evaluation of the between the lions mississippi literacy initiative 2007 2008 · 2011-05-13 ·...
TRANSCRIPT
EVALUATIONOFTHEBETWEENTHELIONS
MISSISSIPPILITERACYINITIATIVE2007‐2008
DeborahL.Linebarger,Ph.D.
Children’sMediaLab
AnnenbergSchoolforCommunicationUniversityofPennsylvania
March19,2009
2|P a g e
2007‐2008
We would like to thank the talented and dedicated efforts of the staff and students in Mississippi and Pennsylvania who helped with this project including Cathy Grace, Stacey Callender, Nikki McCelleis, Jessamine Huffman, Beverly Willis, and Katie McMenamin. In addition, we would like to thank the children, families, teachers, directors, and staff at the child care centers where we were fortunate enough to work. Without their time, energy, and enthusiasm, this project would not have been completed. For additional information, please contact: Dr. Deborah L. Linebarger Director, Children’s Media Lab Annenberg School for Communication University of Pennsylvania 3620 Walnut Street Philadelphia, PA 19104 215.898.7855 (lab) 215.898.2024 (fax) Email: [email protected]
3|P a g e
IntroductionSubstantialevidencedocumentsthesuccessoftheprogramBetweentheLionsinhelpingyoungchildrenacquireearlyliteracyskills,particularlythoseyoungchildrenwhomaybeat‐riskforreadingfailureduetoeconomicdisadvantage.Youngchildrenwhowatchthisprogramhavedemonstratedconsistentgainsacrossalphabetknowledge,phonologicalawareness,phonemicawareness,andfluencythroughsimpleexposure(Linebarger,Kosanic,Greenwood,Doku,2004;Uchikoshi,2006)andexposurecombinedwithclassroommaterials(Linebarger,2006;Prince,Grace,Linebarger,Atkinson,&Huffman,2001).ThestudypresentedhereextendstheresearchontheuseofBetweentheLions,supplementalclassroommaterials,andamentoringprogramtoanewsampleofteachers,classrooms,andchildren.
• Hypothesis1:Thetreatmentgroupwouldoutperformthecontrolgroupatthepost‐test.• Hypothesis2:Themaintenancegroupwouldoutperformthecontrolgroupatthepost‐test.• Hypothesis3:Thereshouldbenoor,atmost,aslightdifferencebetweenthetreatmentgroup
andthemaintenancegroup.
InterventionChildcareclassroomswerecategorizedintothreegroups:Treatment,Maintenance,andControl.
• 9TreatmentclassroomswereencouragedtouseBetweentheLionsintheirclassroomsandeachreceivedasetoftheBetweentheLionsPreschoolLiteracyInitiativeclassroommaterialsaswellastrainingandmentoringdescribedbelow,beginninginNovember2007followingthecollectionofpre‐testdata.Forbothteachersandstudents,thiswastheirfirstyearofparticipationusingBetweentheLions.
• 9MaintenanceclassroomshadpreviouslybeenTreatmentclassrooms,sotheseteachershadreceivedtheBetweentheLionsclassroommaterials,training,andmentoringinapreviousschoolyear.Thechildrenintheseclassroomshadnotpreviouslybeenexposedtotheintervention.TheteacherswereencouragedtocontinueusingBetweentheLionsintheirclassroomsoncethepre‐testdatawascollected.ThreeoftheoriginalmaintenanceclassroomswerenotincludedinthisstudybecauseintwooftheclassroomstheoriginalteachersleftandthenewteachershadnotpreviouslybeeninTreatmentclassroomsthatreceivedtraining/mentoring.AthirdMaintenanceclassroomwaspulledfromtheprojectbecausetheteacherstoppedusingtheBetweentheLionsmaterialsandwasreplacedbyanotherclassroomthathadnotbeenaTreatmentclassroomthepreviousyear.
4|P a g e
• 13ControlclassroomswerenotprovidedwithanyBetweentheLionsmaterials,training,ormentoring,butcouldopttoreceivethematerialsandtrainingfollowingthecollectionofpost‐testdata.TheycouldalsobeputonthelisttobeenteredintotherandomselectionprocesstobecomeTreatmentclassroomsthefollowingyear.
ClassroomMaterialsTreatmentandMaintenanceclassroomsreceivedthefollowingmaterials:
• BetweentheLionsPreschoolLiteracyInitiativeLessonPlans:Aseriesof30theme‐basedweeklylessons,organizedinto5units,withideasandstrategiesforconductingdailyliteracyactivitiesLessonsincludeideasforactivitycentersalongwithwhole‐andsmall‐groupactivities.Additionallessoncomponentsincludeaweeklyplanner,suggestionsforsettinguptheclassroom,AFamilyLetter,theme‐relatedsongsandpoems,andarecommendedbooklist.ThelessonsfollowascopeandsequencedesignedtoaddressallkeyearlyliteracyskillsandarealignedwithstatepreschoolstandardsandHeadStartframeworks.
• DVDs:EachunitisaccompaniedbyaDVDwithsixBetweentheLionsepisodeseditedspecificallyforpreschooluse,foratotalof30episodesplusbonustracks.
• Books:Eachlessonincludesatleasttwoaccompanyingbooks;thediversecollectionof61tradebooksincludesfolktales,contemporarystories,rhymingbooks,alphabetbooks,conceptbooks,andnonfiction.
• SongandPoemCharts:Illustratedsongandpoemchartslinkedtothelessonsaidinteachingchildrenaboutconceptsofprintandthesoundsofspokenlanguage.
• AdditionalClassroomResources:Abinofadditionalmaterialsincludeslettercards,wordcards,picturecards,storyfiguresforaVelcroboard,magneticletters,lionpuppets,andotherbasicsuppliesforusewiththelessons.
TrainingandMentoringImmediatelyafterthepre‐testingiscompleted,teachersintheTreatmentclassroomsparticipateinfivehoursoftrainingprovidedbymentorsfromMississippiPublicTelevision.ThetrainingprovidesanoverviewoftheBetweentheLionscurriculummaterialsandstrategiesforsettinguptheirclassroomstoencourageliteracy.Teachersalsogainanunderstandingofhowthementorswillworkwiththemduringtheintervention.TeachersintheMaintenanceclassroomsarealsoinvitedtoattendthistraining.TeachersinTreatmentclassroomsreceivetwothree‐hourvisitsfromamentoreveryweek,for16weeks,foratotalof96hoursofmentoring.TeachersintheirfirstyearasaMaintenanceclassroomreceivetwothree‐hourvisitspermonthfor16weeks,foratotalof24hoursofmentoring.TeachersintheirsecondyearasaMaintenanceclassroomreceiveonethree‐hourvisitpermonthfor16weeks,foratotalof12hoursofmentoring.Thementorsguideteachersinsettingupandorganizingtheirclassrooms,modelBetweentheLionslessons,andprovidefeedbackonhowteacherscarryouttheselessons.After16weeks,thementorsdonotreturntotheclassroomsuntilpost‐testinghasbeencompleted.Atthepointofpost‐testing,mostteachershavecompleted17to20ofthe30lessons.Asaresult,theyhavenotcoveredalltheletters,andhavenotdevotedasmuchfocusoninitialsoundfluencyoron
5|P a g e
blendingbeginningandendingsoundsandwords,whicharethefocusoflaterunits.Afterpost‐testing,theTreatmentandMaintenanceclassroomscontinueusingtheBetweentheLionslessons.
MethodParticipantsTable1providesdetailedinformationaboutteacherandchildparticipants.TeachersandChildCareCentersThirty‐oneteachersandclassroomsin23differentchildcarecenterswererecruitedtoparticipateinthisstudy.Nearlyequalnumbersofclassroomswerematchedacrossthethreeconditionsusingteachereducation,classroomage,percentageofsubsidies,censusdataonpercentoffamilieslivinginpovertybyzipcode,andcentersize.Whiletherewasasystematicattempttomatchclassroomsacrossconditions,theresultspresentedinTable1indicatethatthereweredifferencesatthestartofthestudyassociatedwithcentersandclassrooms.
DroppedClassrooms.TwoteachersoriginallyintheMaintenanceconditionlefttheirrespectivecentersandwerenotreplacedwithteacherswhoweretrainedtousetheinterventionmaterials.Athirdteacherwhowasassignedtothecontrolconditionwasatacenterwhereotherclassroomswereinthetreatmentormaintenancegroups.Therewassubstantialevidencethatthisteacherimplementedtheinterventiondespitearequestnottodoso.DroppedCenters.Twochildcarecentersweredroppedfromtheoriginalsampleduetotheteacherchangesdescribedabove.
ChildrenTheoriginalsampleconsistedof319childrenattendingpreschoolsandchildcarecentersinMississippi.Ofthistotal,23childrenwhowereinthethreeclassroomsdescribedaboveweredroppedfromthefinalanalyses;therefore,thefinalsampleconsistedof296children(MeanAge=59.93months,SD=16.87months).Matchingbyclassroomcharacteristicsresultedin111childrenin9treatmentclassrooms,95childrenin9maintenanceclassrooms,and90childrenin13controlclassrooms.Justoverhalfofthechildrenwereboys(i.e.,53.3%).Childrenparticipatedintheassessmentsattheirchildcarecenters.All21centersservedchildrenwhowerepredominantlyfromeconomicallydisadvantagedbackgrounds.Nootherdemographicinformationwascollected.
1|P a g e
Table 1. Characteristics of the Sample ***p < .001; **p < 0.01; ** p < 0.0
1Threeclassroomsand23childrenweredroppedfromtheanalysesduetoeitherteacherchangesorimplementationofthecurriculuminacontrolclassroom.
Attribute Description All Treatment Maintenance Control InitialGroupDifferences
OriginalTotal EntireSample 319 RevisedTotal1 296 111 95 90
CenterCharacteristics
TotalNumber 23 7 9 7
%Subsidy 69.0% 67.9% 86.5% 55.8% F(2,293)=67.19***
ClassroomCharacteristics
TotalClassrooms 31 9 9 13
FOURS 156 53 30 73 F(2,153)=0.26,ns
MeanAge(months) 54.1 54.3 54.4 53.9 FIVES 140 58 65 17 F(2,137)=5.86**
MeanAge(months) 65.6 64.4 66.8 65.4
49.3%HS 57.9%HS 35.9%HS 60.4%HS Χ2=69.77***29.1%CDA/AA 13.7%CDA/AA 52.4%CDA/AA 17.1%CDA/AA
19.9%BA 23.2%BA 13.3%BA 22.5%BA
TeacherCharacteristics
Education
1.7%MA 5.3%MA 0MA 0MA BTLLessonsCompleted 13.18 18.35 23.98 0 F(2,293)=1075.48***
ChildCharacteristics
Gender53.4%boys
52.6%boys 47.8%boys 58.6%boys Χ2=2.35,ns
ChildAge 59.6months 62.8 56.1 59.6 F(2,293)=25.97***
7|P a g e
Measures
MeasureswereselectedordevelopedtoassesstargetedskillssupportedthroughtheBetweentheLionsLiteracyInitiativeandtoreflectthekeyearlyliteracyskillsasdescribedbyNeumanandRoskos(2005).Theseskillsarelanguagedevelopment,letterknowledge,phonemicawareness,andprintconventions.Normative,orstandardized,measurestappedintoeachofthesedomainsusingmultipleindices.
