ContentslistsavailableatScienceDirect
Early
Childhood
Research
Quarterly
The
language
of
play:
Developing
preschool
vocabulary
through
play
following
shared
book-reading
Tamara
Spiewak
Toub
a,
Brenna
Hassinger-Das
a,∗,
Kimberly
Turner
Nesbitt
b,1,
Hande
Ilgaz
a,2,
Deena
Skolnick
Weisberg
a,3,
Kathy
Hirsh-Pasek
a,
Roberta
Michnick
Golinkoff
c,
Ageliki
Nicolopoulou
d,
David
K.
Dickinson
eaDepartmentofPsychology,TempleUniversity,1701North13thStreet,6thFloorWeissHall,Philadelphia,PA19122,UnitedStates
bDepartmentofTeachingandLearning,PeabodyCollege,VanderbiltUniversity,230Peabody,230AppletonPlace,Nashville,TN37203,UnitedStates cSchoolofEducation,UniversityofDelaware,206WillardHall,Newark,DE19716,UnitedStates
dDepartmentofPsychology,LehighUniversity,17MemorialDriveEast,Chandler-UllmanHall,Bethlehem,PA18015,UnitedStates ePeabodyResearchInstitute,VanderbiltUniversity,230AppletonPlace,PMB181,Nashville,TN37203,UnitedStates
a
r
t
i
c
l
e
i
n
f
o
Articlehistory: Received29June2017
Receivedinrevisedform8January2018 Accepted29January2018
Availableonline31May2018 Keywords: Vocabulary Low-income Preschoolclassrooms Book-reading Play
a
b
s
t
r
a
c
t
Twostudiesexploredtheroleofplayinavocabularyinterventionforlow-incomepreschoolers.Both stud-iespresentednewvocabularythroughbook-readings.Study1children(N=249;Mage=59.19months)
werealsorandomlyassignedtoparticipateinFreePlay,GuidedPlay,orDirectedPlaywithtoysrelating tothebooks.GuidedandDirectedPlayconditionsinvolveddifferentstylesofadultsupport.Although childreninallconditionsshowedsignificantgainsinknowledgeoftargetvocabularywords,childrenin bothadult-supportedconditionsshowedsignificantlygreatergainsthanchildrenexperiencingFreePlay. InStudy2,classroomteachersimplementedourproceduresinsteadofresearchers.Allchildren(N=101; Mage=58.65months)reviewedhalfthevocabularywordsthroughahybridofguidedanddirectedplay
andhalfthewordsthroughapicturecardreviewactivity.Childrenshowedsignificantpre-topost-test gainsonreceptiveandexpressiveknowledgeforbothsetsoftaughtwords,buttheyalsoshowed signif-icantlygreaterexpressivevocabularygainsforwordsreviewedthroughplay.Theseresultssuggestthat thereareuniquebenefitsofadult-supportedplay-basedactivitiesforearlyvocabularygrowth.
©2018ElsevierInc.Allrightsreserved.
Oneproverbstates,“AllworkandnoplaymakesJacka dull boy”(Howell,1659).Yet,“AllplayandnoworkmakesJackamere toy”(Edgeworth,1825,p.155).Thispoeticpaircapturesnotonly thebenefitsof children’splay but also itslimitations. Timefor playin classroomsis dwindling,as educatorsandpolicymakers emphasizeotheractivitiesbelievedtobemoreeducationally effec-tive(Elkind, 2008;Miller &Almon,2009).However,datashow
∗ Correspondingauthor.Presentaddress:41ParkRow,Room1324,NewYork,NY 10038,UnitedStates.
E-mailaddresses:tamara.spiewak.toub@temple.edu (T.S.Toub),bhassingerdas@pace.edu(B.Hassinger-Das),
Kimberly.Nesbitt@unh.edu(K.T.Nesbitt),hande.ilgaz@bilkent.edu.tr(H.Ilgaz), deena.weisberg@psych.upenn.edu(D.S.Weisberg),khirshpa@temple.edu (K.Hirsh-Pasek),Roberta@udel.edu(R.M.Golinkoff),agn3@lehigh.edu (A.Nicolopoulou),david.dickinson@vanderbilt.edu(D.K.Dickinson).
1 Presentaddress:DepartmentofHumanDevelopmentandFamilyStudies,
Uni-versityofNewHampshire,217PetteeHall,Durham,NH03824,UnitedStates.
2 Presentaddress:DepartmentofPsychology,BilkentUniversity,06800Bilkent,
Ankara,Turkey.
3 Presentaddress:DepartmentofPsychology,UniversityofPennsylvania,425S.
UniversityAve.,Philadelphia,PA19104,UnitedStates.
thatplayfullearningpromoteslanguage,cognitive,andsocialskills thatarecriticalforacademicsuccess(e.g.,Hirsh-Pasek,Golinkoff, Berk,&Singer,2009;Roskos&Christie,2000).Althougharecent review of the research on the developmental benefits of play (Lillardetal.,2013)providedanimportantcritique,theconcerns arenotequallyapplicabletoalltypesofplayorall developmen-taloutcomes(Nicolopoulou&Ilgaz,2013).Tobalanceconflicting notionsoftheroleofplayinchildren’sdevelopmentand educa-tion,weneedtounderstandwhichtypesofplayactivitiesrelateto whichoutcomes.AspartofthebroaderRead-Play-Learnproject, twostudiespresentedhereexploredvariousapproachesto lever-agingplayalongsidemoreestablishedbook-readingmethodsto promotevocabularygrowthinlow-incomepreschoolers.Play pro-videsawaytoscaffoldchildren’svocabularydevelopmentthrough engagementwithwordsinmeaningfulcontexts.Aplayfulreview of newvocabulary itemsdifferssubstantially from thedidactic teachingmethodscommonlyusedinpreschoolclassrooms(Early et al., 2010).Here we focus on vocabulary introducedthrough book-readingsessionsandevaluatewhetherplayfulapproaches activitiestowordlearningassistlow-incomepreschoolersin aug-mentingtheirvocabularyknowledge.We alsoexaminewhether https://doi.org/10.1016/j.ecresq.2018.01.010
playismoreeffectivethanadirectinstructionapproachto supple-mentingthevocabularysupportoccurringwithinbook-reading.
1. Promotingvocabularygrowthinlow-income
preschoolers
Earlyvocabularypowerfullypredictschildren’slaterlanguage development,readingskills,school-readiness,andacademic suc-cess (Dickinson & Porche, 2011; Morgan, Farkas, Hillemeier, Hammer,&Maczuga,2015;Rowe,Raudenbush,&Goldin-Meadow, 2012).Forexample,StorchandWhitehurst(2002)found statisti-callysignificantindirecteffectsofpreschoolorallanguageskills (e.g.,receptiveandexpressivevocabulary)onreadingabilitiesin Grades1–4. Also, rates of early vocabulary growth predictthe structureof children’sbrainsyearslater, withchanges in corti-calregionsrelevanttocontinuedlanguagedevelopment(Asaridou, Demir-Lira,Goldin-Meadow,&Small,2017).
Althoughfosteringvocabularydevelopmentisimportantforall children,it isespecially crucialtoexaminetrajectoriesfor chil-drenfromlow-incomefamilies(Farkas&Beron,2004;Snow,Burns, &Griffin,1998).Thequantityandquality oflanguageinputfor childrenfromlow-incomefamiliesistypicallylessthanfortheir higher-incomepeers(Hart&Risley,1995;Hindman,Wasik,&Snell, 2016;Hoff,2013;Rowe,2008).Bothquantityandqualityofearly languageinputplayalargeroleinlanguagedevelopment(Cartmill etal.,2013;Goldin-Meadowetal.,2014;Hirsh-Pasek,Adamson et al., 2015), and quality often mediates the relation between socioeconomicstatusandlanguageoutcomes(Bracken&Fischel, 2008).Enhancinglanguageenvironmentsinformalchildcare or preschoolinitiatives(e.g.,HeadStart,universalpre-K)isone poten-tialvenueforprovidinglanguagesupporttochildrenotherwiseat risk(Vernon-Feagans,Bratsch-Hines,&TheFamilyLifeProjectKey Investigators,2013),andthatiswhatweaimedtodointhecurrent study.
1.1. Vocabularydevelopmentinearlychildhoodprograms
Muchworkneedstobedonetoensurethatearlychildhood programsareproviding high-qualitydevelopmentalsupportfor allchildren.Numerousstudieshavefoundlow-qualitylanguage andliteracyinstructioninearlychildhoodclassrooms(Dickinson, Hofer, Barnes, & Grifenhagen, 2014; Hindman & Wasik, 2013; Justice,Mashburn,Hamre,&Pianta,2008;Mashburnetal.,2008). Earlywordlearning,inparticular,isnotsufficientlysupportedat manypreschools(NationalEarlyLiteracyPanel,2008;Neuman& Dwyer,2009).RecentdatafromFuller,Bein,Bridges,Kim,and Rabe-Hesketh(2017)indicatethatpre-kindergartenprogrammingthat isacademicallyoriented(i.e.,whichspendssignificantamountsof timeemphasizingorallanguageskillsandpreliteracyskills,aswell asmathconcepts)isassociatedwithgreatergainsinchildren’s lan-guageandmathskillsthanthoseseenforchildreninhome-based careorlessacademic-orientedprogramming.Suchresearch fur-thersuggeststhatincorporatinghigh-qualitysupportforlanguage developmentisakeyingredientinmaximizingthebenefitsofearly education.
