DOI: 10.1126/science.1204526, 964 (2011);333 Science

David K. DickinsonChildrenTeachers' Language Practices and Academic Outcomes of Preschool

This copy is for your personal, non-commercial use only.

clicking here.colleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others

  here.following the guidelines

can be obtained byPermission to republish or repurpose articles or portions of articles

  ): April 25, 2014 www.sciencemag.org (this information is current as of

The following resources related to this article are available online at

http://www.sciencemag.org/content/333/6045/964.full.htmlversion of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/content/333/6045/964.full.html#relatedfound at:

can berelated to this article A list of selected additional articles on the Science Web sites

http://www.sciencemag.org/content/333/6045/964.full.html#ref-list-1, 2 of which can be accessed free:cites 27 articlesThis article

http://www.sciencemag.org/content/333/6045/964.full.html#related-urls3 articles hosted by HighWire Press; see:cited by This article has been

http://www.sciencemag.org/cgi/collection/educationEducation

subject collections:This article appears in the following

registered trademark of AAAS. is aScience2011 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience

on

Apr

il 25

, 201

4w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

o

n A

pril

25, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Apr

il 25

, 201

4w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

o

n A

pril

25, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Apr

il 25

, 201

4w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

34. L. S. Vygotsky, Mind in Society: The Development ofHigher Psychological Processes (Harvard Univ. Press,Cambridge, MA, 1978).

35. A. Diamond, W. S. Barnett, J. Thomas, S. Munro, Science318, 1387 (2007).

36. M. Montessori, Absorbent Mind (ABC-CLIO, Oxford, 1949).37. K. Lloyd, NAMTA J. 36, 165 (2011).38. P. A. Cohen, J. A. Kulik, C. L. C. Kulik, Am. Educ. Res. J.

19, 237 (1982).39. C. R. Greenwood, J. C. Delquadri, R. V. Hall, J. Educ.

Psychol. 81, 371 (1989).40. M. Mastropieri et al., Learn. Disabil. Res. Pract. 16, 18 (2001).41. A. Lillard, N. Else-Quest, Science 313, 1893 (2006).

42. C. A. Kusché, M. T. Greenberg, The PATHS Curriculum(Developmental Research and Programs, Seattle, WA, 1994).

43. N. R. Riggs, M. T. Greenberg, C. A. Kusché, M. A. Pentz,Prev. Sci. 7, 91 (2006).

44. C. Webster-Stratton, M. J. Reid, Infants Young Child. 17,96 (2004).

45. C. C. Raver et al., Early Child. Res. Q. 23, 10 (2008).46. C. C. Raver et al., Child Dev. 82, 362 (2011).47. C. P. Li-Grining, C. C. Raver, R. A. Pess, paper presented

at the Society for Research in Child Development BiennialMeeting, Montreal, Canada, 31 March to 2 April 2011.

48. T. O'Shaughnessy, K. L. Lane, F. M. Gresham,M. Beebe-Frankenberger, Rem. Spec. Edu. 24, 27 (2003).

49. A. F. Arnsten, Science 280, 1711 (1998).50. J. Cacioppo, W. Patrick, Loneliness: Human Nature and

the Need for Social Connection (Norton, New York, 2008).Acknowledgments: A.D. gratefully acknowledges grant

support from the National Institute on Drug Abuse(R01 grant no. DA19685) and the National Institute ofMental Health (R01 grant no. MH 071893) during thewriting of this paper.

Supporting Online Materialwww.sciencemag.org/cgi/content/full/333/6045/959/DC1Tables S1 and S2

10.1126/science.1204529

REVIEW

Teachers’ Language Practicesand Academic Outcomes ofPreschool ChildrenDavid K. Dickinson

Early childhood programs have long been known to be beneficial to children from low-incomebackgrounds, but recent studies have cast doubt on their ability to substantially increase therate of children’s academic achievement. This Review examines research on the role of languagein later reading, describes home and classroom factors that foster early language growth, andreviews research on preschool interventions. It argues that one reason interventions are nothaving as great an impact as desired is because they fail to substantially change the capacity ofteachers to support children’s language and associated conceptual knowledge.