DemographicInformationChildrenandFamiliesChildren’sgenderanddatesofbirthwererecorded.TeachersandChildCareCentersTeachersprovidedinformationregardingtheiryearsofeducationandanydegreestheypossessed.Centersindicatedthepercentageofchildrenandfamiliesateachcenterwhoreceivedsubsidies.
IndicatorsoftheClassroomEnvironmentTheclassroomliteracyenvironmentwasexaminedusingtheEarlyLiteracyandLanguageClassroomObservationTool(ELLCO).TheELLCOmeasuredliteracyandlanguagepracticesandmaterialsinearlychildhoodclassroomsacross4components:theGeneralClassroomObservationandTeacherInterview;theLiteracyEnvironment;theLanguage,Literacy,andCurriculumAssessment;andtheLiteracyActivitiesRatingScale.
1. GeneralClassroomObservationandTeacherInterview:measuresorganization,contents,technology,andclassroomclimateandmanagement
2. LiteracyEnvironment:measuresavailability,content,anddiversityofreading,writing,andlisteningmaterials.
3. Language,Literacy,andCurriculumAssessment:measuresreadingandwritinginstruction,orallanguageuse,culturalsensitivity,andassessmentapproaches
4. LiteracyActivitiesRatingScale:measureshowmanytimesandforhowlongnineliteracybehaviorsoccurredintwocategories,BookReadingandWriting
IndicatorsofLanguageDevelopment
IGDIPictureNamingTask GeneralizedvocabularyknowledgewasevaluatedusingthePictureNamingTask,atoolthatmeasuredchildren’sexpressivelanguageknowledge(PNT,Missall&McConnell,2004).ThePNTisanIndividualGrowthandDevelopmentIndicator(IGDI)usedtotrackpreschoolers’vocabularyacquisitionona
8|P a g e
regularbasisovertime.Childrenwerepresentedwithimagesofobjectsfamiliartopreschoolersoneatatimeandaskedtonamethepicturesasfastaspossibleforoneminute.Categoriesofobjectsusedincludedanimals,food,people,householdobjects,gamesandsportsmaterials,vehicles,tools,andclothing.Psychometricpropertiesforthismeasurewereadequate.Specifically,alternateformsreliabilityrangedbetween.44and.78whiletest‐retestreliabilityoveratwo‐weekperiodwas.69.ConcurrentvalidityestimateswiththePeabodyPictureVocabularyTest–3rdEdition(Dunn&Dunn,2000)andwiththePreschoolLanguageScale–3(Zimmerman,Steiner,&Pond,1992)wereadequate,.53to.79.ThePNTwasalsosensitivetodevelopmentalstatusandgrowthovertime.Childrenidentified21.4picturesatthepretest(SD=6.7).Benchmarkingnormswereprovidedbytheauthors:scoresat59monthsaveraged16.97fortypicallydevelopingchildren;16.51forchildrenfromlowincomebackgrounds;and14.13forchildrenwithidentifieddisabilities(Missall&McConnell,2004).
IndicatorsofLetterKnowledge
PALS‐PreK‐AlphabetKnowledge ThePALSPreKAlphabetKnowledgeTask(Invernizzi,Sullivan,&Meier,2002)wasusedtoevaluatealphabetletterknowledge.ThedevelopersofthePALSincludedthreedifferenttasksthattappedintovariouscomponentsofletterknowledge:1)identificationofthe26UpperCaseletters;2)identificationofthe26LowerCaseletters;and3)identificationofthesoundsassociatedwith23lettersand3digraphs.Childrenarefirstpresentedall26UpperCaselettersinarandomorder.Tobeeligibletoproceedtothesecondtask,identificationofall26LowerCaseletters,thechildmustcorrectlyidentify16UpperCaseletters.TobeeligibletoproceedfromLowerCaseletterstoLetterSounds,thechildmustcorrectlyidentify9LowerCaseletters.Psychometricsareadequatewithreportedreliabilitiesrangingfrom.74to94.Withthistask,wederivedthreetypesofscores:1)thenumberoflettersorsoundsachildcouldcorrectlyidentify;2)thenumberofchildrenineachviewinggroupwhowereabletoidentifyanyLowerCaselettersorLetterSounds(i.e.,onlychildrenwhoreachedacertaincut‐offwereabletoproceedtoLowerCaselettersandLetterSounds);and3)fluencyscores(i.e.,thenumberofsecondsittooktoidentifyoneletterorsound).
1. NumberofLettersorSoundsCorrectlyIdentified.Thetotalnumberofuppercase,lowercase,andlettersoundswererecorded.
2. IdentificationofAnyLowerCaseNamesorLetterSounds.Childrenwerepresentedwiththesetasksiftheywereableto1)identify16ormoreUpperCaselettersand2)9ormoreLowerCaseletters.
3. FluencyScores.Children’sperformanceoneachofthe3subscales(i.e.,UpperCase,LowerCase,LetterSounds)wastimed.Then,thenumberoflettersorsoundsaccuratelyidentifiedwasdividedbythenumberofsecondsittookthechildtocompleteeachtask.Thisproducedaletterorsoundidentificationpersecondrate.AllchildrenwereadministeredtheUpperCasetask;therefore,allchildrenhadafluencyscoreassociatedwithUpperCaseLetterKnowledge.OnlythosechildreneligibletocompletetheLowerCaseLetterKnowledgeandtheLetterSoundstaskswereincludedinthoseanalyses.
9|P a g e
IndicatorsofPhonemicAwareness
IGDIInitialSoundsFluency TheDIBELSInitialSoundFluencytaskisanindividuallyadministeredandtimedmeasureofchildren’sabilitytorecognizeandproducetheinitialsoundinanorallypresentedword,acomponentofphonemicawareness.Theexaminerpresentsfourpicturestothechild,nameseachpicture,andthenasksthechildtoidentify(i.e.,pointtoorsay)thepicturethatbeginswiththesoundproducedorallybytheexaminer.Forexample,theexaminersays,"Thisissink,cat,gloves,andhat.Whichpicturebeginswith/s/?"andthestudentpointstothecorrectpicture.Thechildisalsoaskedtoorallyproducethebeginningsoundforanorallypresentedwordthatmatchesoneofthegivenpictures.Theexaminercalculatestheamountoftimetakentoidentify/producethecorrectsoundandconvertsthescoreintothenumberofinitialsoundscorrectinaminute.
PALS‐PreK–AlphabetKnowledgeAdescriptionofthistaskwasdetailedabove.OnlytheindicesthatwerederivedfromthismeasuretorepresentPhonologicalandPhonemicAwarenessdiscussedbelow.
1. IdentificationofAnyLetterSounds.ThepercentageofchildrenineachviewinggroupwhowereeligibletotaketheLetterSoundstaskwasrecorded.
2. NumberofSoundsCorrectlyIdentified.Thenumberoflettersoundsachildwasabletoidentifycorrectlywasrecorded.
3. LetterSoundsFluency.Children’sperformanceontheLetterSoundssubscalewastimed.Then,thenumberofitemsaccuratelyidentifiedwasdividedbythenumberofsecondsittookthechildtocompleteeachtask.Thisproducedasoundidentificationpersecondrate.OnlythosechildreneligibletoattempttheLetterSoundstaskwereincludedinthoseanalyses.
IndicatorsofPrintConventions
PrintandStoryConceptsTasksThisassessmentwasadaptedfromtheHeadStartFACESSurvey(informationavailableonline:http://www.acf.hhs.gov/programs/opre/hs/faces/instruments/child_instru02/language_story.html)toexaminechildren’sunderstandingofbasicstoryconceptsincludingbookknowledge,printknowledge,andreadingcomprehension.Bookknowledgeexaminedchildren’sfamiliaritywithstorybooksandprintconventionssuchaswherethefrontofthebookis,wheretobeginreading,anddifferentiatingprintfrompictures.Printknowledgeexaminedchildren’sknowledgeofthemechanicsofreadingincludingreadingfromlefttoright,toptobottom,andword‐by‐wordpointing.Readingcomprehensionmeasuredchildren’sknowledgeofastoryplotandrequiredthemtoanswerquestionsbasedonpresentedstorycontent(e.g.,whatissaidgoodnighttoinGoodnightMoon)andwellastogenerateinferences(e.g.,howdoesacharacterfeel)andtomakepredictions(e.g.,whatdoyouthinkhappensnextinthisstory).Differentbookswereusedateachtestingpoint:GoodnightMoonbyMargaretWiseBrownwasusedatPre‐TestandWhere’sMyTeddy?byJezAlboroughwasusedatPost‐Test.Whilemostquestionswerebasedonascoringsystemof(0)incorrectand(1)correct,someofthecomprehensionquestionswereworthupto3points.Eachprintandstoryconstructwassummedtoformthreescoresforanalysis:bookknowledge,printknowledge,andreadingcomprehension.
10|P a g e
CombinedEarlyLiteracySkills
GetReadytoRead!ScreenerThisscreener,consistingof20items,assessedprintknowledge(i.e.,knowledgeofthelettersofthealphabet);bookknowledge(recognitionofhowbooksworkincludingthedifferencebetweenwordsandimages);phonologicalawareness(i.e.,understandingthatspokenwordsarecomposedofindividualsounds);phonics(i.e.,recognitionofthesoundslettersmake);andwriting(i.e.,understandinghowtextshouldlook:lettersgroupedtogetherintowords).Eachitemrequiredthechildtoselectaresponsefromagroupoffourpictures(orfourletters,words,etc.).Example:“Thesearepicturesofabook.Findtheonethatshowsthebackofthebook.”Example:“Findtheletterthatmakesatuhsound.”Example:“Somechildrenwrotetheirname.Findtheonethatiswrittenthebest.”Childrenweregivenascoreofa(1)foreverycorrectanswerprovidedanda(0)foreveryincorrectanswerprovided,withamaximumscoreof20points.Scoresgreaterthan11arepredictiveofreadingsuccessby2ndgrade.
AnalyticalApproachRepeated‐measuresAnalysisofCovariance(ANCOVA)isaprocedurethatcanbeusedtostatisticallycontrolforinitialgroupdifferenceswhenevaluatinginterventioneffectsonoutcomemeasures.Inthesemodels,bothInterventionGroupandChild’sAgewereincludedasfactors.Threecovariateswereconstructedtoextractthevarianceassociatedwithvariablesthatwerefoundtorelatetotheoutcomesofinterestorthatsignificantlyvariedbygroup.Theclassroomliteracyenvironment,theteacher’seducation,andachild’spre‐testperformancewereusedascovariatesintheanalyses.Whenmultipletestswereconductedforeachsetofoutcomes,BonferroniadjustmentsofthealphalevelweremadetoreduceType1errorrates(i.e.,findingasignificantdifferencewhenonedoesnotexist).Fortheseanalyses,onlysignificanteffectsassociatedwithGrouparereportedinthetext(i.e.,Group;WavebyGroup,AgebyGroup).Alongwiththestatisticalsignificancetests,effectsizesarealsoreported.FactorDetails
Group:ThisBetween‐Subjectsfactortestedformeandifferencesamongthethreepossibleinterventiongroups.Therewere3levelsassociatedwiththisfactor.
• TheTREATMENTgroupiscomposedofchildren,teachers,andclassroomswhoparticipatedintheBTL‐LIinterventionforthefirsttime.Therewere111childreninthisgroup.