Part of high-qualitysupport for vocabulary growthinvolves addressingthemany aspectsofwhatit meanstoknowa word well.Vocabularyknowledgecanbeconceptualizedasbeingalong acontinuumfromnotunderstandingaword’smeaningtohaving adeepunderstanding,andvocabularyinterventionsand assess-mentsdifferinthelevelofknowledgetargeted(Christ&Wang, 2011; Coyne, McCoach, Loftus, ZipoliJr, &Kapp, 2009; Hadley, Dickinson,Hirsh-Pasek,Golinkoff,&Nesbitt,2016).Althoughmany effortstoimprovechildren’swordknowledgeandthewordgap focusonpresenting manynewwordsefficiently, childrenshow
greaterdepthofvocabularyknowledgewhentheyengagewith newwordsinextendedandvariedways(Coyneetal.,2009). 1.2. Efficacyofvocabularyinterventions
Researchonvocabularyacquisitionsuggeststhatword learn-ing is facilitated through six principles: (1) frequent exposure, (2) capturingthechild’sinterest,(3) interactiveandresponsive environments,(4)meaningfulcontext,(5)diversityofwordsand language structures, and (6) leveraging of grammatical knowl-edge (Harris, Golinkoff, & Hirsh-Pasek, 2011; Hassinger-Das, Toub,Hirsh-Pasek,&Golinkoff,2017;Konishi,Kanero,Freeman, Golinkoff, &Hirsh-Pasek, 2014; Reed, Hirsh-Pasek, &Golinkoff, 2017).Wasik,Hindman,andSnell(2016)particularlyemphasize thevalueofsystemicexposurestonewwordsandavarietyoftasks thatinvitechildrentotrulyengagewiththewords.
Our workfocusesoncommonactivitiesthat canimplement theseprincipleswithinthepreschoolclassroom:readingand play-ing.Duringtheseactivities,vocabularywordsandtheirmeanings canbehighlightedininterestingand relevantways forchildren throughexplicitverbaldiscussionsandimagesorpropsthatdepict wordmeanings,aswellasgesturesthatprovidenonverbalsupport (Goldin-Meadow&Alibali,2013).Readingandplayingare promis-ingwaystogobeyondrotememorizationandfacilitatechildren’s deeperandlonger-lastingwordknowledge.Furthermore,playing especiallyenablesthekindofadaptiveandresponsiveinteractions thatshouldfacilitatethenaturallearningofvocabulary.
1.3. Sharedbook-readingtosupportvocabulary
Anextensivereviewofpreschoolandkindergarten interven-tions that were designed to build language found that shared book-readingistheinterventionthatmostconsistentlyresultsin vocabulary growth(NationalEarly Literacy Panel, 2008). Meta-analysesthatincludedstudiescompletedafterthecut-offpointof theNationalEarlyLiteracyPanel(Marulis&Neuman,2010;Mol, Bus,de Jong,&Smeets,2008;Mol, Bus,&Jong,2009)continue tofindconsistentevidencethatbook-readinginterventionsyield moderate-to-strongeffectsonpreschoolers’vocabulary.Although arecentsystematicreviewofbook-readinginterventions(Wasik etal.,2016)notedthatgainsaretypicallymodestintermsofthe proportionofwordsthatarelearned,theauthorsalsoarguethat book-readinglendsitselftothetypesofopportunitiesfor expo-suretoandengagementwithwordsthatbestfacilitatelearning. We therefore used book-reading in our program and incorpo-ratedevidence-basedstrategiestoenhancevocabularygrowth.For example,webuiltinmultipleexposurestowordsthroughrepeated readings,questioningaboutwordsand thestory, andprovision ofexplicitdefinitions(Beck&McKeown,2007;Biemiller&Boote, 2006;Coyne,McCoach,&Kapp,2007;Coyneetal.,2009;Wasik, Bond,&Hindman,2006).Wealsoaddedopportunitiesforchildren tosaytargetwords(Sénéchal,1997)andtoobserveanduse ges-turefornon-verbalexpressionofwordmeaning(Goldin-Meadow& Alibali,2013;Rowe,Silverman,&Mullan,2013).Moredetailsonour book-readingapproachcanbefoundinDickinsonandcolleagues (inpreparation).
1.4. Play-basedactivitiestofurthersupportvocabulary
Play might be another particularly effective component for vocabulary interventions (Weisberg, Zosh, Hirsh-Pasek, & Golinkoff,2013).Playisofteninherentlyinteresting,interactive, andmeaningfulforchildren,andtheinterdisciplinaryscienceof learningliteratureindicatesthat learningismaximizedthrough suchmeaningfulandsociallyinteractivelearningenvironmentsin which childrenare activeand engaged(Chi,2009; Hirsh-Pasek,
Zoshet al.,2015;Meltzoff,Kuhl,Movellan, &Sejnowski, 2009). Skill-specificcurriculaoftenembedfocusedacademicinstruction within play, and studies suggest that these curricula are more effectivethanlesstargetedapproachesinpromotingtherelevant academicskills(Jenkins&Duncan,2017;Phillipsetal.,2017).Prior studiesonvocabularygrowth,inparticular,havefoundpositive effectsofcouplingreadingandplaying(Han,Moore,Vukelich,& Buell,2010;Hassinger-Dasetal.,2016;Roskos&Burnstein,2011; Weisberg,Kittredge,Hirsh-Pasek,Golinkoff,&Klahr,2015).Indeed, Lillardetal.(2013)suggestedthatoneofthefewareaswhereplay hadademonstrableeffectwasinlanguagelearning.Whatprior studieshavenotexploredisthecomparativeanduniqueeffectsof differenttypesofpostreadingplayactivitiesonpreschoolers’word learning.Ourresearchaddressesthisgap.
Specifically,our three differentapproaches in Study1 were free playand two approaches involvingadultparticipationand support:guidedplay anddirectedplay. Freeplay, whichgrants childrenfreedomforexploration,allowsthemtoengagein dis-coverylearningandpracticeskillswithnoconstraintsfromadult involvement(Fisher,Hirsh-Pasek,Golinkoff,Singer,&Berk,2010; Johnson,Christie,&Yawkey,1999;Singer&Singer,1990).Pyleand Danniels(2017)foundfreeplaytobethemostcommonformofplay inearlychildhoodclassrooms.Thisapproachtoplaycapitalizeson children’sinterestsandprovidesameaningfulcontextinwhichto encountervocabulary,thustappingintoatleasttwoofthesix prin-ciplesofvocabularyacquisition(Harrisetal.,2011;Hassinger-Das etal.,2017;Konishietal.,2014).
AlthoughweincludedaFreePlayconditioninStudy1tosee whetheritcouldfacilitatevocabularylearning,theliterature sug-gests that free play is not an optimal pedagogy when there is a specificlearning goal (Chien etal., 2010; Fisher, Hirsh-Pasek, Newcombe, & Golinkoff, 2013; Klahr & Nigam, 2004). Instead, adults’subtleassistancethroughguidedplayhasbeenassociated withbetterlearningoutcomes.Guidedplayfallsundertheumbrella of playfullearning (Hirsh-Pasek&Golinkoff, 2011; Hirsh-Pasek etal.,2009;Weisberg,Hirsh-Pasek,Golinkoff,Kittredge,&Klahr, 2016),preservingtheplayfulnessand child-directednessoffree playandthegoal-directednessofformalinstruction(Ash&Wells, 2006;Hanetal.,2010; Rogoff,2003;Roskos,Tabors,&Lenhart, 2004;Toub,Rajan,Golinkoff,&Hirsh-Pasek,2016;Weisberg, Hirsh-Pasek,&Golinkoff,2013;Weisbergetal.,2016).Guidedplaybuilds ontheories emphasizingthe significance of social interactional contextsforchildren’slearning.Accordingtotheseframeworks,a significantportionofchildren’slearninghappensinshared mean-ingful contexts,through apprenticeships that occur againstthe backdropofguidedinteraction(Rogoff,2003;Vygotsky,1978).In thisspirit, adultsusingguided playfor vocabularysupportfind opportunitiestoinfusechildren’sunfoldingplaywithreferences totargetwords,remindersoftheirmeaning,andquestionsthat linkthemeaningtopriorknowledgeandthatinvitethemtoapply andexpandupontheirknowledge.Becauseitexemplifiesallsix principlesofvocabularyacquisition(Harrisetal.,2011; Hassinger-Dasetal.,2017;Konishietal.,2014),weanticipatedthatguided playwouldbeasuccessfulstrategy.
Ourthirdapproachtoplaywasdirectedplay.Here,theadult leadstheplay as anexternal director, rather than servingas a supportiveparticipant.Forexample,directedplaymightinvolve anadultleadingchildreninare-enactmentofastorybookusing figurines.Thisapproachresemblesdirectinstructionmethodsin whichtheadultexplainstheexactmeaningofconceptsand con-trols the flow of behavior (Kirschner, Sweller, & Clark, 2006). Directedplayalsoembodiesalltheprinciplesofvocabulary acqui-sitionexceptpossiblytheinteractiveandresponsiveprinciplesof languagelearning,since thereis notmuch roomfor considera-tionofchildren’scontributions(Harrisetal.,2011;Hassinger-Das etal.,2017;Konishietal.,2014).Directedplaymightbean
effec-tivecombinationofanelementofplayfulnesswithamoredidactic approach.However,thistypeofplayisnotplayfullearning,because childrendonotlead.
Here,weempiricallycomparethebenefitsofthesethree differ-entplayapproachesforpost-readingvocabularylearning.Byfirst testingplayfulvocabularyreviewthroughschoolvisitsby mem-bersoftheresearchteam(Study1)andthenshiftingtodeliveryby classroomteachers(Study2),weexaminethefeasibilityand effec-tivenessofthesemethodsinpracticalsettings.Additionally,our Study2comparisonofaplayfulapproachversusamoredidactic approachfurtherunpackstheuniquebenefitsofaplayfulpedagogy. By starting withtheestablishedactivity ofbook-reading, these studiesprobethepossibleaddedvalueofplayasawordlearning tool.
2. Study1:testofthreeplay-basedapproaches
In Study 1, children experienced an enriched shared book-reading activity followed by either free play, guided play, or directedplay.Weinvestigatedthefollowingresearchquestions: (1)Dochildreninthisinterventionshowimprovementsintheir knowledge of the targeted vocabulary over time? (2) Which approachtoplayismosteffectiveforsupportingsuchvocabulary development?Additionally,sincemaintenanceofgainsispartof successfullearning(Neuman,Newman,&Dwyer,2011),wetook advantageofanopportunityatoneofoursitestoexplorechildren’s retentionofwordknowledgeafteradelayoftwoweeks.
