Longitudinal Success: School Study A Affective Mastery and Motivation Social in Preschool as Predictors ofEarly Early Childhood Research Quarterly

(1)

Contents lists available atScienceDirect

Early Childhood Research Quarterly

Affective and Social Mastery Motivation in Preschool as Predictors of Early School Success: A Longitudinal Study

Krisztián Józsa

^a,∗

, Karen Caplovitz Barrett

^b

aInstituteofEducation,UniversityofSzeged,30-34,Pet ˝oﬁS.Road,Szeged,6722,Hungary

bDepartmentofHumanDevelopmentandFamilyStudies,ColoradoStateUniversity,1570CampusDelivery,FortCollins,CO,80523-1570,USA

a r t i c l e i n f o

Articlehistory:

Received14April2017

Receivedinrevisedform28April2018 Accepted27May2018

Keywords:

motivation masterypleasure avoidance socialskills schoolreadiness schoolsuccess

a b s t r a c t

Recentresearchhasdocumentedtheimportanceofschoolreadinessinyoungchildren.Childrenwho startschoolwithoutbasicskillsoftencontinuetoshowlowerachievementthroughoutschooling.Most currentassessmentsofschoolreadinessfocusonearlymeasuresofacademicskills,suchasliteracyand numeracy.Althoughtheseskillsareusefulinpredictingschoolsuccess,researchsuggeststhatsocioe- motionalandmotivationalfactorsmaybeevenmoreimportant.Moreover,althoughthereisstrong evidencesupportingtheimportanceofsocialandemotionalcompetencies,suchasemotionunderstand- ingandsocialskills,inschoolreadiness,thereisadearthofresearchontheroleofaffective/expressive andsocialaspectsofmastery/competencemotivationinearlyschoolreadinessandachievement.Inthe presentstudy,weusedStructuralEquationsModelingtoexaminetheroleofaffectiveaspectsofmas- terymotivation,socialmasterymotivation,Socio-EconomicStatus(SES),andIntellectualQuotient(IQ)in preschoolinlongitudinallypredictingmathachievement,readingachievement,andsocialskillsduring grades1and2in327Hungarianchildren.Resultsindicatedthatchildren’snegativereactionstofail- ure/challengepredictedallofthesemeasuresofschoolperformance,overandabovetheroleofchildIQ andSES;inaddition,masterypleasurepredictedreading,andpersistenceinpeerinteractionpredicted socialskillsintheearlygrades.Resultscontributetothegrowingliteraturesupportingtheimportance ofmotivationandofachievement-relatedemotionsinschoolreadinessandschoolsuccess.

1. Introduction

Recent research has documented the enormous importance of school readiness for young children (Guernsey, Bornfreund, McCann,&Williams,2014).Childrenwhostartschoolwithoutbasic skills suchas numeracy, emerging literacy,and socioemotional competencehavegreatdifﬁcultycatchingupwithpeerswhohave higherlevelsofschoolreadiness(e.g.,Burchinal,Magnuson,Powell,

&Hong,2015;Snow,2006;Tymms,Jones,Albone,&Henderson, 2009).Althoughinthepast,schoolreadinesshasbeendeﬁnedpri- marilyintermsofpre-academicskillssuchasemergingliteracy andnumeracy,therehasbeenincreasedawarenessoftheimpor- tance of non-academic, behavioral aspects of school readiness, includingsocioemotionalcompetence(e.g.,Denhametal.,2012;

Izardet al.,2001; Rhoades, Warren,Domitrovich,&Greenberg, 2011) and Approaches to Learning (e.g., Meng, 2015). A body of researchdocuments thatsocioemotional competence,includ-

∗Correspondingauthor.

E-mailaddresses:krisztian.jozsa@colostate.edu,jozsa@sol.cc.u-szeged.hu (K.Józsa),Karen.Barrett@ColoState.EDU(K.C.Barrett).

ingemotionknowledge,socialskills,andabilitytoappropriately regulateone’semotions,isassociatednotonlywithfuturesocial andemotionalcompetence,butalsowithacademiccompetence (e.g.,Denhametal.,2012;Izardetal.,2001;Rhoadesetal.,2011).

However,italsohasbecomeclearthat,beyondskills,anotherfac- torthataffectsschoolsuccessischildren’spositiveApproachesto Learning.“ApproachestoLearning”referstoasetofinterrelated behavioralpropensitiesthatfacilitatechildren’slearning,including enthusiasmforlearning;focused,goal-orientedpersistence;and mastery/competencemotivation(e.g.,Hyson,2008; McDermott, Rikoon,Waterman,&Fantuzzo,2012).

Positive Approaches to Learning have been found to help compensateforless-than-optimallearningenvironments(Meng, 2015).Althoughtheseattributessometimesareclassiﬁedaspart of social emotional school readiness, they encompass attitudes andmotivationtowardlearning,ratherthaninterpersonaloremo- tion regulation competencies or skills (e.g., Fantuzzo, Perry, &

McDermott,2004).Theyarethereforequitedifferentfromtheusual characteristicsthatarecharacterizedasSocialEmotionalLearning orSocioemotionalCompetence,suchasemotionunderstanding, socialskills,andabilitytoeffectivelyregulateemotions.

(2)

Disparitiesinallschoolreadinessdimensions,basedonearly environmental differences such as those associated with low Socio-Economic Status (SES), can potentially lead to continued stratiﬁcationofeducationalopportunityandattainment(Duncan, Magnuson,&Votruba-Drzal,2015;Józsa,2016).Asaresult,both theUnitedStatesgovernmentandmanyU.S.statesareworkingto requireschoolstoassessandintervenetoenhanceyoungchildren’s schoolreadiness,andoneofthedomainsofschoolreadinessthat hasbeentargetedisApproachestoLearning.However,although thereisevidencethatcognitiveandinstrumentalApproachesto Learningareimportantinschool-relatedcompetenciesandschool success(e.g.,Fantuzzoetal.,2004;Meng,2015),thereisadearthof researchontheroleofsocialandemotionalaspectsofApproaches toLearning,and, inparticular, theroleof social and emotional aspects of mastery motivation, in school readiness and school success.Beforediscussingthepresentstudyinmoredetail, itis importanttobetterdeﬁnethetopiconwhichitfocuses,mastery motivation.

Whatismasterymotivation,andwhatistheroleofemotion init?Masterymotivationisamultifacetedpsychologicalforcethat pushesindividualstotrytomastertasksorskillsthatareatleast somewhatchallenging(Barrett&Morgan,1995;Morgan,Józsa,&

Liao,2017).Theterm,“multifaceted,”highlightsthemanydifferent domainsofdevelopmentandcontextsinwhichmasterymotivation occurs,aswellasthefactthatmasterymotivationmaydifferacross thesecontextsanddomains(Barrett&Morgan,1995;Hwangetal., 2017;Józsa,2014;Józsa,Kis,&Huang,2017;Wang&Barrett,2013).

Morgan,MacTurk,andHrncir(1995)identiﬁedthreemaindomains formasterymotivation:(1)cognitive,achild’smotivationtopersist atandmastercognitiveandschool-relatedtasks;(2)grossmotor, themotivationtomasterathleticskills;and(3)social,themotiva- tiontomasterinterpersonalrelationswithadultsandwithpeers(a focusofthepresentstudy).Morerecently,WangandBarrett(2013) addedself-masteryasafurtherimportantdomain.