• TheMAINTENANCEgroupiscomposedofchildren,teachers,andclassroomswhoalsoparticipatedintheBTL‐LIintervention;however,theteachershadpreviouslyreceivedinterventiontrainingandmentoringbetween1and3yearsprior.Therewere95childreninthisgroup.
• TheCONTROLgroupiscomposedofchildren,teachers,andclassroomswhodidnotparticipateinanyBTL‐LIinterventiontraining.Therewere90childreninthisgroup.
11|P a g e
Age:ThisBetween‐Subjectsfactortestsformeandifferencesamongdifferentagegroupsofchildren.Thereare2levelsassociatedwiththisfactor.
• TheFOURSgroupofchildreniscomposedof156childrenwhoarebetween46monthsand59months.Onaverage,childreninthisgroupwere54.1months(i.e.,4.5years;SD=3.7).
• TheFIVESgroupofchildreniscomposedof140childrenwhoarebetween60monthsand74months.Onaverage,childreninthisgroupwere65.6months(i.e.,5.0years;SD=3.8).
Wave:ThisWithin‐Subjectsfactortestsformeandifferencesassociatedwithgains(orlosses)frompretesttopost‐test.Thereare2levelsassociatedwiththisfactor.
• ThePRETESTwasadministeredpriortoparticipationinanyoftheinterventionmaterials.
• ThePOST‐TESTwasadministeredattheendofparticipationinanyoftheinterventionmaterials.
12|P a g e
Results
ELLCOClassroomEnvironment2007to2008FoursubscalesfromtheELLCOwereevaluatedforsignificantdifferencesacrossgroups.GeneralClassroomEnvironmentAllclassroomsimprovedfrompre‐testtopost‐test(Figure1),withtreatmentclassesshowingthemostimprovement.Maintenanceclassroomsstartedhigherandremainedatthatlevel2.Controlclassroomsdemonstratedslightgrowthfrompretesttopost‐test.TherewasalsoasignificantdifferencebyGroup3.
2 F(2, 30) = 4.73, p < .05, 3 F(2, 30) = 14.17, p < 0.001
13|P a g e
Figure1.GeneralClassroomEnvironmentSubscaleDifferencesbyGroupAcrossWave
14|P a g e
LiteracyEnvironmentAllclassroomsimprovedfrompre‐testtopost‐test(Figure2),withtreatmentclassesshowingthemostimprovement.Maintenanceclassroomsstartedhigherandremainedsoatthepost‐test4.Controlclassroomsdemonstratedslightgrowthfrompretesttopost‐test.Therewas
alsoasignificantdifferencebyGroup5. Language,Literacy,andCurriculumAllclassroomsimprovedfrompre‐testtopost‐test(Figure3),withtreatmentclassesimprovingmost.Maintenanceclassroomsstartedhigherandremainedsoatthepost‐test6.Controlclassroomsdemonstratedlittlegrowthfrompretesttopost‐test.TherewasalsoasignificantdifferencebyGroup7.
4 F(2, 30) = 4.73, p < 0.05 5 F(2, 30) = 14.17, p < 0.001 6 F(2, 30) = 6.64, p < 0.01 7 F(2, 30) = 24.33, p < 0.001
Figure2.LiteracyEnvironmentSubscaleDifferencesAcrossGroupbyWave
Thissubscaleisintendedasaninventoryoftheliteracy‐specificmaterialsfoundintheclassroom.Thereare
24itemsacross5conceptualdomains:BookArea;BookSelection;BookUse;WritingMaterials;and
WritingDisplays.
Thissubscaleisdesignedtoquantifyorallanguage
facilitation,thepresenceofbooks,diversityintheclassroom,connections
betweenhomeandschool,andapproachesto
assessment.materialsfoundintheclassroom.
Thereare24itemsacross5conceptualdomains:BookArea;BookSelection;BookUse;WritingMaterials;and
WritingDisplays.
Thissubscaleisintendedasaninventoryoftheliteracy‐specificmaterialsfoundintheclassroom.Thereare
24itemsacross5conceptualdomains:BookArea;BookSelection;BookUse;WritingMaterials;and
WritingDisplays.
15|P a g e
Figure3.Language,Literacy,andCurriculumSubscaleDifferencesbyGroupAcrossWave
16|P a g e
LiteracyActivitiesAllclassroomsimprovedfrompre‐testtopost‐test(Figure4),withtreatmentclassesimprovingmost.Maintenanceclassroomsstartedhigherandremainedsoatthepost‐test8(i.e.,thesegroupbywavedifferencesweremarginallysignificant).Controlclassroomsdemonstratedlittlegrowthfrompretestto
post‐test.Therewasalsoasignificantdifference
byGroup9.
ELLCOClassroomEnvironment2006to2008Figures5and6arebasedonasubsetof7teacherswhoweretrainedineither2005(n=3)or2006(n=4)toimplementtheBTLcurriculum.
8 F(2, 30) = 3.13, p < 0.06 9 F(2, 30) = 11.07, p < 0.001
Figure5.LiteracyEnvironmentOverTimefor7TrainedTeachers
Figure4.LiteracyActivitySubscaleDifferencesbyGroupAcrossWave
Thissubscaleisdesignedtoquantifyorallanguage
facilitation,thepresenceofbooks,diversityintheclassroom,connections
betweenhomeandschool,andapproachesto
assessment.materialsfoundintheclassroom.
Thereare24itemsacross5conceptualdomains:BookArea;BookSelection;BookUse;WritingMaterials;and
WritingDisplays.
17|P a g e
OncetrainedtousetheBTLClassroomInterventionmaterials,teachersmaintainclassroomsthatarecharacterizedbyhigh‐qualityliteracymaterialsand,perhapsmoreimportantly,languageinteractionsthatfacilitateliteracydevelopment.
ELLCOClassroomEnvironmentConclusionTeacherswhowerenewtotheBTLClassroomInterventionwereabletocreateclassroomsthatprovidedpreschoolchildrenwith“optimalsupportfortheirlanguageandliteracydevelopment”(p.1;Smith,Dickinson,Sangeorge,&Anastasopoulos,2002).Theseclassroomswerecharacterizedbymoreandhigherqualityreadingandwritingmaterialsandactivitiesaswellasmultipleteacher‐childinteractionsthatareknowntofacilitateorallanguagedevelopmentandearlyliteracyskillacquisition.TeachersinMaintenanceclassroomswereabletomaintainthestructuralcomponentsindicativeofastrongLiteracyEnvironmentfrompreviousyearstothisprojectyear.TheBTLClassroomInterventionhasalwaysbeensuccessfulinhelpingteacherscreateliteracyenvironmentsthatwouldbecapableofsupportingchildren’sdevelopingliteracyabilities.Moreimportantly,thementoringprovidedduringthe2007and2008projectyearindicatedthatteacherswereabletomovebeyondchangingthestructuralfeaturesoftheenvironmentintosystematicallyandconsistentlyalteringboththegeneralclassroomenvironmentaswellastheliteracy‐enrichinginteractionsthatarecrucialtosupportingchildren’soptimallanguageandliteracy‐skilldevelopment.First,theGeneralClassroomEnvironmentsubscaleindicatedthattrainedteachersintroducedgreaterintentionalityinthephysicalorganizationoftheclassroom;providedchildrenwithmultipleopportunitiesforchoiceandfortakinginitiative;andusedmorepositivemanagementstrategies.Thissubscalemoregenerallyevaluatedtheclassroomenvironmentincludingwhetherthisenvironmentcouldsuccessfullysupportgeneralchilddevelopmentandpositiveclassroomexperiencesforbothteachersandchildren.
Figure6.ELLCOSubscalesOverTimefor7TrainedTeachers
18|P a g e
Next,specificenvironmentalfeaturesthatsupportedahighqualityliteracyenvironmentwerescoredusingtheLiteracyEnvironmentsubscale.Thissubscaleprovidesaquickinventoryofthetypesofliteracy‐relateditemsorsupportsthatwerefoundineachclassroom.Bothtreatmentandmaintenanceteacherscreatedliteracyenvironmentsthatincorporateddesignatedbookareas;providedanumberofbooksthatfeaturedvariedtopicsandthatwereingoodphysicalcondition;madethesebookseasilyaccessibletochildren;andprovidedavarietyofwritingtoolsthatwerealsoeasilyaccessible.Theremainingsubscalesassessedthevariousinteractionsthatoccurredinclassroomscenteredaroundlanguageandliteracytopicsincludinginteractionsbetweenteachersandchildrenandbetweenchildrenandthematerialsandactivitiesavailabletothemintheirclassrooms.First,theLanguage,Literacy,andCurriculumsubscalemeasuredateachers’abilitytofacilitateorallanguage;toadoptpositiveandintentionalapproachestobookreading,writing,curriculumintegration,andassessment;andtofacilitatehomesupportforliteracy.Thesetypesofinteractionshelptocreatealanguage‐andliteracy‐richenvironmentthat,inotherresearch,hasbeenlinkedtopositivechilddevelopmentoutcomes.Teachersinboththetreatmentandthemaintenancegroupsconsistentlyscoredbetweenproficientandexemplaryonthesekeyteacher‐childinteractionalitems.Second,thenumberofpositiveLiteracyActivitiesinbothtreatmentandmaintenanceclassroomsalsoincreasedfrompretesttoposttestandweresubstantiallyhigherthancontrolteacherclassrooms.TeachersimplementingtheBTLClassroomInterventionengagedinmoreandlongerfull‐groupandone‐to‐onebook‐readingsessions;modeledwriting;providedwritingassistanceandopportunitiesforwriting;andsetasidetimeforchildrentolookatbooksaloneorwithaclassmate.Takentogether,thesefindingssuggestthatteacherswhohavebeentrainedtodelivertheBTLClassroomInterventionareprimedandabletomakechangestothestructuralfeaturesoftheirclassroomsincludinggeneralclassroommanagementabilities,specificenvironmentalfeaturesthatarenecessaryforliteracysupport,andincreasedopportunitiesforliteracyactivities.Inaddition,thesestructuralchangesalsoresultedinprocessqualitychangesasindexedbyhigher‐qualitylanguage‐andliteracy‐richinteractionsandexperiences.Finally,maintenanceteacherswhohadbeentrainedinpreviousyearswereabletosustainthekindofhigh‐qualityenvironmentneededtofundamentallyshiftyoungchildren’searlylanguageandliteracytrajectories.