Wehypothesizedthat,acrossconditions,childrenwouldhave greater receptive and expressiveknowledge of target wordsat post-testthanatpre-test.Gainswerealsoexpectedtobegreater fortargetwordsthanforwordsthatwerepresentinthestorybut notexplicitlytaught (exposurewords)andfor wordsthat were notincludedatall(controlwords).Wealsohypothesizedthatthe twoconditionsinvolvingadultsupportfortargetedlearning dur-ingplay(GuidedandDirectedPlay)wouldpromotesignificantly greaterreceptiveandexpressivevocabularygrowththanwouldthe FreePlayconditioninwhichtheadultonlyprovidestoys.Further, GuidedPlaywasexpectedtobethestrongestconditionbecauseit incorporatedmoreofthelanguagelearningprinciples.Finally,we hypothesizedthatgainswouldbesustainedafteradelayoftwo weeks.Whileourinvestigationofretentionwasmoreexploratory innature,wesuspectedthattheGuidedPlayconditionwouldshow thegreatestknowledgeaftertheadditionaldelay,sincethelearning washypothesizedtobestrongerinthatcondition.
Echoingothers’effortstobetterunderstandhowintervention programscanimpactsubgroupsofchildrendifferently(Greenberg &Abenavoli,2017;Miller,Farkas,Vandell,&Duncan,2014),we collected additional background information to address possi-ble moderators of the effects of different approaches to play. Specifically,basedonpriorresearchonvariedlanguagelearning trajectories,wegathereddataonchildren’shomelanguagesand EnglishLanguageLearner status(Hindman&Wasik,2013;Hoff, 2013;Milleretal.,2014),andmaternaleducation(Dollaghanetal., 1999;Hindman&Wasik,2013;Qi,Kaiser,Milan,&Hancock,2006; Richels,Johnson,Walden,&Conture,2013).
3. Method
3.1. Participants
AfterobtainingapprovalfromourInstitutionalReviewBoards (IRBs),participantswererecruitedfrom10HeadStartpreschool classroomsinEasternPennsylvaniaandfrom18Pre-Kclassrooms in CentralTennessee. TheHeadStart programservesonly
low-incomefamilies,andTennessee’sprogramprioritizesenrollment ofeconomicallydisadvantagedchildren.
Several weeks before the start of the study, classroom teachers distributed consent forms (in English and Spanish) to children’s families. From the two sites, consent was provided for 258 children (165 in Tennessee). Nine children with-drew from the study and were excluded from analyses. The final sample of 249 included approximately nine children per classroom (46% male; Mage=59.19months, SDage=4.75months,
rangeage=39–66months).
Withineachclassroom,consentedchildrenwerefirstrandomly assignedtocondition.Then,and3–4childrenfromthesame con-ditionwereassignedtoasmallgrouptoparticipateintheactivities together.Mostclassroomsonlyhadonegrouppercondition,and teacherswerekeptblindtothegroups’conditions.Adjustmentsto theseinitialarrangementsweremadeinasmallminorityofcases, basedonconsiderationofteachers’inputonbehavioral incompat-ibility,anaimtohavemixedgendergroups,andattendance.These adjustmentsweremadeblindtochildren’spre-testabilities. Distri-butionacrossconditionwasalmostexactlyequal,with84children inFreePlay,83childreninGuidedPlay,and82childreninDirected Play.Asapreliminarywayofexploringretentionofgainsafteran additionaldelay,weconductedfollow-uptestingattheTennessee site(n=153;7childrenwereabsent).Thisexplorationofretention wasnotanoriginalfocus,aswereourotherresearchquestions; itwasabonussetofdatathatwaspossibleatonlytheTennessee site,andweelectedtoincorporateitasanadditional,secondary exploration.
Allbutthreechildrenpreviouslyparticipatedinapriorstudy in this XXXX project, which compared shared book-reading approaches,followedbyarelatedfreeplaysession.Inthatstudy, children’svocabularywordsand booksweredifferentfromthe onestheyencounteredinthisstudy.Noparticipantshadreferrals orIndividualizedEducationPrograms(IEPs)relatingtocognition orlanguage,and allchildren hadsufficient vision, hearing,and Englishproficiency.Teacherreportindicatedthat14.9%ofthe sam-plewasdesignatedasEnglishLanguageLearners(ELL;n=37),and parentreportindicatedthatSpanishwasthehomelanguagefor themajorityofthosechildren(seeTableA1foradditional informa-tionaboutchildren’slanguages).Raceand ethnicityinformation wasprovidedforallbut5ofthefinalparticipants(55.3% African-American;23.4%Hispanic/Latino,0.8%Asian,13.5%Caucasian,and 7.0%OtherorMultiracial).Parentsreportedhighestlevelof mater-naleducationfor196children:17.3%somehighschool,41.3%high schooldiplomaorGED,18.4%tradeschool,13.3%associate’sdegree, 6.1%bachelor’sdegree,and 3.6%graduatedegree(seeTable A2 foradditionalfamilybackgrounddata).Comparisonacrossthetwo sitesindicatedstatisticallysignificantdifferencesinchildren’sage, ELLstatus,homelanguage,maternaleducation,andrace/ethnicity. Moderationofconditioneffectsbysitewasthereforeexaminedin ouranalyses.
3.2. Materials,intervention,andmeasures 3.2.1. Bookandvocabularywordselection
Twotheme-basedpairsofcommerciallyavailablebookswere selectedafterconfirmingthattheywerenotpartofthe partici-patingclassroom’slibrarycollections:dragon,withTheKnightand theDragon (dePaola,1998)andDragon for Breakfast(McMullen & McMullen, 1990), and farm, with Farmer Duck (Waddell & Oxenbury,1991)andPumpkinSoup(Cooper,2005).
Thedesigninvolvedthreetypesofvocabularywords(seeTable A3):(1) targetwords(10perbook;20 pertheme),which were explicitlytaughtwithspecificstrategiesduringtheintervention activities;(2)exposurewords(3–5perbook;8–9pertheme),which werementionedinthetextofthebookandsometimesduringplay
butneverexplicitlytaught;and(3)controlwords(8pertheme), whichwerenotincludedinanyinterventionmaterialsor activi-ties.Wetestedexposureandcontrolwordstoisolatelearningthat wastheresultofourteachingeffortsandnotduetogeneral vocab-ularygrowth.Theoriginalbookscontainedsomewordsthatwere usedfortargetandexposurewords,andwealsoinsertedadditional wordsintothetexts,asneeded.Inadaptingthetexts,weensured thedesiredfrequencyof1–2usagespertargetvocabularywordand similaritiesacrossbooksintotalnumberofwords,numberofwords perpage,totalnumberofpages,andpictorialrepresentationsof targetwords.Theadaptationsdidnotsignificantlyalterthe story-linesfromtheoriginalbooks;theytypicallydescribedadditional details.
In selecting target, exposure, and controlwords, we sought adult-conversationalwordsofhighutility,asinthe“Tier2” desig-nationbyBeck,McKeown,andKucan(2013).Wordselectionwas alsodrivenbyfeasibilityofchild-friendlydefinitions,semanticand phonologicaldistinctness,difficulty,frequencyofuse(Biemiller, 2010;Chall&Dale,1995),andthelikelihoodofwords’ unfamil-iarity, basedonprior datafrom thisproject. Each setofwords containedconcreteandabstractnouns,verbs,adjectives,and spa-tialprepositions,sincechildren’sword-learningmightvarybased onformclassandconcreteness(Beck&McKeown,2007;Hadley etal.,2016).
3.2.2. Interventionprocedure
DatawerecollectedMarch–May2012.Participatingclassrooms wererandomlyassignedtothedragonorfarmthemeduringaprior study,andassignmentswereswappedforthisphasesothat partic-ipantsencounteredanewtheme,withnewbooksandnewwords. Groupsof3–4childrenassignedtothesameconditionengagedin ouractivitieswithanInterventionSpecialist(IS).TheseISswere recruitedthroughadvertisementsinthecommunity(e.g.,ona uni-versityjoblistingssiteandatcommunitycolleges)whichtargeted peoplewhowereinterestedinchildren’slanguagedevelopment andwhohadpreviouslyworkedwithyoungchildren.AllnineISs werefemaleswithexperienceinearlychildhoodsettingsandhad orwerecompletingrelateddegrees.Somehadpreviouslyworked aspreschoolteachers,andothershadworkedatlocallibrariesor bookstoresreading bookstochildren.ISscompleted2–3 train-ingsessions(3–4heach)inwhichtheyreviewedthescriptsand guidelineswiththeresearchteamandparticipatedinrole-playing exercisestopracticeleadingthevariousbook-readingandplay ses-sions.ISsconductedbook-readingsessionsandallthreetypesof playsessionsintheirassignedclassroom(s).MostISshad8play sessionsofeachtype(FreePlay,GuidedPlay,andDirectedPlay;4 sessionspergroupforeachof2booksinthetheme).
Duringeachoftwoweeks,childrenparticipatedinfour consecu-tivedaysofback-to-backbook-readingandplaysessions.Activities werebasedonthefirstbookinthethemeinweek1andthe sec-ondbookinthesamethemeinweek2,withhalfofthegroupsthat wereassignedtoagiventhemestartingwithonebookandhalf startingwiththeother.ISs’recordsshowedthatchildrenattended thesmallgroupactivitiesanaverageof7.14ofthe8intervention days(SD=1.30,range=3–8days),withnoconditiondifferencesand noopportunitiestomakeupabsences.Thereweretwomain mea-surementperiods:pre-testingoccurredwithinoneweekpriorto thefirstdayofinterventionactivities,andpost-testingoccurred withinoneweekafterthelastdayofactivities.Afollow-upround oftestingoccurredwithasubsampleofchildrentwoweeksafter post-testing.Testingtypicallyoccurredbyclassroom,with2days per classroomper measurement period.Therefore, within each classroom, children fromall conditions were tested withinthe same2-dayperiod,andanydifferencesbetweenclassroomsinthe amountoftimethatpassedbetweenactivitiesandtesting peri-odswerenotconfoundedwithconditioncomparison.Onaverage
23.78days(SD=6.95)elapsedbetweenpre-andpost-testing.Time elapsedwasnotsignificantlyrelatedtopost-testscores.
