18·90).

RESULTS Full-day preschool participants had higher scores than part-day peers on socioemotional development (58.6 vs 54.5; difference. 4.1; 95% Cl, 0.5·7.6; P = .03), language (39.9 vs 37.3; difference, 2.6; 95% Cl, 0.6·4.6; P = .01), math (40.0 vs 36.4; difference, 3.6; 95% Cl, 0.5·6.7; P = .02), physical health (35.5 vs 33.6; difference, 1.9; 95% Cl, 0.5·3.2; P = .006), and the total score (298.1 vs 278.2; difference, 19.9; 95% Cl, 1.2·38.4; P = .04). Literacy (64.5 vs 58.6; difference, 5.9; 95% Cl, -0.07 to 12.4; P = .08) and cognitive development (59.7 vs 57.7; difference, 2.0; 95% Cl, -24 to 63; P ~ 38) were not significant. Full-day preschool graduates also had higher rates of attendance (85.9% vs 80.4%; difference. 5.5; 95% Cl, 2.6·8.4; P = .001) and lower rates of chronic absences (;,10% days missed; 53.0% vs 71.6%; difference, -18.6: 95% Cl, -28.5 to -8.7; P = .001; ;,20% days missed; 21.2% vs 38.8%: difference -17.6%: 95% Cl, -25.6 to -9.7: P < .001) but no differences in parental involvement.

CONCLUSIONS AND RELEVANCE In an expansion of the CPCs in Chicago, a full·day preschool intervention was associated with increased school readiness skills in 4 of 6 domains, attendance, and reduced chronic absences compared with a part-day program. These findings should be replicated in other programs and contexts.

JAMA. 2014:312(20):2126· 2134. doU0.100l/jama.2014.1S376

Copyright 2014 American Medical Association. All rights reserved.

m Editorial page 2101

0 JAMA Report Video and Author Video Interview at jama.com

G Supplemental content at jama.com

Author Affiliations: Institute ofChWd Development. University of Minnesota, Minneapolis (Reynolds, Hayakawa, Lease, Warner-Richter, Englund, Ou); Human Capital Research Collaborative, University of Minnesota. Minneapolis (Reynolds, Richardson, Hayakawa, Lease, Warner-Richter, Englund. Ou. Sullivan): Department of Applied Economics, University of Minnesota, Minneapolis (Richardson): Humphrey School of Public Affairs. University of Minnesota. Minneapolis (Reynolds, Hayakawa. Sullivan).

Corresponding Author: Arthur J. Reynolds. PhD. Institute of Child Development & Human Capital Research Collaborative, University of Minnesota. 51 E River Rd. Minneapolis, MN 55455 (ajr@umn.edu).

jama.com

Downloaded From: http://jamanetwork.com/ by Jonnna Robs on 09/27/20)7

Preschool Full or Half Days and School Readiness

E arly childhood interventions improve educational suc­cess and well-being.1

•2 Participation in high-quality cen­

ter-based programs at ages 3 and 4 years is associated with greater school readiness and achievement, higher rates of educational attainment and socioeconomic status, and lower rates of crime. 1"5

Although publicly funded preschool such as Head Start and state prekindergarten serve an estimated 42% of US 4-year­olds, most provide only part-day services, and only 15% of 3-year-olds enroll. 6 These rates plus differences in quality may account for only about half of entering kindergartners having mastered skills needed for school success.7·8

One approach for enhancing effectiveness is increasing from a part-day to a full-day schedule. In addition to increas­ing the amount of learning time, full-day preschool can increase continuity in learning as children avoid multiple education placements during the day; reduce family stress by increasing time for parents to pursue employment and edu­cation; and promote long-term effects on well-being. Although evidence from prior studies is meager,9 -11 imple­mentation of full-day preschool within a high-quality, evidenced-based model may be particularly cost-effective, especially for children exposed to early adversity.

The Child-Parent Center Education Program (CPC) is a school-based public program with strong evidence ofbenefits.12

Implemented in the Chicago Public Schools since 1967, the pro­gram provides comprehensive education and family services beginning in preschool. Cohort studies have found that par­ticipation has helped eliminate the achievement gap in school readiness and performance; reduced rates of child maltreat­ment, remedial education, and crime; and increased rates of high school graduation and economic well-being. 1

l-15 Ben­efits exceed costs by a ratio of7 to 1.15 However, the preschool day was limited to 3 hours.

A scale-up of the CPC program began in 20U in more di­verse communities. The model was revised to incorporate ad­vances in teaching practices and family services and included the opening of full-day preschool classrooms in some sites.

We investigated whether full-day preschool was associ­ated with higher levels of school readiness, attendance, and parent involvement compared with part-day participation. We also examined variation by age and program attributes.

Methods

The Midwest Expansion of the CPC is a contemporary expan­sion of the original program implemented for a 2012 preschool cohort to be followed up to third grade. Five school districts of various sizes serving a broad spectrum of predominantly low-income families in Illinois and Minnesota agreed to imple­ment CPC and follow the guidelines and requirements. Ap­proval for the project was granted by institutional review boards at the University of Minnesota and participating institutions, including written informed consent.

In 11 of16 Chicago sites, both full- and part-dayprograms were conducted in the same schools. This reJX>rt compares out­comes of children in these programs at the end of preschool.

jama.com

Original Investigation Research

Sample and Design The study included 982 three- and four-year-olds in these 11 schools. (A description of the larger Midwest CPC Expansion is in eAppendix A in the Supplement.) Three of the schools with full-day classes were new CPCs in underrepresented areas; the others were established prior to 1980.

Schools offered full-day preschool primarily because they had space, slots were available, and there was a perceived de­mand. This was not the case in other schools as they all lacked space and had little demand. There was no evidence the schools implementing full-day differed from the 5 schools offering only part-day services in commitment to school improvement. All showed evidence of effective implementation of services, 16 and one was undergoing a reform initiative.

For full-day preschool, children enrolled at age 3 or 4 years for the entire school day (7 hours) were compared with chil­dren in the same schools who participated for part of the day (between 2.75 and 3 hours). Children were not randomly as­signed to full- or part-day, due to the high likelihood of non­adherence by parents and school resistance. Three criteria were used by principals in consultation with the project team to as­sign children to the full-day program: children whowere4 years rather than 3 years of age; parental preference due to employ­ment or education, transportation barriers, or the lack of avail­able care for the other part of the day; and children with greater educational needs. In some cases, existing part-day class­rooms were converted to full-day and families participated who would not have otherwise enrolled. Children in both groups attended preschools days a week for at least 3 months and be­gan no later than January 2013.

Intervention The Midwest CPC intervention was designed to enhance early childhood development in multiple domains of health and well-being. Located within or near elementary schools, the program provides educational and family-support ser­vices between preschool and third grade. Within a structure of comprehensive services (education, family, health, and social services), 6 major components are included17

:

(1) collaborative leadership team led by a head teacher and 2 family coordinators; (2) effective learning experiences (eg, small classes, certified teachers, and literacy-rich instruction); (3) parent involvement and engagement; (4) aligned curriculum across grades; (5) continuity and sta­bility; and (6) professional development system of teacher coaching and site support.

In the effective learning component, the emphasis is on the acquisition ofbasic skills in language and literacy, math, and so­doemotional development through relatively structured but di­verse learning experiences that include teacher-directed, whole­class instruction, small-group and individualized activities, field trips, and child-initiated learning. The parent component is an intensive menu-based approach that includes parenting edu­cation, volunteering in the classroom, attending school events and field trips, furthering education, and receiving horne vis~ its and health and nutrition services, including screening and diagnostics, meal services, and referrals. Professional develop­ment includes online teaching modules.

JAMA November 26, 2014 Volume 312, Number 20 2127

Copyright 2014 American Medical Association. All rights reserved.

Downloaded From: http://jamanetwork.com/ by Jovanna Robs on 09/27/2017

Research Original Investigation

Outcome Measures School Readiness We assessed 7 indicators of school readiness at the end of the preschool year using the Teaching Strategies GOLD Assess­ment System.18 Teaching Strategies is a performance-based as­sessment designed for children from birth through kindergar­ten composed of 66 items measuring mastery on 38 objectives in 9 domains of development. As a widely used assessment in early childhood settings, Teaching Strategies has shown strong reliability and validity in measuring school readiness that is predictive of school achievement and performance (eAppen­dix Bin the Supplement).19-

22 Scores reflect functional perfor­mance in the classroom context that is not directly measured by tests of cognitive skills, yet they are highly correlated with direct assessments. 19

'22 The assessment is also aligned with

state early learning standards (eAppendix Bin the Supple­ment).

We reported outcomes for 6 domains assessed with 49 items as administered by the Chicago Public School District: literacy with u, oral language with 6, math with 7, cognitive development with 10, socioemotional with 9, and physical health with 5 items. Social studies, science, and art were not assessed in most sites. Each item is rated from o, not yet meet­ing objective, to 9, full mastery of objective, as observed by the classroom teacher (eTable 1 in the Supplement provides item descriptions). The mean of the scale is set at the distribution midpoint, which is the expected score for age 36 months. We analyzed raw scores summed across items for the subscales adjusted for age plus the total score for all domains. Measure­ments were taken at the fall baseline (October to November 2012) and mid-May 2013. Dichotomous scores measuring per­formance at or above the national norm also were assessed.19

Meeting the national norm on 4 or more subscales was the set threshold.

Attendance We used 3 indicators from official school administrative rec­ords. Average daily attendance was the percentage of total available days of enrollment that a child was in attendance. Chronic absence was a dichotomous indicator of whether a child missed 10% or 20% ofthe possible school days or more. Average attendance and chronic absence were based on the total number of school days a child was enrolled during the year. Attendance and absences reflect health problems, ill­ness, adverse experiences in the family, and economic fac­tors and predict not only academic achievement but socio­emotional adjustment and health. 23~2s

Parental Involvement We used 3 indicators of participation in children's education. For parent involvement, classroom teachers rated on a 10-point scale the "percent of parents who participated in school events and activities from January to the end of the year:• A rating of 1 indicated that less than 10% of families in the classroom par­ticipated and a rating of10 indicated that 90% or more of fami­lies participated (range, 2-10; median [SD], 6 [2.2]). The rating for each class was assigned to each individual child, which re­duces response bias and halo effects found in individual rat-

2128 JAMA November 26, 2014 Volume 312. Number 20

Preschool Full or Half Days and School Readiness

ings. A dichotomous indicator at or above the mean of6 also was assessed. Previous studies show that ratings by teachers are valid indicators of parenting practices and are a mecha­nism of long-term effects of early intervention.14

•26 As a sec­

ondary measure, parents rated mid-yeartheir own frequency of participation: "So far this year, about how often have you participated in school or center activities?" (range, 0-5 num­ber of activities).

Statistical Analysis Data were analyzed in SPSS (version 22). 27 Findings are reported as marginal means and group differences control­ling for the influence of the following: child's sex, race/ ethnicity, eligibility for subsidized lunches (based on family income), age in months, special education, school-level achievement (attended a school in which 70% or more of third graders met state reading norms), fall baseline perfor­mance (school readiness or attendance), and a dichotomous indicator of the timing of the baseline assessment. These covariates were measured at preschool entry from school administrative records and parent surveys. Continuous and dichotomous outcomes were analyzed as linear or probit regressions in the generalized estimating equations (GEEs) approach, which is an extension of the generalized linear model appropriate for correlated or clustered data.28 Using maximum likelihood techniques, estimates account for clustering of observations by school through the Huber­White-sandwich correction. The GEE approach provides robust estimates of standard errors and accommodates non­normal data. 29·30

Multiple imputation of missing data on Teaching Strate­gies was based on the expectation-maximization algorithm af­ter determining that scores were consistent with the assump­tion of missing at random. 31 A sensitivity analysis was conducted using imputation. Adjusted group differences at the .OS probability level for a 2-tailed test were emphasized. Stan­dardized mean differences (SDs) were also reported with val­ues of 0.20 or higher in the range of clinical or practical significanceY Raw score differences equivalent to one-fifth of a year of growth (2-3 months) in school readiness were con­sidered of practical significance.19 These ranged from 1.5 (physi­cal health) to 4.0 points (literacy). To assess subgroups, pro­gram interaction terms included child age, race/ethnicity, and whether the site was a new CPC. Differences for existing and new sites also were tested. The significance of subgroups was set at .os.

Results

Sample Characteristics Among the 11 sites, 409 children enrolled in full~day classes and 573 in part -day classes. The pattern of participation and data collection for these groups are shown in Table 1. They rep­resent 57% of the original sample of 1724 children who en­rolled in fa1l2012. Excluded children attended partfiday pro­grams in the s other schools not offering full-day. They had similar characteristics as the study sample of age, sex of child,

jama.com

Copyright 2014 American Medical Association. All rights reserved.

Downloaded From: http://jamandwork.com/ by Jovanna Robs on 09/27/2017

Preschool Full or HalfDays and School Readiness

low~income status, and fall baseline performance. The ex~ eluded group had a higher concentration of Latino families.

The characteristics of the full-day and part-day groups in the same school are shown in Table 2. Children were well­matched on fall baseline school readiness, including the mean total score across the 6 subscales (193.2 vs 190.2; difference, 3.0; P = .46; dichotomous, 14.2% vs 16.1%; difference,-1.9; P = .49). Groups were also equivalent on many child and fam­ily background characteristics. These included sex of child, race/ethnicity,low-income status, parent educational achieve-­ment, employment status, and special education. The major difference between groups was age because full-day partici­pation was more likely for 4-year-olds. This difference was taken into account by including age as a covariate in the main analysis as well as baseline performance. Also taken into ac­count was that proportionally fewer full-day participants at­tended high-performing schools (Table 2).

Implementation Adherence and Fidelity Overall, the sites successfully implemented the program re­quirements including establishing the leadership teams, main­taining small class sizes, and providing comprehensive child development and family services. 16•17 All sites met these and related requirements, and 75% of observed classrooms were rated moderately high to high in task orientation (a key pro­gram focus). Four sites experienced delays in opening full­day classrooms but these were fully operating by January.

The overall average rating of implementation fidelity across the 6 elements was 3.9 or moderately high (minimum score, 1, maximum score, 5).l6 The highest was continuity and stability at 4.3 and the lowest aligned curriculum at 3.3. The collaborative leadership score of 4.0 and parent involvement score of 3.9 were also moderately high. Mean classes sizes were 17.8 in full-day and 15.1 in part~day. Although no differences in classroom ratings of student engagement (eg, task orientation and responsiveness) were detected, a greater percentage of math instruction in full­day classrooms was child-initiated compared with part-day classrooms (eTable 2 in the Supplement). A similar pattern occurred for language and literacy.

The total amount of instruction time for the year was 2.2 times greater in full-day classes (936 vs 418 hours; P < .001; eTable 2 in the Supplement). The median duration ofpartici· pation for each group was 165 days (8 months; range, 3-9 months) with 91% in the full-day and 84% in the part-day groups enrolled for at least 6.5 months. Patterns of enroll­ment showed no evidence of crossovers.

Outcomes of CPC Full-Day and Part-Day Participation Table 3 shows the group differences, P values and 95% confi­dence intervals for the same-school full- and part-day groups after adjustment for the covariates.

School Readiness For 4 of the 6 subscales, full-day participants demonstrated higher mean skill mastery than part-day participants. These included language (39.9 vs 37-3; difference, 2.6; 95% CI, 0.6-4.6; P = .01), math (40.0 vs 36.4; difference, 3.6; 95% Cl, 0.5·

jama.com

Original Investigation Research

Table 1. Patterns of Participation of Full-Day and Part·Day Preschool Groups in 11 Schools, Midwest Child-Parent Center Expansion

No. of Children

Study Category Full-Day Part-Day Characteristics at start of study, No.a

CPC preschool 409 573

Classrooms, No. (sessions/dasses) 23 (1) 19 (2)

Original sites, No. 285 529

Expansion sites, No. 124 44

4-y-olds at program entry 351 215

3-y-olds at program entry 58 358

Study participants with datab

Attendance and chronic absence 409 573

?1 measure of school readiness 337 471

Parent involvement (teacher ratings) 409 573

Abbreviation: CPC, Child-Parent Centers. 3 Program group enrolled in the CPC program in 2012-2013 as 3- or 4-year·olds

in 11 schools offering full-day preschool classes. Children attended at least 3 months and were enrolled no later than January. Part·day classes had 2 sesslons per day (mornlng and afternoon).

b Attendance data are from school administrative records; school readiness is from the Teaching Strategies GOLD Assessment; and parent involvement is from teacher ratings at the end of the preschool year. Parent ratings of involvement was a supplemental measure in which 272 and 332 full-day and part-day participants, respectively. provided data by early spring2013.