Masterymotivationis not onlymultifacetedin terms ofthe domainsandcontextsinwhichitisobservedandfostersdevel- opment,butalsothefactthatitincludesbothexpressive/affective aspectsandinstrumentalaspects.Affectiveaspectsincludemildto moderatepositiveandnegativeemotionsthatpromoteapproach andcontinuedmasteryattempts,includingpleasure,interest,and enthusiasmintryingtomasterand/ormasteringchallengingtasks, aswellasfrustration/angeratperceivedobstaclestothatmastery.

Inaddition,affectivemasterymotivationincludesemotionsthat promotewithdrawaland givingup,suchassadnessandshame atlesssuccessfulmastery(and/oranticipationoffailure).Instru- mentalaspectsincludegoal-directedpersistenceandinclinationto controland/orhaveimpactontheenvironment(Barrett&Morgan, 1995;Wang&Barrett,2013).

Keyfeaturesdistinguishingthemasterymotivationapproach fromotherlearningmotivationapproachesareitsfocusonchal- lengingtasksanditsinclusionofmultiplemasterydomains.Thus, althoughsimilar,masterymotivationisdifferentfromtheconcept ofmasterygoalsintheliteratureonachievementgoalorientations (e.g.,Elliot,2005).Inthislatterapproach,peoplehaveadominant approachtolearning–eitheramasterygoaloraperformancegoal orientation.Thosewithmasterygoalorientationsfocusontheir ownstandardsforachievement;incontrast,thosewithaperfor- mancegoalapproachseektoperformbetterthanothers.Although in both mastery motivation and achievement goal approaches, pursuitofmasteryisassociatedwithincreasedstrivingandper- sistence;intheachievementgoalapproach,onepersistsdespite failure;inthemasterymotivationapproach,onepersistsdespite challenge/difﬁculty.

AnotherrelatedapproachtomotivationisSelf-Determination Theory. According to Self-Determination Theory, people have threeimportantneeds− autonomy(volitional/agenticfunction-

ing),relatedness(feelingloved/valued),andcompetence(feeling effective).Thesefundamentalneedsformthebasisformotivated behavior(Deci&Ryan,2000).Accordingtothisapproach,although autonomousstrivingfeelsintrinsic/self-motivated,itactuallycan originate in extrinsic processes such as parenting or teaching approaches(e.g.,Ratelle,Guay,Larose,&Senécal,2004).Thissug- geststhatwhetherthemotivationtrulyisintrinsicorextrinsicin originmaybelessimportantthanwhetheritfeelsvolitionalversus controlled/anxiety-based.Thisisconsistentwiththemasterymoti- vationapproach, whichfocuses onpersistence andemotionsin differentdomainsbutdoesnottrytodeterminetheintrinsicor extrinsicoriginsofthemotivation.

One of the needs speciﬁed by Self-Determination Theory, competence, has been the focus of most research on mastery motivation.Despitethetruebreadthof themasterymotivation conceptualization,mostextantmasterymotivationresearchhas only studied instrumental mastery motivation in the cognitive domain,namelycognitivepersistence.Thisresearchdemonstrated that children’s persistence at challenging cognitive tasks is an importantpredictorofschoolsuccess(e.g.,Gilmore,Cuskelly, &

Purdie, 2003; Mercader, Presentación, Siegenthaler, Moliner, &

Miranda,2017;Mokrova,O’Brien,Calkins,Leerkes,&Marcovitch, 2013).Somestudiesindicatedthatcognitivepersistencewasabet- terpredictorofcognitivedevelopmentthanintelligence(Józsa&

Molnár,2013;Yarrow,Klein,Lomonaco,&Morgan,1975).Achild’s tendencytopersistoncognitivetaskseven when theybecome challengingwouldseemcrucialforsuccessinschoolandbeyond.

However,masterymotivationinotherdomains(suchasthesocial domain), and affective aspects of mastery motivation may be justasimportant.Positiveemotionsand/oranger/frustrationmay motivateautonomousandpersistentmasteryattempts;whereas, negative withdrawal emotions suchas shame and sadnessare expectedtomotivateavoidance/givingup.Moreover,givenevi- dence that socioemotionalcompetencies are crucial aspects of schoolsuccess(e.g.,Denham,Bassett,Zinsser,&Wyatt,2014),it isimportanttostudythedomainsofmotivationalschoolreadi- nessthataremostrelevanttosocioemotionalschoolsuccess.The presentstudythusassessestheroleoftheseless-studiedaspectsof masterymotivationinearlyschoolsuccessinthreedomains:reading,math,andsocialskills.Wearetheﬁrststudytoexaminethe roleofmasterymotivationinthepeercontext(socialpersistence) inlongitudinallypredictingsocialskillsperformanceintheearly yearsofschool,andtheﬁrsttolongitudinallystudytherelationof affectiveaspectsofmasterymotivationtoearlyschoolsuccessin allthreeofthesedomains.

Asalludedtoearlier,researchsupportstheimportanceofsocial competencein schoolreadiness andboth short-termand long- termsuccessinschoolandinlife(e.g.,Denhametal.,2014;Jones, Greenberg,&Crowley,2015;Schonfeldetal.,2015).However,the focusofresearchonsocialcompetenceandschoolreadinesshas beenontheimportanceofteachingsuchskillstochildren.Although effectiveinterventionstoteachsocial-emotionalskillstochildren havebeendeveloped,childrenaredifferentiallysusceptibletosuch environmentalinﬂuencesonsocialcompetence(e.g.,Lianos,2015;

Mortensen&Barnett,2015).Onepotentiallyimportantfactorin children’sresponsestosocioemotionalschoolreadinessinterven- tionscouldbetheirmotivationtointeracteffectivelywithothers.

Yet,theroleofaffectiveandsocialaspectsofmasterymotivation inbothacademicandsocioemotionalschoolreadinesshasbeen under-researched.

Mostresearchonsimilartypesofmotivationhasfocusedon theroleofbeingintrinsicallyratherthanextrinsicallymotivated in school performance, or on whether one has mastery goals versusperformanceachievementgoals.Morerecently,thislatter approachhasconsideredalsowhetheroneismotivatedtoapproach performanceormasterygoalsversuswantingtoavoidnegativeper-

(3)

formanceormasteryoutcomes.Althoughtheseapproacheshave yieldedimportantinsights,therehasbeenrelativelylittleatten- tionineitheroftheseapproachestoaffectiveaspectsofmotivation;

moreover,studyresultshavedifferedwithrespecttowhetherper- formancegoals,especiallyapproachperformancegoals,undermine orenhanceachievement(Vansteenkiste, Lens,Elliot,Soenens, &

Mouratidis,2014).It hasbecome evident,based onthelimited availableresearch,thataffective/expressiveaspectsofmotivation maybe importantpredictors of school performance(Berhenke, Miller,Brown,Seifer,&Dickstein,2011;Vansteenkisteetal.,2014), anditisreasonabletoexpectthatsocialmasterymotivationwould beassociatedwithimprovedsocialcompetence.

Thefocusofthepresentinvestigation,masterymotivationinthe faceofchallenge,isoptimizedbyrelativelyhighlevelsofpositive emotionand relativelylow levelsof negativewithdrawal emo- tionssuchassadnessorshame.Undersuchconditions,children areexpectedtobemotivatedtomaintainfocusonthegoalsthey areattemptingtomaster.Stayingfocusedonone’sgoalsisalso characteristicofanotherschoolreadinesscharacteristic,executive function,soitisimportanttodistinguishmasterymotivationfrom executivefunction.