3.2.2.1. Book-reading. Foreachbook,wecreatedascripttoguide ISs through each 10-min reading, and ISs were instructed to minimizedeviationsfromthescripts.Thescriptoutlined evidence-based strategies (Biemiller & Boote, 2006; Goldin-Meadow & Alibali,2013; Sénéchal,1997; Wasiket al., 2006)for enriching thebook-readingbyelaboratingonthebook’stextand illustra-tions.Eachofabook’s10targetwordswasincludedinthebook’s text. Direct instruction of word meaningswas supplied across all readings; children were encouraged to participate more in laterreadings.Initiallychildrenweregivenachild-friendly defini-tion,commentarythatincludedadditionalconceptualinformation abouttheword’smeaning,encouragedtouseagesture,andwere directedtolookatrelevantpartsofthepictures.Forexample,when thetargetwordfierceappearedinthebook,theISusedherscript toexplain,“Thedragonwantstoscaretheknightsoheis practic-ingmakingafierceface[pointtopicture][growlyscaryvoice].He showshisteethlikethis[makeaface].Youtrytomakeafierceface thatmightscaresomeone.Youguyslookscary!”Scriptsforlater sessionsencouragedchildrentoproducethewordsbyfirstgiving abriefreviewofkeycontentfromthepageandthenpausingat thepointthewordwastobeused.Thebriefreviewsupplied addi-tionalinformationaboutthewords,andpromptsforgestureswere includedwhenrelevant:“Thedragonalsowantedabookonhow tofight.Whatdidhedo?Yes,helookedaroundforabookonhow tofightknights;herummagedaroundinhisboxforabook.Let’s pretendwe’rerummagingaroundtolookforatoy.”Afterthe read-ing,theISreviewedthestoryandwordsusingillustrationsfrom thebookandseparatepicturecardstobrieflyreviewtargetwords. Thesereviewsalsograduallyincreaseddemands onchildrento jointlyconstructthestoryandproducethewordsthemselves:“The dragonalsotriedtoscaretheknightbyshowinghissharpclaws. Whatdowecallananimal’sclaws?Yes,talons.[showpicturecard] Let’sallsaythatnow!Talons.[childrenrepeat]]Whatelsecan animalsdowiththem?[catchfood].”Comprehension-related dis-cussionswereincludedinthescriptsforeachbookreading,with progressivelymorechallengingquestionsasthestorybecamemore familiar.
3.2.2.2. Play. Therewasacollectionoftoysrelatedtoeachbook, withasettingpiece(i.e.,castle,farmhouse),figurines,andprops relevanttothestory.Onecompletesetoftoyswasavailabletothe groupsofchildrenfora10-minplayperiodfollowingeach book-reading.IntheFreePlaycondition,childrenplayedwiththetoys inwhatevermannertheychose,withtheISsupervisingandonly interveningregardingsafetyorconflicts.Bydesign,thiscondition didnot involveactiveadultsupportofvocabulary.Intheother twoconditions,theISactivelyparticipatedintheplayandadded reviewsoftargetvocabularywords.Tominimizetaskdemands, onlyhalfofeachbook’s10targetwordsweredesignatedasfocus wordsonagivenday,soeachwordwasreviewedduringtwoplay sessions.
In theDirectedPlay condition,the IS followedscripted lan-guagetodirectchildreninareenactmentofthebook’sstory.Each scriptincludedthreeinstancesofeachoftheday’stargetwords: abriefdefinitionand twousesas partof theIS’snarrative and re-enactmentdirections(e.g.,“...everyoneinthecastle[was]so surprisedtoseeadragonemergingfromtheegg,”“Emergingmeans tocomeoutofsomething,”and“Canyoumakethedragonemerge from,orcomeoutof,theegg?”).Ifnecessary,ISscouldadaptthe script based ontheirinteractions withchildren, but theywere instructedtofollowthescriptascloselyaspossibleto incorpo-ratethewordsproperly.Childrenwereexpectedtolistentothe IS’snarrativeandplayalong.
IntheGuidedPlaycondition,childrenchosewhattododuring theirplay.TheISdidnotreadfromascriptorrequirechildrento followre-enactmentinstructions.Instead,shewastoldtofollow children’slead,jointheirplay(e.g.,manipulatingthetoysand talk-ingasacharacter),andincorporatetargetvocabularyasnaturally aspossible.Tokeepfrequencyofexposurecontrolledacross adult-supportedplayconditions,theISwasinstructedtoincorporateeach oftheday’sfiveassignedwordsthreetimes,justlikeinDirected Play.In this condition,however,thethree instances includeda definition,aclosed-endedquestion,andanopen-endedquestion. Closed-endedquestionshadeasyanswers,oftenbasedonthe ques-tionitself,theword’sdefinition,orthestory.Open-endedquestions challengedchildrentothinkmoredeeplyabouttheword,andthere wasnosinglecorrectanswer.Theaimwastofindnatural opportu-nitiestoincorporatethewordsintothemorechild-directedplay. Insteadofaverbatimscript,aguidancecardoutlinedkeyelements ofthewords’definitions(e.g.,“comingoutofsomething”for emerg-ing)andsamplequestions(e.g.,“Doesemergingmeangoingintoor comingoutof?”foraclosed-endedquestionor“Canyouthinkof someotherthings[character]canemergefrom?”foranopen-ended question).TheISkepttheassignedwordsinmindandlookedfor opportunitiestobuildonwhatchildrenwerealreadydoingandto segueintoreviewingarelatedvocabularyword.
3.3. Measures
Children were tested one-on-one (20–25min) with both a receptiveandanexpressivetest,inacounterbalancedorder.All wordsassociatedwithagiventhemeweretestedonatleastone,if notboth,ofthesevocabularytests.
3.3.1. Receptivetest
We devised a receptive test similar to the Peabody Picture VocabularyTest(Dunn&Dunn,2007),butwithfewerfoilsand incorporatingthespecificwordsinourstudy.Theexperimenter displayedapagewiththreeimagesinarowandaskedthe partici-panttopointtotheimageshowingarepresentationofaparticular word(e.g.,“Where’sthrone?”).Eachpageincludedthecorrect ref-erent,aconceptualfoil,andathematicfoil(e.g.,throne,folding chair,andcrown,respectively).Allimagesdifferedfromthebooks’ illustrations,thepicturecardsusedduringbookreading,andplay sessiontoys,makingthisatestofextensionandnotjustrecognition offamiliarimages.Positionsofthethreetypesofimageswithinthe rowvaried.Childrenweretestedontheirtheme’swords:20target words(fromthetwobooks),8controlwords,andasmanyofthe 8–9exposurewordsascouldbedepicted(6forfarm,8fordragon). Children’sscoresreflectedthepercentageoftestitemsforwhich theychosethecorrectimage.Internalconsistencyforthereceptive measurewascalculatedfromchildren’spost-testdata,wherewe expectedtoseeabove-chanceresponses.WithCronbach’s˛=.72 forfarmand.61fordragon,resultswereacceptablebutonthelow side,perhapspartiallyduetothe33%chanceofbeingrightona givenitemsimplybyguessingandtothenarrowrangeofscores withinoursample.
3.3.2. Expressivetest
WeusedtheNewWordDefinition Test-Modified(NWDT-M; Hadleyetal.,2016),modeledafterBlewitt,Rump,Shealy,andCook (2009).Childrenwereaskedtoexplainthemeaningofvocabulary wordsusingbothverbalresponseandgestures.Theexperimenter askedchildren onequestion(e.g.,“Whataretalons?”) andthen promptedthemfurther,aftertheirinitialresponse(e.g.,“Canyou tellmeorshowmeanythingmoreabouttalons?”).Tokeeptypical taskdurationunder15min,theexpressivetestcontainedasubset ofwordsfromourstudy—21totalfordragon(13target,4exposure, 4control)and18totalforfarm(8target,6exposure,4control).
There were three exposure words—tip, plan, accidentally—that werenotfeaturedonthereceptivetest,becausetheywerenoteasy todepictvisually.Thesewordswereincludedintheexpressivetest, alongwithasamplingoftarget,exposure,andcontrolwordsfrom otherformclasses(nolessthan67%ofanyformclassofword). Video-oraudio-recordingswereusedtofacilitatelaterscoring. 3.3.2.1. Scoring. A child’sexpressivetaskscore wastheaverage informationunitsthechildprovidedperword.Informationunits wereindicationsofknowledgeabouttheword’smeaning, includ-ingtheuse ofsynonyms, antonyms, part-wholeor hierarchical relations,function,perceptualfeatures,gestures,andmeaningful context(eachcountingfor1point)orbasiccontext(countingfor 0.5pointbecauseitinvolvedanassociationwithoutsemantic infor-mation,suchasonlysaying“SantaClaus”forthewordchimney). Forexample,forthrone,“achairthatqueenssiton”has3 informa-tionunits:“chair”isasynonym,and“thatqueenssiton”provides meaningfulcontextanddescribesfunction.Therangeofpossible scoresbeginsat0(noinformationunitsprovidedforanyitem)but hasnopredeterminedmaximum,sinceparticipantsarenot lim-itedintheamountofinformationthattheycanprovideintheir responsestotheopen-endedprompts.Scoringfortheexpressive taskisdescribedinmoredetailinHadleyetal.(2016).
Fourmasters’studentsineducationservedasexpressivetask coders,andtheydemonstratedatleast90%agreementwiththe doctoralstudentGoldStandardcoderduringtraining.Then,every fifthsetofchildresponsesconsecutivelycodedbyanindividual coderwasalsocomparedagainsttheGoldStandardcoder’sdatato preventcoderdrift.Inthecase ofadiscrepancy,theGold Stan-dard coder’s code was used, and the other coder re-calibrated throughadditionalcodingpracticeandfeedbackbeforecontinuing. Theaveragepercentagreementforthe20%ofchildrenscoredfor reliabilitywas93.2%(Cohen’skappa=0.82),indicatingsubstantial agreement.