ChildLiteracyOutcomesChildrenwereassessedattwotimepoints:fall2007,priortoanyinterventionparticipation,andagaininMarchorAprilof2008,towardtheendoftheacademicyear.Whereavailable,benchmarksandexpecteddevelopmentalrangeshavebeenindicated.TablesareprovidedattheconclusionofeachIndicatorsection.Thesetableslistedallindicatorsassociatedwithalargerconstruct(i.e.,largerconstructsincludedVocabularyKnowledge,LetterKnowledge,PhonemicAwareness,PrintConventions,CombinedEarlyLiteracyTask).Eachtablecontainedacolumnofassessmentsthatmeasuredaspectsoftheoverallindicator.Forinstance,theLetterKnowledgeIndicatorwascomprisedofUpperCaseTotallettersnamed,UpperCaseletternamingrate;AnyLowerCaselettersnamed?;LowerCaseTotallettersnamed;andLowerCaseletternamingrate.Foreachassessmentlistedinatable,therewerecolumnsforeachofthethreeinterventiongroups(i.e.,Treatment,Maintenance,Control).Withineach
19|P a g e
interventiongroup,therewerealsocolumnsthatindicatedwhetherscoreswerethepretestaverageorthepost‐testaverage.Tableswereconstructedusingthefollowinginformation:Allmeansatboththepretestandthepost‐testcontrolledforpre‐interventionknowledge,teachercharacteristicsassociatedwithanindividualchild’sclassroom,andtheclassroomliteracyenvironment.Indicatortableswerecreatedforeachofthe5setsofIndicators:a. IndicatorofLanguageDevelopment:LanguageDevelopmentconsistedofoneindicatorthat
measuredchildren’sexpressivevocabularyknowledge.b. IndicatorsofLetterKnowledge:Therewerefivedifferenceindicesofletterknowledge.Notall
indicatorswereadministeredtoeverychild.Administrationdependedonwhetherchildrenwereabletocompleteenoughitemsontheprevioustask.Specifically:
• AllchildrenwereadministeredtheUpperCaseletternamingtask.ThistaskinvolvedpresentingallUpperCasealphabetletterstoachild.Thechildwasaskedtonameasmanyashe/shecould.Thedatacollectorscoredeachofthe26lettersascorrectorincorrect.Thedatacollectoralsomeasuredhowlong(inseconds)ittookachildtocompletethistask.
• Ifthechildaccuratelyidentified16UpperCaseletters10,he/shewaseligibletoattempttheLowerCaseletterknowledgetask.ThistaskinvolvedpresentingthechildwithallLowerCaselettersinrandomorder.Thechildwasaskedtonameasmanyoftheselettersasshe/hecould.Inadditiontoscoringwhetherornotthechildwasabletoaccuratelyidentifyeachlowercaseletter,thedatacollectormeasured(inseconds)howlongthechildtook.
• Ifthechildwasabletoaccuratelyidentify9ormoreLowerCaseletters10,he/shewaseligiblefortheLetterSoundstask.Thistaskinvolvedpresentingthechildwith23alphabetlettersand3digraphsinrandomorder.Thechildwasaskedtoproducethesoundassociatedwitheach.Inadditiontoscoringwhetherornotthechildwasabletoaccuratelyidentifythelettersounds,thedatacollectoralsomeasured(inseconds)howlongittookforthechildtocompletethetask.
c. IndicatorsofPhonemicAwareness:Therewerethreedifferenceindicesofletterknowledge.Thefirstindex,theIGDIInitialSoundsFluencytask,wasadministeredtoallchildren.ThePALSpreKLetterSoundstaskwasonlyadministeredtothosechildrenwhowereabletoidentify16UpperCaseand9LowerCaselettersaccurately10.Notallindicatorswereadministeredtoeverychild.
d. IndicatorsofPrintConventions:Theseindicatorsincludedbookknowledge,printknowledge,andstorycomprehensionthatwereadministeredwhilereadingabooktogether.
e. CombinedEarlyLiteracyTask:Thismeasurewasadministeredtoallchildrenandwasdesignedtomeasuretheirknowledgeofbookconventions,printconventions,letterknowledge,phonologicalandphonemicawareness,andearlywritingskills.
Table2.Percentageofchildrenwhowereabletoidentifyatleastoneitemcorrectlyoneachtask. Treatment Maintenance Control Pre Post Pre Post Pre PostUpperCase1112 94.7% 97.9% 82.2% 91.1% 89.2% 88.3%LowerCase1314 54.7% 84.2% 26.7% 55.6% 30.6% 61.3%
10 The criteria for proceeding to the next subtest on this overall measure was the same at both the pre-test and the post-test (e.g., to be administered the Lower Case subtest, children at both the pre-test and post-test needed to accurately identify 16 Upper Case letters). 11 Χ2 = 7.33, p < 0.05 12 Χ2 = 6.82, p < 0.05 13 Χ2 = 6.03, p < 0.05 14 Χ2 = 19.68, p < 0.001
20|P a g e
LetterSounds1516 51.6% 76.8% 21.1% 51.1% 25.2% 50.5%
IndicatorsofLanguageDevelopment
IGDIPictureNaming
Thistaskevaluatedyoungchildren’sexpressivevocabularyknowledge.Childrenwereaskedtonameorlabelasmanypicturecardsastheycouldinoneminute.Overall,performancewashighestforchildreninthecontrolgroup(i.e.,15.34)followedbychildreninthetreatmentgroup(i.e.,14.71)andthenchildreninthemaintenancegroup(i.e.,13.48)17;however,theseresultsweremoderatedbychild’sAge18.AllchildrenwhowerecategorizedintotheFOURSagegroupscoredsimilarly.TheperformanceofchildrenwhowerecategorizedintotheFIVESgroupdifferedbygroup:controlchildrenoutscoredtreatmentchildrenwho,inturn,outperformedmaintenancechildren.Follow‐uptestsindicatedthatboththetreatmentandmaintenancegroupsscoredsignificantlylowerthanthecontrolgroup.Differencesbetweenthetreatmentandmaintenancegroupswerenotstatisticallysignificant.
15 Χ2 = 23.79, p < 0.001 16 Χ2 = 18.19, p < 0.001 17 F(2,285)=4.07,p<0.05 18 F(4,285)=2.61,p<0.05
Figure7.IGDIPictureNamingScoresOverallandSplitbyGroupandAge
21|P a g e
Table3.Pre‐TestandPost‐TestMeansfortheLanguageDevelopmentIndicatorbyAgeandGroup
Treatment Maintenance Control
Pre Post Pre Post Pre Post
FOURS
IGDIPictureNaming 13.45 14.40 13.57 13.14 12.95 14.20
FIVES
IGDIPictureNaming 14.26 16.74 13.61 13.59 16.34 17.86
OVERALL
IGDIPictureNaming 13.85 15.57 13.59 13.36 14.64 16.03
22|P a g e
IndicatorsofLetterKnowledgeUpperCaseLetterKnowledge
UpperCaseTotalScores.
Therewasasignificant2‐wayinteractionbetweengroupandchild’sage19(Figure8).ForchildrenclassifiedasFOUR,thetreatmentgroupoutperformedboththecontrolandthemaintenancegroups.ForchildrenclassifiedasFIVE,allgroupsscoredsimilarly;thatis,therewerenosignificantdifferencesintheirknowledgeofUpperCaseletternames.
Figure8.UpperCaseLettersSplitbyGroupandAge
19F(2,288)=3.68,p<0.05
SpringBenchmarkRange:12–21UpperCaseLetters
23|P a g e
UpperCaseLetterNamingSpeed20.Therewasasignificantgroupbywaveinteraction21.Allthreegroupssignificantlyimprovedfrompretesttopost‐test.ChildreninthetreatmentgroupnamedUpperCaselettersthefastestfollowedbycontrolgroupchildrenandthenmaintenancegroupchildren.SeeFigure9.
20Notethattheletternamingspeedrepresentstherateatwhichchildrenwereidentifyingthelettersandisnotnecessarilyareflectionofhowmanyletterswerenamed;thatis,itextrapolatestheratebasedonthetotaltimespentonthetaskandthenumberoflettersnamedcorrectly.Incorrectresponseswerefactoredintotheiroverallnamingratessothatarateof22lettersperminutedoesnotnecessarilymeanthattheycorrectlyidentified22letters. 21F(2,212)=3.89,p<0.05
Figure9.UpperCaseLetterNamingSpeedbyGroupAcrossWave
24|P a g e
LowerCaseLetterKnowledge
AnyKnowledgeofLowerCaseLetters?AlargermajorityofTreatmentchildrenwereabletoidentifyanyLowerCaseLetters(i.e.,84.2%)comparedwithMaintenance(i.e.,55.6%)andControl(i.e.,61.3%)childrenatthepost‐test22.Whenconductingthisanalysisforbothagegroups,moreyoungerchildren(i.e.,FOURs)inthetreatmentgroupcorrectlyidentifiedatleastoneLowerCaseletter(i.e.,80.0%)whencomparedwiththeirpeersinboththecontrol(i.e.,50.9%)andmaintenancegroups(i.e.,50.7%)23.Thedifferencesforolderchildrenweremarginallysignificant24butevidencedthesamepatterns;thatis,86.2%oftheFIVEsinthetreatmentgroupwereabletoaccuratelyidentifyanyLowerCasecomparedwith76.5%oftheFIVEsinthemaintenancegroupand70.7%ofFIVEsinthecontrolgroup.
LowerCaseTotalScores.
Therewasasignificant2‐wayinteractionbetweengroupandage25(Figure10).ChildreninthetreatmentgroupwhowereFOURwereabletoidentifymoreLowerCaseletterswhencomparedwiththeirpeersinboththecontrolandmaintenancegroups.Incontrast,childreninthemaintenancegroupwhowereclassifiedasFIVEsoutperformedtheirpeersinboththetreatmentandcontrolgroups.
22Χ2=19.68,p<0.001 23 Χ2=8.41,p<0.05 24 Χ2=4.40,p<0.11 25F(2,288)=3.05,p<0.05
SpringBenchmarkRange:9–17LowerCaseLetters
25|P a g e
Figure10.LowerCaseLettersSplitbyAgeandGroup
LowerCaseLetterNamingSpeed26.Therewasasignificant2‐wayinteractionbetweengroupandage27(Figure11).ChildrenintheFOURStreatmentgroupidentifiedlowercaselettersmore
quicklythanchildrenintheFOURScontrolandmaintenancegroups.ForbothYOUNGandOLD5s,controlgroupchildrenoutperformedtheirtreatmentandmaintenancegrouppeers.
26 Notethattheletternamingspeedrepresentstherateatwhichchildrenwereidentifyingthelettersandisnotnecessarilyareflectionofhowmanyletterswerenamedcorrectly;thatis,achildmayhavebeenidentifyinglettersatarateof22lettersperminutebutmayhaveonlynamed15letterscorrectly. 27 F(2,71)=3.20,p<0.05
26|P a g e
Figure11.LowerCaseLetterNamingSpeedbyGroupAcrossWave
27|P a g e
Table4.Pre‐TestandPost‐TestMeansforAllLetterKnowledgeIndicesbyAgeandGroup Treatment Maintenance Control
Pre Post Pre Post Pre Post
FOURS
UpperCaseTotal 12.56 17.61 10.88 12.54 11.51 13.91
UpperCaseRate 12.14 23.63 12.07 15.71 11.81 19.18
AnyLowerCase? 33.3% 80.0% 16.4% 50.7% 24.5% 50.9%
LowerCaseTotal 8.44 13.59 7.47 9.19 8.19 10.39
LowerCaseRate 23.91 32.20 17.15 25.39 16.41 28.95
FIVES
UpperCaseTotal 10.54 16.49 12.04 16.47 10.86 17.55
UpperCaseRate 10.59 22.08 12.80 18.41 11.70 22.58
AnyLowerCase? 64.6% 86.2% 70.6% 76.5% 36.2% 70.7%
LowerCaseTotal 5.46 12.06 8.03 12.22 5.52 12.75
LowerCaseRate 18.28 29.40 22.76 24.46 24.37 33.52
OVERALL
UpperCaseTotal 11.55 17.05 11.46 14.50 11.19 15.73
UpperCaseRate 11.36 22.86 12.43 17.06 11.76 20.88
AnyLowerCase? 54.7% 84.2% 26.7% 55.6% 30.6% 61.3%
LowerCaseTotal 6.95 12.83 7.75 10.71 9.29 9.13
LowerCaseRate 21.10 30.80 19.96 24.92 20.39 31.24
28|P a g e
IndicatorsofPhonemicAwarenessIGDIInitialSoundsFluency
Therewasasignificant3‐wayinteractionamonggroup,age,andwave28(Figure12).BothmaintenanceandcontrolgroupchildrenwhowereclassifiedasFOURoutperformedtheirpeersinthetreatmentgroup.ChildreninthemaintenancegroupwhowereclassifiedasFIVEobtainedhigherscoresfollowedbyFIVEsinthetreatmentgroupwho,inturn,scoredhigherthanFIVEsinthecontrolgroup.