Every year, large numbers of children failto complete high school, and the cost toboth the students and society is enor-

mous. It has been estimated that in 2007 in theUnited States, 16% of youth between ages 16and 24 were high school dropouts, and thatevery such student costs roughly $260,000 inlost earnings, taxes, and productivity (1). Read-ing success hinges on acquisition of a cluster oflanguage and print-related competencies in pre-school and the early primary grades (2). It alsois associated with competencies such as math-ematical ability and self-regulation (3), but herewe focus on language.

For many, the seeds of academic failure aresown early and are evident in early reading strug-gles. Despite sustained efforts by educators, anachievement gap persists between the readingskills of children from more and less advantagedhomes (4). Language ability at ages three andfour predicts later reading comprehension throughhigh school (5, 6), and later language ability buildsdirectly on earlier competencies. Differences inchildren’s language ability and associated capac-ity emerge early, relate to social demographic fac-tors, and foreshadow future reading success (Fig. 1).

In this article, I review research indicatingthat a major factor accounting for this reading

gap is the language competencies associated withliteracy. Between birth and school entrance, thereis rapid growth of language and associated com-petencies essential to later literacy (3), and learn-ing gaps associated with these competenciesrelate to social economic status (SES) (7, 8). Fac-tors in homes and classrooms partially accountfor differential language growth, and althoughpreschool programs have had some success inmeeting children’s needs, many have failed to helpteachers’ language-enhancing practices that areneeded to bolster language learning. Without bet-ter understanding of the mechanisms by whichprograms can foster teachers’ support for children’slanguage, we may continue to struggle to createprograms that reliably result in improved learning.

The Emergence of LiteracyAlthough language is a universal human capa-bility, the rate of acquisition of vocabulary (8)and syntax (9) varies and is associated with SES.By age three, there are substantial economic-related differences in receptive vocabulary that persistto age 13 (7) for all groups except African-Americanchildren, for whom this gap continues to growuntil school entry at age five. For example, the av-erage child enrolling in Head Start is roughly oneyear below national norms in receptive vocabulary.

Researchers are making progress in identify-ing how early home experiences account forearly differences in the rate at which childrenlearn new vocabulary and how these experi-

ences are influenced by SES and parenting. Par-ents gesture when communicating even beforetheir infants begin to talk, and the number ofdifferent gestures mothers use with their 14-month-old children is associated with the size ofchildren’s vocabulary at age 54 months. Moth-ers from higher SES backgrounds gesture morefrequently (10). Similarly, the amount of lan-guage that children hear in their first months,from their parents, affects their ability to quicklyunderstand words at 18 months (11), and thisspeed of lexical access predicts vocabulary sizein preschool (12) and as late as age eight (13).Additionally, the amount of verbal communica-tion by mothers when they interact with 30-month-old children also predicts the size of children’svocabularies a year later (8). In this study, whenmothers’ beliefs and knowledge about child de-velopment were taken into account, SES nolonger predicted children’s later verbal ability,indicating that parents from all backgrounds maybe able to provide appropriate language supportto young children but may not recognize thevalue of parenting strategies that support lan-guage learning. Language researchers also haveidentified a number of other features of interac-tions that support language learning through thepreschool years, such as the variety of wordsused and mother’s ability to be responsive tochildren’s efforts to talk and to extend and clar-ify what they say (14).

The pace of language growth before schoolentry also predicts other capacities that undergirdlater academic success. The emergence of self-regulation ability is a capacity linked to lateracademic achievement (3). Phonological aware-ness development, the ability to reflect on thesounds of language, is critical to decoding andalso is associated with language development inthe preschool period (15, 16). Age four lan-guage also predicts the ability to sound out wordsin kindergarten (17) and first grade (6).