3.3.3. Parentquestionnaire
Abriefquestionnaire(availableinEnglishandSpanish)about familybackgroundwassenthomewithparticipants.Familiesof 242participantscompletedallormostofthequestionnaireitems. For the purposes of addressing potential moderator effects for oursecondresearchquestion,weusedparentquestionnairedata tocreatecompositescoresbasedonprincipalcomponent analy-sis(PCA)ofrelevantquestionnaireitems.Forexample,children’s homeliteracyexposurewasa composite ofitems askingifthe parenthasalibrarycard(yes/no),howmanybooksthechildhas (<5,6–10,11–20,or>20),andhowmanytimesperweek some-onereadstothechild(0–1,2–3, 4–5,5–6,oreveryday).Home languagewassimilarly a composite of responsesregarding the proportionsoftimethechildhearsandspeaksEnglishandother languages.Anotherfamilybackgrounditemwasmaternal educa-tion,documentedaseithersomehighschool,highschooldiploma orGED,tradeschool,associate’sdegree,bachelor’sdegree,or grad-uatedegree.Withthemedianfallingatahighschooldiploma(or GED),adichotomoussplitwasbetweenthosewithamaximumof ahighschooldiploma(n=113)andthosewithfurthereducation (n=83).
3.4. Fidelityofimplementation
Researchers(e.g.,graduatestudentsand postdoctoralfellows involved inthe study) coded a subsetof thebook-reading and playsessionsbycountingthepercentageofassignedwordsthat theISreviewedwithinthesession;theythencalculatedan aver-agescorefor each IS. Codersattended at leastone ofeach IS’s book-readingsessionsandprovidedfeedbacktotheISupon com-pletionofthesessionorsoonafterward.Observationsindicated
thateachISreviewedatleast90%ofthe10targetwordsduring thebook-readingsessions,whichwerescripted.Membersofthe researchteamatbothsitesobservedatleastoneGuidedandone DirectedPlaysessionforeachIS(ISswereuninvolvedinFreePlay). ResearcherslistenedforISs’reviewsofthe5assignedwordsfor thatday.AllISsscoredatleast90%onGuidedPlayandonDirected Play,andtheyweregivenfeedbackregardingerrors.Tofurther con-firmfidelityduringGuidedandDirectedPlaysessions,researchers codedrecordingsof4sessionsofeachtypefromeachIS.ForGuided Play,eachISatbothsitesscoredanaverageofatleast90%across 4codedsessions,withhalfoftheISsscoring100%onall4 ses-sions.Thelowestsingle-sessionscorewas80%.ForDirectedPlay, thePennsylvaniasitehadtechnicalproblemsthatleftallbutone IS’srecordingsunusable,butthatoneISscored100%onall4coded sessions.AttheTennesseesite,allbutoneISscoredanaverageof atleast90%,andtheremainingIShadanaverageof80%across4 sessions,duetooneinitialoutliersession.ThatISwasgiven feed-back,andsheimmediatelyimproved,witha90%averageacrossall 8DirectedPlaysessions.
3.5. Dataanalyses
Usingmultilevelmodeling (MLM)toaccountfor interdepen-dency (childrennested withinintervention play groups, nested withinclassrooms),weexaminedtheeffectivenessofeach play approachforimprovingchildren’svocabularyknowledge.Fewer than5%ofchildrenweremissing dataononeofthetwostudy outcomes,andanalyseswereconductedonallavailablecasesfor agivenoutcome.Unlessotherwisenotedbelow,allposthoc pair-wisecomparisonswereconductedusingFisher’sLeastSignificant Difference(LSD)test,andalleffectsizesarepresentedasCohen’s d(Cohen,1988).
4. Results
Table1showsmeanrawscoresandstandarddeviationsforall measuresbytimeforthefullsampleofchildrenandthesubsample thatwasfollowedupafteranadditionaldelay.
4.1. Children’svocabularygainsovertime
Our first research question asked if children gained vocab-ularyknowledge through this intervention. In separatemodels forthetwo vocabularytests,timet wasnestedwithinchildreni
and differences betweentime points were explored (10) (see
Supplementary Materials, Eq. (1)). Collapsing across condition, initial analysesshowed that children’s increase in scoresfrom pre-test to post-test were statistically significant (ps<.01) for bothreceptiveandexpressiveknowledgeoftarget(Breceptive=0.21,
SE=0.01,Cohen’sd=1.34;Bexpressive=0.42,SE=0.03,d=1.34),
expo-sure(Breceptive=0.06, SE=0.01, d=0.56; Bexpressive=0.10, SE=0.02,
d=0.26), and control words (Breceptive=0.05, SE=0.01, d=0.28;
Bexpressive=0.06,SE=0.02,d=0.24).
Testsofbaselineequivalenceamongtarget,exposure,and con-trolwordsindicatedthatchildrenshowedtheleastknowledgeof targetwordsatpretestandthemostknowledgeofcontrolwords (p<.05).Toaccountforpotentialdifferencesatbaseline,ourmodel controlledforvocabularyknowledgeatbaseline.Usingthatmodel, wenextexaminedif children learnedtarget wordsbetterthan exposureandcontrolwords.AsseeninthetoppanelofTable2, pre-topost-testreceptiveandexpressivegainsfortargetwordswere significantlylargerthangainsforexposurewords(dreceptive=−0.32
anddexpressive=−0.47)andforcontrolwords(dreceptive=−0.58and
dexpressive=−0.89). Posthoc pairwise comparisons also indicated
Table1
Study1:receptiveandexpressivevocabularyunadjustedmeans(standarddeviations)byassessmentsessionforfullsampleandsubsamplewithfollow-updata.
Measures Fullsample(N=249) SubsamplewithFollow-updata(n=153)
Pre-test Post-test Pre-test Post-test Follow-up
Receptiveandexpressivevocabularybylevelofinstruction(collapsedacrossconditions)(N=249) Targetwords Receptivevocabulary 0.37(0.13) 0.58(0.17) 0.37(0.13) 0.59(0.17) 0.61(0.18) Expressivevocabulary 0.14(0.21) 0.56(0.48) 0.15(0.24) 0.60(0.50) 0.60(0.49) Exposurewords Receptivevocabulary 0.50(0.20) 0.56(0.22) 0.51(0.18) 0.57(0.23) 0.61(0.24) Expressivevocabulary 0.38(0.42) 0.48(0.26) 0.40(0.42) 0.48(0.44) 0.52(0.50) Controlwords Receptivevocabulary 0.45(0.17) 0.51(0.18) 0.45(0.17) 0.51(0.19) 0.50(0.19) ExpressiveVocabulary 0.19(0.23) 0.25(0.26) 0.21(0.24) 0.25(0.25) 0.29(0.31)
Targetreceptiveandexpressivevocabularybyexperimentalcondition Guidedplay Receptivevocabulary 0.37(0.12) 0.60(0.17) 0.38(0.11) 0.62(0.17) 0.62(0.17) Expressivevocabulary 0.14(0.21) 0.59(0.48) 0.16(0.24) 0.67(0.50) 0.64(0.52) Directedplay Receptivevocabulary 0.37(0.13) 0.60(0.16) 0.38(0.14) 0.63(0.15) 0.66(0.16) Expressivevocabulary 0.16(0.24) 0.64(0.51) 0.17(0.27) 0.70(0.53) 0.66(0.48) Freeplay Receptivevocabulary 0.37(0.13) 0.54(0.18) 0.37(0.13) 0.54(0.18) 0.55(0.21) Expressivevocabulary 0.11(0.17) 0.42(0.41) 0.12(0.19) 0.44(0.44) 0.50(0.47)
Note.ThefullsamplewasdistributedacrossGuidedPlay(n=83),DirectedPlay(n=82),andFreePlay(n=84),andthesubsampleswithfollow-updatahadn=52,n=51,and n=50,respectively.Receptivetaskvaluesindicatetheproportionofitemsthatwereansweredcorrectly,withchancelevelequalto0.33(i.e.,selectoneofthreeimages). Expressivetaskvaluesindicatetheaveragenumberofinformationunitschildrenprovidedperexpressiveitem,andthesescoresrangedfrom0.00to2.63.Forcomparison, therangeforHadleyetal.(2016)was0–2.02.
Table2
Study1:parameterestimates(standarderrors)foreffectsoflevelofinstruction(top panel)andcondition(middlepanel)atpost-testandeffectsizeforconditioneffects (bottompanel).
Parameters Receptive
vocabulary
Expressive vocabulary Levelofinstructioneffects
Level1,levelofinstruction
Intercept,000 0.392(0.023) −0.500(0.202)
Pre-testperformance,100 0.335(0.039)** 0.617(0.042)**
Targetversusexposure,200 −0.062(0.015)** −0.220(0.029)**
Targetversuscontrol,300 −0.102(0.015)** −0.342(0.027)**
Conditioneffects(fullsample) Level1,child Intercept,000 0.180(0.134) −0.748(0.341) Pre-testperformance,100 0.473(0.078)** 1.229(0.124)** Ageatpre-test,200 0.001(0.002) 0.018(0.005)** Attendance,300 0.026(0.008)** 0.031(0.019) Gender(males=0),400 0.007(0.020) 0.024(0.050)
Level2,playgroup
Guidedplayversusdirectedplay,010 −0.010(0.023) 0.035(0.060)
Guidedplayversusfreeplay,020 −0.068(0.023)** −0.116(0.059)*
Level3,classroom
Theme(0=farm),001 −0.028(0.021) −0.212(0.053)**
Cohen’sdeffectsizes
Guidedplayversusdirectedplay,␥010 −0.061 0.069
Guidedplayversusfreeplay,␥020 −0.388 −0.260
Directedplayversusfreeplay −0.341 −0.324 Note.N=249.Standarderrorsadjustedforinterdependencyamongobservations.For testofLevelofInstructioneffects,targetwordsisthereferencegroupforthe com-parison(negativeestimatesindicatethattargetwordshadlargercovariateadjusted post-testscores).Fortestofconditioneffects,guidedplayconditionisthereference groupforthecomparison(negativeestimatesindicatethatchildrenintheguided playconditionhadlargercovariateadjustedpost-testscorescomparedtochildren inthecomparisoncondition).Positiveestimatesforgenderindicatethatfemales madegreatergainscomparedtomales.PositiveestimatesforThemeindicatethat childrenintheFarmthemehadlargerpost-testgainscomparedtochildreninthe Dragontheme.ContrastofDirectedPlay(comparisongroup)versusFreePlaywas estimatedwithpost-hocpairwisecomparison.
