You are on page 1of 16

ImprovingAuthoritativeTeaching

Dr.SigrunK.Ertesvg,AssociateProfessor,CentreforBehaviouralResearch,Universityof Stavanger.Email:sigrun.ertesvag@uis.no Paper presented at the ICSEI Congress 2011 International Congress for School effectiveness and Improvement. Linking Research, Policy and Practice to PromoteQuality inEducation,LimassolCyprusJanuary47th Abstract Thepresentstudyreportsresultfromtwotwoyearinterventionsaimingatimproving teachersauthoritativeteaching.Authoritativeteachingwasunderstoodasatwo dimensionalconstructofwarmthandcontrol.Teachersatschoolsattendingoneoftwo schoolwideinterventionswereincludedinthesurvey.Aquestionnairewasadministeredat threetimes,before(T1),oneyearinto(T2)andattheendof(T3)thetwoyearinterventions. Throughtheapplicationofconfirmatoryfactoranalysis(CFA)andstructuralequation modelling(SEM)thefactorstructureandmeasurementinvariancewereinvestigated.Latent variablegrowthcurveusingMpluswasappliedtoinvestigatetheimprovementin authoritativeteaching.Generally,resultssuggestthattheteachersimprovedincontrol, howevernotinwarmth.Resultsarecomparedtopreliminaryresultsfrompupilsreportsof theirteachersemotionalsupportandmonitoring,whicharerelatedbutnotidentical concepts. Introduction Overtheyearstheteacherroleasbecomeincreasinglymorecomplex.Asaresultmany teachersstrugglemanagingbehaviourinclass(e.g.Midthassel,2006;vanTartwijk,denBrok, Veldman,&Wubbels,2009),thisappliesbothtobeginningandmoreexperiencedteachers (Evertson&Weinstein,2006).Also,creatingapositivelearningenvironmentisofgreat concernforteachersaspupilsrelationshipswiththeirteachershavebeenshowntobe importantpredictorsofacademicandsocialadjustment(e.g.Hamre&Pianta,2005;Hughes, 2002;Roland&Galloway,2002). Asaresponsetothisinternationallyacknowledgedproblem(seeErtesvg,2011for overview),therehasbeenanincreasingfocusfromgovernment,researchersandschoolsin Norwayonimprovingteachersskillsinmanagingbehaviourandclassroomsettings.This studyfocustheevaluationoftwoschoolwideinterventionsofferedtoNorwegianprimary andsecondaryschools.TheRespectprogramisatwoandahalfyearactiontopreventand reduceproblembehaviourthathasbeenofferedtoNorwegianschoolssince2002.The creationofahandbookonclassroommanagementisatwoyearinterventiontoimprove classroommanagementwithanevenlongerhistory.Amainprincipleofbothinterventions isauthoritativeteachingandimprovingteachersclassroommanagementskillisamainaim. WhereastheRespectprogramisaschooldevelopmentprogrammewhichchallengesall sidesoftheschoolactivitywithstructureandculture,theHandbookinterventionisfocusing ontheteacherasclassroommanager.However,bothinterventionsweretargetingthe teacherrole,whichisthecornerstoneofteacherswork. Inthepresentstudy,onewayofconceptualizingrelevant,selected,aspectsofauthoritative teachingunderstoodasatwodimensionalconceptconstitutingwarmthandcontrolis discussed.Apreviousstudy(Ertesvg,2011)discussedthedevelopmentofameasurement forteachersselfreportsofwarmthandcontrolinlongitudinalsamples.Basedonthe

conceptualframeworkforinvestigatingauthoritativeteachingoutlined,thepresentstudy explorestheimprovementofwarmthandcontrolinteachersparticipatinginthetwo interventions.Aspartofalargerstudytoinvestigateboththeeffectandtheeffectivenessof theinterventionsperformedbytheInnovationResearchGroupatCBR,themainaimofthe studyathandwastoexploretheimprovementinteachersauthoritativeteachingand classroommanagementskills.Teachersselfreportsofauthoritativeteachingwerethemain focus.However,asanattempttoshedfurtherlightontheteachersresultpupilsperception oftheirteachersemotionalsupportandmonitoringisincluded.Latentgrowthcurve analysiswsappliedforinvestigatingofchangeinauthoritativeteaching. In2000Taylorandcolleaguesarguedthatalthoughlatentgrowthcurvemodellinghadbeen utilizedinnumerousstudies,fewstudieshadappliedthesetechniquesinprogram evaluation.Tenyearslater,thesituationhasnotchanged.Whencollectinglongitudinaldata, theprimarymodeofanalysishassomevariantofexamininggroupdifferencesatdifferent timepoints.Inalatentgrowthmodel,changeovertime,ismodelledasrandomcoefficients thatcanbeestimatedforeachindividual.Anadditionaladvantageisthatitrepresentsa multivariateapproachtotheanalysisofmultiplewavesofoutcomedata.Thisapproach providesgreaterstatisticalpowerfortestingoverallprogrameffectsthatareconsistentover time,andappropriatelydownweightpotentiallyspuriousprogrameffectsthatareoverved injustoneofmanywavesofoutcomedata(Tayloretal.2000). Authoritativeteachingandteacherpupilrelationships TheconceptofauthoritarianteachingisadaptedfromBaumrinds(1967,1978,1991) parentingstyles.Authoritarianteachersworktobuildrelationshipsofwarmth,acceptance, andopennessontheonehandandtocontrol,establishhighstandards,andhavehigh expectationsofsociallyresponsiblebehaviourontheother.Suchacombinationaimsat preventingproblems,andalsohasthedualpurposeofmanagingbehaviourintheshort termanddevelopingresponsibilityamongpupilsinthelongterm(Hughes,2002;Pellerin, 2005). Inshorttheauthoritativestyle,inlinewithBaumrindstypology,centrearoundthe combinationofthetwodimensionswarmthandcontrol.Walker(2009)arguethatboth dimensionsareneededinordertohandlethecomplexprocessesintheclassroom.The warmthdimensiondealswithteachersinterestintheirpupils,buildingpositive relationships,whilethecontroldimensionfocusonestablishingroutinesfordifferent situationsaswellasmonitoringtheirbehaviour.Thisteacherstyleseemstobebeneficial especiallyforpupilswhoareatriskforpooracademicachievement(Hughes,2002)and aggressivepupils(Meehan,Hughes&Cavell,2003).However,asBakeretal.(2009)argues, authoritativeteachingmaybebeneficialforallpupilsandsuggestthistypeofteacherstyle tobeuniversal. Althoughagrowingbodyofresearchfocusontheconceptoftheauthoritativeteacher, therearestillrelativelyfewstudies.However,theconceptofthewarmdemanderisnotnew (e.g.Kleinfeld,1975;Morrison,1975;Walker,2008).Also,researchfocusingonsimilar conceptse.g.therelationbetweenteacherandpupil,describespositiveresultsforpupils socialadjustmentandachievementwhentherelationwascharacterisedbywarmthand control.Considerableevidencesupportsthebeneficialeffectofacloseteacherpupil