Executivefunctioninvolves children’sself-control over their attention, thoughts, and behavior. It is usually defined as the abilities to: inhibit a well-learned but undesirable response (inhibitorycontrol),holdthoughtsinmindwhileproblem-solving (working memory), and modify strategies toadjust tochanging goals(cognitiveflexibility).Inearlychildhood,thesethreeaspects of executive functions are typically difficult to separate from one another (Griffin, McCardle, &Freund, 2016). Both mastery motivationandexecutivefunctionsarekey,malleableschoolreadi- ness characteristics manifested in goal-directed behavior; yet, they also are different in important ways. Mastery motivation involveschildren’sattitudes/approachtowardchallenge,learning, andperformance;andfeelingsofautonomy,desire/enthusiasmfor mastery,andlow negative,withdrawalemotionduringmastery attempts(e.g.,Barrett&Morgan,1995;Ratelleetal.,2004).Incon- trast,likesocioemotionalcompetence,executivefunctionsseemto beskills,whichdeveloprapidlyduringearlychildhood.

Itisreadilyapparentthatskillscanbetaught,butonemight imagine that motivation,as anattitude/approach rather than a skill,wouldbeintrinsictothepersonandlesssusceptibletoexter- nalinﬂuence.It is truethatmotivation is notimprovedbythe sameprocessesonewouldusetoteachandreinforceskills,such aspracticeandpositivefeedbackaboutcorrectnessofskilledper- formance.In fact,there is evidencethat,rather thanincreasing motivation,tangiblerewardscanundermineit(e.g.,Cameron&

Pierce,1994;Lepper,Greene,&Nisbett,1973).However,research hasindicatedthatinterventionsthattrainteacherstoshowchil- drenthevalueoftryingorthattrainthemtosupportstudents’effort (ratherthanperformance)canincreasemasterymotivation(e.g., Hashmi,Seok,&Halik,2017;Vansteenkiste,Simons,Lens,Sheldon,

&Deci,2004).Incontrast,interventionsinvolvingrepeatedprac- ticeofexecutivefunctionsincreaseexecutivefunctions(e.g.,Bryck

&Fisher,2012;Diamond&Lee,2011;Greenberg&Harris,2012).

Masterymotivationinterventions,thus,typicallyinvolveaffective- and autonomy-supportive teaching that focuses on the learn- ingprocessrather thanoutcomes(e.g.,Sakiz,2017;Schiefele &

Schaffner,2015).Incontrast,executivefunctioninterventionstyp- icallyinvolvepracticingtherelevantskills(e.g.,Bierman&Torres, 2016). Moreover, there is evidence that Approaches to Learn- ingsuchasmotivationandenthusiasmcanmediatetherole of preschoolexecutivefunctionsin laterreading skillsandsocioe- motionalcompetence(Sasser,Bierman,&Heinrichs,2015).It is thereforeimportanttodistinguishthesetwocharacteristicssothat onecanassesstherolesofbothmasterymotivationandexecutive functions.

AreportofStateEarlyLearningGuidelines(ELGs)inFebruary, 2016 indicated that some 48 of 50U.S. states’ ELGs included Approaches to Learning and all included social competence (National Center on Early Childhood Quality Assurance, 2016).

Empirical research highlights the crucial importance of such attributes(e.g.,Berhenkeetal.,2011;Denhametal.,2014;Jones etal.,2015;Mokrovaetal.,2013).Moreover,onerecentlongitudi- nalstudyusingteacherreportsofprimarilycognitive/instrumental motivationalaspectsofApproachestoLearning(attentionalcon- trol,persistenceduringlearningtasks)atkindergartenfoundthat itpredictedbothacademicandsocialcompetenceinprimarilylow income, minority Americanchildren at age9 (Razza,Martin, &

Brooks-Gunn,2015).

Arapidlygrowingbodyofresearchthusdocumentstheimpor- tanceofsocioemotionalcompetenceandofcognitive/instrumental ApproachestoLearning,aspredictorsofacademicandsocioemo- tionalschoolsuccess.However,thepresentstudyfocusesonaless studiedschoolreadinessdomainatthenexusoftheseothertwo, namelysocioemotionalaspectsofApproachestoLearning−affec- tiveaspectsofmasterymotivationandmotivationtomasterpeer relationships.WeexaminetheseimportantApproachestoLearn- inginpreschool-agedchildren,andtheirlongitudinalpredictionof elementaryschoolachievementandsocialskills.

For decades, we have known that emotions are elicited in mastery contexts (e.g., Barrett, 1998; Barrett & Morgan, 1995; Lewis, Alessandri, & Sullivan, 1992; Stipek, Recchia, &

McClintic,1992).Infact,“effectancepleasure”wasakeycompo- nent of White’s(1959)seminal paperoneffectance motivation that inspired most of the early work on mastery motivation.

There is evidence that satisfaction/mastery pleasure is associated with appetitive mastery motivation (mastery approach goals) and fear of failure is associated with performance- avoidance goals (Elliot & Church, 1997; Tanaka & Yamauchi, 2001).Morerecently,though,researchindicatedthatfearoffail- ure is sometimes associated with performance-approach goals as well (Elliot & Murayama, 2008; Zusho, Pintrich, & Cortina, 2005). These ﬁndings suggest that the predominant distinc- tions between performance-approach, performance-avoidance, mastery-approach,andmastery-avoidancegoalsmaynotbesuf- ﬁcienttounderstandtheroleofmasteryandperformance-related emotionsinschoolsuccess.

Moreresearchisneededthatdirectlyassessesaffectiveaspects of motivation, such as negative/withdrawal reactions to failure/challenge and pleasure/satisfaction, rather than considering emotionsassecondaryeffects oftype ofgoalorintrinsicversus extrinsicoriginofmotivation.Inparticular,theroleoftheseaffec- tiveaspects,asdistinguishedfromtheroleof thegeneralgoals thechildhasintheachievementcontext,needstobeexamined inconnectionwithschoolreadinessandschoolsuccess(e.g.,see Barrett&Morgan,1995;Berhenkeet al.,2011; Roskes,Elliot,&

DeDreu,2014;Vansteenkisteetal.,2014).Thepresentstudywill examinetheroleofnegative/withdrawalreactionstofailureand pleasure/satisfactioninmastery,withintheframeworkofmastery motivationtheory.

Mastery motivation and competence, school readiness.

Despiteitsobviousrelevancetoschoolsuccess,masterymotiva- tionhasrarelybeenstudiedasaschoolreadinesscharacteristicin earlychildhood.Someclassicstudiesdemonstratedthatmastery motivationstronglypredictedlatercompetence(e.g.,Messeretal., 1986).Morerecently,Berhenkeetal.(2011)studiedtheconcur- rentrelationbetweeninstrumentalandaffectiveaspectsofmastery motivationand schoolreadiness.Theyfoundthat pride,shame, and persistence positively predictedsocial competence;persis- tenceandshamepositivelypredictedmathskills,andpersistence andshamepositivelypredictedreadingskills.It wasinteresting thatbothprideandshame,observedduringmasterytasks,were

(4)

positivepredictorsofsocialcompetenceandshamewasapositive predictorofbothreadingandmathskills.Thissupportstheidea thataffectiveaspectsofmasterymotivationareimportant,butis contrarytoexpectationintermsofthespeciﬁcroleofshame.How- ever,alimitationofthisstudywasitsconcurrentassessmentof bothmotivationandcompetence.

JózsaandMolnár(2013),inacross-sectionalstudyofthirdand sixthgraders,alsofoundacontemporaneousassociationbetween instrumental mastery motivation and both GPA and achieve- mentinspeciﬁcsubjects.Similarly,WalkerandMacPhee(2011) foundthatpreschoolchildren’sinstrumentalmasterymotivation completelymediatedtheconcurrentpredictionoftheirdevelop- mentallevelfromtheirparents’coercivecontrol.Neitherofthese lattertwostudiesassessedaffectiveaspectsofmasterymotiva- tion,however,andallthree studiesofmasterymotivationonly assessedconcurrent relations betweenmastery motivation and schoolreadiness/success.Additional researchusinglongitudinal assessmentofaffectiveandsocialaspectsofmasterymotivation isneeded,toassessfurthertheabilityofthesemeasurestopredict laterschoolreadiness.