*p<.05.
**p<.01.
forcontrolwordsonbothreceptive(p=.005,d=−0.20)and expres-sivemeasures(p<.001,d=−0.33).
4.2. Comparisonofplayconditions
Our secondresearchquestionfocusedoncomparative bene-fitsofthethree differentplayconditions forchildren’sgainsin vocabularyknowledge.Testsofbaselineequivalenceforreceptive andexpressivepre-testscores,gender,age,andELLstatuswere conductedinseparatemultilevelmodels,witheachvariablebeing predictedbyconditionassignment.Resultsindicatedthatpre-test differencesamongtheconditionswerenon-significant(ps>.10). BasedonthemeansprovidedinTable1,themagnitudeofthe dif-ferencewaszeroacrossconditionsforthereceptivemeasure.For theexpressivemeasure,childrenintheFreePlayconditiontended tohavelowerscoresatpretest(d=0.16forcontrastwithGuided Play and 0.24 for DirectedPlay), while thedifference between GuidedandDirectedPlayconditionswassmallandnon-significant (d=0.09). Baseline vocabulary(100), age(200), gender (300),
attendance(400),andtheme(001)wereincludedascovariates
inmodelsexaminingconditioneffects(010and020)on
vocabu-larygains,whichincorporatedchildreninestedwithinplaygroupj
andclassroomk(seeSupplementaryMaterials,Eq.(2)).
Inthisdiscussionofconditioneffects,wepresentresultsfor tar-getwordsonly,asthosearetheonlywordsthatweresystematically treateddifferentlybycondition.Resultsforboththereceptiveand expressivemeasures(seeFig.1)indicatedthatchildreninGuided PlaymadesignificantlylargergainsthandidchildreninFreePlay, d=−0.39and −0.26,respectively (seeTable2,middle and bot-tompanels).Similarly,posthoccomparisonsindicatedthatchildren inDirectedPlaymadesignificantlylargerreceptiveand expres-sivegainsthanchildreninFreePlay(ps<.015),d=−0.34and0.32, respectively.Therewerenodifferencesbetweenchildren inthe GuidedandDirectedPlayconditionsforreceptiveandexpressive gains,d=−0.06and0.07,respectively.
Toexplorewhetherthecomparativebenefitsofthedifferent playconditionswouldvarybasedonchildcharacteristics,we con-ductedaseriesofanalysestoexamineifchildren’shomeliteracy
Fig.1.Study1vocabularygainsbycondition.Proportionofitemscorrectonthe receptivemeasureofvocabulary(a)andtheaveragescoreperexpressiveitem(b) atthepost-testcontrollingforpretestperformance,attendance,age,gender,and theme.Standarderrorsrepresentedinthefigurebytheerrorbarsattachedtoeach column.*Conditioncontrastsignificantatp<.05.
exposure(compositeofresponsesregardinglibrarycard,number of books,and reads per week), maternal education, home lan-guage(compositeofwhatchildren hearandspeakathome),or ELLstatusmoderatedconditioneffectsforreceptiveorexpressive post-testscores(appropriateinteractiontermswereaddedtoEq. (2)).Analysesinvolvinghomeliteracyexposure,homelanguage, andmaternaleducationwereconductedonlyon80%ofthesample (n=195or196dependingonanalysis),assomechildrenwere miss-ingthoseparent-reportdata.BasedonWhatWorksClearinghouse attritionstandards(2014),examinationofthereductionin sam-plesizesacrossconditionindicatedthatmissingdatarateswere approximately equivalent acrosscondition (less than 0.08 SD). Wealsotestedformoderatingeffectsofchildren’spre-testscores (to address potential Matthew effects), attendance during our activities(toaddressdosage),andsite(toaddresssample, imple-mentation,orotherdifferencesacrosssites).
Theonlystatisticallysignificantresultswereinthemodels test-ingthedichotomousmaternaleducationvariableasamoderator ofreceptiveand expressivevocabularygains(seeTableA4).For receptivevocabulary,maternaleducationmoderatedcondition dif-ferencesbetweenGuided Playand Free Play (B=0.15, SE=0.06, p=.01)andDirectedPlayandFreePlay(B=0.12,SE=0.06,p=.03) butnotbetweenGuidedandDirectedPlay.Childrenwhosemothers hadhighereducationshowedgreatergainsinGuidedPlay com-paredtoFreePlay(p<.01,d=0.73)andinDirectedPlaycompared
toFreePlay(p<.01,d=0.76).Forchildrenwhosemothershadlower education,therewerenoconditioneffects.
Forexpressivevocabulary,maternaleducationmoderated dif-ferencesbetweenGuidedandFreePlay(B=0.34,SE=0.13,p=.01) butneitheroftheothercomparisons.Childrenof motherswith highereducationshowedgreatergainsinGuidedPlaythaninFree Play(p<.01,d=0.62).Therewerenoconditioneffectsforchildren ofmotherswithlowerlevelsofeducation.
4.3. Follow-uptestingresults
Weaddressedthequestionofretentionthroughdatafromthe subsampleofchildrenwhoparticipatedinfollow-uptestingtwo weeksafterthepost-testing.Expandinguponthemodeltotest conditiondifferences(seeSupplementaryMaterials,Eq.(2)),tests offollow-upeffectsincludedanadditionalnestingleveltoaccount forobservationst (100 and200)nestedwithinchildreninested
withinplaygroupj.Ofprimaryinterestwerethecross-level
inter-actionsbetweenconditionandobservation(101,102,201,and
202)(seeSupplementaryMaterials,Eq.(3)).Forthesechildren,
gains and conditioneffects betweenpre- and immediate post-testweresimilartothoseforthefullsamplereportedabove(see Table3).Theresultsfromfurtheranalysesindicatedthattherewas nostatisticallysignificantchangeinchildren’svocabulary knowl-edgefrompost-testtofollow-upforeitherthereceptiveorthe expressivemeasure.Thispatterninthedatameansthatinitialgains madeoverthecourseoftheinterventionwerepreservedforat leastanothercoupleofweeks.Thisalsomeansthateffectsizesfor conditiondifferencesatthefollow-uptimepointwere compara-bletothosereportedaboveforthefullsample’spre-topost-test gains.Morespecifically,forboththereceptiveandexpressive mea-sures,childreninGuidedPlayhadsignificantlyhigherfollow-up scoresthandidchildreninFreePlay,d=−0.28and−0.28, respec-tively.Similarly,childreninDirectedPlayhadsignificantlylarger follow-upscoresthandidchildreninFreePlayforreceptiveand expressivegains,d=−0.33and−0.24,respectively.Therewerealso nodifferencesbetweenchildrenintheGuidedandDirectedPlay conditionsfor receptiveand expressivegains,d=0.04and 0.05, respectively.Moderatorswerenotexaminedforthefollow-updata duetothereducedsamplesizeandthesimilaritybetweenpost-test andfollow-up.
5. Discussionofstudy1
The overarching purpose of this project was to teach low-income preschoolers new words through a combination of book-readingandplay.OurfirstmainresearchquestionforStudy1 waswhetherpreschoolerstaughtwordsthroughbookreadingand playcouldlearnthosetargetwordsbetterthantheywouldlearn exposureandcontrolwords.Our findings,controllingfor many variables,indicatethatthevocabularyinstructionweofferedwas effective:Theinterventionimprovedchildren’sabilitiesto recog-nizenoveldepictionsofthetargetwordsandtoexplain words’ meanings.Thisisavaluablefinding,astransferanddepthof knowl-edgeareoftennotwellsupportedinpre-Kcurricula(Neuman& Dwyer,2009).
Oursecondmainresearchquestionwaswhethervariationsin adultsupportforvocabularylearningduringplaywouldmakea dif-ferenceinwordlearning.Theresultssupportourhypothesisthat childrenintheadult-supportedplayconditions(GuidedPlayand DirectedPlay)wouldoutperformtheirpeersintheFreePlay con-ditiononboththereceptiveandexpressivemeasures.Onelikely reasonforpoorerlearninginFreePlayisthelackofbuilt-in vocab-ularyreviewforthetargetwords.Itwasimportanttoexplorethis typeofplay,however,becauseitmostcloselyapproximatedthe
Table3
Study1:parameterestimates(standarderrors)formodelsestimatingconditiondifferencesforreceptiveandexpressivevocabularygains(post-test&guidedplay=reference groups).
Parameters Receptivevocabulary Expressivevocabulary
Fixedeffects Level1,observation
Intercept,000 0.188(0.204) −0.727(0.574)
Post-testversuspre-test,100 −0.249(0.024)** −0.515(0.055)**
Post-testversusfollow-up,200 0.007(0.024) −0.035(0.055)
Level1,child
Ageatpre-test,010 0.004(0.003) 0.020(0.009)**
Attendance,020 0.022(0.011) 0.024(0.030)
Gender,030 0.016(0.022) 0.050(0.061)
Level3,playgroup
Guidedversusdirected,001 −0.001(0.033) 0.036(0.086)
Guidedversusfree,002 −0.080(0.033)* −0.242(0.086)**
Theme(0=farm),003 −0.007(0.022) −0.003(0.060)
Cross-levelinteractions
Guidedv.directed*postv.pre,101 0.008(0.033) −0.012(0.077)
Guidedv.directed*postv.follow-up,201 0.020(0.033) 0.006(0.077)
Guidedv.free*postv.pre,102 0.073(0.033)* 0.197(0.076)**
Guidedv.free*postv.follow-up,202 −0.001(0.033) 0.090(0.077)
Note.N=153.Standarderrorsadjustedforinterdependencyamongobservations.Observations(Level1)wererepeatedwithinchildren(Level2),whowerenestedwithin PlayGroups(Level3).Theguidedplayconditionisthereferencegroupforthecomparisonofconditiondifference;assuch,negativeestimatesindicatethatchildrenin theguidedplayconditionhadlargerpost-testscorescomparedtochildreninthecomparisoncondition.Post-testisthereferencegroupforthecomparisonofobservation effects;assuch,negativeestimatesindicatethatscoresatpost-testwerelargerthanthecomparisonobservation.PositiveestimatesforThemeindicatethatchildreninthe Farmthemehadlargerpost-testscorescomparedtochildrenintheDragontheme.