relationshiponchildrensacademicandbehaviouraloutcomes(Ertesvg,2009for overview).Therelationshipbetweenteacherandpupilsisimportantfordevelopingagood learningenvironment,aspupilsrelationshipswiththeirteachershavebeenshowntobe importantpredictorsofacademicandsocialadjustment(e.g.Hamre&Pianta,2005; Henricsson&Rydell,2006;Hughes,2002;Roland&Galloway,2002).Further,teachers abilitytomanageandcontroltheirpupilslearningandbehaviourintheclassroomalso predictspupilachievement(Good&Brophy,2007).Controlreflectsconsistentenforcement ofrulesandprovisionofstructuretochildrensactivities.Theconceptsofauthoritative teachingandteacherpupilrelationshiparediscussedinmoredetailinpreviousstudies (Ertesvg,2009,2011). Method Subjectsandprocedures Theteachersamplecontained642teachersattendingoneoutoftwoschoolwide interventions.Twohundredandfortythreeteachersattenschoolsattendedtheproject Thecreationofahandbookonclassroommanagementand399teachersat18schools attendedtheRespectprogramtopreventandreduceproblembehaviour.Inboth interventionsstrengtheningteachersauthoritativeteachingwereanessentialaspectand authoritativeteachingwasamainprincipleofbothinterventions.Theprojectperiodwas twoyearsforbothinterventions.Preliminaryanalysisofdescriptivedatadidnotreveal significantdifferencesbetweenthetwogroupsforanyoftheitems.Giventhis,thedata fromthetwoprojectswerepooledtogether. Pupilsingrade510(1116years)ateachdatacollectionwereincluded.Atthethreetime points(2504/2609/2606)pupilsintheRespectprogramand(1266/1399/1455)pupilsinThe creationofahandbookonclassroommanagementwereincluded.Thenonrespondents weremainlyduetopupilsbeingabsentfromclasswhenthequestionnaireswere administeredortomissingparentalpermissionslips. Forbothteachersandpupilsawebbasedquestionnairewasadministeredatthreewaves, before(T1),oneyearinto(T2)andattheendof(T3)thetwoyearsinterventions.Altogether, 22schoolsparticipated(11schoolsintheRespectprogram,10schoolsinDevelopmentof handbookinclassroommanagement).DatacollectionwasconductedinMayeachyear. Althoughparticipatingindifferentprojects,allschoolswerepartofthesamedatacollection. Giventhistheproceduresfordatacollectionwasthesame.Theteachersconsentwas obtainedbyvoluntaryparticipationbasedonawrittendescriptionoftheprojectaccording tostandardsprescribedbytheNorwegianDataInspectorate.Tosafeguardanonymitya contactpersonateachschooladministeredauniqueidentificationcodeforeachparticipant generatedbythedelivererofthewebbasedquestionnaire.Thecontactpersonwas responsibletokeeptheautogeneratedidentificationcodesstoredunderlockandkeyfor useatthelaterdatacollections.Thispersondidnothaveanyaccesstothecollecteddata. Instruments Teachersauthoritativeteachingwasmeasuredbytwoscales,consistingoffouritemseach (seeTable1).Thescalesweredevelopedbasedonthetheoreticalframeworkofwarmthand controloutlined,foramorecomprehensivediscussionoftheframeworkandinformationof thescalesseeErtesvg(2011).

Thepupilquestionnairecomprisedscalesincorporatinganestimateofthepupils assessmentoftheirteacherssupportandmonitoring.Thekeyinformationrequired concernedpupilsperceptionsofteachersacademicsupport(4item,forthedifferenttime points),teachersemotionalsupport(4item,=0.780.79forthedifferenttimepoints)and teachersmonitoring(6item,=0.830.84forthedifferenttimepoints).Seeappendixfor items.Theinternalconsistencyprovedtobefairlygoodtoverygood(Brown,1970).All scalesarepreviouslydocumented(e.g.Bruetal.,1998;Bruetal.,2002;Thuenetal.,2007; Ertesvg,2009).However,thescaleforemotionalsupporthasbeenusedwithadifferent numberofitemsvaryingfrom46.Inourstudy,afouritemsolutionwaschosen.Thesefour itemshavebeenincludedinallthestudiesbyBruandhiscolleagues.Also,threescales measuringproactivepowerrelatedaggression(4item,=0.810.84forthedifferenttime points),proactiveaffiliationrelatedaggression(4item,=0.830.85forthedifferenttime points)andreactiveaggression(6item,=0.750.78forthedifferenttimepoints)were includedtocontrolforselectionbias.Thescaleshavepreviouslybeendocumentedby RolandandIdse(2001).Allscaleshadafourstepscoringformat;Disagreestrongly, Disagree,Agree,andAgreecompletely.Allscaleswerescored0,1,2,and3respectively. Dataanalyses Teachersample Fortheteachersamplealatentvariablegrowthcurveapproachwereapplied.Ertesvg (2011)providedaframeworkformeasuringauthoritativeteachingasatwodimensional conceptofwarmthandcontrol.Themodelwasfittedtothedatabythemeansofrobust maximumlikelihoodprocedureasimplementedinMplus(Muthen&Muthen,19972008). Themissingdatamethodwasemployed,asthisallowedfortheuseofallobservationsinthe datasetwhenestimatingtheparametersinthemodels.Conventionalanalyseswere conductedusingSPSS(Norusis,2007).Onewayanalysisofvariance(ANOVA)wasappliedto provideinformationonmeandifferencesbetweengroups.Spearmanswasemployedfor correlationsbecauseoftheordinalnatureofthedata. HuandBentler(1999)recommendedusingacutoffvaluecloseto.08forthestandardized rootmeansquaredresidual(SRMR)andsupplementingitwithindicesliketheTuckerLewis Index(TLI)andtheComparativeFitIndex(CFI)withcutoffvaluescloseto.95.Theyalso recommendedincludingtheRootMeanSquareErrorofApproximation(RMSEA)withcutoff valueofabout.06orless.TheRMSEAissupportedbya90%confidenceinterval(90%CI). Pupilsample Aspartofalargerstudy,thepupilreportoftheirteacherssupportandmonitoringwas analysedapplyinganexpandedversionofaselectioncohortdesign(Cook&Campbell, 1979;Shadish,Cook,&Campbell,2002)alsotermedacohortlongitudinaldesignwith adjacentorconsecutivecohorts(Olweus&Alsaker,1991)waschosen.Theextended