Although the speed of language acquisitionis strongly affected by experience, genetic studiesindicate that roughly a third of the variability inlanguage and later reading is determined bygenetic factors (18, 19), leaving substantialopportunities for early interventions to havepositive impacts. Given that early environmentalfactors shape language, children at high risk ofeducational problems should begin to receive

Department of Teaching and Learning, Peabody College,Vanderbilt University, Nashville, TN 37203–5721, USA. E-mail:david.dickinson@vanderbilt.edu

19 AUGUST 2011 VOL 333 SCIENCE www.sciencemag.org964

intervention support in infancy. Progress is beingmade to devise effective approaches for usebetween birth and age three (20), but this paperfocuses on the later preschool years because thatis when most children are in programs.

The findings of child language researchershave guided studies conducted in classrooms. Twostudies examined the impact of language-enhancingvariables in classrooms (such as density of so-phisticated vocabulary, interactive conversations,and density of complex syntax). One study foundevidence that children in classrooms with teacherswho used relatively more complex syntax madegreater fall-spring growth in comprehension ofcomplex syntax (21) than did children in otherclassrooms. Another study examined languageuse across the day in preschool and correlatedthese language environments with language andliteracy through grade four (17). Children’s vo-cabulary, decoding, and comprehension skills inkindergarten and grade four were related to fea-tures of teacher-child conversation in preschool,with grade four outcomes mediated by kindergar-ten skills. Language-focused interventions I dis-cuss later have drawn on these studies and relatedliterature when identifying intervention goals.

Center-Based InterventionsIn 2008, 63% of all children in the United Statesbetween age three and five were enrolled insome form of group care (22). Whereas homefactors account for substantial variance in chil-

dren’s academic outcomes, center-based care canmake meaningful contributions (23). Large lon-gitudinal studies have found beneficial effects ofhigh-quality child care or preschool on children’sacademic performance in the middle elementarygrades (24), with these effects still present at age15 (25). Experimental studies designed to test thecapacity of center-based programs to redress theeffects of poverty have had noteworthy successesthat were still evident at age 21 (26). A programbegun in the preschool years also has showneffects on increased high school graduation ratesat age 20 and lower arrest records at age 18 (27).

Awareness of the impact of the early years ledto the creation of Project Head Start in 1965 andmore recently to state-supported pre-kindergarten(pre-K) programs. In the 1990s, pressure beganto build on Head Start to demonstrate an impacton children’s pre-academic development. Thisemphasis came as a shock to many who hadlong viewed instruction related to literacy as de-velopmentally inappropriate for three and fouryear olds (28). The impact of this pedagogicalstance was evident when a longitudinal studylaunched in 2000 found no significant growth inletter knowledge relative to age norms and onlymodest growth in vocabulary (29). A secondcohort studied in the 2006–2007 academic yearfound some improvement in letter knowledgeand vocabulary (30). Less encouraging resultscame from an experimental study that comparedchildren admitted to Head Start with a group

who applied but were not admitted. It foundonly modest fall-spring growth in print knowl-edge and a very small change in vocabulary forthe admitted children (31).

Evaluations of state-funded pre-K programshave yielded stronger results. In many cases,these programs are staffed by teachers with four-year degrees who are paid the same as ele-mentary school teachers. Evaluators have foundeffects on letter knowledge, phonological aware-ness (32, 33), and language and comprehension(34) in preschool and through the end of secondgrade (35). One study also found a decrease inthe number of children from low-income homesplaced in special education programs and a re-duction in kindergarten retention, two effects thatwill translate into enormous cost savings (33).

Correlational and experimental studies indi-cate that high-quality child care typically is bet-ter for children than what occurs when childrenreceive whatever care their parents can arrange.However, these studies fail to identify criticalfeatures of classrooms that help account for thebeneficial effects.

Intervention EffortsRecently, researchers have sought to intervenein existing programs using curriculum or pro-fessional development to increase the benefits ofprograms beyond those typically seen. They haveencountered considerable challenges.