Figure12.IGDIInitialSoundsFluencySplitbyAgeandGroup
28F(2,288)=3.51,p<.01
FallKindergartenBenchmarks:0‐3AtRisk4‐7SomeRisk8+LowRisk
At Risk
Some Risk Low Risk
29|P a g e
LetterSoundsKnowledge
AnyKnowledgeofLetterSounds?AlargermajorityofTreatmentchildrenwereabletoidentifyanyLetterSounds(i.e.,76.8%)comparedwithMaintenance(i.e.,51.1%)andControl(i.e.,50.5%)childrenatthepost‐test29.Whenconductingthisanalysisforbothagegroups,moreyoungerchildren(i.e.,FOURs)inthetreatmentgroupcorrectlyidentifiedatleastoneLetterSound(i.e.,76.7%)whencomparedwiththeirpeersinboththecontrol(i.e.,43.4%)andmaintenancegroups(i.e.,46.6%)30.Thedifferencesforolderchildrenweremarginallysignificant31butevidencedthesamepatterns;thatis,76.9%oftheFIVEsinthetreatmentgroupwereabletoaccuratelyidentifyanyLetterSoundscomparedwith70.6%oftheFIVEsinthemaintenancegroupand56.9%ofFIVEsinthecontrolgroup.
LetterSoundsTotalScore.
Performancewashighestforchildreninthemaintenancegroup(i.e.,5.73)followedbychildreninthetreatmentgroup(i.e.,5.33)andthenchildreninthecontrolgroup(i.e.,4.08)32.Therewasalsoasignificant2‐wayinteractionbetweengroupandage33(Figure13).ChildreninthetreatmentgroupwhowerecategorizedasFOURsoutperformedtheirpeersinthecontrolandmaintenancegroups.ChildreninthemaintenancegroupwhowerecategorizedasFIVEsoutperformedtheirpeersinthetreatmentgroupwho,inturn,outperformedthosechildreninthecontrolgroup.
29Χ2=18.19,p<0.001 30Χ2=9.70,p<0.01 31Χ2=5.70,p<0.06 32F(2,288)=4.06,p<0.05 33F(2,288)=7.39,p<0.001
SpringDevelopmentalRange:4–8LetterSounds
30|P a g e
Therewasalsoasignificant2‐wayinteractionbetweenGroupandWave34(Figure14).Atthepost‐
test,bothtreatmentviewersandmaintenanceviewersoutperformedtheircontrolgrouppeers
LetterSoundsNamingSpeed.Therewerenosignificantdifferencesbygrouporageforchildren’slettersoundnamingspeed.
34F(2,288)=11.08,p<0.05
Figure8.LetterSoundsSplitbyAgeandGroup
Figure9.LetterSoundsbyGroupAcrossWave
31|P a g e
Table5.Pre‐TestandPost‐TestMeansforAllPhonemicAwarenessIndicesbyAgeandGroup
Treatment Maintenance Control
Pre Post Pre Post Pre Post
FOURS
IGDIInitialSoundsFluency 6.12 8.72 9.26 10.50 8.29 10.31
AnyLetterSounds? 33.3% 76.7% 9.6% 46.6% 20.8% 43.4%
LetterSoundsTotal 3.52 7.86 3.73 5.29 4.71 5.79
LetterSoundsRate 7.40 10.36 6.07 7.73 5.81 9.32
FIVES
IGDIInitialSoundsFluency 8.98 12.69 5.90 16.06 7.47 11.06
AnyLetterSounds? 60.0% 76.9% 70.6% 70.6% 29.3% 56.9%
LetterSoundsTotal 2.84 7.08 5.04 8.84 1.91 3.92
LetterSoundsRate 5.95 10.58 9.41 11.06 5.81 9.32
OVERALL
IGDIInitialSoundsFluency 7.55 10.70 7.58 13.28 7.88 10.68
AnyLetterSounds? 51.6% 76.8% 21.1% 51.1% 25.2% 50.5%
LetterSoundsTotal 3.18 7.47 4.39 7.07 3.31 4.85
LetterSoundsRate 6.67 10.47 7.74 9.40 7.08 8.96
32|P a g e
IndicatorsofPrintConventionsBookKnowledge
TherewasasignificantmaineffectofGroup35(Figure15).Childreninthetreatmentgroupoutperformedtheirpeersinthemaintenanceandcontrolgroups.
PrintKnowledge
Therewasasignificant2‐wayinteractionbetweenGroupandAge36(Figure16).TheFOURStreatmentandmaintenancegroupsobtainedhigherscoresonprintknowledgewhencomparedwithcontrolgroupFOURs.TheFIVEstreatmentandcontrolgroupsoutperformedtheirpeersinthe
maintenancegroup.
35F(2,288)=5.29,p<0.01 36F(2,288)=3.79,p<0.05
Figure11.PrintKnowledgeScoresbyAgeandGroup
Figure10.BookKnowledgeScoresbyGroup
33|P a g e
Comprehension
Therewasasignificant2‐wayinteractionbetweengroupandwave37(Figure17).Childreninboththetreatmentandcontrolgroupsoutperformedtheirpeersinthemaintenancegroupatthepost‐
test.
37 F(2,288)=15.26,p<0.01
Figure17.ComprehensionScoresSplitbyGroup
34|P a g e
Table6.Pre‐TestandPost‐TestMeansforAllPrintConventionsIndicesbyAgeandGroup
Treatment Maintenance Control
Pre Post Pre Post Pre Post
FOURs
BookKnowledge 2.21 2.39 1.85 2.11 1.99 1.58
PrintKnowledge 0.67 1.33 0.58 1.25 0.59 0.74
Comprehension 4.46 5.58 5.22 4.54 4.25 4.91
FIVEs
BookKnowledge 2.45 2.66 2.46 2.37 2.29 2.14
PrintKnowledge 0.81 1.30 0.18 0.98 0.84 1.26
Comprehension 5.03 5.63 5.22 4.54 5.08 5.72
OVERALL
BookKnowledge 2.33 2.53 2.16 2.24 2.14 1.86
PrintKnowledge 0.74 1.32 0.38 1.12 0.72 1.00
Comprehension 4.75 5.61 5.05 4.41 4.67 5.32
35|P a g e
CombinedEarlyLiteracySkillsGetReadytoRead!Screener
Therewasasignificantmaineffectofgroup38.Childreninthetreatmentgroupoutperformedtheirpeersinthemaintenancegroupwho,inturn,outperformedtheirpeersinthecontrolgroup.Thiseffectwasmoderatedbywaveofassessment39(Figure18).Treatmentandmaintenancechildrenscoredhigherthantheircontrolgrouppeersatthepost‐test.
38 F(2,288)=3.74,p<0.05 39 F(4,285)=5.42,p<0.001
StrongGRTR:12–16points
Nat’lHeadStartMean=8.52
Figure12.GetReadytoReadSplitbyAgeandGroupScoresatorabove11arepredictiveoflaterreading
success
36|P a g e
Table7.Pre‐TestandPost‐TestMeansfortheCombinedEarlyLiteracyIndicatorbyAgeandGroup Treatment Maintenance Control
Pre Post Pre Post Pre Post
FOURS
GetReadytoRead 9.84 11.78 9.01 10.06 9.30 9.79
FIVES
GetReadytoRead 9.18 11.80 8.77 12.52 9.17 11.29
OVERALL
GetReadytoRead 9.51 11.79 8.89 11.29 9.24 10.54
37|P a g e
ResultsSummaryTableOnthenexttwopages,twotablescontaineffectsizeestimatesforeachoutcome.Aneffectsizeisanobjective,standardized,andmetric‐freeindexofthepracticalsignificanceofaresult.Itreflectsthemagnitude,orsize,ofgroupdifferences(Hedges,2008).Thistypeofinformationcanhelpresearchersandpolicymakersdeterminewhetheraparticulardifferencebetweentwogroupsisbigandmeaningfulorwhetherthedifferenceisactuallyanartifactofalargesamplesize.Standardizedeffectsizesreflectthenumberofstandarddeviationunitsthatseparatetwogroups.Astandarddeviationreflectsthedispersionofchildren’sscoresaroundagroupmeanusinganindexoftheexpectedvariationaroundthatmean.Asmallstandarddeviationindicatesthatchildren’sscoresarecloselyclusteredaroundthemeanvaluewhilealargestandarddeviationindicatesthatthespreadoftheirscoresisrelativelywide.About68%ofchildren’sscoreswillfallbetweenonestandarddeviationaboveandonestandarddeviationbelowthemeanwhile95%ofchildren’sscoreswillfallbetweentwostandarddeviationsaboveandtwostandarddeviationsbelowthemean.Inthisstudy,standardizedeffectsizesarereportedasawaytocontextualizethemagnitudeofdifferencesinanequivalentfashionacrossmeasuresorparticipants.Cohen’sd(Cohen,1988)wasselectedbecauseitisoneofthemostwidelyusedeffectsizeindicesintheliterature.Whenmakingcomparisonsinvolvingtwogroupsofchildrenwhoparticipatedintwodifferenteducationalinterventions(i.e.,InterventionAorInterventionB),obtaininganeffectsizeof1.0(withInterventionAchildrenoutperformingInterventionBchildren)indicatesthatInterventionAchildrenscored,onaverage,astandarddeviationhigherthanInterventionBchildren.Tables8and9belowcontaineffectsizesassociatedwithtwodifferentcomparisons:1)overallbygroupand2)withineachagelevelbygroup.Positivevaluesindicatethateitherthetreatmentgrouporthemaintenancegroupscoredhigherthanthecontrolgroup.Negativevaluesindicatethatthecontrolgroupscoredhigherthanthetreatmentgrouporthemaintenancegroup.Forexample:
1. TheeffectsizedescribingthedifferencebetweentreatmentandcontrolgroupmeansforUpperCaseTotalis0.47.Thiseffectsizeindicatesthatchildreninthetreatmentgroupscoredjustundera½standarddeviationhigherthantheircontrolgrouppeers.
2. TheeffectsizedescribingthedifferencebetweenmaintenanceandcontrolgroupmeansforUpperCaseTotalis‐0.17.Thiseffectsizeindicatesthatchildreninthecontrolgroupscored.17standarddeviationunitshigherthantheirmaintenancegrouppeers.
Cohen(1988)suggestedbenchmarksforinterpretingwhetheraneffectsizeismeaningfulortrivialand,ifmeaningful,howbiganeffectsizewas(i.e.,small,moderate,orlarge).Onthenextpage,thereisalegendwiththesebenchmarksaswellasacolorkeyindicating,at‐a‐glance,whicheffectsweresmall,moderate,orlarge.