Table 1 Item means (M), standard deviation (SD), Skweness (Ske), and Kurtosis (Kurt) for scales measuring warmth and control

Item
Warmth 1. I work actively to create good relationships with my pupils 2.I show interest in each pupil 3. I often praise my pupils 4. I show the pupils that I care about them (not only when it comes to academic work) Control ( 5. I have established routines/rules for how the pupils are supposed to act when they change activity/workplace etc. 6. I have established routines/rules for how the pupils are supposed to act in plenary teaching sessions 7. I have established routines/rules for individual work 8. I am closely monitoring the pupils behaviour in class Rating format is 0-5, where 0 = never and 5 = very often.

M
4.46 4.46 4.25 4.50

Time 1 SD Ske
.61 .63 .72 .67 -.71 -.84 -.56 -1.08

Kurt
-.47 .12 -.33 .58

M
4.42 4.49 4.30 4.48

Time 2 SD Ske
.66 .62 .73 .66 -.94 -1.33 -.79 -1.07

Kurt
.89 4.2 .38 .76

M
4.45 4.46 4.31 4.52

Time 3 SD Ske
.64 .61 .71 .63 -.87 -.75 -.63 1.06

Kurt
.35 -.16 -.45 -.51

3.55 4.19 3.99 4.32

.91 .79 .79 .70

-.28 -.81 -.79 -1.0

-.01 -.61 1.58 .94

3.71 4.33 4.17 4.37

.86 .72 .72 .68

-.48 -.88 -.64 -.90

.69 .65 .63 1.03

3.78 4.39 4.24 4.41

.86 .67 .70 .64

-.42 -.80 -.66 -.86

-.03 -.23 .25 1.09

versionisdescribedthoroughlybyDanOlweusandcolleagues(e.g.Olweus,2005;Olweus& Alsaker,1991). Thesubjectsbelongedtoacohortinthattheywereattendingaparticulargradelevelatthe sametimeandwereapproximatelythesameage.Thecohortswerenamedafterthegrade leveltheywereinatTime1(T1),Cohort5(age11)atT1wascomparedwithCohort4at Time2(T2)whichatthattimehadreachedthesameage(11)asCohort5atT1(seeTable2). Similarly,Cohort3(age11)atTime3(T3)werecomparedwithCohort5atT1.Thesame kindsofcomparisonweremadebetweentheothercohortsatT1andcohortsin correspondinggradelevelsatlaterpointswhentheyreachedthesamegrade/age.The comparisonwasappliedforallmeasuresusedinthestudy.
Table2 Overviewofcohortatthedifferenttimepoints(T1T3)


T1

Gradea
5 5 5

Grade
6 6

Grade
7 7 7 7 7

Grade 8

Grade
99

Grade
10

T2

6 6 6

T3

8 8

9
9

10 10

10

a.Sameshapedfiguresindicatethesamecohort.Cohortsthatonlymedareincludedonceareidentifiedbyno figure.

Cohortdesigndoesnotautomaticallyruleoutselection(Cook&Campbell,1979).Therefore, thedesignwillgainadditionalstrengthifthedatashowsthatgroupsinandoutsideof interventiondonotsystematicallydifferonreliablemeasuredthirdvariables.Obviouslysuch measurementscannotruleoutselectionthreatsassociatedwithunmeasuredvariables;it canonlyruleoutsomeparticulartypesofselectionbias.Giventheriskofnegative relationshipsbetweenaggressivepupilsandtheirteachers,thereisreasontobelievethat aggressionwillaffectpupilsreportsofclassroomleadershipnegatively.Therefore,the cohortswereanalysedforthreemeasuresofaggression:reactiveaggression,proactive affiliationaggressionandproactivepoweraggressiontoinvestigatewhetherdifferencesin anyofthethreeaspectsofclassroomleadershipmightbeduetodifferencesinthelevelsof aggressionbetweenthecohortsatT1andsubsequentcohorts. Missingdata Attritioniscommoninlongitudinalstudies(Jelicicetal2009),andthiswasalsothecasefor thisstudy.Theattritionwaspartly,duetoarollingdesign.Teacherswholefttheschool droppedoutofthestudy.Newteachersstartingtoworkattheschoolwereincludedatthe secondandthirdwaves,asthesampleateachtimepointwasteacherscurrentlyworkingat theschool.Theteachersdropoutandinclusionwere,mostlikely,notrelatedtothe programme.Itseemsreasonabletobelievethatteachersleftschoolsforreasonsotherthan theschoolsparticipationinaschoolwideintervention.Forthisreason,themissingnessmay beconsideredtoberandom(MAR)(Buhietal.2008).Whendataaremissingatrandom,