*p<.05.
**p<.01.
styleofplayusedinmanyclassrooms.Intermsofplaywithadult support,wefoundnostatisticallysignificantdifferencesbetween theDirectedandGuidedPlaygroupsonvocabularygains.Although wehadexpectedsomeadditionalbenefitinGuidedPlay,theless child-directedplaywithintheDirectedPlayconditionwassimilarly effectiveforvocabularysupport.
Anintriguingfindingisthefactthatthesetwoplayconditions’ effectsweremoderatedbythevariableofmaternaleducation.That is,children’svocabularygainsinbothGuidedand DirectedPlay werebeneficialonlyforthesubsetofchildrenwhosemothershad greaterthanahighschooldiploma.Childreninourstudywhose mothershadlesseducationmightnothavehadasmuchexperience withlearningthroughadult-supportedcontextsandthereforedid notbenefitasmuchfromadultsupport.Consistentwiththis inter-pretation,findings froma recent article (Hirsh-Pasek,Adamson etal.,2015)showvariationsevenwithinalow-incomesamplein thequalityofmother-infantinteractionsduringbook-readingand playactivities.Inthatsample,ameasureoffluencyand connect-ednessbetweenmotherandinfantwassignificantlyandpositively correlatedwithmaternaleducation.Giventhatmaternal educa-tionmoderatedchildren’slearninghere,forStudy2wesoughtto developanewapproachthatmightbettersupportmorechildren bycombiningelementsfromthetwoadult-supportedconditions.
Asfortheissueofretention,childrenatthesitewherefollow-up testingwasconductedretainedtheirnewfoundknowledgeafteran additionaltwoweeks,andthesameoverallpatternemerged,with arelativebenefitofGuidedandDirectedPlayconditionscompared totheFreePlaycondition.Thisresultillustratesthestrengthof ourapproach.Ifchildrendonotretainvocabularygains,they can-notbenefitfromeithertheknowledgeofthosespecificwordsor thesnowballeffectinwhichknowledgeofsomewordscan pro-moteunderstandingofothernewwordsandconcepts(Neuman etal.,2011).Thesedataarefromonlyoneofthetwosites,andit wasbeyondthescopeofthisstudytoconductextended follow-uptesting(i.e.,afteranadditional3–6months)ortoexploreany longer-termeffectsongeneralvocabularygrowth;however,these initialfollow-upresultssuggestthatusingadult-supportedplayfor vocabularyreviewisapromisingavenueforfurtherpursuit.
6. Study2:playversusnon-playvocabularyreview
activities
Wehavearguedthatitwastheplay-basedadultsupportfor word-learningthatledchildrenintheGuidedandDirectedPlay conditionstoshowgreatergains thanchildrenin theFreePlay condition showed. However, we have not yet tested the pre-cise role of the playfulcontext for vocabularyreview. It could bethecasethatchildrenshowedbettervocabularygainsinthe adult-supportedplayconditionssimplybecausetheadultsensured additionalvocabularyreviewinthoseconditions;theplayful con-textforthatreviewmightnothavebeenimportant.Iftheplayful learningapproachisuniquelyeffective,thenchildrenshouldlearn wordsreviewedin aplay contextbetterthantheylearn words reviewedinanengagingcontextthatalsohasadultsupportbut usesmoredidacticmethods.Study2wasdesignedtoaddressthese issues. Also,in Study2,ourISteamtrained classroomteachers toimplementtheactivitiesthemselves,allowingustoinvestigate whetherteacherscouldadoptourprogrameffectively.Thisshift increasestheecologicalvalidityoftheresearchandthe implica-tionsofourresults.
TheprimaryresearchquestionsinStudy2were(1)Do chil-drenshowimprovementsinvocabularythroughacombinationof book-readingandadult-supportedplaythatarebothdeliveredby classroomteachers?and(2)Dochildrenlearnmorethroughthe combinationofbook-readingandadult-supportedplayfullearning sessionsthantheydo throughthecombination ofbook-reading andamoredirectteachingapproachtosupplementaryvocabulary reviewthatgivescomparableexposuretothetargetwords?
TherewereavarietyofkeydifferencesinthedesignsofStudy1 andStudy2.ForStudy2,inadditiontothewithin-subjects compar-isonofpre-andpost-testdata,weusedawithin-subjectsdesign tocompareplayandnon-playreviewactivitiesassupplementsto thebook-reading.Therewasnobetween-subjectscondition com-parisoninStudy2.WealsotransitionedfromhavingISsdeliverthe interventioninsmallpull-outgroupstohavingclassroom teach-ers delivertheactivitieswithIS support.Teachersrantheplay sessionswithsmallgroups ofparticipantsbutconducted book-readingswiththeirfullclass.Further,inStudy2,weusedonlythe
dragontheme,andclassrooms(N=8)wereeachassignedoneof twobooks.Thiswasmoremanageableforteachers,astheyhadthe fulltwoweekstocompletetheactivitiesforonebook.
Thenumberoftargetwordsperbookwasincreasedfrom10to 16,andweremovedexposurewordsfromthedesigntofocusonthe within-subjectscomparisonsbetweenthetwopost-read review activitiesandbetweenthetargetwordsandthe8controlwords. FortheplaysessionsinStudy2,sincetheGuidedandDirectedPlay approacheswereeacheffectiveinStudy1,wemergedelements ofbothstylesintotheplayapproachforthisstudyandkeptthe 10-mindurationforeachofthe4playsessionspergroup.Words notreviewedintheplaysessionswerereviewedinsteadduring acomparisonactivity:Teachersconducted6picturecardreview sessionswiththeirfullclass,andthesesessionstendedtolastup to5mineach.
7. Method
7.1. Participants
Wepartnered withthesameprograms servingpreschoolers fromlow-incomefamiliesinPennsylvaniaandTennessee,andwe followedsimilarIRBandrecruitmentprocedures.Fourclassrooms fromeachstatetookpartinthestudy,andallparticipatingteachers werefemale.Therewereapproximately12consentedchildrenper classroom(N=101,54.5%male;Mage=58.65months,SDage=5.84 months,rangeage=40–67months).Teacherreportsindicatedthat
13.9%ofthechildrenwereclassifiedasEnglishLanguage Learn-ers(ELL;n=14).ParentreportsconfirmedthatSpanishwasspoken inthehomesofmostofthosechildrenandalsointhehomesof manychildren not designatedasELL. Anadditional6.9% (n=7) weremissingbothteacherandparentreportdata.Raceand eth-nicityinformationwasnotcollectedattheTennesseesite(n=52) duetologisticalconstraintsbutwasprovidedthroughparent ques-tionnairesfor34ofthe49childrenatthePennsylvaniasite(20.6% African-American;44.1%Hispanic;35.3%Caucasian).Highestlevel ofmaternaleducationwasreportedbyparentreportfor84 chil-dren:10.7%somehighschool,40.5%highschooldiplomaorGED, 13.1%tradeschool,15.5%somecollege,14.3%associate’sdegree, 1.2% bachelor’s degree, and 4.8% graduate degree.Other back-grounddataareavailableinTablesA1andA2.Comparisonsacross thesitesshowedsimilarpatternsasthoseseenin Study1, and moderationofconditioneffectsbysitewasagainexamined. 7.2. Materials,intervention,andmeasures
7.2.1. Bookandvocabularywordselection
Foreachofthetwodragon-themedbooks,wordswereselected inamannersimilartoStudy1,but withafocusonlyonnouns andverbsintheinterestofsimplifyinganalreadycomplexdesign. Wordsfromtheoriginalorrevisedtextwereused,andnewwords wereadded,asneeded,toreach16targetand4controlwordsper book(seeTableA5).TargetwordsweresplitintosetsAandB,with8 wordsineachsetandonlyonesetbeingemphasizedduringagiven readingorplayactivitytominimizeoverload.Giventhe within-subjectsdesignforaddressingquestion2,SetsAandBwerefurther dividedinto4Read+Playwords(i.e.,taughtduringreading and play)and4Read+PictureCardwords(i.e.,taughtduringreading andpicturecardreview).Therefore,allchildrenweretaughtall16 wordsinthecontextofbookreading,andtheywerealsotaughthalf thosewordsduringplayandtheotherhalfduringthepicturecard reviewactivity.Tofacilitatecomparisonsofvocabularygainsacross groupsofwordsinthisdesign,Read+Playwords,Read+Picture Cardwords,andControlwordsineachbookhadequaldistributions ofabstractnouns,concretenouns,andverbs.
Table4
Study2:wordreviewsperconditioninwithin-subjectsdesign.