Gilmoreetal.(2003)studydidassesslongitudinallytheability ofinstrumentalaspectsofmasterymotivationtopredictschool- relatedskills. Theydemonstrated that,for girlsonly, parentally reportedmasterymotivationpredictedIQandspellingandreading achievementsixyearslater.Mokrovaetal.(2013),likeGilmoreat al.,studiedthelongitudinalpredictionofkindergartenacademic skills(language and math). Theydidnot ﬁnda role of gender, butlikeGilmoreandcolleagues,foundthatinstrumentalaspects ofmasterymotivation(persistence)longitudinallypredictedboth languageandmathskills.Alloftheseﬁndingsaresuggestivethat masterymotivationmaybeanimportantpredictorofschoolreadi- nessandsuccess.However,theydonotlongitudinallyaddressthe roleofaffectiveorsocialaspectsofmasterymotivationinpredict- inglaterschoolreadinessorsocial/emotionalschoolsuccess,even thoughextantcross-sectionalresearchsuggeststhatthesemaybe important.

1.1. Thepresentstudy

Thepresentstudy,thus,addressesimportantgapsinthelit- erature:First,wehadpreschoolteachersrateaffectiveaspectsof children’smasterymotivation,aswellasinstrumentalaspectsof masteryofpeerrelationships(socialpersistencewithpeers).Thus, weaddressedthemultifacetednatureofmasterymotivationby examiningaffectiveaswellasinstrumentalaspectsandbyexam- iningadomainofinstrumentalmasterymotivation,peermastery, whichhad notyetbeenstudiedinrelationtoschoolreadiness.

Second,wefollowedchildrenlongitudinallytograde2andthen useddirectchildassessmentsmadeatchildren’sschoolsbytrained examiners(whowerenaivetothepreschoolratings)tomeasure notonlymathandreadingskills,butalsosocialskillsatgrades1 and2.Thus,ourstudygoesbeyondtheexistingliteraturebyobtain- ingaschool-relevantmeasureofaffectiveandsocial-instrumental aspectsofmasterymotivationonarelativelylargeandrepresenta- tivesampleofHungarianchildren,followingthemlongitudinally, andstudyingalargersetofschoolachievementandschool-related skills,asassessedbyindependentand well-trainedmembersof theresearchteam.Wehypothesizedthatsocialmasterymotiva- tionwouldpredictsocialcompetence,butthatbothnegativeand positiveaffectivemasterymotivationwouldpredictallthreeschool readinessoutcomes.

1.2. TheHungarianschoolsystem

Thepresent studywas conductedin Hungary;thus, tofully understandthecontextfortheﬁndings,itisimportanttoknow

Table1

ReliabilitiesandDescriptiveStatistics.

Measure NofMV Cronbach’s␣ Min. Max. M SD

SES 4 .86 1.00 4.75 2.58 .76

Preschool

NRF 5 .80 1.00 5.00 2.89 .86

MP 5 .89 1.00 5.00 3.67 .80

SPC 4 .89 1.00 5.00 3.60 .81

IQ 3 .77 64.97 132.95 100.00 15.00

Grade1

Reading 3 .71 32.18 94.09 75.31 11.11

Math 4 .81 12.71 95.30 58.12 17.96

Socialskills 4 .89 1.86 5.00 4.01 .54

Grade2

Reading 3 .79 36.27 91.53 77.27 8.51

Math 4 .86 17.33 97.83 62.94 18.19

Socialskills 4 .89 1.86 5.00 4.22 .55

Note.MV=ManifestVariables;NRF=Negativereactiontofailure;MP=Masteryplea- sure;SPC=Socialpersistencewithchildren;SES=Socioeconomicstatus.

somebasics abouttheHungarianschoolsystem.Hungarianlaw guarantees free preschool and kindergarten (all called kindergarten) for all children. Thereare three years of kindergarten, beginningatage3yearsandcontinuinguntilage6,withsome ageﬂexibility(Józsa,Török,&Stevenson,2018).Thesegradesare labeledKindergarten1(K1),K2,andK3.Itismandatoryforchildren tospendatleast4hoursperdayinkindergartenbeginningatage3 (Hungarian Government,2011);moreover,mostofthemattend allday.In 2014,97%offour-year-oldchildren attendedkinder- garten(OECD,2016).TheNationalCoreProgramofKindergarten Education(Hungarian Government,2012)deﬁnestheprograms forallHungariankindergartens.Thisnationalcurriculumcanbe supplementedwithlocalcurriculaandprograms.

ThepresentstudybeganinK2,whichchildrenenteredatage 4-5yearsofage,althoughbythetimeofthestudy,mostwere5 yearsofage.Thus,althoughchildrenwere“inkindergarten,”the contextandageofthesechildrenwascomparabletochildrenin theﬁnalyearofpreschoolintheUnitedStates.

2. Method

2.1. Participants

Atotalof327Hungarianchildrenbeganthelongitudinalstudy atapproximatelyageﬁve (M=62.05,SD=4.43months).Approx- imately 49%of thesample wasfemale.Theaveragenumber of yearsofparentaleducationwasapproximately10.7yearsforboth fathersandmothers(SD=1.79forFathers’,SD=2.03forMothers’).

Childrenwereenrolledinpubliclyfundedkindergarten(preschool) programsayearbeforethefirstdatacollection.Wecollecteddata insixsmalltomiddlesizecitiesinthesoutheastpartofHungary.A totalof12children(0.4%ofthesample)waslosttofollow-upduring thelongitudinalstudy.Thesechildrenwerenotsignificantlydiffer- entfromthoseinthefinalsampleintermsofanyofthepredictive variables(Alltvalues(325)<.52,p>.61).

2.2. Measures

2.2.1. DimensionsofMasteryQuestionnaire(DMQ)

Table 1 presents themeasures assessed in this study along withdescriptivestatisticsandreliability.Affectiveandsocialmas- terymotivationweremeasuredusingtheTeacherreportofthe DimensionsofMasteryQuestionnaire(DMQ;Morgan,1997;Morgan, Busch-Rossnagel,Barrett,&Wang,2009).TheDMQassessesmas- terymotivationbyhavingparentsorteachersratetheirperceptions of thechild’s behavior (orschool-aged children rate theirown

(5)

behavior)inmasterycontextsonaLikertScalefrom1-5(notat alltypicaltoverytypical).

Priortoitsuseinthisstudy,theDMQwastranslatedintoHun- garian,then back-translatedintoEnglish andany discrepancies resolved.Then, more than 10,000,mostly typically developing, childrenratedthemselvesand/orwereratedbytheirparentsand teachersinavalidationstudy(Józsa&Molnár,2013).Inthepresent study,theHungarianversionoftheitemsassessingsocialpersis- tencewithchildren(SPC)andmasterypleasure(MP)wereused.

Examplesofitemsmeasuringtheselatentvariablesare:“Triesto getincludedwhenotherchildrenareplaying”and“Smilesbroadly afterﬁnishingsomething”,forSPCandMP,respectively.

The DMQ alsoincludes items that index shameor sadness- related behavior inresponse tounsuccessful masteryattempts.