incompletedataarisenotfromthemissingvaluesthemselves,ratherthemissingnessisa functionofsomeotherobservedvariableinthedata. Bothinterventionscontainedproceduresforintroducingnewteacherstothetheoretical groundingandprinciplesoftheinterventionandincludingthemintheworkattheschool. Giventhis,ismadesensetoincludenewteachersinthesample. However,attritionisnottheonlytypeofmissingnessinlongitudinalstudies.Threetypesof missingdatamaybeathreattolongitudinalstudies(Buhietal,2008).Inadditionto attrition,missingnesscanoccurasresultofpeoplenotrespondingtotheinvitationto participate,usuallyreferredtoasresponserate.Also,missingnessduetoteacherswho participatingbutdonot,fordifferentreasons,answerallitems,normallydescribedasitem nonresponse.Thelatterwasnotamajorprobleminthestudyathandas9395%ofthe teachersrespondedatthedifferenttimepoints.Ofthosewhodidnotrespondtoallitems, mostdidnotrespondtooneortwo. Theoverallresponseratewere74%,71%,and59%respectivelyforthedifferenttimepoints. ThisresponserateisinlinewiththeresponserateinpreviousstudiesamongNorwegian teachers(e.g.Munthe,2001;Midthaasel,Bru&Idse,2000). Attritioncanbeseparatedinteacherswithdrawingfromthestudyforreasonsotherthan leavingschoolandthosemissingbydesignforthosenotincludedinthesampleatalltime points,i.e.thosenotworkingtheschool.Apreviousstudycontainingalargersampleof teachersattendingthetwointerventions(Ertesvg,2011)indicatedthatlessthan10%ofthe teachersdroppedoutofthestudyduetootherreasonsthanleavingtheschool.Teachers beingsickatthetimeofdatacollection,parentalleave,leaveofabsence,ignoring subsequentinvitationswasamongthereasonsforthistypeofmissingnessamongteachers. ThelowestnumberofteachersparticipatedatT3.Theremaybeseveralreasonsforthis,for example,newteachersmayhavethoughttheirresponseswerenotsoimportantsincethey werenewtotheintervention. Results Teachersimprovementinauthoritativeteaching. Descriptivestatisticofteachersselfreportsfortheeightitemsandalldatacollectionsare presentedinTable1.Theresultsindicatedanormalunivariatedistribution,giventhatmost skewnessandkurtosisvaluesfellwithintherangeof1.0to+1.0. Inthepresentstudy,onewayofconceptualizingrelevant,selected,aspectsofauthoritative teachingunderstoodasatwodimensionalconceptconstitutingwarmthandcontrolis discussed.Basedontheconceptualframeworkforinvestigatingauthoritativeteaching outlined(Ertesvg,2011)thepresentstudyexploretheimprovementofwarmthandcontrol inteachersparticipatinginthetwointerventions.Additionally,toshedfurtherlightonthe improvementofteachersselfreportedauthoritativeteaching,preliminaryanalysisofpupils reportsoftheirteacherssupportandmonitoring,willbeinvestigated.Teachersupportand monitoringisrelated,butnotidenticalconceptstoauthoritativeteaching.

Table3 FitIndicesforModelsoftheWarmthandControl. Control Time1 Time2 Time3 Warmth Time1 Time2 Time3 Twofactormodel Time1 Time2 Time3 Combinedmodel(T1T3) Growthmodel 2 1.829 2.357 2.162 2.357 1.715 4.871 26.234 35.040 30.962 438.859 586.870 CFI 1.00 1.00 1.00 1.00 1.00 0.99 .99 .98 .98 .96 .96 TLI 1.00 1.00 .99 1.00 1.00 0.97 .99 .97 .98 .95 .95 SMRS .015 .024 .011 .024 .014 .018 .031 .033 .031 .061 .062 RMSA(90%CI) .000(.000.111) .012(.000.119) .015(.000.109) .012(.000.119) .000(.000.101) .070(.000.153) .035(.000.066) .049(.022.074) .047(.010.075) .030(.026.035) .031(.026.035)

Note.CFI=ComparativeFitIndex;TLI=TuckerLewisIndex;SRMR=StandardizedRootMeanSquared Residual;andRootMeanSquaredErrorofApproximationsupportedby90%

Teacherswerenestedwithinschoolswhichenabledthetestingofatwolevelmodelof factoranalysis.Designeffectslowerorcloseto2didnotsupportatwolevelmodelas.Also, thenumberofschoolinthesamplewasrelativelylow.Therefore,aonelevelmodelwas applied. Growthmodel.Intheanalysesinterceptisalatentvariableexpressingtheindividual baselinescore.Slopeisalatentvariablethatexpressestheindividualchangefrom baselinetoposttest.Theinterceptandslopehavemeanvalueswhichareexpressionsof averagebaselinescoresandaveragechangescores. ThegrowthmodelprovidedgoodfitaccordingtotherecommendationsofHuandBentler (1999)(2(244)=327.13,p=.000;CFI=.96;TLI.95;SRMR=.062;RMSEA=.03190%CI =.026.035).Ofconceptualreasons(Baumrind,1991;Hughes,2002)webelievethereisa relationshipbetweenteachersscoreonwarmthandcontrol.Itisassumedthatteachers scoringhighoncontrolmayalsoscorehighonwarmth,atthestartoftheintervention.Also, itwasassumedthattheacherschangeinwarmthisaffectedbytheirchangeincontrol