Read+play Read+picturecard
Numberofwords 8(4SetAand4SetB) 8(4SetAand4SetB) Readingsessions1–4 (6reviews)
Multi-strategyreview in2pre-readingsand2 readings(SetA Readings1&3orSetB Readings2&4)and brieflydefinedduring other2readings
(6reviews) Multi-strategyreview in2pre-readingsand2 readings(SetA Readings1&3orSetB Readings2&4)and brieflydefinedduring other2readings Playsessions1–4 (6reviews)Days1&2:
multi-strategyreview during1pre-playand1 playsession(SetADay 1orSetBDay2)and brieflydefinedduring1 pre-play(theother Day)Days3&4:used meaningfullyin2 pre-playsandgiven multi-strategyreview in1playsession(SetA Day3orSetBDay4) Picturecardsessions
1–6
(6reviews) Multi-strategyreview duringeachsession
Totalreviews 12 12
Note.Natureofvocabularyreviewfor16targetwordsovertwoweeks.Allword sets(AandBofRead+Playwords;andAandBofRead+PictureCardwords)were reviewed12times:6timesduringreadingsessionsand6timesduringeitherplay sessionsorpicturecardsessions.Reviewsconsistedofeitherabriefdefinition,useof thewordinameaningfulcontext,ormultiplestrategies(e.g.,definitions,gestures, pictures),asoutlinedintheguidancematerialsforteachers.
7.2.2. Interventionprocedure
DatawerecollectedinApril2013.Althoughtheteachersand
studentshadbeenpartofaninterimphaseoftheRead-Play-Learn
project(between Study1 and Study2)that involved thesame
books,eachteacherwasassignedtowhicheverbookhadnotbeen
usedin her classroomduring that phase.Therefore, while
par-ticipantsalready had someexposureto thedragon theme,the
story and words for this studyhad not been presented in her
classroompreviously.Thetwo-weekinterventionconsistedof
mul-tipleactivities,allledbytheclassroomteacher:4book-reading
sessionsthatreviewedRead+PlayandRead+PictureCardwords
with the wholeclass (i.e., 12 participantsplus up to
approxi-mately8otherstudents);4smallgroupplaysessionsthatreviewed
onlytheRead+Playwordswitheachgroupof3–4study
partici-pants;and6whole-classpicturecardsessionsthatreviewedonly
theRead+Picture Card words(seeTables 4 and5 and detailed
descriptionsbelowforinformationonequivalenceanddifferences amongreviewsacrosscondition).Tominimizethelevelofproject interference with regular classroom functioning and maximize implications for broader application, we intentionally choseto allowfornaturalvariabilityinsomeelementsofintervention deliv-eryratherthantotightlycontrolallelements.Inthatvein,teachers scheduledtheseactivitiesbasedontheirclassroomneeds,in con-sultationwiththeISwhowascoachingthem.Ingeneral,aftera givenreadingandbeforethenextone,teacherstypicallyconducted aplaysessionwitheachsmallgroup,suchthateachchild experi-encedreadingsandplaysessionsinanalternatingpattern.Picture cardreviewsessionswerealsospreadacrossthetwoweeks, inter-spersedamongtheotheractivities,atatimewhenthefullclass couldbepresentandattentive.
Eachteacher’sassignedISobservedherreadingandplay ses-sions,contributedin-sessionpromptsorclarifications(ifdesired), andprovidedbrieffeedbackafterwards.Basedonteacherreport, average participant attendance was3.75 of the4 reading
ses-Table5
Study2:examplesofwordreviews.
Typeofsession Typeofreview Example
Pre-readingforreadings 1–4
Multi-strategyreview “Thedragonusedhisnostrilstoblowfireandscaretheknight.Nostrilsarelittle holesinyournose.[Showpicturecardasyousaydefinition.]Canyousaythat word?[Kidsrepeat.]Canyoupointtoyournose?[Pointtoyourownnose].Do youhavenostrils,too?Yes!Youusethemtobreatheair,notfire!”
Duringreadings1–4 Multi-strategyreview “Nostrilsarelittleholesinyournosethatyouusetobreathe.[Pointtopicture
asyousaydefinition.]Thedragonblowsfireoutofhisnostrils,butwebreathe airoutofours[pointtoyournostrils].Canyousaynostrils[kidsrepeat]and pointtoyours?[kidspoint]”
Duringreadings1–4 Briefdefinition “Nostrilsarelittleholesinyournosethatyouusetobreathe.[Pointtopicture
asyousaythis.]”
Pre-playfordays1–2 Multi-strategyreview “Heisangryatthedragon.HebangshisfootdownHARDtoshowhe’sangry,
doesn’the?Whatdowecallthat?Yes,heisstampinghisfoot.Canyousay stamping?[Childrensayword.]Showmehowyoudoit.[Childrendoit.]”
Pre-playfordays1–2 Briefdefinition “Heisangryatthedragon,soheisstampinghisfoot.Thatmeansheisbanging
hisfootdownhard.”
Duringplaydays1–2 Multi-strategyreview [Incorporatewordreviewbasedonchildren’splay,atleast3times.]
“Theeggiscracking!Somethingisemergingfromtheegg.” “Whatisitcalledwhensomeonecomesoutofsomething?”
Pre-playfordays3–4 Meaningfuluse
(duringtoydistribution)
[Handingoutthedragonfigurine]“Don’tforgetaboutthefirethatcomesout ofhisnostrils!”
Duringplaydays3–4 Multi-strategyreview
(incorporatebasedon children’splay)
Useeachofthreereviewstrategies:(1)adefinition(e.g.,“Whatmayhem!That iswhenthereisalotofmessandtrouble.”),(2)aclosed-endedquestion(e.g., “Ismayhemcalmoralittlecrazy?”),and(3)anopen-endedquestion(e.g., “Howismayhemdifferentthancalmandpeaceful?”).
Picturecardreviews1–6 Multi-strategyreview “Mayhemiswhenthereisalotoftroublehappening.”[Onlaterdays,try
havingchildrenguessthewordwhenyouprovidethedefinition,beforeyou showthepicture]
“Canyoushowmewhatthatlookslike?”[Gesture;kidsrepeatgesture.]
sions(SD=0.55,range=2–4days)and3.78ofthe4playsessions
(SD=0.60,range=1–4days).Oneclassroomteacherdidnotprovide
attendancedataforherstudents(n=12).
7.2.2.1. Book-reading. Teacherstaughtall16targetwordsoverthe
courseof4readingsessions(seeTable4).SetAwasthefocusduring
sessions1and3andSetBwasthefocusduringsessions2and4.The wordsinthefocussetwerefirstreviewedinapre-readingportion inwhicheachoftheday’s8targetwordswastaughtthrougha scriptedmultistrategyreview,includingdefinitions,referencesto theword’suseinthestory,apicturecardshowingthewordina differentcontextfromthestory,havingchildrenguessorrepeatthe word,andhavingchildrenusegestures(seeTable5foranexample). AsinStudy1,duringthebook-readingportionofthesession, theteacherusedascriptthatoutlinedtheuseofsimilarstrategies forreviewingtheday’sfocuswords(i.e.,SetA orB)when they wereencounteredinthebook’stext.Thescriptalsoexplainedhow tobrieflydefinetheothersetofwords.Forexamples,seeTable5. Storycomprehensionpromptsrelated toresearchquestionsnot includedinthispaperwerealsoincorporatedintothescript. 7.2.2.2. Play. AsinStudy1,onecompletesetofbook-relatedtoys wasavailabletothegroupsofchildrenduring4small-groupplay sessionsapproximately 10min long.Guidancecards (similarto those usedbyISsin Study1’sGuided Playcondition) provided teacherswithsuggestionsforthevariousportionsofdifferentplay days,aswellasgeneraltips.Theteacherwasinstructedtojoin children’splayduringallsessionsandincorporatetheday’s desig-natedtargetwordsintotheplayassmoothlyaspossible,according togeneralguidelines(seeTable4).
7.2.2.3. Days1&2. UnlikeinStudy1,playsessionsdidnot usu-allyfollowimmediatelyafterbook-readings.Therefore,theteacher startedplayDays1and2 withapre-playvocabularyandstory reviewusingillustrationsfromthebook.Onagivenday,halfofthe Read+Playwordswerethemainfocusofthepre-playreview(e.g.,
the4Read+PlaywordsfromSetA),andtheotherhalfwereonly brieflydefinedduringthispre-playportion(e.g.,the4Read+Play wordsfromSetB).Whenthewordwasafocusword,theteacher wasinstructedtousemultiplestrategiestoelaborateontheword’s meaning,promptchildrentosaytheword,andtrytousegesture toillustratemeaning.OnDay2,whenthewordwasnotafocus word,thepre-playreviewwasjustabriefdefinition(seeTable5 forexamples).
Theplayactivityitselfincorporatedsomeoftheadult-provided re-enactmentstructurethatwasintegraltotheDirectedPlay con-ditioninStudy1.However,whiletheISsinStudy1usedascriptto deliverprecisedirectionstochildren,hereteacherswereaskedto gentlysuggestideasthroughquestionsanduseofthetoyswhile respectingchildren’schoices,asintheGuidedPlayconditionin Study1.Theguidancecardssuggestedhowtoplayfullyrelatetarget wordstowhatchildrenmightbedoingandprovidedbrief defini-tionsandexamplequestionstopromptchildren’swordusage(see Table5).Thegoalwastoincorporateeachoftheday’s4assigned words(i.e.,fromSetAonDay1,SetBonDay2)atleastthreetimes, usingmultiplestrategies.
7.2.2.4. Days3&4. OnDays3and4,theteacherbeganby pre-sentingchildrenwithachoiceofstoryre-enactmentorunrelated playscenarios(i.e.,beach,birthdayparty).Offeringchildrenchoices wasa wayto givethem control, akinto theGuided Play con-ditioninStudy1,butalsoprovidesomestructure,whichmight havebeenabeneficialcomponenttotheDirectedPlaycondition ofStudy1.Therewereguidancecardsspecifictoeach scenario. Thepre-playportionforthesedaysdidnotincludestoryreview, sincethechildrenhadalreadyencounteredthestoryrepeatedlyby thatpoint.Instead,theguidancecardhadsuggestionsforbriefly usingall8Read+Playwordsinameaningfulwayduringtoy dis-tribution.Then,thereweresuggestionsfortheteacherforhelping childrenbegintoplay(e.g.,“Howiseveryonegoingtogettothe playground?”).ThesedayswerethussimilartoStudy1’sGuided Playcondition,butwithsuggestionsprovidedtoteachersfor