Thereisevidencethatwithdrawalemotionssuchasshameorsad- nesscan undermine continuedfocusand mastery(e.g.,Barrett, 1998),sothesenegativeemotionsseemparticularlyproblematic to school success. Some of the items we used to index negativereactionstofailure/challengeoriginallyweredevisedtobe reversedand includedin instrumentalscalesoftheDMQ;however,priorresearchindicatedthattheseitemsdidnothavehigh loadingsfromthecorrespondinginstrumentalmasterymotivation factors.Instead,theyrelatestronglytothenegative/withdrawal emotionitemsandseembestcharacterizedasnegative/withdrawal reactionstochallengeand/orfailure (e.g.,DeVellis, 2003; Feifei

&Tanner,2013).Wethereforeincludedtheseitemsasmanifest variablestomeasurea latentconstructthatwecalled“negative reactionstofailure/challenge”(NRF).Anexampleitemmeasuring NRFis“Looksdownorawaywhentriesbutcannotdosomething”.

Foreachofthesescales,theconstructwasthelatentvariable,and theitemsforthatscaleservedasmanifestvariables.So,itemswere notaggregatedintoatotalscale;rathereachitemservedasaman- ifestvariablefor therelevantscale (latentvariable). AsTable1 shows,allDMQscaleshadhighreliabilityinthepresentsample (Cronbachalphasall>.80).

2.2.2. IQ

TheRavenColouredProgressiveMatricesIQtest(Raven,Raven,&

Court,1998),administeredatpreschool,hasthreesetsofmatrices A,AB,andB.Thethreesetsareconsideredasthreesubscalesof thetest:A(11items,Cronbachalpha:.89),AB(12items,Cronbach alpha:.90),andB(12items,Cronbachalpha:.88).Theoverallrelia- bilityforthe35itemsis.92.Thethreesubscalescoreswereusedas manifestvariablesinthisstudy,measuringthelatentIQvariable.

EachIQscoreisstandardized,withameanof100andaSDof15.

2.2.3. Socioeconomicstatus

Fourmanifestvariables—mothers’andfathers’educationlevel andteacherandgovernmentratingsofdisadvantage—wereusedto assesschildren’ssocioeconomicstatus(SES).Mothers’andfathers’

levelofeducationweremeasuredasnumberofyearsofschool.Dis- advantagedstatusofthechildwasratedbythepreschoolteacher and,separately,bythelocalgovernment,ona4-pointscale(very disadvantaged,somewhatdisadvantaged,somewhatadvantaged, and veryadvantaged).The governmentrating is usedtoassess needforgovernmentsupport(e.g.,freelunch).Thecorrelationof mothers’andfathers’levelofeducationwas.58(p<.001).Thedis- advantagevariableswerealsostronglycorrelatedwitheachother (r=.70,p<.001).Thetwodisadvantagevariablesweremoderately highlycorrelatedwithmothers’education(r=.50,p<.001;r=.49, p<.001)andalsomoderatelycorrelatedwithfathers’education (r=.37,p<.001;r=.36,p<.001),respectively.

2.2.4. Schoolachievementtests

Table1 alsopresentsdescriptive statisticsand reliability for theGrade1and2achievementassessments.Allschoolachieve-

mentscoresarebasedona100-pointscore.Themathandreading achievement tests werebased on theHungarian NationalCur- riculum(HungarianGovernment,2011)andweredevelopedby evaluationexpertsintheﬁeld.Thetestsweredifferentin1stand 2ndgrades.Themathtestsincludedfoursubtests,andthereading teststhree.Thesesubtestswerethemanifestvariablesinthisstudy formeasuringthelatentvariablesofmathandreadingachieve- ment.Theoverallreliabilitiesforthesetestswerehigh(Cronbach alphas−Grade1:Math.96,Reading.97;Grade2:Math.95,Reading .98).

2.2.5. Socialskills

In additiontomeasuresofachievementinmathematicsand reading,weobtainedscoresonsocialskills.Thesocialskillstest ispartoftheDIFERschoolreadiness/schoolsuccesstestbattery thatiswidelyusedacrossHungary.DIFER,whichisadministered atthechild’sschool,isanacronym forHungarianwordsmean- ing“Diagnosticassessmentsystemsfor development”(cf.Nagy, Józsa,Vidákovich,&Fazekasné,2004,2016).Thestandardizedtests comprising DIFER include seven pre-academic skills, which are administeredindividuallyinthreeface-to-facesessionswiththe examiner(E).Tocreatethesocialskillsscaleforeachofthesethree individualsessions,theexaminerrateschildren’scommunication andsocialinteractionswithEon1-5Likertscales.Anadditional social skills scale is obtainedin a peerinteraction setting that includesfourchildrenwhoareinthesameclassatschool.Inthis peercontext,Erateschildrenontheircommunication,socialinter- actionwithE,andsocialinteractionwithpeers,using1-5Likert scales.Inter-raterreliabilitiesforthesocialskillscalesfromthese foursessionswereabove.75foreachsessionand.84forthetotal score(Nagy,Józsa,Vidákovich,&Fazekasné,2016).Thesocialskills ratingsineachofthefoursituationswereusedasmanifestvariables inthisstudy.Theratingswerecarriedoutinthesamemannerin1st and2ndgrade;however,theexaminersdoingthemweredifferent atthetwoages.

2.3. Procedure

TheDMQwascompletedbythepreschool(K2)teachersinearly October,nearthebeginningofK2,butafterteachershadgottento knowthechildren.TheIQtestwasadministeredatK2bytrained graduatepsychologystudents.Thedatacollectiontookplaceinthe ﬁrstquarteroftheK2academicyear,inOctober.Parentswereasked abouttheirlevelofeducation.

The achievement tests were completed by students during classes inGrade 1and Grade2 and resultswere scoredby the researchandevaluationteam.Themathtestsweregradedbyone teacherfromtheresearchandevaluationteam,and thereading testswereevaluatedbyadifferentteacher.Socialskillswererated bytrainedpsychologygraduatestudents(whoweredifferentat grade1andgrade2).Eachsituationlasted10–15minutes.Thedata collectionswereinMayinbothgrades.Therewereabout2.5years betweenthe1stand2nddatapoint,and3.5yearsbetweenthe1st and3rddatapoint.

3. Results

3.1. Preliminaryanalyses

3.1.1. Intercorrelationsamongvariables

Themainpurposeofthisstudywastoexaminetheabilityof preschool(K2)masterymotivation(masterypleasure(MP),negativereactiontofailure(NRF),andsocialpersistencewithchildren (SPC))topredictgrade1andgrade2reading,mathematics,and socialskills,aftertakingintoaccountIQandsocioeconomicstatus

(6)

Table2

IntercorrelationsfortheVariables.

Measure 1 2 3 4 5 6 7 8 9 10

1. NRFinPreschool –

2. MPinPreschool -.41 –

3. SPCinPreschool -.44 .49 –

4. IQinPreschool -.25 .22 .18 –

5. SES -.19 .19 .17 .37 –

6. ReadinginGrade1 -.29 .29 .25 .29 .29 –

7. ReadinginGrade2 -.25 .25 .16 .25 .34 .49 –

8. MathinGrade1 -.32 .26 .30 .36 .38 .62 .53 –

9. MathinGrade2 -.21 .17 .11^# .32 .29 .53 .53 .68 –

10. SocialskillsinGrade1 -.41 .37 .39 .22 .34 .34 .46 .46 .39 –

11. SocialskillsinGrade2 -.34 .35 .32 .20 .34 .41 .49 .48 .46 .84

Note.NRF=Negativereactiontofailure;MP=Masterypleasure;SPC=Socialpersistencewithchildren;#=Notsignificant.Unmarkedcoefficientsaresignificantatp≤.01.