Those who have used curriculum as the pri-mary leverage point have had limited success inthe improvement of language and literacy skills.A meta-analysis of 13 major curriculum-basedinterventions studies failed to find any evidenceof effects on language or print-based outcomes(36). An example of these challenges comesfrom my own recent work. My research groupintroduced a curriculum designed to foster en-hanced language in a Head Start program andcompared fall-spring growth between practice-as-usual and two intervention conditions (37).Few notable effects were found, only some ofwhich were positive, and there was no con-sistent pattern of results.

Research assistants videotaped classrooms soas to assess fidelity of implementation and tran-scribed and coded for fidelity of implementationand language use. Fidelity of implementation waswell below the level we hoped to achieve. Forexample, our efforts resulted in only 38% fidelityto core intervention instructional elements duringbook reading versus 16% of these elements in thecontrol groups. The frequency of talk about wordmeanings, a specific target of the intervention,was generally low and varied dramatically. Acrossfive settings and about an hour of transcribeddata from 44 classrooms, teachers averaged 67.3comments about vocabulary (such as referenc-ing words and word meanings). This averagemay seem high, but because we coded eachutterance by children and teachers, a single

Birth to age threeBasic vocabulary and syntax.

Lexical processing capacities develop.

Enjoyment of books and word play.

Language used to form relationships.

Three to KindergartenAcademic vocabulary acquired.

Advanced syntax develops.

Understand stories read aloud.

Phonological awareness emerges.

Letter knowledge solidified.

Emergent forms of readingand writing appear.

SES–based gaps emerge

Clear gap in receptive vocabulary by age 3.

Speed of processing differences linked to language exposure by 18 months.

SES and race-related gaps increase

Complex syntax use diverges by SES.Vocabulary gap widens for African-Americans.Classrooms fail to narrowlanguage-related gaps.

Upper Elementary

Extensive academic vocabulary.

Broad conceptual knowledge from varied academic domains.

Understand and use syntax found in academic writing. Comprehend and producenarrative and expository texts. Reflect on words and theirmeanings.

SES-related gap in academic skillsReading comprehension lags for children with low language skills.

The vocabulary and related conceptual knowledge gap stabilizes but is not reduced.

Fig. 1. Developmental progression in language acquisition and emergence of achievement gapsassociated with income and race.

www.sciencemag.org SCIENCE VOL 333 19 AUGUST 2011 965

SPECIALSECTION

conversation about the meaning of one wordcould easily include six or more utterances. Therewas great variability; one teacher had only sixutterances talking about words, whereas anotherhad 268, suggesting that individuals’ discoursepatterns heavily influence their approach to con-versing with children.

This is despite the low fidelity to desired in-structional elements of book reading, which wasthe setting in which our team has found evidencethat curriculum affected language use. One fac-tor that seemed to be important was that the in-tervention curriculum provided scripted guidancethat encouraged talk about vocabulary and dis-

cussions about the meanings of stories. Use ofthis guidance was reinforced in professional de-velopment sessions and by classroom coaches. In-dicators of language climate suggested that therewere some effects of the intervention on use ofsophisticated words (intervention, 14.2; versus con-trol, 6.9) and the total number of sophisticatedwords used (28 versus 13.4). However, I suspectthat our intervention failed to achieve sufficient-ly high density of such talk to have had substan-tial effects on word learning. Careful review ofvideotapes of classrooms also leads our teamto believe that teachers in the intervention hadproblems teaching lessons about science-related

content. Ongoing analyses suggest that theseshortcomings in part reflect limitations in teach-ers’ grasp of the material they were to teach (Fig. 2).When the teacher attempts to explain the con-cept of fading, she states that the white shirt willfade, confusing fading with what happens whendark clothes lose color and fade and the re-sulting dye stains other light-colored clothes. Shemay fully understand the concept, but this diffi-culty in communicating it is illustrative of otherinteractions we observed, in which teachers failedto communicate needed knowledge clearly.