6.7; P = .02), socioemotional development (58.6 vs 54.5; dif­ference, 4.1; 95% Cl, 0.5-7.6; P = .03), and physical health (35.5 vs 33.6; difference, 1.9; 95% Cl, 0.5·3.2; P = .006). Results for literacy (64.5 VS 58.6; difference, 5.9; 95% Cl, -0.07 to 12.4; P = .08), and cognitive development (59-7 vs 57.7; difference, 2.0; 95% Cl, -2.4 to 6.3; P = .38) were not statistically significant.

For rates of mastery at or above the national average, 4 of the 6 subscales showed differences. Full-day participants had higher rates ofliteracy (85.1% vs 74.6%; difference, 10.5; 95% Cl,1.5-19.4; P = .03), math (84.4% vs 72.3%; difference 12.1; 95% CI, 5-3-18.9; P = .001), socioemotiona1(73.4%vs56.o%;differ­ence, 17.4; 95% CI, 0-35.0; P = .05), and language develop­ment (81.2% vs 61.7%; difference, 19.5; 95% CI, 4.5-34.6; P = .01). Although literacy showed a positive association for mastery at the national average, findings for physical health were limited to mean differences. Standardized mean differ­ences were 0-57 for language, 0.46 for socioemotional, 0.42 for physical health, 0.41 for math, 0.37 for literacy, and 0.16 for cognitive development.

In addition, the full-day group had a significantly higher rate of mastery on the total readiness metric, for 80.9% were at or above the national average on 4 or more subscales com­pared with 58.7%ofthepart-day group (difference, 22.2; 95% CI, 5.8-38.5; P = .ooS). The standardized mean difference of 0.65 was relatively large. Mean differences also were signifi­cant (298.1 vs278.2; difference, 19.9; 95% CL 1.2-38.4; P = .04; standard mean difference, 0.33).

These findings translate to percentage change differ­ences associated with full-day preschool of16.7% (at or above norm in math) to 37.6% (total score; eFigure in the Supple­ment). Converting the observed raw score differences to

JAMA November 26, 2014 Volume 312. Number 20 2129

Copyright 2014 American Medical Association. All rights reseJVed.

Downloaded From: http://jamanetwork.com/ by Jovanna Robs on 09/27/2017

Research Original Investigation Preschool Full or Half Days and School Readiness

Table 2. Characteristics of Same-School Child-Parent Center Full-Day and Part-Day Groups at Fall Baseline. 2012-2013

Characteristicsb Demographics, No. (%)

Girl

Black

Hispanic

Special education status<

Age on September 1, 2012, mean (SO), mo

Parent survey<'

Mother completed high school

Eligible for fully subsidized meals"

Single-parent f.:~mily status'

Mother employed full- or part-time'

Child-Parent Center Sample, No. (%)a

Preschool full-Day Part-Day (n = 409) (n = 573)

216 (52.8) 295 (51.2)

363 (88.8) 523 (93.0)

31 (7.6) 40 (7.0)

19 (4.6) 22 (3.8)

51.6 (5.4) 45.8 (6.5)

215 (79.9) 253 (78.1)

367 (89.7) 529 (92.3)

177 (65.1) 218 (65.7)

186 (53.7) 231 (48.0)

PVatueforfuU-Dayvs Part-Day

.55

.02

.80

.63

<.001

Attended a school with a high percentage of students meeting state reading norms 80 (15.4) 188 (28.9)

.61

.17

.93

.19

<.008

Fall baseline, mean (SO), score

literacy subscale

Math subscale

Socioemotional development

Total scale

<:National norm on >4 subscates

Fall baseline assessed after October

a The sample included participants who enrolled in full-day or part-day preschool in the same llsites. P values show the significance of mean or percentage for group differences. fall baseline scores were adjusted for age. The threshold for grade 3 state reading norms was 70% or higher on the Illinois State Achievement Test. The sample had valid values for 1 or more outcome indicators.

b Data on child and family characteristics were collected from school administrative records with the exception of low-income status. which was a combination of administrative records and parent reports.

months of expected improvement during the year, full-daypre­school was associated with about a third of a year (3·4 months) of improvement in all domains except cognitive deve1op­ment {1·1.5 months).

Attendance Compared with part-day, full-day participation was associ­ated with a higher rate of average daily attendance (85.9% vs 80.4%; difference, s.S; 95% CI, 2.6-8.4; P = .001) and a ]ower rate of chronic absences (53.0% vs 71.6%; difference, -18.6; 95% Cl, -28.5 to -8.7; P = .001) as well as absences defined at 20% or more days missed (21.2% vs 38.8%; difference, -17.6; 95% Cl, -25.6 to -9.7;P < .001). Standardized mean differences were around -0.50. This corresponds to percentage reductions in chronic absences associated with full~day preschool of26.0% to45-4%.

Parental Involvement No significant differences were detected for teacher (3.95 vs 4.6S;difference, -0.7;95%CI, -L7~3.0;P = .17)andparentrat­ingsofschool involvement(2.54 vs 2.51; difference, 0.03; 95% CI, -0.54·0.61; P = .92).

2130 JAMA November 26. 2014 Volume 312, Number 20

35.3 (16.3)

23.5 (8.9)

40.2 (11.8)

193.2 (57.4)

48 (14.2)

201 (53.4)

33.9 (16.4)

22.6 (9.2)

39.2 (14.7)

190.2 (64.7)

76 (16.1)

266 (58.0)

.20

.16

.26

.46

.49

.20

c Children who have an individual education plan under Individuals with Disabilities Education Act (IDEA).

d Parent education. single-parent family status. and employment were garnered from parent surveys. Sample size for parent survey was 272 in the full-day group and 332 in the half-day group. The sample sizes for the fall baseline assessments were 337 for the full-day group and 471 for the half-day group.

" Ertgibility was defined at 130% of the federal poverty line or lower. 1 There were 272 parents whose children were in the full-day group and 332 in

the half-day group.

Sensitivity Analysis The pattern of findings for full-day vs part~day preschool was found with or without multiple imputation ofTeaching Strat­egies (see eTables 4 and 5 in the Supplement). With fully im~ puted scores (17.7% imputed for spring scores), full-day in the same schools was positively associated with the total score (296.7 vs277.7; difference, 19.0; 95% Cl, 0.2-34.8; P = .02; stan~ dard mean difference, 0.31) and 5 of the 6 subscales, includ~ ing literacy (64.1 vs 58.3; differences, 5.8; 95% Cl, 0.3~11.2; P = .04; standard mean difference, 0.33), math (39.8 vs 363; difference, 3.5; 95% CI, 0.9·6.1; P = .008; standard mean dif· ference, 0.37).and physical health (35-3Vs33.6;difference, 1.7; 95% CI, 0.6-2.8; P = .003; standard mean difference, 0.29; eTable 4 in the Supplement). Moreover, alternative specifica­tions of GEE and related approaches showed a similar pattern of findings.

Subgroup Differences We found few differences in estimates ofCPC full-day pre~ school on outcomes by race/ethnicity, age, and CPC status (new vs established). Table 4 shows the results for select continu· ous outcomes. We used the fully imputed and continuous out~

jama.com

Copyright 2014 American Medical Association. All rights reserved.

Downloaded From: hftp://jamanetwork.com/ by Jovanna Rohs on 09/27/2017

Preschool Full or Half Days and School Readiness Original Investigation Research

Table 3. Child-Parent Center Same-School Fuii-Dayvs Part-Day Preschool: Adjusted Marginal Meansa

Group, No. 00 Full-Day Part-Day Preschool Preschool Difference Standard Mean

Outcome (n • 409) (n • 573) {95%CI) PValue Difference School readiness skills a The sample includes 808-982

literacy, raw score. 64.5 58.6 5.9 (-0.7 to 12.4) .08 0.33 children from 11 sites offering 12 items full-day preschool. Coefficients are

<!:National norm 287 (85.1) 351 (74.6) 10.5 (1.5 to 19.4) .03 0.37 from linear or probit regression

Language, 6 items 39.9 37.3 2.6 (0.6 to 4.6) .01 0.34 analysis (generalized linear models via maximum likelihood)

<!:National norm 274 (81.2) 291 (61.7) 19.5 (4.5 to 34.6) .01 0.57 transformed to marginal effects, Math, 7 items 40.0 36.4 3.6 (0.5 to 6.7) .02 0.38 and they are adjusted for child sex,

<!:National norm 284 (84.4) 341 (72.3) 12.1 (5.3 to 18.9) .001 0.41 race/ethnicity, age (months), subsidized lunch status, special

Cognitive 59.7 57.7 2.0 (-2.4 to 6.3) .38 0.16 education, school-level development, achievement. and fall baseline 10 items

performance (school readiness or ~National norm 237 (70.3) 301 {64.0) 6.3 (-16.2 to 28.8) .99 0.22 attendance). For school readiness, a

Socioemotional 58.6 54.5 4.1 (0.5 to 7.6) .03 0.34 dichotomous indicator for a laterfall development, 9 items assessment also was included. The P

~National norm 247 (73.4) 264 (56 O) 17.4 (0 to 35.0) .OS 0.46 value is the probability level of the

Physical health, 35.5 33.6 1.9 {0.5 to 3.2) .006 0.32 adjusted mean or percentage difference. Standard errors. and 5 items

<!National norm 277 (82.2) 323 (68.6) 13,5 (-1.0 to 28.1) .07 0.42 thusPvalues, are adjusted for variation among program sites by

Total score, 49 items 298.1 278.2 19.9 (1.1 to 38.4) .04 0.33 the Huber-White-sandwich with 6 subscales correction. Three hundred

~National norm on 273 {80.9) 276 <san 22.1 (5.8 to 38.5} .008 0.65 thirty-seven children were in the >4 subscates full-day and 471 in the part-day

Attendance Child-Parent Center.lhe possible Average daily, % 85.9 80.4 5.5 (2.6 to 8.4) .001 0.41 (not actual) ranges for continuous

<!:10%Absences, d 217 {53.0) 410 (71.6) -18.6 (-28.5 to -8.7) .001 -0.50 outcomes were literacy (0-108). language (0-54), math (0-63),

;: 20% Absences, d 87 (21.2) 222 (38.8) -17.6 (-25.6 to -9.7) <.001 -0.53 cognitive development (0-90), Parental participation socioemotional development score (0-81), physical health (0-45), total

Teacher ratings 3.95 4.65 -0.70 (-1.7 to 3.0) .17 -0.38 score (0-441), average daily

High Involvement, ~6 124 (30.3) 254 (44.3) -14.0 (-38.3 to 10.3) .23 -0.37 attendance (1%-100%). parent involvement in school (teacher,

Parent report, spring 2.54 2.51 0.03 (-0.54 to 0.61) .92 0.02 1·10), and parent involvement cases

comes to optimize power. Notably, differences in mean atten­dance (14.4 percentage points) significantly favored children in new sites (95% CI, 11.6-17.2) as did chronic absences at 22.1 percentage points (95% CI, -33.9 to -10.3; P <.001 for both mea­sures). Teaching Strategies GOLD scores were similar by site status and age, although the pattern of findings favored 3-year­olds. The only difference for parent involvement was that com­pared with part-day, full~day in established sites had signifi~ cantly higher parent-reported involvement than in new sites (0.3 vs -1.10; difference in difference, -1.3; 95% Cl, -2.2 to -0.38; p = .005).

Discussion

The current study shows that full-day preschool in the Mid­west CPC program was associated with higher scores in 4 of6 domains of school readiness skills-language, math, socio­emotional development, and physical health-increased at­tendance, and reduced chronic absences by 26% to 45% over part-day services. The greater amount of time spent in pre­school was associated with 17% to 38% increases in children meeting national norms on 4 of6 subscales-language, math,

jama.com

{parent. 0-5).

socioemotional development, and literacy-and gains in school readiness of 3 to 4 months. Only for cognitive development were there no group differences detected. Full·day preschool appears to be a promising strategy for school readiness. The size and breadth of associations go beyond previous studies. g-n

The positive assodation offull~daypreschool also suggests that increasing access to early childhood programs should con· sider the optimal dosage of services. In addition to increased educational enrichment, full-day preschool benefits parents by providing children with a continually enriched environ­ment throughout the day, thereby freeing parental time to pur­sue career and educational opportunities. By offering an­other service option, full~day preschool also can increase access for families who may not otherwise enroll. These findings also support the prevention goals of Healthy People 2020.33

The relation between full~day preschool and school readi­ness found in this report is consistent with prior dosage stud­ies examining early reading and math achievement.9 -n·34 For example, a report of the federally sponsored Early Childhood Longitudinal Study found that length of day in center-based preschool was positively associated with reading and math skills at kindergarten entry, especially for low-income children. 9

No differences were found for social behavior, however, and

JAMA November 26, 2014 Volume 312. Number 20

Copyright 2014 American Medical Association. All rights reserved.

Downloaded F.-om: http://jamanetwork.com/ by Jovanna Robs on 09/27/2017

2131

Research Original Investigation Preschool Full or Half Days and School Readiness

Table4. Adjusted Mean Differences at the End of Preschool Between Same-School Full-Day and Part-Day Preschool for New and Established Sites andbyAgea

full-Day and Part-Day Full-Day and Part-Day Preschool Difference in Preschool Difference,

New Sites Established Difference 4-Year-Oids 3-Year-Olds Difference Outcome (n-= 168) (n • 814) (95%CI) PValue (n = 566) (n = 416) (95%(1) PValue School readiness skills

Uteracy, raw score. 6.5 5.8 -0.7 (-6.8 to 8.2) .82 4.7 10.1 5.4 (1.4 to 9.3) .01 12 items Language, 6 items 2.1 2.7 0.6 (-2.7 to 1.5) .60 1.7 1.1 -0.6 (-2.6 to 1.4} .57

Math, 7 items 3.7 3.5 -0.2 (-3.2 to 3.7) .90 2.6 3.5 0.9 (-0.5 to 2.2) .21

Cognitive development, 1.6 2.2 0.6 (-4.9 to 3.7) 10 items

.79 3.1 1.0 -2.1 (-5.3 to 1.0) .19

5ocloemotional 2.9 32 0.9 (-4.6 to 2.9) development, 9 items

.65 2.1 3.0 0.9 {-4.5 to 6.2) .75

Physical health, 5 items 0.9 2.0 1.1 (-2.8 to .6} .19 2.9 2.5 -0.4 (-1.9 to 1.0) .55

Total score, 49 items, 17.6 19.9 2.3 (-23.1 to 19.2) .38 12.5 31.3 18.8 (-4.5 to 42.1) .11 6subscales At or above the national 2.6 16.6 14.0 (-34.4 to 6.3} .19 13.2 30.3 17.1 (-4.7 to 29.9) .11 no~m on ~4 subscales, %

Attendance

Average daily 17.3 2.9 14.4 (11.6 to 17.2) <.001 5.6 4.1 -1.5 {-4.2 to 1.1) 26 attendance,% Chronic absences, ~20% -35.6 -13.5 -22.1 (-33.9 to -10.3) <.001 16.0 18.6 2.6 (-11.8 to 6.6} .58 days,%

Parental partitipation

Parent involvement in 0.2 -0.4 0.6(-l.Oto2.1) school (teacher ratings)

.46 -0.4 -1.8 -1.4 (-3.2to 0.4) .12

High involvement -7.2 -6.2 -1.0 {-35.5 to 33.9) .96 -10.1 -33.1 -23.1 (-23.0 to 68.0) .40 (~6 score),%

a The sample is fully imputed and includes 982 children from 11 sites offering fufl·day preschool. The Difference in Difference is the mean difference of the difference between each respective subgroup. CoeffiCients are from linear or probit regression analysis (generallzed linear models via maximum likelihood) transformed to marginal effects, and they are adjusted for child gender, race/ethnicity. age (months), subsidized lunch status, special education.

results of the observational study were consistent across a range of analyses. In a randomized controlled trial of Head Start pro­grams in Chicago, full-day preschool at ages 3 years, 4 years, or both was associated with nearly double the gains in school readiness compared with part-day preschool. 11

To our knowledge, this study is the first to extend the outcomes of full-day preschool to higher attendance and lower chronic absences. Unlike previous studies, we also documented relatively large associations with socioemo­tional development and physical health. As a comprehen­sive evidence-based program, CPC's demonstrated quality is higher than most other interventions. These findings are a1so consistent with those in the Chicago Longitudinal Study and other projects showing both immediate and long­term associations of preschool intensity measured in total days or years of attendance. 35·38

Although the program was associated with significant gains in 5 of the 6 domains for raw scores or rates of mastery at or above the national average, not all scores were improved. The large percentage of each group that was absent IO% to 20% or more of days enrolled as well as the limits of measuring a com­prehensive set of outcomes may have contributed. No differ­ences were detected for either indicator of cognitive develop­ment. This may be due to the instructional focus of the program on specific skills in language, numeracy, and behavior rather

2132 JAMA November 26, 2014 Volume 312, Number 20

school-level achievement, and fall baseline performance (school readiness or attendance). For school readiness, a dichotomous indicator for a later fall assessment also was included. The P value is the probability level of the adjusted mean or percentage difference. Standard errors, and, thus, P values. are adjusted for variation among program sites by the Huber-White-sandwich correction.

than general cognition (eg, thinks symbolically). Moreover, the subscale may not reflect the wide range of skills and ap· preaches to learning that encompass the broad concept of cog­nitive development. Although literacy readiness scores were not different between groups, differences favoring full-day pre­school were detected for the dichotomous indicator of meet­ing the national norm and in the fully imputed model ofmeans. These estimates translate to educationally meaningful differ­ences (standard mean difference, 0.37 or a 4-month gain). Physical health showed mean differences but no differences in the rate at or above the national average. Results were also educationally meaningful (standard mean difference, 0.42 or a 3·4 month gain.)