(SES).However,beforeexecutingtheseprincipalanalyses,inter- correlationswererun,and strengthofcorrelations ofpreschool (K2)variableswithgrade1variableswascomparedtocomparable correlationswithgrade2variables.Inaddition,relativestrengthof zeroordercorrelationsofvariouspredictorswithparticularout- comevariableswasalsoassessed,toseeifourmodelexpectations seemedreasonabletopursue.AsTable2shows,alloftheinde- pendentvariables showedsigniﬁcantlow tomoderate(0.2-0.3) correlationswithreadingperformance.Thiscorrelationwasnega- tivefornegativereactiontofailure,andpositiveforalloftheother variables.Fisher’sz-testscomparingthestrength ofthecorrela- tionsofpredictorswithgrade1versusgrade2readingindicated nosigniﬁcantdifferences.

Formathematics,asimilarcorrelationalpatternemerged.The independentvariables’relationstofirstandsecondgradeperfor- mancewereofsimilarstrength,withoneexception:thecorrelation betweensocialpersistencewithchildrenandsecondgrademathe- maticsperformancewaslower,andwasnotsignificant.Thisisthe onlycorrelationthatwasnotsignificantintheentirematrix.

Thesocialskillsoutcomevariableshowedacorrelationalpat- ternsimilartothoseofreadingandmathematics,inthatrelative correlationswithgrade1andgrade2weresimilar.However,there weredifferencesamongpredictorsinstrengthofcorrelationwith thisvariable.Socioeconomicstatusshowedasigniﬁcantlystronger correlationwithsocialskillsthanIQdid(grade1:z=2.03,p=.02, grade2:z=2.37,p<.01).Masterymotivationvariables(mastery pleasure,negativereactiontofailure,socialpersistencewithchildren)werealsomorestronglyrelatedtosocialskillsthanIQwas (grade1:MPz=2.31,p=.01,NRFz=3.02,p<.01,SPCz=2.57,p

<.01;grade2:MPz=1.99,p=.02,NRFz=1.86,p=.03,SPCz=1.78, p=.04).TheseﬁndingssupportourexpectationthatIQwouldnot stronglypredictsocialskills.

Thestrongestcorrelationswerefoundamong thedependent variables, supporting the concept that all are measuring early schoolsuccess, and that schoolsuccessis relativelystable. The correlationbetweenthetwogradesforreadingwas0.49,formath- ematics0.68,andforsocialskills0.84.Socialskillsweresimilarly relatedtomathematicsandreading,thecoefﬁcientsbeingaround 0.4.Thecorrelationbetweenmathandreadingwas0.63ingrade 1,and0.52ingrade2.

Ingeneral,wecanconcludethatthereweresigniﬁcant,small- to-moderate correlations between independent and dependent variables.Therelationsofindependentvariablestograde2vari- ablesandtograde1variableswereofsimilarstrength,indicating thattherelationshipofpreschoolvariablestotheschoolskillsdid notweakenwithtime.Themainanalysesassessedhowourmodel oflatentpreschoolvariablestogetherpredictedschoolrelatedskills atgrade1andgrade2,andwhethertherelationofpreschoolpre- dictorstograde2performancewasindirect(throughgrade1school performance).

Table3

ModelFitsoftheMeasurementModels.

Model ² df p RMSEA CFI TLI SRMR

Preschool

5factors^a 437.81 195 <.01 0.060 0.932 0.920 0.059 3factors^b 1041.91 182 <.01 0.120 0.754 0.717 0.087 1factor 1987.90 209 <.01 0.161 0.505 0.453 0.132 Grade1

3factors^c 92.43 41 <.01 0.061 0.971 0.961 0.047 1factor 522.06 44 <.01 0.182 0.729 0.661 0.126 Grade2

3factors^c 82.03 41 <.01 0.055 0.979 0.972 0.036 1factor 687.93 44 <.01 0.212 0.673 0.591 0.131 Note.a=factorsarenegativereactiontofailure,masterypleasure;socialpersistence withchild,socioeconomicstatus,IQ;b=factorsaremasterymotivationtotal,socioeconomicstatus,IQ,c=factorsaremath,reading,socialskills.Eachfactorloadingwas signiﬁcant(p≤.01).

3.1.2. Measurementmodels

Thefocus ofourstudy wasonhowpreschool(K2) affective andsocialmasterymotivationpredictedschooloutcomes,overand aboveanyrolesofIQandSES.Structuralequationmodeling(SEM) wasutilizedtoexaminetheabilityof5latentpredictorvariables:

masterypleasure,negativereactiontofailure,socialpersistence withchildren,IQ,andSES,topredictthreeschoolskilllatentvari- ables:readingachievement,mathachievement,andsocialskills.

Everylatentvariablewasmeasuredby3-5manifestvariables,and thesameconstructsweremeasuredatgrade1andgrade2(see Table1).

Mplus, version 7.31, was used for data analysis (Muthén &

Muthén,2010).Themodelwastestedutilizingfullmaximumlikeli- hoodestimates.Toanalyzemodelfit,basedonBrown’s(2006)and Schreiber,Stage,King,Nora,andBarlow’s(2006)work,wecon- sideredthefollowinggoodness-of-fitindicators:␹²,RMSEA,CFI, TLIandSRMR.FollowingHuandBentler(1999),weregardedthe modelashavinganacceptablefitifthefollowingconditionswere met:RMSEAandSRMR≤0.06,CFIandTLI≥0.90.

Before conducting the principal analyses using the planned latentandmanifestvariables,weusedconﬁrmatoryfactoranaly- ses(CFA)totestthefactorialstructureofthemeasurementmodels forlatentandobservedvariables(seeTable3).Wetestedthemea- surementmodelsforthepredictivevariables,andalsoforthe1st gradeand2ndvariables.Allofthemanifestvariablesﬁt wellin thesemodels.

Incase of thepredictivevariableswe testedthree measurementmodels.Basedontheliteratureweexpectedﬁvepredictive variables:negativereactiontofailure,masterypleasure;socialpersistencewithchild,socioeconomicstatus,andintelligence.Wealso testedathreefactormeasurementmodel,wherethemasterymoti- vation manifest variables (negativereaction to failure, mastery pleasure;socialpersistencewithchild)allwerepredictedbythe samelatentvariable.Finally,wetestedaonefactormeasurement

(7)

modelwhereallofthepredictivemanifestvariablesbelongedto thesamelatentvariable.Neithertheonefactornorthreefactor modelfitthedatawell(seeTable3).However,asexpected,thefive factormodelfitthedatawell.Moreover,thefivefactormodelfitthe datasignificantlybetterthantheonefactor(␹²diff(14)=1540.48, p≤.01),andthethreefactormodels(␹²diff(3)=605.9,p≤.01).

Wetestedthesametwomeasurementmodelsfordependent variablesin ﬁrstandsecondgrades.In oneofthesemodelswe assumedthreelatentvariables(math,reading,socialskills).Inthe othermodelalloftheschoolmanifestvariables werepredicted byonelatentvariable(schoolsuccess).Thethreefactormodelsﬁt thedatawellinbothgrades,buttheonefactormodelsdidnot.

Moreover,thethreefactormodelsﬁtthedatasigniﬁcantlybetter thantheonefactormodelsbothinthe1stgrade(␹²diff(3)=429.63, p≤.01),andthe2ndgrade(␹²diff(3)=605.9,p≤.01).

Basedontheliterature,inconjunctionwiththeﬁtstatisticsof themeasurementmodels,we thereforeusedﬁvelatentpredic- tivevariablesandthreelatentoutcomevariablesfortheprincipal analysesinthisstudy.