38|P a g e
Table8.EffectSizeEstimatesforIndicatorsofLanguageDevelopmentandLetterKnowledge FOURS
(46–59months)FIVES
(60–74months)OVERALL
LanguageDevelopment
IGDIPictureNaming
Treatmentvs.Control d=0.10 d=‐0.44 d=‐0.34Maintenancevs.Control d=‐0.05 d=‐0.94 d=‐0.90
LetterKnowledge
UpperCaseTotal Treatmentvs.Control d=0.47 d=0.15 d=0.19
Maintenancevs.Control d=‐0.17 d=0.30 d=‐0.16
UpperCaseNamingSpeed
Treatmentvs.Control d=0.92 d=‐0.47 d=0.15Maintenancevs.Control d=‐.016 d=‐0.60 d=‐0.27
LowerCaseTotal Treatmentvs.Control d=0.35 d=‐0.07 d=0.17
Maintenancevs.Control d=‐0.17 d=0.19 d=‐0.11
LowerCaseNamingSpeed
Treatmentvs.Control d=0.33 d=‐0.39 d=‐0.02Maintenancevs.Control d=‐0.07 d=‐0.42 d=‐0.27
EffectSize Interpretation
0.10to0.30 Small
0.30to0.50 Moderate
0.50andAbove Large
< ‐0.10 Controlgroupscoredhigher
Between‐0.10and0.10 Trivial
39|P a g e
Table9.EffectSizeEstimatesforIndicatorsofPhonemicAwareness,PrintConventions,andGetReadytoRead FOURS
(46–59months)FIVES
(60–74months)OVERALL
PhonemicAwareness
IGDIInitialSoundsFluency Treatmentvs.Control d=‐0.23 d=0.20 d=0.00
Maintenancevs.Control d=0.02 d=0.62 d=0.28
LetterSoundsTotal
Treatmentvs.Control d=0.08 d=0.19 d=0.45Maintenancevs.Control d=‐0.08 d=0.45 d=0.34
LetterSoundsNamingSpeed Treatmentvs.Control d=0.04 d=0.08 d=0.08
Maintenancevs.Control d=‐0.10 d=0.30 d=0.02
PrintConventions
BookKnowledge Treatmentvs.Control d=0.56 d=0.34 d=0.46Maintenancevs.Control d=0.19 d=0.21 d=0.18
PrintKnowledge
Treatmentvs.Control d=0.41 d=0.01 d=0.14Maintenancevs.Control d=0.28 d=‐0.56 d=‐0.38
Comprehension
Treatmentvs.Control d=0.31 d=‐0.03 d=0.13Maintenancevs.Control d=0.00 d=‐0.31 d=‐0.36
CombinedEarlyLiteracy
GetReadytoRead
Treatmentvs.Control d=0.64 d=0.13 d=0.42Maintenancevs.Control d=0.00 d=0.20 d=0.23
EffectSize Interpretation
0.10to0.30 Small
0.30to0.50 Moderate
0.50andAbove Large
< ‐0.10 Controlgroupscoredhigher
Between‐0.10and0.10 Trivial
40|P a g e
BridgingtheGapContextualizingtherelationsbetweentheinterventionandkeyearlyliteracyoutcomesofinterestcanbeaccomplishedbycomparingchildperformancewithbenchmarksanddevelopmentalranges,whenavailable.Anotherwaytocontextualizethefindingsinvolvescomparingthemagnitude,orsize,ofgroupdifferences(Hedges,2008;Hill,Bloom,Black,&Lipsey,2008).Effectsizeshelpresearchersandpolicy‐makersdeterminewhetheraparticulargroupdifferenceispracticalormeaningfulorwhetherthedifferenceisactuallyanartifactofalargesamplesize.Policy‐RelevantPerformanceGapsEstablishingcriteriathatallowresearchersandpolicy‐makerstojudgetheeffectivenessofaneducationalprogramorinterventioniscritical.Typically,minoritychildrenandchildrenfromlowincomefamiliesevidencesubstantialdeficitsintheirearlylanguagedevelopmentandliteracyskillswhencomparedwiththeirmoreadvantagedpeers.Forinstance,HartandRisley(1995)foundthattheamountandqualityoftalkthatparentsengagedinwiththeirinfantsandtoddlersfrom6monthstothreeyearspredictedthesizeofthesechildren’svocabulariesatage3.Specifically,childrenofprofessionalparentshadvocabulariesthataveraged1,116words;childrenfromworking‐classfamilieshadvocabulariesthataveraged749words;andchildrenfromlow‐income(or“welfare”)familieshadvocabularieshalfthesizeofchildrenfromprofessionalfamilies(i.e.,525words).Thisdisadvantagecontinuedtofollowthesechildrenthroughtheirtransitiontoformalschooling(Walker,Greenwood,Hart,&Carta,1994);itgrewlargerovertheyears(i.e.,Mattheweffects;Stanovich,1986);andwasnearlyimpossibletoovercomeevenwithsubstantialinterventionefforts.Nationalestimatesofpreschoolers’earlyliteracyabilitiesweretakenfrompublishedreportsusingdatafromtheEarlyChildhoodLongitudinalSurvey–BirthCohort(i.e.,ECLS‐B;NCES,2006).Theunderlyingearlyliteracyskillsthatweremeasuredinboththatstudyandthepresentstudyincludedindicatorsofletterknowledge,printconventions,andacombinedearlyliteracyskillstask.Becausetheactualmeasuresusedineachstudydiffered,itwasnecessarytostandardizedalloutcomedata.Oncestandardized,comparisonsbetweenbothstudiesusingacommonframeworkbecamepossible.Recallthatthestandardizedeffectsizerepresentsthenumberofstandarddeviationunitsthatseparatethemeansoftwogroups.
1. LetterKnowledge:IntheECLS‐B,letterknowledgewasexaminedbyhavingchildrenidentifybothletternamesandlettersounds.Inthepresentstudy,threesubtestscoresassociatedwiththePALSpreKAlphabetKnowledgetaskwerestandardizedandaggregatedtoformoneindicatorofLetterKnowledge(i.e.,UpperCase,LowerCase,andLetterSoundNaming).
2. PrintConventions:IntheECLS‐B,printconventionsweremeasuredthroughaseriesofquestionstargetingyoungchildren’sunderstandingofwhatprintrepresentsandhowitworks(e.g.,howtoorientthebook,discriminatingprintfrompictures,readinglefttoright).Inthepresentstudy,twosubtestsderivedfromthePrintandStoryConceptstaskswerestandardizedandaggregatedtoformoneindicatorofPrintConventions(i.e.,bookknowledgeandprintknowledge).
3. GeneralEarlyLiteracyAbility:IntheECLS‐B,earlyliteracyskillswereevaluatedusingacombinationofletterrecognition,receptiveandexpressivelanguage,lettersounds,andearlyreading.Inthepresentstudy,theGetReadytoReadscoreswerestandardizedandcombined
41|P a g e
withtheindicatorsofLetterKnowledgeandPrintConventionstomatchtheprocedureusedbyintheECLS‐Bmethods.
Table10.EffectSizeEstimatesContrastingNationalEstimateswithStudyEstimates IdentifiedGapsBetweenDifferentGroupsofChildren SizeoftheGapa BTL:Tvs.C BTL:Mvs.C
LetterKnowledge
EAvs.AAachievementgap 0.25
EAvsnon‐EAachievementgap 0.39
LowSESvs.MiddleSESachievementgap 0.39
LowSESvs.HighSESachievementgap 0.98
0.62 0.18
PrintConventions
EAvs.AAachievementgap 0.21
EAvsnon‐EAachievementgap 0.35
LowSESvs.MiddleSESachievementgap 0.37
LowSESvs.HighSESachievementgap 0.96
0.69 0.24
CombinedEarlyLiteracyTask
EAvs.AAachievementgap 0.25
EAvsnon‐EAachievementgap 0.39
LowSESvs.MiddleSESachievementgap 0.40
LowSESvs.HighSESachievementgap 1.00
1.51 0.66
Note.EA=EuropeanAmerican;AA=AfricanAmerican;non‐EA=AfricanAmerican,Hispanic,AmericanIndianandAlaskaNative;SES=socioeconomicstatus;T=Treatment,M=Maintenance,C=ControlaAllmetricsinthistablerepresentstandardizedeffectsizes(i.e.,Cohen’sd)
42|P a g e
EffectSizesandtheAchievementGapAsindicatedinthetablesabove,theBTLClassroomInterventionprovidedapowerful,effective,andengaginginterventionthathelpedeconomicallydisadvantagedpreschoolersbridgetheachievementgapbetweentheirperformanceoneachofthethreeindicatorsandthescoresoftheirmoreadvantagedpeers.TheBTLClassroomInterventionwasdesignedtohelpreducetheachievementgapstypicallyfoundbetweenminorityandmajoritysubgroupsofchildrenandbetweenchildrenfromlowerSESandmiddleSESfamilies.Recently,researchershaveproposedthattheeffectsizesassociatedwithaninterventionoreducationalreformshouldbecomparedwiththesizeofknownachievementgapsinordertojudgewhetheraninterventionismeaningfulorworthimplementing(Hilletal.,2008;Konstantopoulos&Hedges,2008).Acrossthethreeindicatorsofkeyearlyliteracyskills,theestimatedimpactoftheBTLClassroomInterventionwasremarkablyconsistentanduniversallylargerthanthedocumentedgapsbetweenEuropeanAmericanandAfricanAmericanpreschoolersorbetweenpreschoolersfromlowSESfamiliesandfrommiddleSESfamilies.PuttingtheLastPiecesofthePuzzleTogetherEstablishingthataneffectexistsandissimilarinmagnitudetothesizeofthedifferencebetweentwopopulationsofinterest(e.g.,lowSESvs.middleSESchildren)isimportant.Comparingthesetwoeffects(i.e.,comparingthemagnitudeofthedifferencesinfavoroftheinterventionwiththemagnitudeofthedifferencesbetweenlowSESandmiddleSES)helpstoestablishthattheinterventioncannormativelyresultinaneffectthatissimilartoorbiggerthantheobservedachievementgap.Thenextstepinthepuzzleisdeterminingwhethertheobservedgainsmeetthecriterionlevelassociatedwithaparticularskill.Didparticipatingintheinterventionhelppreschoolersobtainscoresthatwereeitheratorabovetheestablishedbenchmarksorinsidetheexpecteddevelopmentalrangesassociatedwithaparticularskill.Forexample,GetReadytoReadscoresinpreschoolthatareatorabove11pointshavebeenfoundtobepredictiveoflaterconventionalreadingsuccess.Atthestartoftheintervention,allchildrenonaveragescored9.21.Atthepost‐test,thosewhoparticipatedintheinterventionscoredabovethe11‐pointbenchmark(i.e.,treatment=11.79;maintenance=11.29).Across20comparisons,preschoolersinthetreatmentgroupreachedestablishedbenchmarks83.3%ofthetime;preschoolersinthemaintenancegroupreachedtheestablishedbenchmarks77.8%ofthetime;andchildreninthecontrolgroup
43|P a g e
reachedtheestablishedbenchmarks66.7%ofthetime.SeeFigure19.