The current study is the first to assess full-day CPC pre· school. The positive association between full-day and school readiness should be seen in the context of changes in the in­tervention from that evaluated previously. First, 6 elements are emphasized: effective learning experiences, collabora­tive leadership, parent involvement and engagement, aligned curriculum, continuity and stability, and professional devel­opment. The previous model emphasized only the first 3 and with a lower degree of intensity. The Midwest expansion also introduced a professional development system of coaching, provided program support by site mentors, and imple­mented curriculum alignment and parent involvement plans

jama.com

Copyright 2014 American Medical Association. All rights reserved.

Downloaded From: http://jamanetwork.com/ by Jovanna Robs on 09/27/2017

Preschool Full or Half Days and School Readiness

in collaboration with principals. These elements inform the interpretation of results and provide a documented frame­work for replication and expansion. For example, the City of Chicago has announced plans to implement this model to serve an additional260o children using Pay-for-Success financing.

The study has at least slimitations. First, the measures as­sessed a limited range of outcomes. Although not a purely ob· jective measure of school readiness skills, Teaching Strate­gies is a performance-based assessment of mastery. Further advantages are that the assessment is aligned to state stan­dards, it includes all domains ofleaming key to school readi­ness, results are used to improve instruction, and it has evi­dence of predictive validity. Moreover, performance-based and direct assessments correlate highly with each other.21

•22

•39 The major disadvantage is the possibility of bias in ratings since teachers were not blind to children's intervention status. Two factors counteract this limitation. First, teachers receive train­ing on the assessment to increase accuracy and help reduce ratings bias. Teaching Strategies is routinely administered by schools and was not specific to this study. Second, if the lack ofblinding about intervention status introduced bias in favor of children in full-day classrooms, it would have been ex­pected to be observed at the baseline assessment, 2 months into the year. However, group differences on the assessment were equivalent.

Second, a significant amount of data for Teaching Strate­gies were missing, which may have affected the reliability and stability of estimates. That findings were similar across a range of imputations minimizes this threat to validity.

The third limitation was that even with the history of prior program implementation, full-day preschool in the CPCs was being implemented for the first time. Delays in staffing and the extra time needed to establish the full-day structure of opera­tions were unavoidable. This suggests that the findings may be conservative compared with implementation after the start-up period.

Original Investigation Research

Fourth, although groups were similar at baseline and analy­ses accounted for many school, child, and family attributes, it is possible that unmeasured factors contributed to find­ings. Consequently, results should be interpreted cautiously. Random assignment, although not possible in our study, can more easily rule out potential confounding variables or those that are difficult to measure (eg, motivation or attitudes}. The inclusion of the most relevant covariates identified in prior studies reduces this threat however.14

•35•36 That the full·day group had a higher concentration of 4-year-olds was ac­counted for by the inclusion of age and baseline performance as covariates. To the extent that this compositional differ· ence was not fully adjusted in the modeL findings may be con­servative because the pattern of associations favored younger children and they had a lower rate of participation in full-day preschool. The fact that the fall baseline assessment oc· curred 1 to 2 months into the year after the program began implementation also mitigates against the influence of un­measured factors. Findings of prior CPC studies support this interpretation.35.36.4°

Finally, the findings may have limited generalizability be­yond urban contexts and to programs different than CPC. De­spite the expansion to new underrepresented areas, most fami· lies were low-income and ethnic minority. That the associations in new sites were largely equivalent to those in established sites suggests a moderate degree of external validity.

Conclusions

In an expansion of the CPC program in low-income Chicago communities, a full-day preschool intervention was associ­ated with increased school readiness skills in 4 of 6 domains, attendance, and reduced chronic absences compared with a part-day program. These findings need to be replicated in other programs and contexts.

ARTICLE INFORMATION

Author Contributions: Dr Reynolds had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

CPC expansion intervention in collaboration with the project partners. No other disclosures were reported.

approval of the manuscript; and decision to submit the manuscript for publication.

Disdalmer: The views expressed in this article are those of the authors and do not necessarily represent the views of the sponsors and funders of this project. Study concept and design: Reynolds, Hayakawa,

Warner-Richter. Acquisition, analysis, or interpretation of data: All authors. Drafting oft he manuscript: Reynolds, Hayakawa, Lease. Ou, Sullivan. Critical revision of the manuscript for important intellectual content: Richardson. Hayakawa. Warner·Richter, Englund. Statistka/ arwfysis: Reynolds, Richardson. Hayakawa. Englund. Obtained funding: Reynolds. Administrative, technical, or material support: Reynolds. Hayakawa. Lease, Warner-Richter. Englund, Ou, Sullivan. Study supervision: Reynolds, Englund.

COnflict oflnterest Disclosures: All authors have completed and submitted the lCMJE Form for Disclosure of Potential Conflicts of Interest. Dr Reynolds reported that he developed the Midwest

jama.com

Funding/Support: Preparation of this report was supported by grant U41lB110098 from the US Department of Education's Investing in Innovation Fund and the following contributors: J. B. and M. K. Pritzker Famfly Foundation, McCormick Foundation, Boeing Corp, Evanston Community Foundation, Finnegan Family Foundation, Lewis· sebring Family Foundation, Foundation65, ~orthwestern University. Elizabeth Beidler Tisdahl Foundation. Target Corp. W. K. Kellogg Foundation, Doris Duke Charitable Trust, Foundation for Child Development. McKnight Foundation, Greater Twin Cities United Way, St Paul Foundation, Minneapolis Foundation. and the Joyce Foundation. Support also was provided by grant R01HD034294 from the National Institute of Child Health and Human Development.

Role of the Funder/Sponsor: The funders had no role in the design and conduct oft he study; collection. management. analysis, and interpretation of the data; preparation. review. or

Additional COntributions: We thank the Office of Early Childhood Education in the Chicago Public School District and 17 CPC schools (1 begins in kindergarten) for their extensive collaboration in the Midwest CPC Expansion Project. Special thanks go to Elizabeth Mascitti-Miller. chief officer of the Office of Early Childhood Education: Chrisopher Rosean. executive director of the Office of Early Childhood Education: Jaclyn Vasquez. CPC manager, Chicago Public Schools; and Serah Fatani. research manager. We thank Barbara Bowman, Paula Cottone. and Sonja Griffin for their leadership in developing the project. Anita President. Desiree Booker, Gwendolyn Jackson, Helen Haley, and Anne Gaddis have also made valuable contributions as site mentors. We thank our project partners. including Christine Maxwell, Linda Hamburg, and Anna Jerabek at the Erikson Institute, Erika Hunt and Usa Hood at the Center of the Study of

JAMA November 26, 2014 Volume 312, Number 20 2133

Copyright 2014 American Medical Association. All rights reserved.

Downloaded From: http://jamanetwork.com/ by Jovanna Rohs ou 09/27/2017

Research Otiginal Investigation

Education Policy at Illinois State University, and Donna Spiker. Erika Gaylor, and Katherine Fergusion at SRI International. for their contributions to the project. All of these contributors were compensated by the University of Minnesota Midwest CPC expansion grant. The following organizations received sub awards to the prime award to the university: Chicago Public School District, SRI International. Erikson Institute, Illinois State University, Evanston-Skokie School District 6S, McLean County Unit District 5, St Paul Public School District. and Arrowhead Economic Opportunity Agency Inc with the Virginia Public School District in Minnesota. We thank the HCRC team of Art Rolnick, Judy Temple, Nicole Smeri!!o, Allyson Candee. Ellen Lepinski, Allie Giovanelrl. Meg Soli, and Christina Montli, none of whom received compensation beyond their regular salaries.

Correction: This article was corrected for2 typographical errors on November 25, 2014.

REFERENCES

1. Braveman P. Gottlieb l. The social determinants of health: it's time to consider the causes of the causes. Public HeDith Rep. 2014;129(suppl2):19-31.

2. Manning M. Hamel R, Smith C. A meta-analysis of the effects of early developmental prevention programs in at-risk populations on non-health outcomes in adolescence. Child Youth ServRev. 2010;32(4):506-519.

3. Karoly LA. Kilburn MR. Cannon JS. Early Childhood Intervention: Proven Results, Future Promise. Santa Monica, CA: RAND Corp; 2005.

4. Campbell FA. Ramey CT. Pungello E, Sparling J, Miller-Johnson S. Early childhood education: young adult outcomes from the Abecedarian project. Appl DevSci. 2002:6(1):42-57.

5. Schweinhart U, Montie J, Xiang Z, Barnett WS, Belfield CR. No res M. Lifetime Effects: The High/Scope Perry Preschool Study Through Age 40. Ypsilanti, Ml: High/Scope Educational Research Foundation; 2005.

6. Barnett WS, Carolan ME, Fitzgerald J, Squires JH. The State of Preschoo/2012: State Preschool Yearbook. New Brunswick. NJ: National Institute for Early Education Research: 2012.

7. US Department of Education, National Center for Education Statistics. Conditions of education: kindergarten entry status: on-time. delayed-entry, and repeating kindergartners: 2013. https://nces.ed.gov/programs/coe/pdf/coe~tea

.pdf. February 15,2014.

8. Heckman JJ. Skill formation and the economics of investing in disadvantaged children. Science. 2006;312(5782): 1900-1902.

9. loeb 5, Bridges M, Bassok D, Fuller B. Rumberger RW. How much is too much? the influence of preschool centers on children's social and cognitive development. [con Educ Rev. 2007: 26(1):52-66.

10. Valenti JE. Tracey DH. Full-day. half-day, and no preschool effects on urban children's first-grade reading achievement. Educ Urban Soc. 2009;41(6): 695-711.

2134 JAMA November 26,2014 Volume312, Number20

11. Robin KB. Frede EC, Barnett WS.Is more better? the effects of full-day vs half-day preschool on early school achievement. New Brunswick. NJ: National Institute for Early Education Research: 2006. h ttp://nieer.org/resources/research /lsMoreBetter.pdf. January 23, 2012.

12. Reynolds AJ. Success in Eorly Intervention: The Chicago Child-Parent Centers Program and Youth Through Age 1S. Uncoln: University of Nebraska Press; 2000.

13. Reynolds AJ, Temple JA. Robertson DL, Mann EA. long-term effects of an early childhood intervention on educational achievement and juvenile arrest: a 15-year follow-up of low-income children in public schools. JAMA. 2001:285(18): 2339-2346.

14. Reynolds AJ, Temple JA, Ou SR, Arteaga lA. White BAB. School-based early chYdhood education and age-28 well-being: effects by timing, dosage. and subgroups. Science. 2011:333(6040): 360-364.

15. Reynolds AJ, Temple JA, White BA, ou S, Robertson Dl. Age-26 cost-benefit analysis of the Child-Parent Center early education program. Child Dev.2011:82(1):782-804.

16. Human Capital Research Collaborative. Implementation Fidelity of Midwest CPC Expansion: Yeors 1 ond 2. Minneapolis·. University of Minnesota; April2014.

17. Human Capital Research Collaborative. Program requirement and guidelines, Midwest Expansion of the Child-Parent Center Program, Preschool to Third Grade. Minneapolis~ Human Capital Research Collaborative; 2012. http://humancapitalrc.org /midwestcpc. March l, 2014.

18. Teaching Strategies Inc. Teaching Strategies GOLD Assessment System: Technical summary. Summary Findings of a Study Conducted by the Center for Educational MeDsurement & Evaluation. Charlotte·. University of North Caronna at Charlotte; 2011.

19. Lambert R, Kim D. Burts D. Technical Manual for the Teaching Strategies Gold Assessment System. 2nd ed. Charlotte: University of North Carolina Center for Educational Measurement & Evaluation; 2013.

20. Soderberg JS. Stull S, Cummings K. et al. Inter-rater reliability and concurrent validity study of the Washington Kindergarten Inventory of Developing Skills (WaKlDS). http://depts .washington.edl-1/cqel/IRCV.php. August 14. 2014.

21. Lambert R, Kim D. Burts D. Evidence for the Association Between Scores From the Teaching Strategies Gold Assessment System and Information from direct Assessments of Child Progress . . Charlotte: University of North Carolina Center for Educational Measurement & Evaluation; 2013.

22. lambert RG, Kim D. Burts DC. Using teacher ratings to track the growth and development of young children using the Teaching Strategies GOLD Assessment System.J Psychoed Assess. 2014;32 (1):27-39.

23. Balfanz R, Byrnes RV. Chronic Absenteeism: Summorizing What We Know From Notionally

Preschool Full or Half Days and School Readiness

Available Data. Baltimore, MD: Center for Social Organization of Schools. Johns Hopkins University; 2012.

24. Ou S, Mersky JP, ReynoldsAJ, Kohler KM. Alterable predictors of educational attainment. income, and crime: Findings from an inner-city cohort. Soc Serv Rev. 2007;81(1):8S-128.

25. Ou S, Reynolds AJ. Predictors of educational attainment in the Chicago Longitudinal Study. Sch Psycho/ Q. 2008;23(2): 199-229.

26. Reynolds AJ, Ou SR. Paths of effects from preschool to adult well-being: a confirmatory analysis of the child· parent center program. Child Dev. 2011;.82(2):555-582.

27. SPSS Inc. Advanced Statistics 22. New York, NY: IBM;2013.

28. Hoffman JP. Generalized LINEAR MODELS: An Applied Approach. New York, NY: Pearson; 2003.

29. Diggle P. Heagerty P, liang K. Zegler S. Analysis of Longitudinal Dota. 2nd ed. New York, NY~ Oxford; 2002.

30. Hubbard AE. Ahern J, Fleischer Nl. et al. To GEE or not to GEE: comparing population average and mixed models for estimating the associations between neighborhood risk factors and health. Epidemiology. 2010:21(4):467-474.

31. little RJ, Rubin DB. Statistical Analysis with Missing Data. New York. NY: Wiley: 1987.

32. Lipsey MW, Wilson DB. Practical Meta-Analysis. Thousand Oaks, CA: Sage; 2001.

33. Koh HK, Blakey CR. Roper AY. Healthy People 2020: a report card on the health of the nation. JAMA. 2014;311(24h247S-2476.

34. Yves H, Maltais C. Thompson K. Effects of a full-day preschool program on 4-year-old children. Early ChildRes Pract. 2007:9(2):1-20.

35. Reynolds AJ. One year of preschool intervention or two: does it matter? Eor/yChild Res Q.199S;10(1):1-31.

36. Arteaga f. Humpage S, Reynolds AJ, Temple JA. One year of preschool or two: is it important for adult outcomes? Econ Educ Rev. 2014;40:221-237.

37. Ramey a, Bryant OM, Wasik BH, Sparling JJ, Fendt KH, laVange LM. Infant Health and Development Program for low birth weight, premature infants~ program elements, family participation. and child intelligence. Pediatrics. 1992;.89(3):454-465.

38. Hill Jl, Brooks-Gunn J, Waldfogel J. Sustained effects of high participation in an early intervention for low-birth-weight premature infants. Dev Psycho/. 2003;39(4):730-744.