(Walker,2009).Generally,theresultsindicatedanincreaseinteacherscontrol(=.46,p= .004).However,contrarytowhatwasexpectedtheteachersreportedadecreaseinwarmth (=.38,p=.012).However,resultsindicatedthataccelerationinthelevelofcontrolover timewasparalleledbyanaccelerationinlevelofwarmth.Thisimpliesthatteachers perceptionsofimprovementincontrolskillstendedtoalsoreportincreaseinwarmth(= .68,p=.000).Also,theresultsindicatedthatthegreaterthelevelofcontrolatbaseline,the greatherthebaselinelevelofwarmth(=.85,p=.000) Pupilsreportsofteachersupportandmonitoring ResultsoftheANOVAfordifferencesbetweentimepointsforthedifferentgradelevels (table4)indicatedthattherewerenosignificantdifferencesbetweentimepointsfor secondaryschoolpupils(grade810)neitherforemotionalsupportnormonitoring.For primaryschoolpupilsasignificantdifferencebetweentimepointsforemotionalsupport (grade6and7)andmonitoring(grade5)werefound.Althoughtheresultsdidnotrevealed significantdifferencebetweentimepointsforallgradelevels,table5presentsresultsforall gradelevelsandtimepointtoprovidefullunderstanding. Table
Table4 ResultsofAnalysisofVariance(ANOVA)forGradeLevels510forTeachersEmotionalsupportand TeachersMonitoring. Teachers Teachers EmotionalSupport Monitoring df F p df F p Grade5 2 0.272 .762 2 3.030 .049 3.367 .035 2 1.359 .257 Grade6 2 8.188 .000 2 1.532 .217 Grade7 2 1.807 .164 2 1.040 .354 Grade8 2 0.062 .940 2 0.814 .443 Grade9 2 0.205 .815 2 0.654 .520 Grade10 2

Generally,theresultsindicatednochangeinpupilsreportsoftheirteachersemotional supportandmonitoringfromT1toT2andT3.Althoughsomeoftheresultsindicated significantchange,effectsizes,notsensitivetosamplesize,indicatednochangeformost gradelevelsforbothsupportandmonitoring.ExceptforGrade7pupilsreportingadecrease inemotionalsupportwithadvalueof.29,nonoftheresultsindicateddifferencesbetween timepointsaccordingtoCohens(1988)classification.However,itshouldbenotedthat someoftheresultse.g.Grade5pupilsreportsofmonitoring,werecloseto.20thatCohen considerasmalleffect.

Table5 NumberofSubjects(n),MeanScore(M),StandardDeviation(SD),andEffectSizes(Cohensd)forthe TwoAspectsofClassroomLeadershipatDifferentTimePoints(T)forGrades510. Teachers Teachers EmotionalSupport Monitoring M SD M SD Grade5 T1Cohort5(n=333) 2.34 0.57 2.49 0.54 T2Cohort4(n=374) 2.33 0.63 2.43 0.61 T3Cohort3(n=370) 2.31 0.67 2.38* 0.59 dT1T2 0.02 0.10 dT1T3 0.05 0.19 Grade6 T1Cohort6(n=399) T2Cohort5(n=340) T3Cohort4(n=379) dT1T2 dT1T3 Grade7 T1Cohort7(n=392) T2Cohort6(n=477) T3Cohort5(n=355) dT1T2 dT1T3 Grade8 T1Cohort8(n=741) T2Cohort7a(n=728) T3Cohort6a(n=647) dT1T2 dT1T3 Grade9 T1Cohort9(n=663) 1.69 T2Cohort8(n=712) 1.71 T3Cohort7a(n=702) 1.70 dT1T2 dT1T3 Grade10 T1Cohort10(n=643) 1.72 0.79 1.72 T2Cohort9(n=567) 1.75 0.81 1.68 T3Cohort8(n=711) 1.76 0.78 1.69 dT1T2 0.04 dT1T3 0.05 Onbothmeasuresanincreaseinscoreindicatesanimprovement. p<.05*p<.001significantlydifferentfromcorrespondinggradeatT1 0.71 0.74 0.70 0.06 0.04 0.81 0.79 0.72 0.03 0.01 1.78 1.79 1.75 0.68 0.70 0.65 0.01 0.04 1.89 1.94 1.87 0.73 0.73 0.75 0.07 0.03 1.96 2.01 1.97 0.65 0.63 0.65 0.09 0.02 2.05 1.93* 1.83* 0.72 0.75 0.83 0.16 0.29 2.11 2.03 2.07 0.64 0.72 0.75 0.13 0.06 2.23 2.12* 2.25 0.67 0.66 0.71 0.17 0.03 2.34 2.26 2.30 0.63 0.67 0.63 0.12 0.06

10

Aggressionwereinvestigatedinordertoexamineforpossibledifferencesbetweenthe interventiongroupsandthenoninterventiongroups,atthesamegradelevelwithregardto reactiveaggression,powerrelatedproactiveaggression,andaffiliationrelatedproactive aggression.Forexample,anincreaseinthereportsofteacherssupportandmonitoring wereduetoselectionbias,wewouldexpecttheinterventiongroupstoreportalowerlevel ofaggressionthannoninterventiongroups.ResultsofANOVApresentedintable6indicated significantdifferencesinreactiveaggressionbetweencohortsinGrades8,9and10.Also, significantdifferencesinpowerrelatedproactiveaggressionbetweencohortinGrade7,8,9 and10aswellasinaffiliationrelatedproactiveaggressioninGrade9cohorts. TheresultsofposthocanalysisofsignificantANOVAresultsarepresentedinTable7. Althoughsignificantdifferencesbetweensamegradecohortsatdifferenttimepointswere foundforallthreetypesofaggression,theeffectsizewerelowerthanthe.20Cohen(1988) consideredasmalleffect.Generally,thecontrolcohortatT1tendedtoreporthigherlevelof aggressionthanthesubsequentinterventioncohorts.However,duetotheloweffectssize thecontrolcohortsarenotconsideredtobedifferentthantheinterventioncohortregarding aggression.Giventhisthereisnoreasontobelievethatresultsforteacherssupportand monitoringisaffectedbydifferencesinlevelofaggressionbetweensamegradecohorts.