3.2. Principalanalyses

Ourhypothesesweretestedusingaseriesoffouranalyses.For everymodeltested,thehypothesiswasthatpreschoolvariables predictﬁrst-andsecondgrade schooloutcomes.We separately examinedpredictionofmath,reading,andsocialskills,andthen alsotesteda modelpredictingall threeof theseschoolsuccess variables.¹

Onthebasisoftheliterature,weassumedthatIQ(Crosnoe&

Benner,2015)andSES(Duncanetal.,2015)wouldpredictreading andmathachievement,soweincludedthemaspredictorsofthese twooutcomesinadditiontothefocallatentvariables involving affectivemasterymotivation.Wedidnotexpectsocialpersistence withpeerstopredictthesecognitiveoutcomes,sothisvariablewas omittedfromthesetwomodels.²Incontrast,weexpectedallthree masterymotivationlatentvariables,especiallysocialpersistence withchildren,andSEStopredictsocialskills(e.g.,Denhametal., 2014;Fishbeinetal.,2016),butwedidnotexpectIQtodoso,so thiswasomittedfromthemodel¹.Ourapproachwastouseonly thesimplermodel,withaprioripredictorsinit,astheﬁnalmodel unlessaddingtheotherpredictorsigniﬁcantlyimprovedthemodel.

Ineach case, ourhypothesized modelﬁt atleastslightly better thanthealternativemodel.However,weincludedall5preschool measuresinanadditionalmodelpredictingallthreeoutcomes.See Figs.1–4fordetails.

Wehypothesizedthatpreschoolvariableswouldpredictfirst gradeperformance,which,inturn,wouldpredictsecondgradeper- formance.However,wefirsttested whetherpreschoolvariables directlypredictedbothfirst-andsecond-gradeschooloutcomes.

Resultsindicatedthat,inallcases,thebestﬁtmodelheldwhen

1Weranmultilevellinearmodelstotesttheeffectofstudents’classesonthe results.Theanalyseswithstudentsnestedinclassroomsshowednonestedeffects ofthepredictors.Thenestingdidnotaccountfornorqualifytheeffectsofpredictors;

therefore,wedidnotincludetheclassroomvariableintheSEMs.

2WealsorantheSEMmodelswithbothsocialpersistenceandIQincludedinthe separatemath,reading,andsocialskillsmodels,butsocialpersistencedidnotcon- tributetopredictingmathorreading;norwerethemodelswithsocialpersistence notablybetterfitting,(math:RMSEA=.050,CFI=.941,TLI=.934,SRMR=.057,and ²(359)=650.47,p≤.01;reading:RMSEA=.054,CFI=.929,TLI=.919,SRMR=.060, and²(306)=600.25,p≤.01).Similarly,IQdidnotcontributetopredictingsocial skills;norwasthemodelfitimprovedsignificantlybyincludingIQ(RMSEA=.051, CFI=.954,TLI=.947,SRMR=.056,and²(328)=611.16,p≤.01),soweonlydiscuss herethesimplermodels.Thesimplermodelsfitthedatabetterinthecaseofmath (²diff(120)=193.48,p≤.01)andreading(²diff(110)=182.35,p≤.01);therewas nosignificantdifferencebetweenthetwomodelfitsinthecaseofsocialskills(²diff

(70)=76.87,p=.27).

Fig. 1.Standardized coefﬁcients for the Math Model. Latent constructs are shown in circles. NRF=Negative reaction to failure; MP=Mastery pleasure;

SES=Socioeconomicstatus;MATH1=Mathachievementingrade1;MATH2=Math achievementingrade2.

Fig. 2.Standardized coefﬁcients for the Reading Model. Latent constructs are shown in circles. NRF=Negative reaction to failure; MP=Mastery pleasure; SES=Socioeconomic status; READ1=Reading achievement in grade 1;

READ2=Readingachievementingrade2.

preschool variables indirectly predictedsecond-grade variables, throughﬁrstgradevariables.Forthisreason,weonlydiscussthese predictedmodels.

3.2.1. MathAchievement

Themodelpredictingmathachievementfrompreschoolmas- terymotivation,IQ,andSESfit well(see Table4).Children’s IQ (␤=.26, p <.01)and SES(␤=.32, p <.01)significantlyand pos- itivelypredictedmathachievement,whilenegativereactionsto failurewasasignificant,negativepredictor(␤=−.16,p<.05).Mas- terypleasuredidnotsignificantlypredictmathachievement.First grade mathperformance stronglypredicted second-grademath performance(␤=.80,p<.01;Fig.1).

(8)

Table4

ModelFitsoftheMeasurementModels.

Model ␹² df RMSEA CI CFI TLI SRMR

1. Math 501.15 262 0.053 0.046−0.060 0.941 0.932 0.055

2. Reading 464.67 217 0.059 0.052−0.066 0.923 0.910 0.061

3. Socialskills 534.29 258 0.057 0.050−0.064 0.953 0.945 0.058

4. Total 1383.83 825 0.046 0.041−0.050 0.935 0.929 0.070

Note.Eachfactorloadingissigniﬁcantatp≤.01.AllmodelsuseSES,IQ,negativereactionstofailure(NRF),andmasterypleasure(MP)aspredictors.Socialskills2andTotal 2modelsalsoincludesocialpersistencewithchildren(SPC).CI=90%conﬁdenceintervalsfortheRMSEA.

Fig.3.Standardized coefﬁcients forthe SocialskillsModel. Latentconstructs areshownincircles.NRF=Negativereactiontofailure;MP=Masterypleasure;

SPC=Socialpersistencewithchild;SES=Socioeconomicstatus;SK1=Socialskills ingrade1;SK2=Socialskillsingrade2.

3.2.2. ReadingAchievement

Thereadingmodelalsofitwell(seeTable4),butasomewhat differentpatternemerged. IQdidnotsignificantly predictfirst-

gradereadingachievement,whileSES(␤=.35,p<.01)andmastery pleasure(␤=.20,p<.01)bothhadsigniﬁcant,positivecoefﬁcients.

Negativereactionstofailure’scoefficientwassignificantandneg- ative,asit wasformathematics(␤=−.19, p<.05).Therelation betweenfirst-andsecondgradereadingwasstrong(␤=.60,p<.01;

Fig.2).

3.2.3. SocialSkills

Thethirdmodelpredictedsocialskills,andagainﬁtwell(see Table4).Itwasassumedthatsocialpersistencewithpeerswouldbe animportantpredictorofthedevelopmentofsocialskills(Barrett

&Morgan, 1995), and,infact,theSPCvariablesignificantlyand positivelypredictedfirst-gradesocialskills(␤=.21,p<.01).The coefficientforSES(␤=.31,p<.01)wasalsosignificant.Negative reactionto failurenegatively predicted social skills(␤=−.23, p

<.01),butmasterypleasuredidnotsigniﬁcantlypredictsocialskills inthismodel.Therelationshipbetweenﬁrst-andsecond-grade socialskillswasquitehigh(␤=.88,p<.01).

3.2.4. TotalModel

In thefourth model,we included reading, mathematicsand social skillsas dependentvariables inone complexmodel, and preschoolpredictorsincludedallﬁveavailablevariables(Fig.4).

IQsigniﬁcantlycontributedtopredictingmathematics(␤=.23,p

<.01)andreading(␤=.16,p<.01)inthismodel.SESsigniﬁcantly predictedallthreedependentvariables(Math␤=.36,p<.01,Read- ing␤=.40,p<.01,Socialskills␤=.37,p<.01).Negativereactions to failure signiﬁcantly and negatively predicted all three vari-

Fig.4.StandardizedcoefﬁcientsfortheTotalModel.Latentconstructsareshownincircles.NRF=Negativereactiontofailure;MP=Masterypleasure;SPC=Socialpersis- tencewithchild;SES=Socioeconomicstatus;MATH1=Mathachievementingrade1;MATH2=Mathachievementingrade2;READ1=Readingachievementingrade1;

READ2=Readingachievementingrade2;SK1=Socialskillsingrade1;SK2=Socialskillsingrade2.