39. Reynolds AJ, Englund MM. Hayakawa C. et al. Assessing the Validity of the Minnesota School Readiness Indicators. St Paul: Minnesota Department of Education; 2011.

40. Reynolds AJ, Temple JA, Ou SR. et al. Effects of a school-based. early childhood intervention on adult health and well-being: a 19-year follow·up of low-income families. Arch Pediotr Ado/esc Med. 2007;161(8):730-739.

jama.com

Copyright 2014 American Medical Association. All rights reserved.

Downloaded From: http://jamanetwork.com/ by Jovanna Robs on 09/27/2017

Lessons from Research and the Classroom: Implementing High-Quality Pr~-K that Makes a Difference for Young Children

Jim Minervino Ready On Day One

Sept~mber 2014

BILL& MELINDA GATES foundation

Introduction ........................................................................................................................... 1

Quality in Center-Based Early Learning: High-Level Findings and Trends ........................... 3

Early Learning: The New Fact Base and Cost Sustainability ................................................ 8

The Essential Elements of High-Quality Pre-K: An Analysis of Four Exemplar Programs .. 21

~~~~{·roa· 'u• :c~r·o,n··. ~!~\L' .-i ~-'

In the United States today, a degree beyond high school paves the way for young people to support themselves, engage in their communities, and achieve their dreams. The Bill & Melinda Gates Foundation supports innovative approaches in K-12 public schools and higher education to ensure that students graduate from high school ready to succeed in postsecondary education and beyond.

Yet the path to a postsecondary degree begins before a child enters high school. High-quality early learning helps children enter kindergarten ready to learn and prepared to thrive in elementary school and beyond. Since 2005, the foundation has worked with public, private, and community partners to strengthen early learning in Washington State.

EARLY LEARNING IN WASHINGTON STATE We support policy development, research, and program implementation to improve the quality of early learning and strengthen the connections with K-12 schools, including:

• Evidence-based home visiting programs that pair at·risk families with trained professionals to provide support for pregnant mothers and families with infants and young children;

• Statewide implementation of Early Achievers, which offers coaching, support, and incentives for early learning providers to improve program quality;

• WaKJDS, which includes a whole-child, observational assessment and increased collaboration among early learning educators, families, and kindergarten teachers as children enter kindergarten; and

• PreK-3rd investments in Washington State school and educational service districts and their early learning partners to align quality instruction and promote deep engagement of educational leaders across the early learning/early elementary continuum-so that the gains young learners make are sustained and deepened as they advance into upper elementary grades and beyond.

Through close collaboration among policymakers, providers, and early childhood advocates, Washington State is building an early learning system that is focused on improving quality and helping young children enter kindergarten ready to learn.

Washington State is not alone. Over the past two decades, dozens of states have expanded early learning opportunities by providing pre-K to three- and four-year olds, launching home visiting services to expectant and new parents, and offering training and support to help child care providers increase the quality of their care. Since 2002, the percentage of four-year-old children participating in Head Start and state-funded pre-K increased from 26 to 38 percent. 1

While access to pre-K and other early learning programs has increased, a growing body of research shows that the quality of these programs has a significant impact on student outcomes. At the same time, the definition of what makes a program "high-quality" is far from settled.

1 Barnett, W.S., Carolan, M.E., Squires, J.H., Clarke Brown, K. (2013). The State o,{Preschoo/ 2013: State preschool yearbook. New Brunswick, NJ: National Institute for Early Education Research. http://nieer.org/sites/nieer/files/yearbook2013_executivesummary.pdf. Barnett, W.S., Robin, K.B., Hustedt, J.T., Schulman, K.L. (2003). The State of Preschool: 2003 State preschool yearbook. New Brunswick, NJ: National Institute for Early Education Research. http://nieer.org/sites/nieer/files/2003%20unsecured%20executive%20summary.pdf

EARLY LEARNING THAT STICKS Beginning in 2012, the foundation commissioned a series of research papers to better understand the changing early learning landscape. Specifically, these papers addressed whether pre-K programs produced gains in student achievement that persisted into the early elementary grades; whether high-quality programs could be cost sustainable; and what program features contribute to strong interactions between teachers and children, the central ingredient in high-quality programs.

The result is the research presented here, along with a companion piece on the early childhood workforce. These three papers are based on published research, program evaluations, and extensive consultation with experts in the early childhood field. The research was iterative. As a result, conclusions about the components of high-quality­including the importance of B.A. degrees for lead teachers-evolved based on new information. The papers are presented sequentially to show the evolution of the project over approximately 18 months, and should be read together.

• Quality in Center-Based Early Learning: High-Level Findings and Trends (January 2013)

• Early Learning: The New Fact Base and Cost Sustainabi/ity (September 2013)

• The Essential Elements of High-Quality Pre-K: An Analysis of Four Exemplar Programs (January 2014)

The research shows that select pre-K programs are producing significant and sustained gains for young children at a lower cost-per-child than previously attained. How these programs succeed is not a mystery. Exemplar programs focus on improving the interactions between teachers and children. They share a set of common features that foster high~quality instruction, support educators and young learners, and benefit from a positive enabling environment.

For some in the early childhood education field, these findings may not seem surprising. After all, research dating to the 1960s has found that high-quality early learning has long-term benefits for children, especially those from low­income families. However, the valuable lessons learned from high-quality programs can and should inform efforts to expand access to pre-K and help all children enter school ready to learn.

NEXT STEPS Looking ahead, the foundation will focus on listening and learning from the funders, advocates, teachers, parents, and early childhood leaders who have worked on these issues for many years. We intend to share our research broadly and collaborate with a wide range of partners to build on the work already underway to expand high-quality pre-Kin Washington State and nationally.

ACKNOWLEDGEMENTS The foundation is grateful to the early childhood researchers and program administrators listed below who shared their expertise and provided helpful input on this project. Inclusion in these acknowledgements does not imply agreement with all of the findings and conclusions of the research papers.

• Margaret Burchinal, Ph.D., Senior Scientist, FPG Child Development Institute, Research Professor, Psychology, University of North Carolina at Chapel Hill

• Steven Dow, Executive Director, CAP Tulsa

• Greg Duncan, Ph.D., Distinguished Professor, University of California, Irvine

• Gail E. Joseph, Ph.D., Associate Professor, Educational Psychology, University of Washington

• Anne Mitchell, President, Early Childhood Policy Research

• Ellen S. Peisner-Feinberg, Ph.D., Senior Scientist, FPG Child Development Institute, Research Associate Professor, School of Education, University ofNorth Carolina at Chapel Hill

• Bob Pianta, Ph.D., Dean, Curry School of Education, University of Virginia

• John R. Pruette, M.Ed., Executive Director, Office of Early Learning (Pre-K- Grade 3), North Carolina State Board of Education/Department of Public Instruction

• Jason Sachs, Ed. D., Director, Early Childhood, Boston Public Schools

• Ellen Wolock, Ed.D., Director, Division of Early Childhood Education, New Jersey Department of Education

Oualitv in Center.,Based Earlv Learnina: • J ~

High-Level Findings and Trends

Jim Minervino, Ready On Day One January 20132

The purpose ofthis paper is to present high-level findings and trends from a review of the literature and discussions with experts on the quality of early learning in center-based early care. This review focuses primarily on children from low-income families. 3

This work addresses seven questions:

• What level of quality must early learning4 deliver in order to achieve improved school-readiness outcomes for children?

• Is there an upper threshold beyond which increasing the quality of early learning stops producing increasing results?

• Do the immediate benefits of high-quality early learning always fade-out?

• Does having children spend more time in early learning programs yield any benefits?

• Do children in low-income families benefit more from high-quality early learning than their more advantaged peers?

• What are the features and environments that constitute high-quality early learning?

• Is the relationship between the quality of early learning programs and school readiness linear? Does it matter? What are the implications for increasing school readiness in Washington?

HIGH-LEVEL FINDINGS AND TRENDS

Research examining the effects of quality early learning and its impact on school readiness is very active. Over the past five years, the amount of high-quality research and program evaluations impacting this field has been substantial. It is not an overreach to say that a review of the literature from even as recently as 2007 would be out­of-date.

2 Recently published references have been added in some parts of this paper. The conclusions have not changed since the original January2013 paper.

3 Low-income families are defined as those families with total family income from all sources equal to 185 percent-or-less of current federal poverty guidelines. This income level coincides with federal eligibility requirements for low-income children to receive free- or reduced-price lunch.

4 For the purposes of this paper, the tenn ·'early learning" stands for all center-based care with an emphasis on children who will attend kindergarten the following year.

There appear to be three levels of early learning quality, each driving different levels of school readiness. Qualitatively, these ranges are poor-/low-, moderate-, and high-quality.' 6

Poor- or low-quality programs can have a negative impact on children's school readiness or, at best, no impact on school readiness. The detrimental effects (or no effect) of poor- or low-quality center-based early learning for children living in poverty has been established for preschool children, as well as fur younger children. The literature is fairly definitive on this point. (Vortruba-Drzal et al., 2004; Loeb et al., 2004; Yandell et al., 1990; Phillips eta!., 1994; NICHD, 1997).

Moderate-quality programs generally have little impact on children's school readiness. It is believed that most children in the U.S. in center-based programs are in moderate-quality care. For children from low-income families and those who are at-risk of not being prepared for kindergarten, this level of quality has been consistently shown to have very little impact on children's school readiness. It must be concluded that, while this level of care may increase school readiness for a child here and there, the incremental number of children moved from not being school-ready to being school-ready appears to be very small. Moderate-quality care maintains the status quo, thereby perpetuating low rates of kindergarten-readiness among low-income children-despite the fact that providing moderate-quality care is not an inexpensive proposition.

High-quality is the minimum necessary to reliably increase children's school readiness. Where the quality of center-based early learning is high, increases in school readiness generally follow. In experimental research on high-quality programs, effect sizes typically ranging from d = 0.05 to 0.30. In state program evaluations where quality delivery was high, effect sizes have been reported beyond this range and, at times, significantly beyond this range. 7

Therefore high-quality is the minimum level of quality that early learning must deliver in order to achieve improved school-readiness outcomes for children.

Having children spend more time (dosage) in high-quality settings also appears to yield improved academic outcomes for children. 8

There is no evidence of a very high quality threshold beyond which further improvements in early learning quality yield diminishing school readiness returns. Available evidence suggests that "higher-quality is better" at all points along the care-quality continuum. (Burchinal eta!., 2010; Blau et al., 2000). This is considered a settled fact and not an active area of exploration or research.

5 The field has struggled to identify specific cut-points on specific measures to delineate between these qualitative levels of quality. While there is a growing consensus in the field as to what features of care constitute poor-, moderate-, and high-quality, quantitatively measured and derived cut-points or thresholds have yet to be fixed.

6 See The Essential Elements of High-Quality Pre-K paper for more infonnation on the elements that comprise high-quality early learning.

7 The effectiveness of high-quality state and city early learning programs is discussed in a subsequent paper: Early Learning: The New Fact Base and Cost Sustainability.

8 Note that at-risk children spending multiple years in high-quality early learning is a rare occurrence today, not just in Washington State, but nationwide.

High-quality early learning has shown significant impacts not only on children's school readiness but on later life outcomes as well. Most ofthe evidence covering high-quality early learning and longer-term outcomes has focused on relatively high­cost, comprehensive, difficult-to-scale (expensive) interventions (e.g., Abecedarian, Perry Preschool). These have been widely covered and deeply researched, including studies following participants from these programs into adulthood.

While initial follow-up research indicated that program effects disappeared within a few years of these interventions (fade-out), long-term studies show participants having significantly improved life outcomes-from lower rates of incarceration to higher earnings and life satisfaction when compared to peers who did not participate in these programs. Furthermore, recent results from promising state and city early learning programs indicate that high­quality early learning can transcend fade-out through first grade and beyond.'

There is debate among funders and policymakers about whether high-quality early learning programs are cost­prohibitive I non-sustainable relative to the long-tenn funding appetites of public and private funders. 10

While having children in high-quality programs can facilitate gains in school-readiness, increasing the total time (dosage) children spend in high-quality care may produce significantly improved rates of school-readiness. 11

Recent well-designed, rigorous studies-as well as meta-analyses-have examined the impact of dosage of instruction on children's learning in full- versus half-day kindergarten (Hahn, et al., 2014; Cannon, et al., 2006; Walston, & West, 2004). The findings have shown that children who attend full-day kindergarten programs learn more than their half-day counterparts (Cooper, Allen, Patall, & Dent, 20 I 0; Lee, Burkam, Ready, Honigman, & Meisles, 2006). In a more recent study, Ramey, Ramey, and Stokes (2009) found a similar pattern of effects for full­day versus half-day pre-kindergarten. Children in the full-day programs demonstrated double the literacy gains compared to children who were in half-day programs (Wasik, et al, 2013). 12

Recent results from Educare conclude that more years of Educare attendance are associated with better school readiness and vocabulary skills (Yazejian, et al., 2012)." It should be noted that higher dosage of early learning programs of lower-quality do not appear to result in school-readiness gains for young children.

In summary, the combination of high-quality care delivered over multiple years shows promise as an accelerant of school-readiness rates.

9 The effectiveness of four exemplar high-quality state and city early learning programs~including effects transcending fade­out-is discussed in a subsequent paper: Early Learning: The New Fact Base and Cost Sustainability.

10 The cost of establishing and maintaining high-quality early learning programs is discussed in a subsequent paper: Early Learning: The New Fact Base and Cost Sustainability.

1 1 Research also notes that the prevalence of children-especially children in low-income families-receiving multiple years of high-quality care is a fairly rare occurrence. See The Essential Elements of High-Quality Pre-K paper for more on dosage.

12 ·'Intervention dosage in early childhood care and education: It's Complicated" (OPRE Research BriefOPRE 2013-15). Washington, DC: Office of Planning, Research and Evaluation, Administration for Children and Families, U.S. Department of Health and Human Services.

13 Most research looking at low-income children spending time in high-quality early learning shows that two years of early learning yield better outcomes than one year. However, the gains from two years of early learning are less than 2x the gains from one year of early learning.

Children in low-income families benefit more from high-quality early learning than do children in moderate- and high-income families. The majority of studies-though not all-as well as most experts examining this issue find that children in low­income families benefit the most from high-quality early learning (versus children from moderate- and high-income families) (Burchinal et al., 2000; Duncan et al., 2003; Gormley, 2007; Dearing et al., 2009). 14

While there is a lack of consensus as to what features and environments constitute high­quality early learning, there appears to be an emerging consensus around key principles and elements of high-quality programs. This consensus includes recognition that both process quality (experiences children have in early care settings) and structural quality (conditions that must be present, such as appropriate teacher/child ratios) are essential to high­quality care.

Aspects of process quality, especially the frequency, quality, and content focus 15 of teacher-child interactions, guided by proven, well-implemented curriculum, are seen as features of care that have a disproportionately large effect on school-readiness outcomes. 16

It appears that the relationship between care quality and school readiness is not linear. Related to quality of care thresholds, the shape of the care-quality I school-readiness curve impacts the size of school-readiness gains that can be achieved for a given unit of care­quality improvement. Since research shows that low- and moderate-quality early learning programs do little or nothing to improve school readiness, it appears that readiness impacts accelerate when they reach the threshold of high-quality. As a result, programs that improve from low- to moderate-quality may show little evidence of improved outcomes for children until programs reach the high-quality threshold.

Improvements in the quality of early learning programs in Washington State must be significant and sustained to result in measureable school-readiness gains. As average program quality starts increasing, initial improvements in school-readiness rates will be negligible. Only when a significant number of programs achieve high-quality will meaningful increases in school-readiness rates result. Therefore, efforts to improve the quality of early learning programs in Washington State must be sustained overtime.

CONCLUSION High-quality early learning is the minimum level of quality required to meaningfully increase children's school readiness. This level of care has also shown significant impacts not only on children's school readiness, but on later life outcomes as well. There is no evidence of a very-high-quality threshold beyond which further improvements in early care quality result in diminishing school-readiness returns.