Somewhat similar results have been reportedby others. Justice and colleagues asked 135

Fig. 2. Teacher explanations of science-related phenomena in two class-rooms. This teacher uses a nonstandard variant of English known as African-American Vernacular English. Teachers in this program varied in the degree to

which they used such forms, and we are studying its impact on children’slearning. Initial analyses found no evidence that variability in a teacher’s useof this form was related to children’s language or literacy growth in preschool.

And so I go down to the washer, and she says, "while you in there, throw Love's shirt in there with you?" But Love's shirt is what color?

White. Her shirt is white, and my shirt is black. So if I was to put my black shirt in the water with Love's white shirt, Love's white shirt going to do what?

What's the word?

Fade, it's going to fade.

White! Fade!

It will dry. All the air around it will help it evaporate and dry.

It melted and made a water spot.

What happened to my snowball?

It melted and made what?

You can't wear it no more.It melted Water...

Becauseit's wet. It will dry.

"I can't wear this anymore," Samir said. Why not?

If I leave ithere, what will happen?

Conversation during a unit on color. This lesson includedshirts as part of an imaginary process of putting clothes in the washing machine.

Teacher read a book in which a character put a snowball in his pocket that melted.

She put a snowball in an apron and left it for a time then discussed the result.

Science-related conversation in a pre-K classroom

Science-related conversation in a Head Start classroom

19 AUGUST 2011 VOL 333 SCIENCE www.sciencemag.org966

teachers in pre-K classrooms to use a 36-weekscripted curriculum focused on literacy instruc-tion (38). They introduced the program in a pro-fessional development session, supported teacherswith a Web site that included videos of effectivepractice, and provided all the books and materialsthat were required. Teachers delivered the requiredinstructional procedures with very high levels offidelity, but the quality of instruction was gen-erally low. For example, high-quality use strat-egies known to foster language growth wereseen in only 8% of the teachers, whereas low-quality use was found in 40%.

One problem faced by the kind of interven-tions I have discussed may be that they askteachers to change too many aspects of theirpractice at one time. Other interventions arenarrowly targeted to specific skills such as vo-cabulary or phonemic awareness. They have amuch stronger record of success (2, 39). In par-ticular, book reading has been the focus of manysuccessful interventions (2). These successes sug-gest that by limiting the range of settings andstrategies, teachers make educationally meaningfulchanges in how they read and discuss books.

Some interventions have sought to improveclassroom instruction using professional devel-opment and coaching to deepen teachers’ knowl-edge of early literacy and to foster use of effectivestrategies in implementing teachers’ current cur-riculum. These efforts have had more successthan comprehensive approaches. An interventiondelivered to 10 Head Start teachers focused onlyon book reading for a full year and includedprofessional development, modeling, and on-sitecoaching support from expert coaches. It led tosubstantial changes in teaching and impressivelanguage outcomes (40). A different larger-scaleeffort lasted 15 weeks, included 88 teachers, andprovided a mixture of professional development,web-based information, and coaching. Teachersreceived seven coaching sessions, with half of theteachers having coaches visit them in their class-rooms, whereas the other half had coaches whoviewed videotapes and provided coaching fromoff-site. Children in both conditions showed signif-icant growth in print knowledge relative to con-trol classrooms, but there was no significant growthin language. This result was related to the fact thatresearchers found no greater improvement in useof recommended language-supporting practicesin intervention relative to control group teachers.