Generally,thepupilsreportsoftheirteachersemotionalsupportandmonitoringdidnot supporttheteachersperceptionofchangeincontrolandwarmth.
Table6 ResultsofAnalysisofVariance(ANOVA)forGradeLevels510forReactiveAggression,ProactivePowerand ProactiveAffiliationAggression. Grade5 Grade6 Grade7 Grade8 Grade9 Grade10 df 2 2 2 2 2 2 Reactive Aggression F .458 1.334 .109 5.100 4.927 4.440 ProactivePower Aggression df F p 2 2 2 2 2 2 .188 .954 3.049 6.931 3.670 5.532 .828 .386 .048 .001 .026 .004 ProactiveAffiliation Aggression df F p 2 1.476 .229 .414 .661 2 .113 .893 2 2.755 .064 2 3.367 .035 2 2.552 .078 2

p .633 .264 .897 .006 .007 .012

11


Table7 MeansandStandardDeviationsforReactiveAggressionandPowerrelatedProactiveAggressionforGrades thatReportedSignificantDifferencesbetweenTimePoints. ProactivePower ProactiveAffiliation ReactiveAggression Aggression Aggression M SD M SD M SD Grade7 T1Cohort7(n=391) 0.49 0.67 T2Cohort6(n=467) 0.38* 0.63 T3Cohort5(n=350) 0.41 0.72 dT1T2 0.17 dT1T3 0.12 Grade8 T1Cohort8(n=742) 1.23 T2Cohort7(n=725) 1.12** T3Cohort6a(n=651 1.16* dT1T2 dT1T3 Grade9 T1Cohort9(n=663) 1.34 0.73 0.70 0.82 0.43 T2Cohort8(n=706) 1.23* 0.78 0.63 0.82 0.36* 0.74 0.59* 0.74 0.34* T3Cohort7a(n=698) 1.23* dT1T2 0.15 0.09 dT1T3 0.15 0.14 Grade10 T1Cohort10(n=640) 1.32 0.76 0.72 0.87 T2Cohort9(n=567) 1.25 0.79 0.73 0.87 T3Cohort8(n=708) 1.20* 0.80 0.59* 0.80 0.01 dT1T2 0.09 dT1T3 0.15 0.16 Onallmeasuresandecreaseinscoreindicatesaloverlevelofaggression. p<.05*p<.001significantlydifferentfromcorrespondinggradeatT1 0.74 0.70 0.65 0.10 0.13 0.73 0.54 0.71 0.43* 0.68 0.41** 0.15 0.10 0.72 0.62 0.67 0.16 0.19

Discussion Themainpurposeofthestudywastoinvestigateteachersselfreportsofauthoritative teachingunderstoodasatwodimensionalconceptofwarmthandcontrolthroughouttwo interventionsaimingatimprovingtheirauthoritativeteaching.Theresultsindicatedanover allincreaseinteachersperformanceofcontrol.Thiswasaccordingtoapriori considerations.Incontrast,theteachersreportedadecreaseinperformanceofwarmth. Thiswasnotexpected.Furthermore,improvementincontrolwasassociatedwith improvementinwarmth.Pupilsreportsoftheirteachersemotionalsupportandmonitoring suggestednoimprovementformostsamegradecohortsexceptforGrade7reportinga decreaseinemotionalsupportthroughouttheprogramme. Theresultssuggestedthatteacherswhoreportedimprovementincontrol,alsotendedto reporthigherscoresonwarmth.Apossibleexplanationforthismaybethatwhenteachers haveestablishedroutinesforworkintheclassroom,theyfeelmorecertaintoengagewith buildingrelationshipswiththeirpupils.ThisisinlinewithRosenholtz(1989).Moreover,the baselinelevelofcontrolaffectedbaselinelevelofwarmth,indicatingthatteachershighon

12

controlalsoseemtobehighonwarmth.Itshouldbenotedthatdescriptivedataindicated thatteachersreportedhighonthetwodimensionsofauthoritativeteachingevenbefore theyattendedtheprogram.Whetherthesehighscoresarerealoriscausedbyanunrealistic viewononesowncapabilities,isimpossibletodetectfromthepresentstudy.Ertesvg (2011)foundthatteachersattendingthetwointerventionsathandscoredsignificantly higherthanaverageforteachersingeneralforbothcontrolandwarmthbeforetheprogram started.Also,apreviousstudybyMidthasselandBru(2001)providedsimilarresults. However,increasedreflectionthroughoutaninterventioncausedamorerealisticview,and thustheteachersthestudyofMidthasselandBrureportedlowerscoresontheposttest. Thisisapossibleexplanationfortheunexpecteddecreaseinmaineffectofwarmth.A qualitativestudyinprogress,includedinthelargerstudy,indicatethatcontrolmeasuresare moreeasilyundertakenthanwarmthmeasures.Apossibleexplanationisthatcontrol measuresaremoretangibleandwhereaswarmthmeasuresaremoreabstractanddemands insightsandconfidencebytheteacher(P.Roland,inprogress). Thelackofincreaseinteachersemotionalsupportandmonitoringreportedisnotinline withpreviousstudyoftheRespectprogram.Afiveyearfollowupstudyindicatedsome improvementintheirteacherssupportthroughouttheprogramformostgrades.However, theimprovementintheirteacherssupportandmonitoringcontinuedactiveprogramperiod wasover(Ertesvg,2009).TheRespectprogramisacomplex,manyfacetedprogram.Given this,itmaybetooearlytofullyevaluatetheeffectofsuchacomplexprogram.Ontheother hand,althoughnotaneffectsizeindicatinganeffect,pupilstendedtoreportlowerlevelof supportandmonitoringatT2andT3comparedtoT1.Thehandbookonclassroom managementprojectwasonlytargetingtheteachersroleintheclassroom. Thelackofimprovementinpupilsreportsofsupportandmonitoringpartlygivesupportto theteachersselfreportsinwarmthandcontrol.Thenegativetrendinpupilsreportsof emotionalsupportisinlinewithteachersreportofdecreaseinwarmth.However,the pupilsdonotreportimprovementinmonitoringasdoteachersincontrol.Oneshouldbear inmindthattheconceptsofemotionalsupportandmonitoringisrelated,notidenticalto theconceptsofwarmthandcontrol.Thisisapossibleexplanationforthedifferencesin results.Also,themethodologicalapproachesaredifferent.Furtheranalysis,e.g.growth curveanalysisofthepupilreports,mayshedlightonthedifferencesbetweenteacherand pupilsreports.However,itcannotberuledoutthatthelackofimprovementinteachers emotionalsupportandmonitoring,aswellastheteacherreportsofdecreaseinwarmthis dueto(low)qualityofimplementation(Domitrovich,etal.,2008). Limitationsandmethodologicalimplications Thereareseverallimitationstothestudy.Anobviousshortcomingisthelackofacontrol group.AsaresulthighfocusonauthoritativeteachingbothfromtheMinistryofEducation aswellasresearch(Kunnskapsdepartementet,2008,2009;Nordahl,Gravrok,Knutsmoen, Larsen,&Rrnes,2006),manyNorwegianschooltakeonsomesortofinterventionaimingat strengtheningthepupilteacherrelationshipand/orauthoritativeteaching.Giventhis,itcan bearguedthathardlyanyschoolwillconstituteatruecontrolcondition.Consequently,the presentstudycomparedtwodifferentinterventionsaimingatstrengtheningauthoritative teaching.