Given the relationship between children's school readiness and the quality of early learning children receive, it is essential that: I) Low-quality programs be avoided or immediately improved to moderate-quality; 2) Moderate­quality programs be improved out of the moderate-quality range and into the high-quality range and; 3) High-quality programs continuously improve to yield greater school-readiness outcomes.

14 Separately, there is research indicating children from low-income families may realize larger cognitive gains when their pre-K classmates come from higher-income families (Schechter, 2007; Henry, 2007). Other research concludes children entering pre-K with more advanced language skiUs benefit more from sharing a classroom with advanced peers {Mashburn, 2009).

15 Content focus includes literacy, math, etc.

16 More detail provided in two papers: Early Learning: The New Fact Base and Cost Sustainability and The Essential Elements of High-Quality Pre-K.

REFERENCES Blau, D. M. (2000). The production of quality in child-care centers: Another look. Applied Developmental Science, 4(3), 136-147.

Burchinal, M.R., Roberts, J.E., Riggins, Jr., R., Zeisel, E.N. & Bryant, D. (2000). Relating quality of center-based child care to early cognitive and language development longitudinally. Child Development, 71(2), 339-357.

Burchinal, M., Vandergrift, N., Pianta, R., & Mashburn, A. (2010). Threshold analysis of association between child care quality and child outcomes for low-income children in pre-kindergarten programs. Early Childhood Research Quarterly, 25, 166-176.

Cannon, J. S., Jacknowitz, A. & Painter, G. (2006), Is full better than half? Examining the longitudinal effects of full-day kindergarten attendance. Journal of Policy Analysis and Management, 25: 299-321.

Cooper, H., Allen, A. B., Patall, E. A., & Dent, A. L. (2010). Effects of full-day kindergarten on academic achievement and social development. Review of Educational Research, 80(1), 34-70.

Dearing, E., McCartney, K., & Taylor, B. A. (2009). Does higher quality early child care promote low-income children's math and reading achievement in middle childhood? Child Development, 80(5), 1329-1349.

Duncan, G. J. (2003). Modeling the impacts of child care quality on children's preschool cogoitive development. Child Development, 74(5), 1454-1475.

Gormley, W. T. (2008). The Effects of Oklahoma's Pre-K Program on Hispanic Children*. Social Science Quarterly, 89(4), 916-936.

Lee, V. E., Burkam, D. T., Ready, D. D., Honigman, J., & Meisels, S. J. (2006). Full-Day versus Half-Day Kindergarten: In Which Program Do Children Learn More?. American Journal of Education, 112(2), 163-208.

Loeb, S., Fuller, B., Kagan, S.L. & Carrol, B. (2004). Child care in poor communities: Early learning effects of type, quality, and stability. Child Development, 75(1), 47-65.

Phillips, D. A., Voran, M., Kisker, E., Howes, C., & Whitebook, M. (1994). Child care for children in poverty: Opportunity or inequity?. Child Development, 65(2), 472-492.

Ramey, C. T., S. L. Ramey, & B. R. Stokes. "Research evidence about program dosage and student achievement: Effective public prekindergarten programs in Maryland and Louisiana." The promise of Pre-K (2009): 79- I 05.

Yandell, D. L., & Corasaniti, M.A. (1990). Variations in early child care: Do they predict subsequent social, emotional, and cogoitive differences?. Early Childhood Research Quarterly, 5(4), 555-572.

Votruba-Drzal, E., Coley, R.L. & Chase-Lansdale, L. (2004). Child care and low-income children's development: Direct and moderated effects. Child Development, 75(1 ), 296-3 I 2.

Walston, J., & West, J. (2004). Full-Day and Half-Day Kindergarten in the United States: Findings from the Early Childhood Longitudinal Study, Kindergarten Class of 1998-99. NCES 2004-078. National Center for Education Statistics.

Wasik, B. A., Mattera, S. K., Lloyd, C. M., & Boller, K. (2013)./ntervention Dosage in Early Childhood Care and Education: It's Complicated (No. 7804). Mathematica Policy Research.

Yazejian, N., & Bryant, D. M. (2012). Educare implementation study findings-August 2012. Retrieved from the University of North Carolina Chapel Hill, Frank Porter Graham Child Development Institute website: http://www.fpg.unc.edu/sites/default/files/resources/reports-and-policy-briefs/FPG% 20Demonstrating% 20Results.

Jim Minervino, Ready On Day One Dr. Robert Pianta, University of Virginia September 2013

INTRODUCTION Early learning holds enormous promise for young children and especially for young children in low-income families (Burchinal, 2000; Duncan, 2003; Magnuson, 2007; Howes, 2008; Dearing, 2009). However, most investments in early learning have not resulted in both significant and sustained academic gains for children. Where gains have been achieved, too often these gains have faded either in large part or completely within a few years (Barnett, 1995; Puma, 2012). And while a number of programs have transcended academic fade-out (e.g., Abecedarian, Perry Preschool), the cost-per-child of these programs is beyond what policyrnakers at the federal, state, and local levels can spend and sustain.

However, over the past few years, a new fact base has emerged. The result is a number of early learning programs that demonstrate significant and sustained gains for young children at a lower cost-per-child than previously attained.

This paper focuses on this new fact base and the role high-quality early learning plays in driving and sustaining gains for children-4:arly learning that sticks. Following this, from the academic literature, there is discussion of the most important elements of high-quality early learning programs-those program features that, when present, substantially increase the odds of early learning that sticks. Following this is an analysis of the cost-per-child of successful early learning programs and whether sufficient dollars may already be in the system.

TAXONOMY This analysis focuses on the approximately 5 million U.S. children, ages 3-5, from low-income families (The Annie E. Casey Foundation, 2013). 17 Other parts of the taxonomy are footnoted. 18

17 In this paper, low-income families are defined as those families with total family income from all sources equal to 185 percent­or-less of current federal poverty guidelines. This income level coincides with federal eligibility requirements for low-income children to receive free- or reduced-price lunch. Note: Head Start and most state- and city-run early learning programs typically have eligibility guidelines for total family income ranging from 100 percent to 200 percent of federal poverty guidelines. The term "lowest-income'' families or children refers to the 3 million children living in families at I 00 percent-or-below federal poverty guidelines.

18 Unless otherwise noted, "children" refers to low-income children as defined above. ''Early learning" refers to center-based programs. "Early learning" and '"pre-K" may be used interchangeably. Many sources use the tenns ''programs" and ''interventions" to refer to early learning offerings. This paper uses the term "programs.'' Unless otherwise noted, "gains" or "outcomes" refer to academic/achievement gains that children have made that increase their chances of being kindergarten­ready and on a trajectory to be ready for future grades. "Sustained gains" or "early learning that sticks'' describes achievement gains made at ages 3-5 that, mostly or in significant part, persist at least through 3rd grade.

REVIEWING THE LANDSCAPE While there have been many experimental successes, there have been few real-world, scaled-up early learning successes showing persistent gains for children. There have been even fewer successes when applying any measure of cost-sustainability. Furthermore:

I. It is estimated that half oflow-income children are not ready for the first day of kindergarten (Isaacs, 20 12). 2. Head Start academic gains have been very small, have not persisted (Puma, 2012), and cost-per-child is high. 19

3. There has been an inability to shift the substantial, existing dollars in the system to drive better outcomes for children. 20

THE NEW FACT BASE Rather than trotting out the "usual suspects" to show early learning's persistent impacts (e.g., Abecedarian, Perry Preschool}-none of which are cost-sustainable at scale--it is useful to look to the states, counties, and cities which have served as R&D labs for what works and what is cost-sustainable.

Some programs have operated long enough to measure gains, see these gains resist fade-out, and prove to be either cost-sustainable or cost-sustainable with modifications. The majority of performance data and measurement referenced here is very recent: 2010-2013. There is also a significantly better understanding ofthe program elements that contribute to early learning that sticks and how to deliver these in a way that is cost-sustainable. A discussion of those program elements, costs, and a cost-sustainability evaluation appears later in this paper.

State/City Programs with Outcomes That Stick21

Reviewing high-quality program evaluations of federal, state, and city-run early learning programs, there are at least four that work, stick, and are (or could be made) cost-sustainable:

1. New Jersey (Abbott Pre-K) 2. Boston Pre-K22

3. Maryland (Extended Elementary Education Program (EEEP) and "Judy Centers") 4. North Carolina (More at Four)

19 Head Start's initial design point was being a safe place offering nurturing care where low-income parents could leave their children while these parents worked. Especially in the last decade, Head Start's purpose has expanded to include early learning and developmentally appropriate achievement outcomes.

20 Total 2012 annual federal and state spending on early learning for all children ages 3-5 (not including kindergarten) is estimated at $21-$27 billion. Were these dollars spent exclusively on lowest-income children, all3 milJion of these children in the U.S. could receive high-quality early learning at cost-sustainable levels (in this case $7,000-$9,000 per-child).

21 Requirements for selection include: 1) Actual programs, not experiments; 2) Program up and running effectively for at least 7 years; 3) Running at scale across a state, county, or large city; 4) Measurability built-in and achieved with high-quality research design; 5) Evidence of marked improvements in achievement for low-income children; 6) Evidence that effects persist through­or beyond 3rd grade; 7) Use observation, quality ratings, and cycle those into coaching and professional development.

22 Boston Pre-K is open to all children but is overwhelmingly comprised of children from low-income families.

These programs demonstrate the following impacts:

Achievement Effect Sizes

Impact at a Glance

Children in Measurable Poverty Persistence Larger Year

_!ltro."_gll,,, ______ gains____ ~~ .... t~- .. __ J>rlm"'J'!IllJ'•cts _ -------- -------·- __ . ____ .(d)_ __ _

New Jersey 0.4013* 5th Grade

Boston 0.44-0.62 3rd Grade

Maryland24 • 4th Grade

Yes

Yes

Yes

1999

1998

2001 (Judy

See text

See text

See text

·--· ___ CentersL ___ ·------.-------·----------- ---·----Moderate-North Carolina 3rd Grade Yes See text

------·---·----·-·-·-~-Large ---·----·-----·--------------~_()_()_1__ _________ ····---------------··-

New Jersey (Abbott Pre-K). Target: Children in high-poverty areas. Large gains have been achieved in classroom quality and quality of instruction. Academic achievement gains at the end of the Abbott year show effect sizes averaging d=0.40. Students spending two years in Abbott Pre-K saw 50 percent larger gains at kindergarten entty than did students spending one year in the program. Subsequent program evaluation shows gains stick with Abbott Pre-K participants through 5th grade (latest measured so far) with effect sizes averaging d=0.24. Long-term effects are equivalent to a+ 10 percentile boost in state test scores. Abbott Pre-K is focused on quality improvement with an emphasis on quality of interactions in the classroom (Barnett, 2013).

Boston Pre-K. Target: Children in Boston. Two-thirds of enrolled children live in poverty. Boston Pre-K shows increases in children's end of year vocabulary d=0.44, early reading d=0.62, and numeracy d=0.59. These are among the largest impacts measured to date of any public pre-kindergarten operating at scale. Moderate improvements in working memory d=0.24 and impulse control d=0.28 may also contribute to academic gains. Larger than average gains have been made by children in poverty and by children whose primary language is Spanish. Boston Pre-K impact measured through 3rd grade (latest measured so far) shows math, literacy, and language skills of participants considerably more advanced than those of same-age children who did not attend Boston Pre-K. On Massachusetts' 3'' grade MCAS English Language Arts, 43 percent of program participants scored proficient or advanced compared to 34 percent of non-participants. Boston Pre-K is focused on the importance of quality interactions between teachers and children and a developmentally appropriate educationally-focused curriculum (Weiland, 2013).

Maryland (Extended Elementary Education Program (EEEP) and "Judy Centers.") Target: Primary focus on improving kindergarten readiness among children in poverty. EEEP has shown significant improvements in kindergarten readiness. Statewide, kindergarten "full readiness" is up 33 percentage points in ten years (Maryland State Department of Education, 2013).

23 Average effect size of vocabulary, literacy, and mathematics gains.

24 Long-term program impact currently under study.

Maryland Kindergarten Readiness: 2001-2012

100,.---------------i 00~----------~.~.---..,,-- ~ v

sol------,_//~./'~~ .. ,...,......- e

10~---------~~~------ ~

••f-----i" ..... ~"'-------­.. ~ ··~~~~------------------­

IlL~.~ ~f--~~~~r------------

30 1----·'ii_._·--;;::.-~---,jil:!!-------'""~. 20 ~---------"',-.::-..---..-- - Full

IO l--:':----r---,-----...,•,-·"'"''"'_=_··-~_"-41i-::-- """"' Approaching

o '-::cm~~"'_.,."'_"""_-,"'-..,·::•::-;-·---"'_·''_-_·~-:~:=::--'-'"·'~-=----,:':::=:-=-'··c::-·""'c::=:,-:;:'::-- :: ~=~:'g Olltll 04'05 01108 11112 12113

Sc.hooiYear

Sour~: M~rn:l SUte DepoJrtmeM of Eduution

NQtes.: ffgwes may not total 100% due to reouncl.ng. ~ "TtMdline" inditates 1he over•U I.IPWi•d pl'ogtt'l.tion of M.aryland'l [ompcnitt futl ri'Adil\l;!n k.>veh.

Long-term program impacts are currently under study. Maryland has made rapid and large, across-the-board gains on all measures of academic achievement through 4th grade.

North Carolina (More at Four). Target: 4-year-old children in poverty, some lower-middle-income children, and children with other risk factors. 25 More at Four's program evaluation finds significant academic achievement differences (moderate to large effect sizes) between program participants and non-participants26 with effects persisting into 3rd grade. Participation in the program is associated with higher sustained math and reading test scores for poor children, but not for non-poor children.

These successful programs all have a number of factors in common. They have all systematically improved outcomes for children by improving the elements of high-quality early learning that matter most: teaching quality, observation, measurement, feedback, coaching, professional development, curriculum based on standards that connect to kindergarten and beyond, and structure that supports these improvements.

HIGH-QUALITY "High-quality" has become mantra for early learning practitioners and policymakers over the past few years. However, the mantra around high-quality begs the question: What is high-quality? For many, the term has become a substitute for the difficult work of determining which program elements must be high-quality in order to have impact and resist fade-out, while being cost-sustainable.

A review of the literature on the elements of high-quality early learning is presented in the next section of this paper.

25 It is estimated that 90 percent of More at Four children qualify for free or reduced-price lunch. Other risk factors include low English proficiency, student disability, or chronic health condition.

26 Forthcoming research by Dr. EHen Peisner-Feinberg.

Important Elements of High-Quality"

Teachers Far and away, teachers account for the majority of students' achievement gains in early learning. Increasing teaching quality is the highest-impact investment that can be made.

Quality of Teacher-Child Interactions

The quality of teacher-child interactions is the mechanism responsible for learning in early care and education settings (Hamre, 2007; Mashburn, 2008).

There is now strong empirical evidence regarding a variety of teaching practices that can, and should, be the focus of classroom observations intended to measure and enhance teacher performance. The key ingredient of any classroom or school environment, with regard to learning and development, is the nature and quality of interaction between adults and children. Through careful research, significant headway has been made in descnbing and conceptualizing what teachers do in the classroom that results in learning. These can be organized into three broad domains of teaching practice that are linked to positive student outcomes: I) SociaVEmotional Support; 2) Organization/Management Support; and 3) Instructional Support.

Mounting evidence suggests that attending to each of these domains in classroom observations helps to define the impact of classroom experiences on student performance. Most importantly, empirical evidence suggests that when teachers use these types of practices, students learn more (Burchinal, 2010).

Quality oflnstructional Support

More specifically, the quality of instruction is the primary mechanism responsible for cognitive and achievement gains in early learning settings (Howes, 2008; Mashburn, 2008). There is now evidence showing teacher-child interactions are generally positive and emotionally supportive. The same evidence shows that classrooms are reasonably organized. What is most concerning is that the level of instructional quality-the kind of cognitive demands teachers embed in their interactions with children--is very low. The histogram below depicts this clearly.28

Moderate Quality H!gh Quality

Cfcu:s: Scoret

What's notable in this graph is the low level of developmentally appropriate instructional support present in pre-K classrooms at this level of scale and scope (including Head Start and state-funded pre-K). 29 This low level of instructional support means that teachers' interactions with children in pre-K are likely less frequent, and when they do occur, are low on content, with little or no attention to application

27 This section distills the elements of high-quality early learning from the academic literature available at publication. A subsequent paper, The Essential Elements of High·Qua/ity Pre~K, describes the essential elements from the point of view of exemplar early learning programs achieving superior results with young children.

28 Findings drawn from observations of over 10,000 pre-K classrooms across the U.S., using the Classroom Assessment Scoring System (CLASS}--the gold standard for observations and for accountability in Head Start and several state programs.

29 There is no difference in observed interactions (in any of the three domains) with regard to teacher experience or educational level. In fact teachers with an A.A./CD A on average look slightly better than those with a B.A./M.A. possibly because of the practical focus of the training.

(problem solving, thinking, reasoning), vocabulary, concepts, and understanding. This has huge (negative) downstream implications for children's perfonnance in vocabulary, comprehension, and math- and reduces the chance children will be grade-ready or meet standards in future years. 30

It is important to note, and will be discussed later, that there are professional development models (online coaching, web-based video library, online and in-person courses) that have been tested and replicated in randomized controlled trials (RCTs) that make teachers effective with great consistency. A number of practices and tools are needed to support these improvements. Surprisingly, many ofthese practices and tools are further along than most people realize.

Systematizing Quality: Observation, Measurement, and Teacher Feedback via CLASS