These professional development-based ef-forts may have been successful in part becausethey achieved a threshold of instructional qualityneeded to bolster children’s learning. Analysis ofmultiple data sets has found that there is not alinear relationship between quality and learning;a threshold may need to be reached (41). Currentwork at Vanderbilt being done with 13 local pre-Kclassrooms supports this notion. This EarlyReading First–supported effort is using the samecurriculum that failed to show effects in the

study conducted in Head Start study, but wehave had coaches in teachers’ classrooms weekly.Multiple measures of classroom quality indicatehigh levels of instruction, and ongoing coding ofvideotapes of conversations has revealed pat-terns of talk in these classrooms that are sub-stantially richer than those from our prior effort(Fig. 2). Indeed, I have had to create new codesfor conceptually rich talk not previously seen. Aregression discontinuity analysis found small tomoderate positive gains for all outcomes and asignificant effect on print knowledge (42). Ourteam will increase the power of this study byadding one more cohort this year. This experi-ence suggests that the combination of curriculum,intensive coaching, and well-qualified teachers canresult in substantial improvements across the year.

ConclusionThe evidence is now in that early language abil-ity lays the groundwork for later literacy andthat the speed of acquisition of these abilities isassociated with powerful demographic factors.Children from at-risk homes fare better in high-quality centers, and interventions designed forthree- and four-year old children have some-times improved alphabetic knowledge andphonological awareness, but the success of theseinterventions in bolstering language abilities hasbeen sporadic. Changing teacher practices re-lated to language use is proving to be nearly ashard as raising children’s performance levels. Wesuspect that the coming years will be a timewhen researchers look more closely at interac-tions in classrooms and strive to create profes-sional development, coaching, and curricula thatresult in substantial improvements in teachers’methods of fostering language learning.

References1. L. Fiester, R. Smith, “Early warning! Why reading by the

end of third grade matters” (Annie E. Casey Foundation,Baltimore, MD, 2010).

2. G. J. Duncan et al., Dev. Psychol. 43, 1428 (2007).3. G. J. Duncan et al., Dev. Psychol. 43, 1428 (2007).4. B. D. Rampey, G. S. Dione, P. I. Donahue, “NAEP 2008

Trends in Academic Progress (NCES 2009-479)” (NationalCenter for Educational Sciences, U.S. Department ofEducation, Washington, DC, 2009).

5. NICHD Early Child Care Research Network, Dev. Psychol.41, 428 (2005).

6. L. Verhoeven, J. van Leeuwe, A. Vermeer, Sci. Stud. Read.15, 8 (2011).

7. G. Farkas, K. Beron, Soc. Sci. Res. 33, 464 (2004).8. M. Vasilyeva, H. Waterfall, J. Huttenlocher, Dev. Sci.

11, 84 (2008).9. M. Vasilyeva, H. Waterfall, J. Huttenlocher, Dev. Sci.

11, 84 (2008).10. M. L. Rowe, S. Goldin-Meadow, Science 323, 951

(2009).11. N. Hurtado, V. A. Marchman, A. Fernald, Dev. Sci. 11,

F31 (2008).12. A. Fernald, A. Perfors, V. A. Marchman, Dev. Psychol.

42, 98 (2006).13. V. A. Marchman, A. Fernald, Dev. Sci. 11, F9 (2008).14. E. Hoff, Dev. Rev. 26, 55 (2006).15. M. Silven, E. Poskiparta, P. Niemi, M. Voeten, J. Educ.

Psychol. 99, 516 (2007).

16. J. Metsala, in Handbook of Early Literacy Research,S. B. Neuman, D. K. Dickinson, Eds. (Guilford, New York,2011), vol. III, pp. 66–84.

17. D. K. Dickinson, M. V. Porche, Child Dev. 82, 870 (2011).18. N. Forget-Dubois et al., Child Dev. 80, 736 (2009).19. N. Harlaar, M. E. Hayiou-Thomas, P. S. Dale, R. Plomin,

J. Speech Lang. Hear. Res. 51, 688 (2008).20. A. L. Mendelsohn, B. P. Dreyer, C. A. Brockmeyer,

S. B. Berkule-Silberman, L. M. Morrow, in Handbook ofEarly Literacy Research, S. B. Neuman, D. K. Dickinson,Eds. (Guildford, New York, 2011), vol. III, pp. 279–294.