13

Thenestedstructureofdata,i.e.teachersandpupilswithinschools,enabledinvestigationof multilevelanalysisofdata.However,thedesigneffectsfortheteachersampledidnot indicatedifferencesamongschools.Giventhatapreviousstudy(Ertesvg,2011)indicated schooldifferencesinarepresentativecrosssectionalstudy,futurestudiesshouldelaborate onthenestedstructureofthedata.Aconceptualmodelofimplementation(Greenberg, Domitrovich,Graczyk,&Zins,2005)suggeststhatimplementationdifferbetweenschools dueto,amongotherthings,differencesinqualityofimplementation.Althoughresultsofthe studyathanddidnotindicateschooldifferences,thisshouldbefurtherinvestigatedin futureresearch. Itistooearlytofullydrawtheconclusionsaboutthesignificanceofthisstudy.Theanalysis andresultsisonestepalongapathtounderstandingthehowauthoritativeteachingmaybe improved,Thisstudyaddstotheknowledgeoftheimprovementoftheteacherpupil relationshipandmanagingbehaviourinclassingeneral,particularlytheknowledgeof authoritativeteaching.Thisappliesbothtopracticalissuesrelatedtothedevelopmentand improvementofinterventionsaimingatimprovingauthoritativeteaching,andtothe methodologicallyaspectsofinvestigatingauthoritativeteaching. Methodologicallyaswellaspractically,missingness/dropoutduetoteacherleavingschool duringandwholeschoolinterventionovermorethanoneyearisagreatchallengebothto researcherandinterventiondevelopers.Thisstudyillustratesanddiscussesthese challenges.However,theimplicationsofmissingnessinlongitudinalshouldbefurther investigated. Referenses Baker,J.A.,Clark,T.P.,Crowl,A.,&Carlson,J.S.(2009).Theinfluenceofauthoritative teachingonchildren'sschooladjustment:Arechildrenwithbehaviouralproblems differentiallyaffected?SchoolPsychologyInternational,30(4),374382. Baumrind,D.(1967).Childcarepracticesateceedingthreepatternsofpreschoolbehavior.. GeneticPsychologyMonographs,75,4388. Baumrind,D.(1978).Parentaldisciplinarypatternsandsocialcompetenceinchildren.Youth andSociety,9,239276. Baumrind,D.(1991).Parentingstylesandadolescentdevelopment.InJ.BrooksGunn,R. Lerner&A.C.Peterson(Eds.),TheEncyclopediaofAdolescence.NewYork:Garland. Birch,S.H.,&Ladd,G.W.(1997).Theteacherchildrelationshipandchildren'searlyschool adjustment.JournalofSchoolPsychology,35,6179. Brown,F.G.(1970).Principle'sofeducationalandpsychologicaltesting.Illinois:TheDryden PressInc. Bru,E.,Boyesen,M.,Munthe,E.,&Roland,E.(1998).PerceivedSocialSupportantSchool andEmotionalandMusculosketalComplaintsamongNorwegian8thGradeStudents. ScandinavianJournalofEducationalResearch,44,339356. Bru,E.,Stephens,P.,&Torsheim,T.(2002).Students'perceptionofclassmanagementand reportsoftheirownmisbehaviour.JournalofSchoolPsychology,40(4),287307. Buhi,E.R.,Goodson,P.,&Neilands,T.B.(2008).OutofSIght,NotOutofMind:Strategies forHandlingMissingData.AmericanJournalofHealthBehaviour,32(1),8392. Clausen,M.(2002).Unterrichtsqualitt:EinefragederPerspktive?[Instructionalquality:A questionofperspectives?Mnster:Waxmann.