(9)

Table5

VarianceExplainedintheModels(R²).

Model Grade1 Grade2

Math Reading Socialskills Math Reading Socialskills

1. Math .37 – – .64 – –

2. Reading – .36 – – .36 –

3. Socialskills – – .38 – – .78

4. Total .41 .50 .43 .62 .43 .77

ables(Math␤=−.18,p<.05,Reading␤=−.24,p<.01,Socialskills

␤=−.28,p<.01).Masterypleasurepositivelycontributedtopre- dictingreading(␤=.20,p<.01)andsocialskills(␤=.21,p<.01), andSPCpredictedsocialskills(␤=.19,p<.01).

Wesummarizethemodels’explainedvarianceinTable5.The testedmodelsexplainedabout40%ofvarianceinallofthethree schooloutcomesvariablesingrade1.Inthecaseofreading,the explainedvariancewasaboutthesameingrade2asingrade1, approximately40%.However,it washigher in2ndgradeinthe caseoftheothertwodependentvariables.Theexplainedvariance isabout60%formathinallofthethreemodels,and78%forsocial skillsinallofthefourmodelsin2ndgrade.

4. Discussion

Theresultsofthisrelativelylarge,longitudinalstudysupport theutilityofexaminingsocioemotionalaspectsofmasterymoti- vation,suchassocialpersistencewithpeersandaffectiveaspects of mastery motivation, as school readiness dimensions. Social persistencewithpeerssigniﬁcantlypredictedsocialcompetence intheearlygrades,and social competenceisknowntopredict positiveschoolandlifeoutcomes(e.g.,Jonesetal.,2015).More- over,negativereactionstofailure(NRF)predictedallthreeschool achievementmeasures,over and abovethecontributionof SES andIQ,andmasterypleasurepredictedboth readingand social skills.Infact,althoughnegativereactionstofailureconsistently predictedallmeasuresofearlyschoolskillsandachievement,IQ onlypredictedmathandreadingachievement.OurNRFvariable doesnotseemtobebestunderstoodasfrustrationtolerance,in thatitisnotstronglycorrelatedwithotherDMQitemsmeasur- ingangry/frustratedresponsestomastery.Itismorelikelyrelated toinclinationtobecomesad and/orashamedand,thus, togive up/withdrawfrommasteryattempts.Moreover,weviewitasa malleablepropensity,ratherthanastemperament,astraditionally conceptualized,andthereissomeevidencethatitcanbemodiﬁed (Chia,Keng,&Ryan,2016).However,furtherresearchisneededto assesswhetheritis,aspredicted,responsivetointervention.

Oftheotherpredictors,onlySESpredictedallthreemeasures, replicatingtheknowntendencyforchildren whose parentsare lesseducatedandwhohavefewermaterialresourcestoperform atlowerlevelsinschool.Moreover,thepreschoolvariablespre- dictedthegradetwoschoolskillsindirectly,throughtheirimpact onﬁrstgradeskills.Theresultssuggestthatchildren’stendencyto respondnegativelytoandtoavoiddifﬁcultachievementtasksmay underminenotonlytheirconcurrentperformanceinschool,but alsofutureskills.Thishighlightstheimportanceofmeasuringsuch negativereactionsearlyindevelopmentandinterveninginearly childhoodclassroomand/orhomeenvironmentstoaddressthem beforetheycanimportantlycontributetonegativedevelopmental cascades.

This ﬁnding complements earlier evidence that positive Approaches to Learning, such as mastery motivation, can help compensateforless-than-optimallearningenvironments(Meng, 2015).Itappearsthat,inaddition,negativeApproachestoLearning, intheformofnegativewithdrawalemotionsinresponsetounsuc- cessfulmasteryattempts,canaddtothenegativeimpactoflowSES,

furtherunderminingschoolsuccessinmath,reading,andsocial competence.Theseattitudesandmotivationtowardlearning,may beasimportantasSocial EmotionalLearningorSocioemotional Competencecharacteristics,suchasemotionunderstanding,social skills,andabilitytoeffectivelyregulateemotions,incontributing behaviorallytoschoolreadinessandschoolsuccess.

Our ﬁndings alsoprovidefurthersupportfor theconclusion that disparities in Socio-Economic Status (SES), can undermine schoolsuccess, potentiallyleading tocontinuedstratiﬁcationof educationalopportunityandattainment(Duncanetal.,2015;Józsa, 2016),andthatApproachestoLearningcanimpacttheseoutcomes.

TheyalsosupporttheconclusionthatIQdifferencesarenotthe mostimportantfactorinsuchdisparities,providinghopethatenvi- ronmentalsupportforotherfactors,suchasmasterymotivation, mighthaveanimportantandlong-lastingimpactonsuchdiffer- ences.

Unfortunately, mastery motivation, especially affective and socialaspectsofmasterymotivation,isalmostneveradequately assessedcurrently.Althoughisnowclearthatnon-academicschool readinessskills,suchasexecutivefunctionsandsocial-emotional competence,playimportantrolesinschoolsuccess(e.g.,Denham etal.,2014;Bierman&Torres,2016),lessattentionhasbeenpaidto motivationalinﬂuencessuchasenthusiasm,persistence,andother aspectsofmastery/competencemotivation,whichmaybejustas important(e.g.,Hyson,2008;McDermottetal.,2012).Theseminal report“FromNeuronstoNeighborhoods”speciﬁcallyhighlighted masterymotivationasakeydomainofdevelopment(Shonkoff&

Phillips,2000).Thepresentresearchsuggeststhatnegativereac- tionstofailure,masterypleasure,andsocialmasterymotivation areparticulardomainsofmasterymotivationthatalsoareworthy ofassessmentasschoolreadinessvariables.

Findingsfromourstudyarepotentially especiallyimportant given the consistent ﬁnding from decades of research, across a number of cultures, that mastery motivation decreases from elementary school through high school. Many have suggested thatmasterymotivationmightbeunderminedbycurrentschool environments,whichtypicallyemphasizeexternalsignsofperfor- mancesuchas grades,ratherthan focusingontryinghard and improving (e.g.,Józsa &Morgan, 2014; Józsa, Wang,Barrett, &

Morgan,2014;Gottfried,Marcoulides,Gottfried,&Oliver,2013;

Harter,1981;Lepper,Corpus,&Iyengar,2005).Thus,ifmastery motivationisanimportantpredictorofsuccessinschool,itwill beimportanttoimplementearlyinterventionstobettersupport itsdevelopmentduringtheschoolyears.Thereisevidencethat earlyinterventionscanincreasemasterymotivation(e.g.,Hashmi etal.,2017; Vansteenkisteetal.,2004),soifinterventionswere implementedintheearlygrades,thiscouldpotentiallydecrease thedeclineandincreaselaterachievement.

The present study highlights the need to include social/emotional aspects of mastery motivation in both school readiness assessments and universal interventions, to try to changethispotentiallyharmfuldecline. Ithighlightsthepoten- tialforinterventionsthatfostermasterymotivationtoincrease schoolperformance, evenin children whoare atrisk forlower performanceduetolowerSES(e.g.,Hashmietal.,2017;Steuer, Rosentritt-Brunn,&Dresel, 2013; Vansteenkisteet al.,2004).In