Even with young children, the quality of teacher-child interactions and the quality of instruction can be observed and reliably measured with assessments such as CLASS (Hamre, 2007; Pianta 2008a; Pianta, 2009). CLASS is being systematically integrated into pre-K delivery systems throughout the United States and other countries. It recently became the gold standard for quality measurement in Head Start (programs can be required to reapply for Head Start funding because of low CLASS scores) and in a number of state early learning programs. These policy initiatives have created a market for quality, in which CLASS is the standard. Thus it has potential to be a powerful lever for improvement. 31

All these resources (reliability training, professional development, quality assurance) are self-sustaining and fit in the cost-structure of existing funding streams and delivery models.

Improving Quality through Coaching and Professional Development Using Technology

Numerous studies have documented that professional development (PD) spending on teachers has 1ittle~to-no impact on teacher practice or outcomes for children (Y oon, 2007; Garet, 2001 ). 32

Conversely, results show that professional development for teachers that allows them to see and label effective interactions results in improved quality of instruction and improved academic outcomes for children (Clements, 2008; Hill, 2008; Pianta, 2008b; Powell, 201 0). In addition to observing teacher behaviors that matter most for student learning, such observations allow for the design and testing of professional development models that produce those teacher behaviors. This reinforces the notion of focus on defined, observable, and valid examples of effective teaching as a starting point.

As CLASS-specific definitions of interactions provide a target for professional development, three modes of CLASS-based professional development have been created:33

• Online Coaching- Ongoing analysis/feedback on teacher-child interactions. Scales well and works.

• Web-Based Video Library- Analysis of others' interactions to see exemplars.

• In-Person and Online Course -Improves teachers' knowledge and analytic skills.34

30 Associations between teacher-child interactions and children's gains in school readiness on standardized tests show both linear effects (e.g., more is better) and nonlinear, or threshold effects. With respect to threshold effects, classrooms in which teachers score a three or above on CLASS [Range l-7] (classrooms in which the teachers' interactions show signs of cognitive demand and conceptual focus) there is a stronger and significant association with positive outcomes. Importantly, below a score of three, there is no association between teacher behavior and child learning-it's as if these classrooms add no va1ue. These are the majority of pre-K classrooms.

31 Scale-up of CLASS is occuning through a private company, Teachstone, devoted to delivery, scale, quality control, and self­sustaining programs for training observers to score reliably (assessment) and professional development (improvement) of teacher-child interactions.

32 Estimates on the amount of this spending range from $2,500-$9,000 per-teacher, per-year.

33 It appears that the use of technology may increase teacher effectiveness while reducing PD costs-per-teacher.

Each of these modes has been tested and found effective and replicated in RCTs. As with many other elements of quality in early learning. the skills derived from teachers experiencing this professional development benefit children in poverty more than other children. 35 Research and practice show that early-in-career teachers benefit even more from this professional development than do other teachers.

CLASS and Outcomes for Children

Evidence exists from more than 10,000 classroom observations36 demonstrating that young children whose teachers rate higher on CLASS are learning at a faster rate. Results consistently show small-to-moderate effect sizes (d~ between .I 0 and .20). Instructional and emotional quality predict more positive achievement and improved social outcomes. There are stronger effects for certain groups of children (effect sizes- 0.50), including children from low-income families and children born to mothers with low levels of education. Effects of pre-K interactions persist into later grades. 37 38

CLASS is an important development in the quest to improve teaching quality and outcomes for children. However, as a single measure of quality, it cannot account for all aspects of high-quality early learning.

Proven Curriculum

Improving teaching quality and implementing CLASS is essentiaL but not enough. This improvement must be combined with proven curricula focused on building the right skills in early learners (e.g., literacy, math, behavioral). Curricula should embed optimal classroom practices within it and teachers must be well-trained on the chosen cuniculum to ensure it is implemented as designed. This combination of teaching quality, proven cunicula,

and teachers well-trained on curriculum may yield the biggest gains in student achievement, and the gains most

likely to stick. 39 Great curricula matters.40

34 The CLASS-based course is essentially a standard college course designed for delivery by faculty. There is a standardized manual for instructors and a set of online videos they access. It focuses on building teachers' skills in describing and observing interactions. In an RCT involving trained instructors (two days) across 10 higher education sites (community colleges and state universities), teachers who took the course improved in Instructional Support by more than 1.5 CLASS scale points, on average, at the end of the course, and this effect was still significant one year later. An online version of this course is being piloted as a MOOC, with 10,000 users signed up.

35 Teachers with online coaches grew more sensitive in interactions with students and increased students' engagement in instruction. Improved language stimulation techniques are especially beneficial to high-poverty classrooms. Children with trained teachers made greater gains in tests of early literacy, had lower levels of problem behavior, and demonstrated higher levels of expressive language.

36 Designs isolate effects controlling for other influences, including family demographics, prior performance, and teacher/school effects.

37 Most notably: 1) Children in pre-K classrooms offering higher levels of Instructional Support displayed better language skills at the end of the kindergarten year; 2) Kindergarten Instructional Support scores made an independent contribution to gains in children's language and math abilities; 3) A one-point difference in obseJVed instructional supports appears linked to shifts in child outcomes; 4) Even into first grade, the academic achievement benefits of being in a classroom that rates high on Instructional Support disproportionately benefits children from homes where their mothers have low levels of education.

38 A virtuous circle is possible: standardized quality measures (like CLASS) now being built into pre·K and Head Start create "market signals" demanding that quality scores improve, leading to repurposing of coaching and professional development dollars that will lead to increases in quality, resulting in better outcomes for children in Head Start and pre-K.

39 Engel (2013) illustrates that most early grade (K-3) curricula are not well-designed and not connected to early learning standards. They further note that kindergarten teachers spend most of their time teaching skills most children already know and do not build on gains children made in pre-K. Engel believes this plays at least some role in the fading out of early gains children make.

Early Learning Standards

Over the last decade, almost all states have developed early learning standards or have improved on existing early learning standards (Barnett, 2012). Many states are now focused on connecting these standards to their K-12 systems to ensure a more seamless transition for children.

Standardized Outcomes Measures for Children

Federal- and state-funded early learning programs are coming under increased scrutiny and being pushed toward more accountability. Most states are responding by adopting systematic and proven ways to measure kindergarten­readiness and future-grade readiness that map to early learning standards. 41 These measurements are providing guidance to teachers and administrators as to where kindergarten-readiness gaps exist. In addition to these new measures, states are using a combination of teacher and student observation, teacher-administered tests, and program evaluation scores as input by which achievement outcomes are being measured.

Structural Quality

Pre-K programs exhibit two categories of quality: process quality and structural quality. Process quality refers to elements of quality, such as teacher-child interactions, and the quality of instruction-issues covered earlier in this paper. Structural quality refers to classroom characteristics such as group size, teacher·child ratios, teacher and staff education/training/certification, and length ofthe early learning day.

Understanding which elements of structural quality most impact child outcomes is critical in two ways. First, it increases the chances that desired outcomes will be achieved. Second, elements of structural quality are far and away the most significant cost drivers in early learning. Those costs are discussed in the next section.

Summarizing the academic literature on elements of structural quality in early learning:

1. Pre·K and kindergarten class sizes above 20 students are genera11y associated with poorer outcomes for children (Barnett, 2004).42

2. Likewise, teacher-child ratios above 2:20 (one lead teacher, one aide, and 20 children in a classroom) are associated with poorer outcomes for young children. Almost all high-quality early learning programs, including all of the programs featured in this paper with outcomes that stick, have teacher­child ratios of2:20 or better.43

Class size and teacher-child ratios tie to the earlier discussion about teaching quality. These elements make it possible (necessary, but not sufficient) for teachers to have the time available for high-quality teacher-child interactions and high-quality instruction, while allowing children to explore and play either independently or in small groups.

1. The data on the importance of teacher degree attainment and certification is murkier. As far as lead teachers are concerned, credible research supports the "B.A. is required for high-quality teaching

40 Pre-K curricula such as Dr. Doug Clements' (University of Denver) "Building Blocks" (numeracy/math) have proven themselves in RCTs and such examples are embedded in successful early learning programs.

41 Most states are (appropriately) choosing to view kindergarten-readiness via a comprehensive set of domains. These domains generally include health and physical well-being, social and emotional skiJls, and cognitive skills. Increasingly, states are further segmenting these domains into finer classification (e.g., cognitive skills, including approaches to learning, language, and literacy).

42 Even after controlling for factors that might correlate with large class sizes (e.g., family income in pre-K area, teaching quality, etc.).

43 Excepting Boston Pre-K which has a teacher-child ratio of2:22. New Jersey's Abbott pre-K operates with a 2:15 teacher-child ratio.

hypothesis" (Bueno, 2010), as well as the "B.A. is not required and is very weakly correlated with better outcomes than teachers with lesser credentials can achieve" hypothesis (Early, 2007). 44 45

2. The formal education and degree attainment required of teacher aides/paraprofessionals is much lower. 46

3. High-quality programs, including the ones highlighted in this analysis, typically offer 6-6.5 hours of early learning per day, five days a week, 180 days a year.

Given that the majority of the costs of providing early learning are the costs of teacher/aide salaries and benefits, the required elements of structural quality--even assuming a teacher-child ratio of 2:20 and a lead teacher with an associate degree---place a floor on how low cost-per-child costs can be driven in early learning while still achieving outcomes that stick for children.

Costs {All Costs Are Per-Child, Per-Year) Well-known, effective programs, not cost-sustainable, not broadly-delivered:47

Educare48 $11,000- $23,000

Abecedarian $16,000 - $40,000

Perry Preschool $20,000

Large programs of mixed quality have broader cost ranges:

Head Start (e.g., WA) $9,000

U.S. State-Run Pre-K Avg. $7,80049

W A State-Run Pre-K (ECEAP) $7,000

Full-Time Licensed Care (WA) $5,000-$12,00050

44 While the academic literature may be split, all of the programs reviewed in this paper require lead teachers to have a B.A. Boston Pre-K goes a step further, requiring their teachers to have an M.A. within five years of commencing teaching.

45 Rather than focusing on degree attainment, it may be most important that early learning teachers be trained and proficient in establishing social-emotional warmth, creating high-quality teacher-child interactions, and delivering high-quality instruction based on established standards and proven curricula. There is widespread recognition that teacher preparation programs need to be strengthened (regardless of degree earned). Ongoing research suggests components of teacher preparation (e.g., revamped courses and online pre-setvice coaching) can be improved and yield more effective teachers-though these components have yet to achieve scale.

46 A Child Development Associate (CDA) is standard and requires a high school diploma/GED, 120 hours of formal early childhood education training, and 480 hours of professional experience with pre-K children ages 3-5.

47 AU costs in 2012 dollars unless otherwise specified.

48 Educare's cost per child per year varies according to location and includes extensive, comprehensive setvices and longer hours of care than most other early learning programs.

49 The definitive study is Levin and Schwartz, 2007. Costs are in 2013 dollars.

50 Early learning is less expensive in rural areas, but more likely to be traditional day care without an instructional component.

[ i ;,; ·:I ··.'.h' •d' · '·i Cl'()!

Aforementioned early learning programs that work, stick, and are (or could be made) cost-sustainable: 51

New Jersey (Abbott Pre-K) $12,000-$14,90052

Boston Pre-K $12,00053

Maryland (EEEP Initiative) $9,800

North Carolina (More at Four) $8,500

Cost Analysis: Sustainability and Implications

Given the previously discussed elements of quality required for early learning success, and given what federal, state and local governments are already spending on early learning, a cost-sustainability target averaging $8,000-$10,000 per-child, per-year is feasible."

The table below shows early learning program cost options, assuming a 48-state average compensation and benefits package, with various assumptions as to lead teacher degree and class size.

Six Hour Pre-K Program: Estimated Annual Per-Child Costs in 2013 Dollars5556

Teacher Qualifications

B.A. I

B.A. Il 57 58

A.A.

_ $10,05_0

..... _$8,950

..... $_7,<)50

51 At a cost-sustainable range of $8,000-$1 0,000/year/child.

Class Size

-- _$9!2_00 _

$8,2_00

--~8,250

$7,4()0

52 Wage and benefit structure in ~20 states would support a New Jersey (Abbott)-type program at the top of the $8,000-$1 0,000/year/child cost-sustainability target.

53 Wage and benefit structure in -30 states would support a Boston Pre-K-type program within the $8,000-$10,000/year/child cost-sustainability target.

54 In some states, early learning teachers with B.A.'s are paid on the same salary and benefits scale as K-3 or K-12 teachers (e.g., New Jersey, Boston/Massachusetts). In high-cost, high-teacher salary/benefits states, solely the cost of one teacher and one aide in a classroom can exceed $6,000 per-child, per-year.In other states, early learning teachers with B.A.'s are paid anywhere from 10 to 40 percent less than K-12 teachers. Early learning teachers with associate degrees or lesser degrees are rarely paid more than $20/hour outside of high-cost states and cities.

55 Source: Gault, 2008. Estimates converted into 2013 dollars.

56 For 180-day school year, aide@ $15/hour inclusive of benefits, compensation- 60 percent+ of program costs.

57 B.A. required but paid at 12 percent discount to K-3 staff with B.A. degrees.

58 Research and practice indicate that when the pay gap between K-12 teachers with a B.A. and early learning teachers with a B.A. reaches somewhere in the neighborhood of25 percent or more, early learning teachers often leave the field for jobs in the K-12 system. Large pay disparities between the early learning system and the K-12 system also undermine time and effort invested in the coaching and professional development of early learning teachers-at least as far as children in early learning reaping the benefits of having higher-quality instruction.

The Dollars (in Large Part) Are in the System

As noted earlier. federal, state, and local governments are already making substantial investments in early learning totaling $21-27 billion annually. 59 Were those dollars to be strictly focused on early learning for the 3 million lowest-income children in the U.S., $7,000-$9,000 per child would be available.60 Though repurposing these dollars is no easy matter, the amount of existing dollars in the system would comprise all or a large portion of the early learning cost-sustainability target of $8,000-$10,000 per child. This repurposing would also cover all costs to systematically implement all of the observation, measurement, teacher feedback, proven curricula, establishing and connecting standards, and measuring child outcomes noted in this paper. 61

There are two immediate implications: There is substantial money already in the system and the bulk of it could be redirected to deliver high-quality early learning to children in low-income families." A focus on different practices (i.e., the essential elements of high-quality early learning) will get better results. Redirected dollars should focus on: 1) Systematic improvements in teaching quality via observation, measurement, teacher feedback, coaching, and professional development; 2) Use of proven curricula; 3) Improved teacher preparation focused on improving the quality of teacher-child interactions and quality of instruction; 4) Measuring outcomes for children and using that data to improve individual child outcomes and to inform overall instruction. Any additional money injected into the system should catalyze or go directly toward this same focus.

Progress in Early Learning: The New Fact Base at a Glance ·- ·-·-

Then (Pre 2010) Now (2013)

Expensive pre-K programs that don't scale. Cost-sustainable pre-K programs ($8,000-$10,000) with high potential for scaling.

Academic gains in pre-K are not sustained. Selected exemplar pre-K programs showing academic gains through elementary grades.

No consensus on how to create and deliver high-quality Field coalescing around the elements of high-quality that pre-K. drive the best outcomes for children.

Low accountability. More systematic, evidence-based, data-driven approaches. Room for innovation.

Insufficient focus on teacher-child interactions and Teachers matter most. Focus: teacher-child interactions quality of instruction, generally not measured. and high-quality instruction, broad adoption of CLASS

in-progress.

Lack of effective coaching and professional Coaching!PD, targeting instruction, using online development (PD), no models to get to scale. coaching, video, and in-person and online coursework.

- --Not enough focus on pre-K student outcomes. Increased use of data to measure outcomes, kindergarten

readiness assessments post-preschool, measuring Head Start grantees and dropping those that are low-quality.

Inconsistent adoption of whole child, comprehensive Early learning standards now the norm, moving quickly early learning standards. to connect to kindergarten standards.

Lack of proven curricula to boost student achievement. Proven curricula in literacy and numeracy exist (adoption of proven curricula remains low).

---

59 Head Start and Early Head Start comprise about $8-9 billion of this total.

60 Lowest-income in this case being children in families at 100 percent-or-below federal poverty guidelines.

61 Startup costs would add an estimated $400-$700 per child for the first year only.