Figure13.BenchmarkComparisonsAcrossGroups
44|P a g e
DiscussionDidtheBTLClassroomInterventionImpactTeachers’Literacy‐RelatedBehaviorsandtheOrganizationofTheirClassrooms?
TeachersusingtheBTLClassroomMaterialsandreceivingthementoringsupportevidencedsignificantchangesacrossallfoursubscalesoftheELLCO.TheLiteracyEnvironmentsubscaleinvolvedanassessmentoftheclassroom’slayoutandcontentsincludingavailability,content,anddiversityofreading,writing,andlisteningmaterials.TheGeneralClassroomEnvironmentsubscalemeasuredtheorganizationoftheclassroom,achild’sopportunitiesforchoiceandself‐initiative,appropriateclassroommanagementstrategies,andanoverallpositiveclimate.TheLiteracyActivitiessubscalemeasuredthenumberandlengthoffull‐groupandone‐to‐onebookreadingsessionsaswellaswhetherchildrenwerewritingontheirownorwithassistanceaswellaswhetherteachersweremodelingpositivewritingbehaviors.TheLanguage,Literacy,andCurriculumsubscalemeasuredorallanguagefacilitation,presenceofandapproachestoreading,writing,andcurriculumintegration,andrecognizingdiversityandbridgingthehome‐schoolenvironments.Maintenanceclassroomswereabletosustainhigh‐qualityliteracyenvironmentsfrompreviousyearstothisprojectyearandtreatmentclassrooms,withextendedmentoringandsupport,wereabletoachievehigh‐qualityliteracyenvironments.Changeswerefoundfromthemostsuperficialstructuralcomponentstoincreasedquantityandhigherqualityenrichinginteractionsthathavebeenpreviouslylinkedtochildren’soptimalgrowthinliteracyandothercognitiveandsocialdomains.Todeterminewhetherthesepositivechangestranslatedintogainsonkeyearlyliteracyskills,children’sabilitiesacrossawidearrayofearlyliteracymeasuresweremeasured.
DidtheBTLClassroomInterventionImpactChildren’sEarlyLiteracySkills?
Childrenbetween46monthsand59monthswhoseteacherswerenewtotheBTLClassroomInterventionobtainedhigherscoreswhencomparedwiththeirmaintenanceandcontrolgrouppeersonnearlyalloutcomesacrossIndicatorsofLanguageDevelopment,LetterKnowledge,PhonemicAwareness,PrintConventions,andaCombinedEarlyLiteracySkillsScreener.Specifically,4yearoldchildrenidentifiedmorePictureNames,UpperandLowerCaseletters,andLetterSounds.TheirabilitytonameUpperandLowerCaseLettersandLetterSoundswasfasterthantheirpeersinthemaintenanceandcontrolgroups.Theywerebetterabletodemonstratethemechanicsofreading(i.e.,PrintKnowledge)andtoidentifykeyactualandinferentialstorycontent.Finally,theyscoredhigherthantheirpeersonacombinedearlyliteracyskillsscreenercalledtheGetReadytoReadachievingscoresthatarepredictiveoffuturereadingsuccess.Developmentaldifferencesare15%OlderchildrenwhoseteachershadpreviouslyparticipatedintheBTLClassroomIntervention(i.e.,MaintenanceGroup)obtainedhigherscoreswhencomparedwiththeirtreatmentandcontrolgrouppeersonIndicatorsofLetterKnowledgeandaCombinedEarlyLiteracySkillsScreener.Specifically,childrenwhowereolderthan5¼yearswereabletoidentifymoreUpperandLowerCaselettersandLetterSounds,tonameUpperCaselettersandLetterSoundsmorequickly,andtoaccuratelyansweritemsontheGetReadytoReadscreener.Whiletherewereseveralinstanceswhenolderchildreninthe
45|P a g e
controlgroupoutperformedtheirolderpeersinthetreatmentgroup,themaintenancegroup,orboth,therewerenoconsistentpatterns,suggestingthatthesedifferencesweremoreidiosyncraticthansystematic.Changingteacherbehaviorisachallengingundertaking.Inthisintervention,teacherswereprovidedwithextensivementoringandsupport,highqualitybooksandmaterials,andaneasy‐to‐usesupplementalliteracycurriculum.Asnotedabove,childrenwhowereineitherBTLClassroominterventiongroupevidencedhigherscoresonavarietyoftasks.Understandingthepatternsoffindingswasenhancedwhenachild’sagewasconsidered.Youngerchildrenobtainedhigherscoreswhentheywereinclassroomswhereteacherswereusingtheinterventionforthefirsttime.Itispossiblethatthesechildrenareattendingahigh‐qualitychildcareenvironmentforthefirsttime.Assuch,becausethesechildrenstartedtheprojectwithscoresindicatingaseriousriskoflaterreadingfailure,itislikelythatanyintervention(particularlyonethatisknowntobebotheffectiveandengaging)isgoingtohavepowerfuleffectsontheseyoungchildren’sskills.Thesechildrenmaybeexperiencingaliteracy‐andlanguage‐enrichingenvironmentonaregularbasisforthefirsttime.Infact,effectsizes(i.e.,anindexofthemagnitudeorsizeofaneffect)fortheyoungerchildrenaveragedbetween0.34and0.44,indicatingthatparticipationintheinterventionproducedchangesthatwereroughly1/3rdto2/5thofastandarddeviationhigherthantheircontrolgrouppeers.Meaningfulchangecanalsobethoughtofastheamountofvarianceinchildoutcomesthatwasattributabletotheintervention.Thisinterventionaccountedforbetween10.9%and19.4%ofthevarianceinchildoutcomes.Unlikeyoungerchildren,olderchildrendidbetterwhentheywereinclassroomswithteacherswhohadpreviouslyparticipatedintheproject.Intheseclassrooms,teacherswereabletomaintainahighdegreeofqualityfromoneyeartothenext.Thisqualityandstabilitymayhaveprovidedanenvironmentthatsupportedthesechildren’sskillsparticularlybecausetheyhavealreadyspentatleastayearattheircurrentchildcarecenterandhaveexperiencedayearwithoutanyextralanguage‐orliteracy‐enrichingexperiences.Themagnitudeorsizeoftheeffectforolderchildrenaveraged0.54,arelativelysizableeffectaccountingfor29.4%ofthevarianceinchildoutcomes.Itislikelythattheinterventioneffectsareespeciallypronouncedbecauseteacherswhohadpriorexperiencewiththeinterventionhaddevelopedacertainfacilitywiththematerialsandwerebetterabletohelpdeveloptheirolderlearners’skills.Further,olderlearnerswhohavenothadthebenefitofanyotherliteracyinstructionare,comparedwiththeiryoungerpeers,atamarkeddisadvantage.ItmaybethatteachersnewtotheBTLinterventionwerestrugglingnotonlywithchangingthegeneralliteracyenvironmentintheclassroombutalsotryingtoincorporatethespecialneedsofolderpreschoolerswhohaveanextrayearofdisadvantagebehindthem.Theaveragemagnitudeoftheeffectfoundforolderchildrenisparticularlyexcitingasitrangesbetween10%and20%higherthantheeffectsfoundforyoungerchildren.Thismeansthatexperiencedteachersweresuccessfulathelpingtheseolderchildrencatchuptotheiryoungerpeersaswellastostandardsorbenchmarksthatchildrenthisageneedtoobtaintoensureconventionalreadingsuccess.ThisstudycontributesadditionalevidenceregardingtheeffectivenessofusingBTLalone(i.e.,puretelevisionexposure;Linebargeretal.,2004;Uchikoshi,2006)andincombination(i.e.,exposureandsupplementalclassroommaterials;Linebarger,2006;Princeetal.,2001)tosupportyoungchildren’sburgeoningearlyliteracyskills.Moreimportantly,BTLhasconsistentlyhelpedyoungchildrenwhoareatsubstantialriskforlaterreadingfailureacquirethekeyearlyliteracyskillsneededforschoolandlaterlifesuccess.Thepowerofthisinterventionisafunctionofitsengagingcharacters,stories,andsketchesincombinationwithearlyliteracycontentthatiscarefullyinterwoventhroughoutthetelevisedcontentaswellasinfusedintoteachermaterialsandclassroommanipulatives.Childrenfromlow‐incomeandminoritybackgroundsspendmoretimewatchingtelevisionandreportthattheexperienceismore
46|P a g e
relaxingandofmorevalueincomparisontochildrenfrommiddleincomeandmajoritybackgrounds.Assuch,BTLrepresentsapowerfultoolforthesechildrenwhentheyareacquiringthekeyearlyliteracyskillsnecessaryforlearningtoread.
OtherThoughtsRegardingtheBTLClassroomIntervention
Inyearspast,effectsassociatedwithviewingBTLorparticipatinginaclassroominterventionweremoderatedbyachild’sinitialriskstatus(e.g.,Linebargeretal.,2004;Princeetal.,2001).Toexplainthemoderatedeffects,analysesassociatedwiththechild’sriskforlaterreadingfailurewerecomputed.Inthecurrentsample,32.6%ofpreschoolerswereconsideredatsignificantriskforlaterreadingfailure;30.7%ofpreschoolerswereconsideredatmarginalrisk;and36.7%ofpreschoolerswereconsiderednottobeatrisk.ItisnotclearwhetherMississippipreschoolers’literacyskillsaregenerallyimprovingorwhetherthesamplerecruitedforthe2007‐2008projectwaslessatriskthanaverage.Futureevaluationsshouldcontinuetomonitorthisemergingtrend.Regardlessofthereason,itisencouragingthatthepercentageofchildrenatriskforlaterreadingfailurehasdroppedfromnearlyallbutthreechildrenat‐risk(Princeetal.,2001)tojustaboutone‐thirdofchildrenat‐risk.RecommendationsforFutureResearchTherearesomeresearchissuesthatshouldbeaddressedwhenconductingadditionalinterventionstudies.Selectionandsamplingstrategiesareimportantissuestoanyresearchdesign,particularlyonethatusesaquasi‐experimentalframework.Becauseitishighlyunlikelythatchildrencanberandomlyassignedtogroups(i.e.,childrenareinclassroomsandclassroomsareassignedtocondition),itisimperativethatclassroomsinvolvedinastudyarerandomlyassignedtoatreatmentorcontrolcondition.Whilepreparingtheavailabledataforanalyses,itwasdeterminedthat,despitebesteffortstomatchcontrolandtreatmentclassroomsonavarietyofdemographicvariables(i.e.,povertyrates,location,targetage,teachereducation),therewerestillsomevariablesthatdifferedsignificantlyacrossthegroups(e.g.,child’sage,pre‐testELLCOscores).Althoughthesevariableswerestatisticallycontrolled,itisimportanttointerprettheresultspresentedherewithcaution.Itispossiblethatdifferencesareduetoothervariablesthatweremeasuredhere(e.g.,olderchildreninthetreatmentgroupmaynaturallyhavestrongerliteracyskills;classroomswithabetterliteracyenvironmentmayprovidechildreninthoseclassroomswithanadvantage)orothervariablesthatwerenot(e.g.,parentaleducation).Tostrengthentheresearchdesignandvalidatethefindingsofpreviousstudies,thisstudy,andfuturestudies,itisimportanttomakeaconcertedefforttokeeptheprocessasrigorousaspossible.Inadditiontochangesinthewaytheresearchisconducted,additionalchangesmaybenecessaryinthewaysinwhichtheinterventionisdelivered,particularlyforteacherswhohavepreviouslyparticipatedintheintervention.Itmaybethatfirstyearteachersreceivedextensivementoringandsupportwhilesecondyearteachersreceivelesssupport.Further,thelargestchangesonELLCOscoresareassociatedwiththegeneralliteracyenvironment(e.g.,displays,books).Theseenvironmentalvariableswerequicklyandsubstantivelychangedduringthefirstyearofintervention.Themorechallengingbehaviorstochangearethoserelatedtothelanguage,literacy,andcurricularenvironmentincludingthequantityandqualityoflanguage‐andliteracy‐promotingstrategies.Itispossiblethatencouragingthesetypesofinteractionswilltakemorementoringsupportthaniscurrentlypossibleoravailable.Onewaytoexaminethispossibilityittocodementorfieldnotesassociatedwithbothtreatmentandmaintenanceclassrooms.Currently,UPennstaffareworkingonthiscodingscheme.