14

Cook,T.D.,&Campbell,D.T.(1979).Quasiexperimentation.Chicago:RandMcNally. Cohen,J.(1988).StatiscitcalPowerAnalysisfortheBehaviouralciences(2nded.).Hillsdale: LaurenceErlbaum. Domitrovich,C.E.,Bradshaw,C.P.,Poduska,J.M.,Hoagwood,K.,Buckley,J.A.,Olin,S.,et al.(2008).MaximizingtheImplemenaionQualityofEvidenceBasedPreventive InterventionsinSchools:AConceptualFramework.AdvancesinSchoolMentalHealth Promotion,1(3),628. Ertesvg,S.K.(2009).ClassroomleadershipTheeffectofaschooldevelopment programme.EducationalPsychology,29(5),515539. Ertesvg,S.K.(2011).Measuringauthoritativeteaching.TeachingandTeachereducation, 27(1),5161. Evertson,C.M.,&Weinstein,C.S.(2006).ClassroomManagementasaFieldofInquiry Handbookofclassroommanagement:Research,practice,andcontemporaryissues. (pp.315).Mahwah,NJUS:LawrenceErlbaumAssociatesPublishers. Good,T.L.,&Brophy,J.(2007).Lookinginclassrooms(10thed.).Boston:AllenandBacon. Greenberg,M.T.,Domitrovich,C.E.,Graczyk,P.A.,&Zins,J.E.(2005).TheStudyof ImplemetnationinSchoolBasedPreventiveInterventions:Theory,Research,and Practice:U.S:DepartmentofHealthandHumanServices,SubstanceAbuseand MentalHealthServicesAdministration,CentreforMenatalHealthServices. Hamre,B.K.,&Pianta,R.C.(2005).Caninstructionalandemotionalsupportinthefirst gradeclassroommakeadifferenceforchildrenatriskofschoolfailure?Child development,76(5),949967. Henricsson,L.,&Rydell,A.M.(2006).ChildrenwithBehaviourProblems:TheInfluenceof SocialCompetenceandSocialRelationsonProblemStability,SchoolAchievement andPeerAcceptanceAcrosstheFirstSixYearsofSchool.InfantandChild Development,15,347366. Hu,L.T.,&Bentler,P.M.(1999).Cutoffcriteriaforfitindexesincovariancestructure analysis:Conventionalcriteriaversusnewalternatives.StructuralEquationModeling, 6(1),155. Hughes,J.N.(2002).AuthoritativeTeaching:Tippingthebalanceinfavorofschoolversus peereffects.JournalofSchoolPsychology,40(6),485492. Jelii,H.,Phelps,E.,&Lerner,R.M.(2009).Useofmissingdatamethodsinlongitudinal studies:Thepersistenceofbadpracticesindevelopmentalpsychology. DevelopmentalPsychology,45(4),11951199. Kleinfeld,J.(1975).EffectiveTeachersofEskimoandIndianStudents:SchoolReview. Kunnskapsdepartementet(2008).St.meld.31(20072008).Kvalitetiskolen. Kunnskapsdepartementet(2009).St.melding11(20082009)Lrerenrollenog utdanningen. Kunter,M.,&Baumert,J.(2006).Whoistheexpert?Constructandcriteriavalidityof studentsandteacherratingsofinstruction.LearningEnvironmentsResearch,9,231 251. Meehan,B.T.,Hughes,J.N.,&Cavell,T.A.(2003).Teacherchildrelationshipsas compensatoryresourcesforaggressivechildren.Childdevelopment,74,11451157. Midthassel,U.V.(2006).Creatingasharedunderstandingofclassroommanagement. EducationalManagementAdministration&Ledership,34(3),365383.

15

Midthassel,U.V.,&Bru,E.(2001).Predictorsandgainsofteacherinvolvementinan improvementprojectonclassroommanagement.ExperiencesfromaNorwegian projectintwocompulsoryschools..EducationalPsychology,21(3),229242. Midthassel,U.V.,Bru,E.&Idse,T.(2000).Theprincipalsroleinpromotingschool developmentactivityinNorwegiancompulsoryschool,SchoolLeadership& Management,20(2),247260. Morrison,T.L.(1974).Controlasanaspectofgroupleadershipinclassrooms:Areviewof research.JournalofEducation,Boston,156(4),3864. Munthe,E.(2001).ProfessionalUncertainty/Certainty:how(uncertain)areteachers,what arethey(un)certainaboutandhowis(un)certaintyrelatedtoage,experience gender,qualificationsandschooltype?EuropeanJournalofTeacherEducation,24(3), 355368. Muthn,L.K.,&Muthn,B.O.(19982007).MplusUser'sGuide(5thed.).LosAngeles,CA: Muthn&Muthn. Nordahl,T.,Gravrok,.,Knutsmoen,H.,Larsen,T.M.B.,&Rrnes,K.(Eds.).(2006). Forebyggendeinnsatseriskolen.Oslo:Sosialog helsedirektorartet/Utdanningsdirektoratet. Norusis,M.J.(2007).SPSS15.0guidetodataanalysis..Chicago,IL:PrenticeHall,Inc. Olweus,D.(2005).Ausefulevaluationdesign,andeffectsoftheOlweusBullyingPrevention Program.Psychology,Crime&Law,4(11),389402. Olweus,D.,&Alsaker,F.D.(1991).Assessingchangeincohortlongitudinalstudywith hierarchicaldata.InD.Magnusson,L.Bergman,G.Ruidinger&B.Torestadm(Eds.), Problemsandmethodsinlongitudinalresearch(pp.107132).NewYork:Cambridge UniversityPress. Pellerin,L.A.(2005).ApplyingBaumrind'sparentingtypologytohighschools:towardsa middlerangetheoryofauthoritativesocialization.SocialScienceResearch,34,283 303. Roland,E.,&Galloway,D.(2002).Classroominfluencesonbullying.EducationalResearch, 44(3),299312. Roland,E.,&Idse,T.(2001).AggressionandBullying.AggressiveBehavior,27,446462. Roland,P.(inprogress).Hvordanbeskriverlrereskolensimplementeringav Respektprogrammet?[Theteachers'perseptionoftheschoolsimplementationofthe Respectprogramme].UniversityofStavanger,Stavanger Rosenholtz,S.J.(1989).Teachers'Workplace.NewYork:ColumbiaUniversity,Teacher college. Shadish,W.R.,Cook,T.,&Campbell.(2002).Experimentalandquasiexperimentaldesigns forgeneralizedcausalinference.Boston:HoughtonMifflinCompany. Taylor,B.,Graham,J.W.,Cumsille,P.,&Hansen,W.B.(2000)ModelingPreventionProgram EffectsonGrowthinSubstanceUse:AnalysisofFiveYearsofDatafromthe AdolescentAlcholPreventionTrial.,PreventionScience,1,(4),183197. vanTartwijk,J.,denBrok,P.,Veldman,I.,&Wubbels,T.(2009).Teachers'practical knowledgeaboutclassroommanagementinmulticulturalclassrooms.Teachingand TeacherEducation,25(3),453460. Walker,J.M.(2008).Lookingatteacherpracticesthroughthelensofparentingstyle.Journal ofExperimentalEducation,76(2),218240. Walker,J.M.T.(2009).AuthoritativeClassroomManagement:HowControlandNurturance WorkTogether.TheoryIntoPractice,48(2),122129.

16

You might also like