62 Based on overwhelming evidence that high-quality early learning has the largest positive impacts on children in low-income families.

REFERENCES

Barnett, W. S. "Long-term effects of early childhood programs on cognitive and school outcomes." The future of children (1995): 25-50.

Barnett, W. S. "Class Size: What's the Best Fit?" NIEER Preschool Policy Matters (2004).

Barnett, W.S., Carolan, M.E., Fitzgerald, J., & Squires, J.H. (2012). The state ofpreschool20I2: State preschool yearbook. New Brunswick, NJ: National Institute for Early Education Research.

Barnett, W. S., Jung, K., Youn, M., & Frede, E. C. (2013). Abbott Preschool Program longitudinal effects study: Fifth grade follow-up. New Brunswick, NJ: National Institute for Early Education Research, Rutgers-The State University of New Jersey.

Bueno, M., Darling-Hammond, L., & Gonzales, D. (20 I 0). A Matter of Degrees: Preparing Teachers for the Pre-K Classroom. Education Reform Series. Pew Center on the States.

Burchinal, M.R., Roberts, J.E., Riggins, Jr., R., Zeisel, E.N., & Bryant, D. (2000). Relating quality of center-based child care to early cognitive and language development longitudinally. Child Development, 71(2), 339-357.

Burchinal, M., Vandergrift, N., Pianta, R., & Mashburn, A. (2010). Threshold analysis of association between child care quality and child outcomes for low-income children in pre-kindergarten programs. Early Childhood Research Quarterly, 25(2), 166-176.

Clements, D. H., & Sarama, J. (2008). Experimental evaluation of the effects of a research-based preschool mathematics curriculum. American Educational Research Journal.

Dearing, E., McCartney, K., & Taylor, B. A. (2009). Does higher quality early child care promote low-income children's math and reading achievement in middle childhood? Child Development, 80(5), 1329-1349.

Duncan, G. J. (2003). Modeling the impacts of child care quality on children's preschool cognitive development. Child Development, 74(5), 1454-1475.

Early, D. M., Maxwell, K. L., Burchinal, M., Alva, S., Bender, R. H., Bryant, D., ... & Zill, N. (2007). Teachers' education, classroom quality, and young children's academic skills: Results from seven studies of preschool programs. Child Development, 78(2), 558-580.

Engel, M., Claessens, A., & Finch, M.A. (2013). Teaching students what they already know? The (Mis) Aligrunent between mathematics instructional content and student knowledge in kindergarten. Educational Evaluation and Policy Analysis, 35(2), 157-178.

Garet, M., Porter, S., Andrew, C., & Desimone, L. (2001 ). What makes professional development effective? Results from a national sample of teachers. American Educational Research Journal, 38(4), 915-945.

Gault, B., Mitchell, A. W., & Williams, E. (2008). Meaningful Investments in Pre-K: Estimating the Per-Child Costs o.fQuality Programs. Pre-KNow Research Series. Pre-KNow.

Hahn, R. A., Rammohan, V ... (2014). Effects of Full-Day Kindergarten on the Long-Term Health Prospects of Children in Low-Income and RacialiEthnic-Minority Populations: A Community Guide Systematic Review. American Journal of Preventive Medicine, 46(3), 312-323.

Hamre, B. K., & Pianta, R. C. "Learning opportunities in preschool and early elementary classrooms." (2007). School readiness and the transition to kindergarten in the era of accountability., (pp. 49-83). Baltimore, MD, US: Paul H Brookes Publishing.

Hill, H. C., Blunk, M. L., Charalambous, C. Y., Lewis, J. M., Phelps, G. C., Sleep, L., & Ball, D. L. (2008). Mathematical Knowledge for Teaching and the Mathematical Quality of Instruction: An Exploratory Study. Cognition and Instruction, 26:4,430.

Howes, C., Burchinal, M., Pianta, R., Bryant, D., Early, D., Clifford, R., & Barbarin, 0. (2008). Ready to learn? Children's pre-academic achievement in pre-kindergarten programs. Early Childhood Research Quarterly, 23(1), 27-50.

Isaacs, J. B. (2012). Starting School at a Disadvantage: The School Readiness of Poor Children. The Social Genome Project. Center on Children and Families at Brookings.

Levin, H. M., & Schwartz, H. L. (2007, March). What is the cost of a preschool program?. In National Center for the study of Privatization in Education. Symposium conducted at the meeting of the AEFA Annual Conference, Baltimore, Maryland.

Magnuson, K. A., Ruhm, C., & Waldfogel, J. (2007). Does prekindergarten improve school preparation and performance?. Economics of Education Review, 26(1), 33-51.

Maryland State Department of Education (2013). The 2012-2013 Maryland School Readiness Report-Children Entering School Ready to Learn.

Mashburn, A. J., Pianta, R. C., Hamre, B. K., Downer, J. T., Barbarin, 0. A., Bryant, D., ... & Howes, C. (2008). Measures of classroom quality in prekindergarten and children's development of academic, language, and social skills. Child Developmellt, 79(3), 732-749.

Pianta, R. C., La Paro, K. M., & Hamre, B. K. (2008a). Classroom Assessment Scoring System (CLASS) pre-K version. Baltimore: Brookes.

Pianta, R. C., Mashburn, A. J., Downer, J. T., Hamre, B. K., & Justice, L. (2008b). Effects of web-mediated professional development resources on teacher-child interactions in pre-kindergarten classrooms. Early Childhood Research Quarterly, 23(4), 431-451.

Pianta, R. C., & Hamre, B. K. (2009). Conceptualization, measurement, and improvement of classroom processes: Standardized observation can leverage capacity. Educational Researcher, 38(2), 109-119.

Powell, D. R., Diamond, K. E., Burchinal, M. R., & Koehler, M. J. (2010). Effects of an early literacy professional development intervention on head start teachers and children. Journal of Educational Psychology, 102(2), 299.

Puma, M., Bell, S., Cook, R., Heid, C., Broene, P., Jenkins, F., ... & Downer, J. (2012). Third Grade Follow-Up to the Head Start Impact Study: Final Report. OPRE Report 2012-45. Administration for Children & Families.

The Annie E. Casey Foundation, KJDS COUNT Data Center, http://datacenter.kidscount.org/ Estimate interpolated for children ages 3-5 in families with incomes less than 150% of the federal poverty limit (FPL) and less than 200% FPL.

Weiland, C., & Yoshikawa, H. (2013). Impacts of a prekindergarten program on children's mathematics, langnage, literacy, executive function, and emotional skills. In Press. (Subsequently published: Child Development, 84(6), 2112-2130).

Yoon, K. S., Duncan, T., Lee, S. W. Y., Scarloss, B., & Shapley, K. L. (2007). Reviewing the Evidence on How Teacher Professional Development Affects Student Achievement. Issues & Answers. REL 2007-No. 033. Regional Educational Laboratory Southwest (NJI).

! he [ssential Elements of High~Quality Pre~K: An Analysis of Four Exemplar Programs

Jim Minervino, Ready On Day One January 2014

OVERVIEW This paper examines four state or local pre-K programs that are proven to drive sustained academic outcomes at cost-effective or near cost-effective levels for children in low-income families. 63 These programs include Boston Pre-K, New Jersey's Abbott pre-K,64 North Carolina's Pre-K (formerly More at Four), and Maryland's Extended Elementary Education Program (EEEP) and "Judy Centers."

This paper is a written summary delineating which elements of high-quality programs are necessary for early learning that sticks. 65

KEY FINDINGS66

High-quality early learning that sticks is the outcome of doing many things well. And it's hard to do all of the essential elements well. 67 A high-quality early learning system is akin to a long, multi-step supply chain where failure at any point along the chain significantly weakens the entire system. Each element must be executed very well and often in concert with other essential elements. At the highest level, this "doing many things well" requirement results in a high degree of difficulty and is a key reason why high-quality early learning that sticks is so infrequently seen.

The essential elements of quality that must be done well are featured in this paper. In addition, other contributors to high-quality programs are presented. While these other contributors might not be absolutely essential to achieving

63 In this paper, low~income families are defined as those families with total family income from all sources equal to 185 percent­or-less of current federal poverty guidelines. This income level coincides with federal eligibility requirements for low-income children to receive free- or reduced-price lunch. Note: Head Start and most state- and city-run early learning programs typically have eligibility guidelines for total family income ranging from I 00 percent to 200 percent of federal poverty guidelines. The term "lowest-income" families or children refers to the 3 million children living in families at 100 percent-or-below federal poverty guidelines.

64 Officially, New Jersey no longer refers to Abbott pre-K by that name. However the program is still referred to as Abbott pre-K by administrators and teachers.

65 The term "ear!( learning that sticks" describes achievement gains made at ages 3-5 that, mostly or in significant part. persist at least through 3 grade.

66 Through on-site visits and phone intetviews with these exemplars it became clear that, while the philosophy regarding the essential elements of high-quality programs are similar across the exemplars, specific choices within each essential element {e.g., specific curricula, professional development programs) differ across these exemplars and, in fact, this is an important finding of this paper.

67 For the purposes of this paper, the essential elements of high-quality pre-K that make learning stick wi11 often be shortened to simply "the essential elements.''

high-quality, these elements are sometimes found (at differing levels of maturation) in early learning programs that stick 68

The essential elements described in this paper are not novel, and to those involved in early learning, are unlikely to surprise. But in addition to recognizing the high degree of difficulty in executing all of the essential elements well, there is one finding that stands over the others: Seeing is believing and understanding.69

By definition, this "seeing" dooms all attempts to fully convey the essential elements in written form. The elements can be named and described and examples of best practices and processes can be documented, but to capture the essence of many of the essential elements-and more importantly to show others what doing each element "well" looks like--the essential elements must be shown in action rather than just described on paper. In that sense, while this paper highlights and describes the essential elements, it does not (and cannot) show them in action.

ESSENTIAL ELEMENTS OF HIGH-QUALITY PRE-K THAT STICKS As previously noted, high-quality early learning that sticks is the outcome of doing many essential things well. The remainder of this section discusses these essential elements. 70

1. Political leadership and/or judicial mandate and ongoing cover are necessary ingredients for creating, scaling up, providing resources, and maintaining high-quality early learning programs that stick. Two of the four exemplars-New Jersey and North Carolina-have pre-K systems operating under judicial mandate. As recently as 20 I 0, all four exemplars had strong political support for high-quality early learning-with the most important leadership coming from state governors (in the case of three state exemplars) and from Boston's mayor (in the case of Boston Pre-K). Support in the state legislatures and city councils of the exemplars has been important, but that importance pales in comparison to the impact of a highly-supportive governor or mayor.

At least three of the exemplars are benefitting from ongoing political leadership. New Jersey was poised to expand the Abbott Pre-K Program in 2013, but budget pressures have delayed that expansion. Still, expansion is supported by New Jersey's governor and increased funding is expected in future years. Maryland and Boston have continued to expand their early learning programs (spending and slots) at a modest rate, despite overall budgetary pressures.

This paper devotes a relatively large amount of space to political leadership and judicial mandate because this leadership is an absolutely essential ingredient for the creation and continuity of high-quality early learning.

2. A compe111ng vision coupled with strong leadership exhibited by the senior early learning leaders of these exemplar programs is an essential element in their success. Specific sub~elements include:

• Designing (or redesigning) an early learning program that includes all of the essential elements of high-quality.

• Forethought in program design and supporting elements so that the essential elements work in concert with one another and reinforce one another to create a true early learning system.

• Creating and driving a culture of high expectations (at all levels in the system down to and including children) and continuous improvement with a focus on end outcomes for children.

• Hiring and retaining senior leaders who share the same vision and are effective at all levels in the system.

• The ability to communicate compellingly all of the above to appropriate political leaders, as well as to those working within the early learning program. An added bonus is the ability to communicate all of the above to parents of children within the early learning system.

68 These other contributors are also included so that the reader can see the list of elements that fall short of making the list of absolute, must-have, essential elements of high-quality early learning.

69 "Seeing is believing and seeing is understanding" will often be shortened to "seeing."

70 The previous paper, Early Learning: The New Fact Base and Cost Sustainability, notes the importance of teacher-child interactions being positive and emotionally supportive and classrooms being reasonably organized. That same paper notes this is generally what is found in practice. For that reason, this paper takes these as a given and does not call these out as additional essential elements of high-quality early learning.

• Maintaining a strong and positive relationship with the political leaders (typically governors and/or mayors) who provide funding and their own leadership for early learning. 71

• Giving programs the time required to implement the essential elements of high quality before evaluating their success. 72

3. Teachers delivering high-quality instruction is a key differentiator between early learning that sticks and early learning that, more than likely, will not stick. Tomes have been written about high-quality instruction in pre-K, so for now, this paper includes a sampling of what the exemplars have to say on the issue:

Teachers with high expectations in a system with high expectations. Asking questions with a mix of simple, single-word answers and longer, more complex answers. Children doing more of the talking-and-doing than the teacher. Teachers guiding at the start of an activity and then facilitating continued student learning. Teachers stretching children's thinking by listening, re;flecting, and asking questions, such as, "How do you know that?" Asking compare and contrast questions, letting children fonn hypotheses, test them out, and report out to the class. Children gathering information using a variety of techniques (e.g., find in the classroom, create--draw, photograph, build~in the classroom and outdoors). Children reporting out on what they find via a story, a song, or a play. Teachers who understand pre-K is about leaJ71ing but also much more~sensitive and warm interactions, tapping into and extending children's natural curiosity, instilling a love of/earning by making learning feel like play, and instilling self-confidence that each child is an excellent learner.

If the words and phrases written above describing teachers delivering high-quality instruction seem to trivialize and not capture well what good teaching looks like, it is because there is no substitute for seeing high-quality instruction in action (live in-person or video). There is also no substitute for seeing and discussing the components of high-quality instruction with skilled teachers, coaches, and evaluators. Seeing is believing and understanding.

The following essential elements of high-quality early learning are related to structural elements and enablers that support finding, developing, and retaining the great teachers who drive superior outcomes. 73

4. All exemplar programs have two adults in the classroom-one lead teacher and one paraprofessionaVaide­at all times. 74 When individual children need attention (e.g., spilled paint, walk to the bathroom, Band-Aid) the paraprofessional can attend to a child's needs while group activities continue uninterrupted. This leads to more high-quality teacher-child interactions and better outcomes for children.

5. All exemplar programs have maximum class size of 22 children or fewer and adult-to-child ratios ranging from 2:15 to 2:22. Adult-to-child ratios at the lower end ofthe range are particularly advantageous for classrooms where a significant number of English language learners (ELLs) are present and/or where a significant number of children with special needs are present.

71 The exemplars note that changes in political leadership generate the most significant changes in funding, structure, and political support for early learning in their respective states and cities.

72 At least three exemplars indicated the first two years of implementation of new learning standards, new curriculum, new assessments, etc., were difficult, and it is unlikely that any significant gains in student outcomes were achieved in those years. Programs implementing the essential elements ofhigh~quality pre~K should take this into account when designing overall program evaluations. Leaders at all levels must be willing to wait and provide cover for these programs until their effectiveness is ready to be measured.

73 A safe and stimulating classroom with developmentally appropriate materials tied to the curriculum and instruction is necessary, but has not been given its own section in this paper. "Dosage" is also an essential structural e1ement that is discussed later in this paper.

74 A minority of classrooms have three adults in the room. The additional adult is often a student studying early childhood education and working as another aide in the room. In instances where paraprofessionals are on-track to receive a degree and/or credential, they may be allowed to function as lead instructor for a small amount oftime per day and always with a fully~ credentialed lead teacher in the room.

6. Lead teachers with a B.A. plus suitable early learning credential, paid at same level as K-3 teachers. While the previous paper, 75 based on the academic literature in the field, called into question the need for pre-K teachers to have a B.A., all of the exemplars require teachers to have a B.A. as well as early learning-specific credentials and/or experience. Furthermore, all leaders of these exemplar programs noted the B.A. requirement was not a close call for them. The exemplars' B.A. requirement stems from a number of factors:

• First, the "plasticity" required to teach 20+ four-year-olds for 6 hours/day-including the ability to orchestrate the children and classroom, meld standards with curriculum, continually assess young children in 5-7 domains, take-up professional development and apply it well, and combine all of this to deliver consistently high-quality instruction-is significant.

• Second, the exemplars are focused on the trajectories of young children-in particular, setting up children to succeed in reading, writing, math, science, and critical thinking. On average, teachers who have B.A.'s have larger vocabularies, more extensive language skills, better understand the building blocks that get children to higher trajectories, and can translate these into high-quality teacher-child interactions.

• Third, while there are certainly effective pre-K teachers in the U.S. without a B.A., the exemplars report that the odds of successful outcomes for children are higher when teachers have a four-year degree. And the presence of a B.A. and the education that stands for is a simple proxy to apply.

• Fourth, the exemplars believe there is a correlation between higher educational attainment and higher expectations. In other words, teachers with a superior education (B.A. as a proxy) will set higher expectations for themselves and for the children they teach.