47|P a g e
Finally,childrenwhoseteachersweresupplementingtheirregularinstructionwiththeBTLclassroominterventiondemonstratedgainsacrossmostearlyliteracymeasureswiththeexceptionofvocabularyknowledge.Initson‐airprogram,BTLfocusesheavilyoncode‐relatedskillacquisition.Charactersspendtimesoundingwordsoutandreadingaloud,smallsegmentsincludewordsthatmorphintootherwordsinthesamewordfamiliesorwiththesamevowelorconsonantsounds.Thegreateremphasisonphonologicalandphonemicawarenessskillsmakesthesewordpropertiesmoresalientandmayleavelittletimeleftoveror,alternatively,littlecognitivecapacityleftovertoencodenewwordsforboththeircode‐relatedpropertiesandtheirvocabularyororallanguageproperties.Itwouldbeexpectedthatovertime,aschildrenarerepetitivelyexposedtothecode‐relatedpropertiesofwords,theywouldbeabletodevotelesstimetothesecodepropertiesandmoretimetotheconceptualunderstandingofthewords.
ToSumItUpIthaslongbeenknownthathighqualityearlychildhoodeducationprogramshelpyoungchildrenexperiencingsignificantandchronicpovertyanddisadvantagebridgethegapbetweentheirreadinessforschoolandtheirmoreadvantagedpeers’readinessforschool(e.g.,PerryPreschoolProject,AbecedarianProject).Theseprogramsaretypicallycomposedofservicesdesignedtocomprehensivelysurroundchildrenandtheirfamilieswith,amongotherservices(e.g.,healthandsocialservices),cognitivelystimulatingtoysandmaterialsaswellaspositiveandsustainedlanguage‐,literacy‐,andprosocial‐promotingexperiencesandinteractions.ThespecificearlyliteracyachievementgapspresentedinTable10indicatethatAfricanAmericanpreschoolersareunderperformingtheirEuropeanAmericanpeersbyapproximately1/4thofastandarddeviationacrossthe3indicatorsofearlyliteracyachievementwhilechildrenfromlowSEShomesareunderperformingtheirpeerslivinginmiddleSESfamiliesby2/5thofastandarddeviation.Whilethesegapsmayseemrelativelysmallinpreschool,thereissubstantialevidencethattheyarepersistent,resistanttointervention,andwideningaschildrenprogressthroughschool.Walkeretal(1994)foundthatearlylanguagedeficitsidentifiedatage3andlinkedtofamilySESpredictedlanguagedevelopment,verbalability,andacademicachievementthroughouttheearlyelementaryschoolyears.Stanovich(1986)labeledthisphenomenonastheMattheweffect40,proposingthatchildrenwhohadmoreandpositiveearlyliteracyexperiencesaremorefrequentlyandintensivelyrewardedfortheseearlyaccomplishmentswhilechildrenwholackthesecumulativeexperiencesandsuccessesfindreadinglessenjoyable,struggletomakesenseofwhattheyarereading,andareoftenunabletobenefitfromandevenutilizeneweducationalexperienceseffectively.Essentially,childrenwhoseearlyliteracyachievementsareslowedordelayedprogressivelydecline,childrenwhoseearlyliteracyachievementscomequicklyandfrequentlyprogressivelyimprove,resultinginever‐wideningdifferencesbetweentheirreading,school,andlifetrajectories.TheBTLClassroomInterventiondescribedandevaluatedinthisstudyhasthepotentialtoprofoundlyalteryoungeconomicallydisadvantagedchildren’searlyliteracyachievements,bridgingthegapbetweentheseachievementsandtheirfasterpeerswhoareeitherEuropeanAmericanorlivinginmiddleandupperSEShomeswhile,atthesametime,placingthemonamorepositiveliteracytrajectory.Whiletherearenocurrentlong‐termstudiesofwhetherthesechildachievementsaremaintained,extrapolatingfromothersuccessfulearlyliteracyinterventionswithsimilarshort‐termeffectssuggeststhatthesegainswillbemaintainedandwillhelptheseat‐riskpreschoolerstoswitchfromthe“poorgetpoorer”trajectorytothe‘richgetricher’trajectory.Theevidenceisquiteclearthat
40TheMattheweffectreferstoapassageintheBibleattributedtoJesuswherehewaspresentingaparableassociatedwithtalents:“Foruntoeveryonethathathshallbegiven,andheshallhaveabundance:butfromhimthathathnotshallbetakenawayeventhatwhichhehath.”Matthew25:29
48|P a g e
thisintervention(i.e.,materialsaccompaniedbysustainedandintensivementoringsupport)substantiallyaltersteachers’behaviors.Thesebehaviors,inturn,createdailyenvironmentsforchildrenthatincludemoreandhigherqualitylanguage‐andliteracy‐promotinginteractions.Thesechangeshavebeenmaintaineduptothreeyearslater(e.g.,Figure6).Favorablechangesinclassroomenvironmentsandteacherbehaviorsarecloselylinkedinthisstudy,aswellasinpreviousstudies,topositivechangesandacceleratinggrowthofat‐riskpreschoolers’earlyliteracyskills(Linebarger,2006,2007;Linebargeretal.,2004;Princeetal.,2001).Basedonthesesubstantial,pervasive,andconsistentclassroomandteachereffects,itishighlylikelythatchildrenexposedtothisinterventionaspreschoolerswillcontinuetobenefitnotonlybyengaginginliteracyexperiencesandinteractionsprovidedbytheirteachersbutalsobyactivelychoosingto“select,shape,andevoketheirownenvironments”(p.381;Stanovich,1986).Specifically,exposuretoandactiveparticipationintheBTLClassroomInterventionhasthepotentialtoshifttheseat‐riskchildren’strajectoriestomirrormorecloselythetrajectoriesofchildrenwhoareacademicallysuccessfulbyprovidingdevelopmentallyappropriateandhighlyengagingcontentthat,throughaseriesofself‐reinforcingexperiencesandevents,supportstheirburgeoningearlyliteracyskillsand,perhapsevenmoreimportantly,increasestheirdesiresandmotivationstocontinuouslyandactivelysolicitnewliteracy‐specificaswellaseducationally‐generalexperiences(Stanovich,1986).
49|P a g e
ReferencesCohen,J.(1988).StatisticalPowerAnalysisfortheBehavioralSciences(2ndEd.).Mahwah,NJ:
Erlbaum.
CTB/McGraw‐Hill.(1990).Developingskillschecklist.Monterey,CA:CTB/McGraw‐Hill.
Dunn,L.M.,&Dunn,L.M.(2000).PeabodyPictureVocabularyTest—III.CirclePines,MN:AmericanGuidanceService.
Hart,B.,&Risley,R.T.(1995).MeaningfuldifferencesintheeverydayexperienceofyoungAmericanchildren.Baltimore:PaulH.Brookes.
Hedges,L.V.(2008).WhatAreEffectSizesandWhyDoWeNeedThem?ChildDevelopmentPerspectives,2(3),167‐171.
Hill,C.J.,Bloom,H.S.,Black,A.R.,&Lipsey,M.W.(2008).Empiricalbenchmarksforinterpretingeffectsizesinresearch.ChildDevelopmentPerspectives,2(3),172‐177.
Invernizzi,M.,&Sullivan,A.,&Meir,J..(2002).Phonologicalawarenessliteracyscreeningforpreschool(PALS‐PreK).Teachers'Manual.Charlottesville,VA:UniversityPrinting.
Konstantopoulos,S.,&Hedges,L.V.(2008).Howlargeaneffectcanweexpectfromschoolreforms?TeachersCollegeRecord,110,1613–1640
Linebarger,D.L.,(2006).TheBetweentheLionsAmericanIndianLiteracyInitiativeResearchComponent:AReportPreparedfortheUnitedStatesDepartmentofEducation.Philadelphia,PA:AnnenbergSchoolforCommunication,UniversityofPennsylvania.
Linebarger,D.L.,Kosanic,A.Z.,Greenwood,C.R.&Doku,N.S.(2004).EffectsofviewingthetelevisionprogramBetweentheLionsontheemergentliteracyskillsofyoungchildren.JournalofEducationalPsychology,96(2),297‐308.
Missall,K.N.,&McConnell,S.R.(2004).TechnicalReport:Psychometriccharacteristicsofindividualgrowthanddevelopmentindicators–PictureNaming,Rhyming&Alliteration.Minneapolis,MN:CenterforEarlyEducationandDevelopment.Retrievedfrom:http://ggg.umn.edu/techreports/dissemination.html#TechRep.
Prince,D.L.,Grace,C.,Linebarger,D.L.,Atkinson,R.,&Huffman,J.D.(2002).BetweentheLionsMississippiliteracyinitiative:AfinalreporttoMississippiEducationalTelevision.ReportpreparedforMississippiEducationalTelevisionandWGBHEducational
50|P a g e
Foundation.Starkville,MS:TheEarlyChildhoodInstitute,MississippiStateUniversity.
Stanovich,K.E.(1986).Mattheweffectsinreading:Someconsequencesofindividualdifferencesintheacquisitionofliteracy.ReadingResearchQuarterly,21,360‐406.
Uchikoshi,Y.(2006).Earlyreadinginbilingualkindergartners:Caneducationaltelevisionhelp?ScientificStudiesofReading,10,89‐120.
Walker,D.,Greenwood,C.R.,Hart,B.,&Carta,J.(1994).Predictionofschooloutcomesbasedonearlylanguageproductionandsocioeconomicfactors.ChildDevelopment,65,606‐621.
Whitehurst,G.(2001).TheNCLDGetReadytoRead!screeningtooltechnicalreport.AreportpreparedfortheNationalCenterforLearningDisabilities.NewYork,NY.
Zimmerman,I.L.,Steiner,V.G.,&Pond,R.V.(1992).PreschoolLanguageScale–3.SanAntonio,TX:ThePsychologicalCorporation.
51|P a g e
ThecontentsofthisdocumentweredevelopedunderacooperativeagreementbetweentheUSDepartmentof
Education,theCorporationforPublicBroadcasting,andthePublicBroadcastingSystemfortheReadytoLearn
Initiative,PR#U295A050003.However,thesecontentsdonotnecessarilyrepresentthepolicyofthe
DepartmentofEducationandyoushouldnotassumeendorsementbytheFederalGovernment.