21. J. Huttenlocher, M. Vasilyeva, E. Cymerman, S. Levine,Cognit. Psychol. 45, 337 (2002).

22. National Center for Educational Sciences, Digest ofEducation Statistics (U.S. Department of Education,Washington, DC, 2009).

23. National Institute of Child Health and HumanDevelopment Early Child Care Research Network,Child Dev. 71, 960 (2000).

24. K. Sylva et al., in Handbook of Early Literacy Research,S. B. Neuman, D. K. Dickinson, Eds. (Guilford, New York,2011), vol. III, pp. 97–117.

25. D. L. Vandell, J. Belsky, M. Burchinal, L. Steinberg,N. Vandergrift, NICHD Early Child Care Research Network,Child Dev. 81, 737 (2010).

26. F. A. Campbell et al., Early Child. Res. Q. 23, 452 (2008).27. A. J. Reynolds, S. R. Ou, J. W. Topitzes, Child Dev. 75,

1299 (2004).28. D. K. Dickinson, Educ. Res. 31, 26 (2002).29. N. Zill et al., “Head Start Performance Measures Center

Family and Child Experiences Survey (FACES 2000):Technical Report” (U.S. Department of Health andHuman Services, Administration for Children andFamilies, Head Start, Washington, DC, 2006).

30. L. Hulsey, N. Aikens, Y. Xue, L. B. Tarullo, J. West,“ACF-OPRE Report: Data Tables for FACES 2006”(U.S. Department of Health and Human Services,Administration for Children and Families, Office of Planning,Research and Evaluation, Washington, DC., 2010).

31. U.S. Department of Health and Human Services, “HeadStart Impact Study. Final Report” (Administration forChildren and Families, Washington, DC, 2010).

32. W. T. Gormley Jr., T. Gayer, D. Phillips, B. Dawson,Dev. Psychol. 41, 872 (2005).

33. C. T. Ramey, S. L. Ramey, B. R. Stokes, in The Promise ofPre-K, R. Pianta, C. Howes, Eds. (Brookes, Baltimore, MD,2009), pp. 79–105.

34. M. Lipsey, D. C. Farran, C. Bilbrey, K. G. Hofer, N. Dong,“Intitial Results of the Evaluation of the TennesseeVoluntary Pre-K Program” (Peabody Research Institute,Vanderbilt University, Nashville, TN, 2011).

35. E. Frede, K. Jung, W. S. Barnett, A. Figueras, “The APPLESBlossom: Abbott Preschool Program Longitudinal EffectsStudy (APPLES): Preliminary Results through 2nd Grade”(National Institute for Early Education Research, 2009).

36. C. L. Darrow, “Language and literacy effects of curriculuminterventions for preschools serving economicallydisadvantaged children,” paper presented at theCampbell Collaborative Joint Symposium, George MasonUniversity, Fairfax, VA, August 2011.

37. D. K. Dickinson, A. P. Kaiser, M. Roberts, K. G. Hofer,C. L. Darrow, “The Effects of Two Language-FocusedPreschool Curricula on Children’s Achievement throughFirst Grade,” paper presented at the Society for Researchon Educational Effectiveness, Washington, DC, March 2011.

38. L. M. Justice, A. J. Mashburn, B. K. Hamre, R. C. Pianta,Early Child. Res. Q. 23, 51 (2008).

39. L. M. Maulis, S. B. Neuman, Rev. Educ. Res. 80, 300 (2010).40. B. A. Wasik, M. A. Bond, A. Hindman, J. Educ. Psychol.

98, 63 (2006).41. M. Burchinal, N. Vandergrift, R. Pianta, A. Mashburn,

Early Child. Res. Q. 25, 166 (2010).42. S. J. Wilson, “Program evalutation of an early reading

first project using the regression discontinuity design,”paper presented at the Society for Research inEducational Effectiveness, Washington, DC, March 2011.

10.1126/science.1204526

www.sciencemag.org SCIENCE VOL 333 19 AUGUST 2011 967

SPECIALSECTION