• Fifth, regarding early learning credentials, exemplars note that K-12 teacher certification is inadequate for understanding early childhood development and how very young children learn. At least two of the exemplars pay the full cost to send pre-K teachers, effectively, back to school to get the training they need. As with a B.A., exemplars report requiring early learning credentials for pre-K teachers is not a close call.

• Sixth, lead teacher compensation must be set at or very near K-3 teacher compensation in a teacher's respective state. 76

Finally, as with the other essential elements presented in this paper, implementing one element-in this case hiring teachers with B.A. degrees-is not a silver bullet. High-quality early learning that sticks is the outcome of doing many things (the essential elements) wel1. 77

7. Dosage. Three of the four exemplars offer pre-K that runs 6-6.5 hours/day, for 180-205 days/year. The other (Maryland) offers full-day (6.5 hours/day, 180 days/year) and part-day (3 hours/day, 180 days/year) options. It is clear from the exemplars and consistent with research findings that within high-quality pre-K programs the dosage required is related to the size of the achievement gap that must be closed for each low-income child.

For low-income children who enter pre-K already on a trajectory to be kindergarten-ready, a high-quality part­day option may be sufficient. For most low-income children, at least one year in full-day, high-quality pre-K is needed to be kindergarten-ready.

For low-income children for whom English is not spoken at home, children with special needs, and children who are significantly below age-level competency in one or more domains, it is likely that two years ofhigh­quality, full-day pre-K is ideal and, in fact, may be necessary for most of these children to be kindergarten­ready on time.

75 See Early Learning: The New Fact Base and Cost Sustainability.

76 Three of the four exemplars mandate pay for all pre-K teachers, regardless of setting, at the same level as their K-8 counterparts statewide. These exemplars say pay equality is absolutely required for early learning that sticks and for retaining great teachers. However, one exemplar (Maryland) requires pay equality but pre-K teachers receive a benefits package that is significantly less generous than K-8 teachers in that state. Boston Pre-K is currently running an experiment whereby I 0 community-based providers (14 classrooms} are running the program but paying teachers about 20 percent less than district-run pre-K.

17 It is reasonable to infer that teachers with B.A. degrees in systems absent some of the other essential elements will find it more difficult to have impact than similarly educated teachers in systems having these elements present.

In instances where two years of high-quality pre-K is not available, there are two alternatives for children significantly behind their peers. The first is supplemental instruction (additional time, often one-on-one) for these children during the school year. The second alternative is a summer "wraparound program" that adds another 360 hours (60 days, 6 hours/day) ofhigh-quality learning in the summer before these children enter kindergarten.

The cost of supplemental instruction is highly-variable. The cost of a summer wraparound program typically runs about $4,000 per child.

The next section adds another five essential elements into the mix. In order to get the most impact from these five elements, they must work in concert with one another. Early learning programs must consciously design each individual element with the other four in mind.

8. All of the exemplars have had pre-K early learning standards for many years. The factors that appear to separate somewhat useful standards from excellent standards that drive better outcomes for children include: fewer standards rather than trying to be too comprehensive (fewer standards are more likely to be understood well and be used); standards covering multiple domains (e.g., health, social-emotional, cognitive); and standards that connect pre-K to kindergarten and beyond.

9. All of the exemplars have thought deeply about curriculum and believe it is a powerful tool when it is chosen well (i.e., it is research-based and can show proven results), tightly coupled with their early learning standards, connected to their system of professional development, and adhered to with very high fidelity. 78 In exemplar programs, the quality of teacher-child interactions is high and the content focus ofthose interactions is guided by the curriculum in use.

There are two different philosophies demonstrated by exemplars' approach to choosing curricula: a centralized, top-down selection and standardization on curriculum used throughout a pre-K system (Boston and Maryland);" or district- or provider-level choice among a group of three to five curricula vetted by the state (New Jersey and North Carolina). The curriculum selection model used in each program corresponds somewhat with each exemplar's pre-K delivery system and the power dynamics within those systems.80 Overall, the exemplars have been successful with two different approaches to curriculum choice. 81

While the exemplars have been successful in melding their curriculum with their early learning standards, they point out that training teachers thoroughly on new or revised curriculum is a time-consuming process-yet a process absolutely necessary and one that can never be shortchanged. 82 These exemplars operate with the knowledge that high-fidelity to curriculum can only be achieved through deep training, continuous monitoring, and the provision of actionable feedback directly to teachers. Ongoing professional development (PD) regarding standards, curriculum, child assessment, and improving the quality of instruction are important parts of the mix.

78 Exemplars recognize that strong curricula executed well--combined with high-quality instruction-is a very potent force for driving superior outcomes for children.

79 Boston Pre-K uses a self-modified version of"Opening the World of Learning'' (OWLs) for literacy and ·'Real Math Building Blocks" for math and numeracy. Maryland has created a custom curriculum.

80 The centralized curriculum approach used by Boston and Maryland corresponds with the fact that pre-K in both places is predominately (Maryland) or completely (Boston) delivered through the public school system. The ''choice of three to five" approach corresponds with New Jersey's and North Carolina's mixed-delivery ofpre-K that includes public schools, community providers, and Head Start. Exemplars with mixed delivery systems noted that the reality on the ground was that community providers were already using curriculum before becoming state/district pre-K providers, and the limited choice model recognizes this reality.

81 It is noteworthy that none of the exemplars use a "free choice of research-based curricula" approach. Exemplars emphasized the difficulty and complexity of this approach regarding administration and implementation of this approach with high fidelity and high quality.

82 Boston Pre-K emphasizes that a new curriculum takes, on average, three years to implement fully with high fidelity.

10. Professional development (PD) is a cornerstone of each exemplar's early learning system and, in this regard, the exemplars share similar characteristics. Primarily:

• The belief that the best proxy for a high-bandwidth teacher-one who is focused on self-improvement, able to understand, synthesize, and apply data, and quickly apply lessons learned from PD-is a teacher with a B.A.

• PD is most effective when a culture of high expectations for teachers and students is present. This type of culture views PD as an ongoing, never-ending cycle of self-improvement.

• PD that is focused on teacher-child interactions, quality of instruction, and student outcomes is most effective. This is true whether development is focused on pedagogy or content and is most useful when it addresses gaps in the quality of instruction, past/present student assessment, fidelity to curriculum and, if applicable, early learning standards.

• PD in the form of observation and one-to-one in-classroom (e.g., watch this, now apply immediately) coaching is most effective.

• Teachers value informal PD at least as much as-and sometimes more than-formal PD. Teachers in all exemplars note the high value of informal coaching and mentoring they have received from colleagues. Related to this, teachers in single-site, single-classroom settings (where no colleagues are present) highlight the lack of informal PD as the item most holding back their development and effectiveness.

• The "full-day, one-to-many, sitting down" format of formal PD is generally viewed as low- or no­value among administrators and teachers in the exemplar programs. This type of training is still occasionally used, but has been mostly discontinued or de-emphasized.

• The exemplars believe the jury is still out regarding the value of online, self-paced PD. No one seems to have •·cracked the code" on how to deliver/receive this in a way that has high-impact.

Exemplars note something obvious but worth relaying here: Investments in professional development are most valuable when they are effective (as described above) and when teacher retention is high. 83

1 1. For the exemplars, assessments and independent program evaluations are essential contributors to their success. Assessment occurs at every level: statewide/citywide program, district, teacher, and student. At the pre­K level, these assessments are almost exclusively formative, have high instructional validity, and by definition, are used to help actors in the system improve performance and outcomes for children. Independent program evaluation looks at overall program quality, outcomes, and effectiveness for children.

While many pre-K systems conduct assessments, exemplar programs consume them. Consistent with cultures of ongoing self-improvement and a focus on outcomes for children, exemplars make aggressive use of assessments to inform administrators, teachers, and students about what is working well and what needs to be fixed, while identifying teachers and students who need more help to improve.

Likewise, a systematic plan for program evaluation-including monitoring program implementation-is critical. The essential elements of high-quality early learning don't matter if implementation of these elements is poor. The exemplars recognize this and use program evaluation data to drive continuous improvement at all levels.

Finally, exemplars are focused on follow through. Using assessments and program evaluations to understand issues is one thing, but acting on the data is less common. Exemplar programs prioritize making the time, space, and resources to work directly with teachers and children who need extra support.

12. Overall, exemplars make aggressive use of data (e.g., assessment and program evaluation data noted above) and have built data systems to supply their early learning systems with the most important and actionable data available. This creates a data-rich environment where data is studied, drives recommendations, and most importantly, is used to drive action and superior outcomes for children.

83 Numerous studies have been done on the primary factors contributing to teacher retention. A unique factor specific to pre-K is the sizable compensation gap between pre-K teachers and K-12 teachers. All of the exemplars stressed the importance of compensating pre-K teachers near or at the same level as K-12 teachers in their systems.

13. The exemplars noted in this paper excel on all ofthe five elements noted above: standards, curriculum, professional development, assessment, and data/data systems. Furthermore, each exemplar has tied these together in a system where each supports and reinforces the others. This is not accidental. Instead, it is the result ofleadership, careful planning, and execution of strategies designed with each of the five elements in mind.

The remaining essential elements of high-quality early learning that sticks impact subsets of the pre-K population, but are nonetheless important features of exemplar programs.

14. English Language Learners (ELLs) (where applicable). Two exemplars-New Jersey's Abbott Districts and Boston Pre·K-have a large number of children from families where Spanish is the primary language spoken in the home. Both programs note the importance of having a specific strategy for these children, and both programs have chosen the same basic strategy: instruction that is 80-100 percent in a child's native language at the start of the school year; transitioning to 20-80 percent by the end ofpre-K; and then 100 percent English at the start of the following year (kindergarten)."

To execute this strategy, ideally, a bilingual teacher is leading these classrooms. Ifthat is not an option, a bilingual paraprofessional can bridge the gap between Spanish and English.

English language learners speaking languages other than Spanish continue to be a challenge for all of the exemplars.

15. Special Education. Exemplars try to mainstream students requiring special education as much as possible. This may take the form of placing these students immediately into mainstream pre-Kif they are judged ready, 85

placing students in mainstream pre-K but pulling them out for one-to-one instruction to help them catch up or stay caught up with their peers, or placement in a special education version of pre-K.

New Jersey's Abbott Districts have been the most aggressive (but not irresponsibly so) in mainstreaming children into kindergarten who were candidates for special education when entering pre-K. By spending the extra time, effort, and money--as much as $70,000 per child per year-New Jersey has reduced special education at kindergarten and beyond and has also reduced grade retention of these children significantly. 86

A REVIEW OF OTHER PROGRAM COMPONENTS While the 15 essential elements of high-quality early learning are shared across the four exemplar programs, other attributes are less common. Although these features very well may have an impact on program outcomes, they are not present in all four exemplars.

I. Setting87

2. Comprehensive health services88

3. District and school administration and site directors89

84 Exemplar programs have shown (as has Educare) that reaching these children at age 3 with instruction in their native language and transitioning to English in the following year results in a larger number of English language learners ready for kindergarten.

85 This may include embedding instruction on individualized goals within the mainstream classroom.

86 From the APPLES Evaluation of New Jersey pre-K: "'We also found that Abbott pre-K attendance reduced the likelihood of in­grade retention by 40 percent and reduced the necessity of special education placement by 31 percent." These reductions result in significantly lower per-student costs for all subsequent years and can recapture a good portion of the dollars spent on high­quality pre-K.

87 At the highest level, there are two delivery models for early learning programs that stick. One model is entirely situated in and run by the existing public school system (Boston). The second model is a mixed delivery system including public schools and community-based centers {for-profit and non-profit). This second model is run in the other three exemplars.

88 Exemplars generaHy have at least basic health screening and screening for learning disabilities.

89 Specifical1y, exemplars report that having district and school administration (especially principals in pre-K-8 public school systems) and site directors who understand child development and the essential elements of high-quality pre-K is a giant plus.

4. Significant outreach to parents, families, and the comrnunity90

5. Parent education 6. Reforming teacher education programs 7. Paraprofessional education and significant training 8. Space91

9. Common planning time for pre-K teachers to increase sharing of best practices 10. Quality Rating and Improvement System (QRIS)92

11. High-quality kindergarten" 12. Evidenced-based kindergarten transition practices 13. Classrooms with a mix oflow-income children with middle- or high-income children94

14. Program accreditation95

15. Hiring and training "interventionists"96

CONCLUSIONS This paper analyzed four exemplar pre-K programs and presented the program elements these exemplars share in common.

As noted in the previous paper, the cost to run these exemplar programs ranges from $8,500-$9,800 per child per year in North Carolina, and Maryland97 While the costs to run Boston Pre-K and New Jersey (Abbott) are higher, most states could run Boston's program and a fair number could run New Jersey's program within an $8,000-$10,000 per child per year cost envelope. 98

In the context of the essential elements presented here, the cost of these exemplar programs is important for at least two reasons. First, this cost range addresses a question policymakers and program designers often ask: Is there some level of funding required to deliver high-quality early learning that sticks? Second, exemplar program cost data

However, most of the exemplars report this is lacking or uneven within their programs. This is an ongoing fiustration, but exemplars are succeeding despite this.

90 All of the exemplars do outreach, but efforts are fairly scattered. In other words, exemplars are succeeding without doing high­quality outreach. A number of the exemplars are working to improve their outreach efforts and believe this will have a significant, positive impact on outcomes for children.

91 Finding appropriate space for pre-K programs is an issue for those delivered through the public school system and for programs that had a very fast ramp-up required by law. Finding space for supplemental instruction is also an issue.

92 While not citing QRIS as an essential element ofpre-K success, all exemplars have QRlS running or under development.

93 All exemplars lamented the current quality of their kindergarten programs relative to their pre-K programs. Some of the exemplars-most notably Boston Pre-K-are directly involved in up-leveling the quality of kindergarten and the early years of their K-12 systems. Not surprisingly, great teachers are the centerpiece regardless of grade.

94 Boston and New Jersey are strong advocates for this. They believe this boosts the vocabulary of low-income children in pre-K.

95 Boston Pre-K pays $6,000 per classroom per site for an average of three years to have each pre-K and kindergarten site accredited by NAEYC. Boston believes this forces added discipline to get their pre-K sites (all within the public school system) to high-quality. None of our other exemplars do this.

96 This is a promising idea pioneered by Boston Pre-K. Interventionists are used to spot and address problems-health, language, social-emotional, learning disabilities-immediately upon children's entry into pre-K. Boston likens this to ;'triage" in that those with the most pressing needs get immediate help and attention.

97 See Early Learning: The New Fact Base and Cost Sustainability.

98 The cost per child per year of Boston Pre-K is approximately $12,000; New Jersey (Abbott) is $12,000-$14,900, with the primary cost driver being relatively high wages and benefits in those locations. If lead pre-K teachers were paid at their states' K-3 pay scale, 30 states could run Boston Pre-K and 20 states could run New Jersey's (Abbott) program.

provides a yardstick against which to judge the costs and benefits of other existing or proposed early learning programs. 99

What is noted at the outset of this paper is worth repeating: High-quality early learning that sticks is the outcome of doing many essential things well. And it's hard to do all of these essential elements well. However, it can be done, and the exemplars provide examples of how this can be done cost-effectively at scale resulting in superior outcomes for children.

99 For example, at an average cost per child that falls squarely in the range of the exemplars, Head Start lacks a number of these essential elements, academic outcomes have been very small, and those outcomes have not persisted. Recently, Head Start has taken a number of steps to address these issues, including increasing classroom quality and measuring quality via CLASS, defunding grantees judged to be operating below expectations, and rewriting Head Start Performance Standards. When complete, Head Start's new standards will embrace most if not all of the essential elements presented in this paper.