Exposure to Effective Instruction and College Student Persistence暴露于有效的教學和學生的持久性

ExposuretoEffectiveInstructionandCollegeStudentPersistence:

AMulti-InstitutionalReplicationandExtension*

ErnestT.Pascarella

MarkH.Salisbury

TheUniversityofIowa

CharlesBlaich

WabashCollege

*TheresearchonwhichthisstudywasbasedwassupportedbyagenerousgrantfromtheCenterofInquiryintheLiberalArtsatWabashCollegetotheCenterforResearchonUndergraduateEducationatTheUniversityofIowa.

Correspondingauthor:ErnestPascarella,ernest-pascarella@

ABSTRACT

Thisstudyanalyzedamulti-institutionalandlongitudinaldatasettodeterminetheimpactofexposuretoeffectiveinstructiononfirstyearpersistence–definedasre-enrollingforthesecondyearofcollegeatthesameinstitution.Netofimportantconfoundinginfluences,exposuretoeffectiveinstructionsignificantlyincreasedthelikelihoodthatthestudentwouldre-enrollforthesecondyearofcollege.Theeffectwasmediatedprimarilythroughstudentsatisfactionwiththequalityoftheoveralleducationalexperienceattheinstitution.Thesefindingshaveimplicationsfortheroleoftheclassroomexperienceinstudentpersistenceinhighereducation.

ExposuretoEffectiveInstructionandCollegeStudentPersistence:

AMulti-InstitutionalReplicationandExtension

Thebodyofcorrelationalandexperimentalevidencedemonstratingthepositive,andperhapscausal,linkbetweenvariousdimensionsofeffectivecollegeclassroominstructionandbothcourse-specificlearningandmoregeneralmeasuresofcognitivegrowthisextensive(seePascarella&Terenzini,1991,2005,forasummaryofthisevidence).Inthispaperweanalyzealongitudinalandmulti-institutionaldatabasetotesttherobustnessofapreviousfindingfromasingleinstitutionsamplewhichsuggestedthatexposuretoorganizedandclearclassroominstructionmayhaveapositivenetimpactonfirst-yearstudentpersistence—operationallydefinedastheprobabilityofreturningtoaninstitutionforthesecondyearofpostsecondaryeducation(Pascarella,Seifert,&Whitt,2008).Thismaybeofsomeconsiderableimportanceforinstitutionalpolicysinceitwouldmeanthatclassroominstructionalpracticesstemmingfromlearnablefacultyskillsmaycontributetoaninstitution’sabilitytoretainstudents.Ourresultsgenerallyreplicatethispreviousfindingandsuggestthattheneteffectonpersistenceofexposuretoorganizedandclearinstructionisthesameacrossinstitutionaltype(researchuniversities,regionalinstitutions,communitycolleges,andliberalartscolleges)andthesameforstudentswithdifferentlevelsoftestedprecollegeacademicpreparation.

EffectiveClassroomInstruction

Literallyhundredsofcorrelationalstudieshavelinkedstudentperceptionsofteacherbehaviorssuchascourseorganizationandpreparation,instructionalclarity,teacherexpressiveness,andfeedbacktostudentstovariousmeasuresofcourse-levelknowledgeacquisitionandcontentmastery.Anumberofcomprehensivenarrativeormeta-analyticreviewsofthisextensivebodyofresearchhavebeenconducted(e.g.,Braskamp&Ory1994;Cashin,1999;d’Apollonia&Abrami,1997;Feldman,1997;GreenwaldandGillmore,1997;Marsh&Dunkin,1997;McKeachie,1997).AdistillationofthesesynthesesbyPascarellaandTerenzini(2005)suggeststhreegeneralconclusionsaboutstudentperceptionsofteacherbehaviorsandinstructionalpractices:(1)theseperceptionsaremultidimensional,(2)theyarereasonablyreliableandstable,and(3)theyhavemoderatepositivecorrelations(e.g.,.30to.50)withvariousmeasuresofcourse-levellearningsuchascoursegradeandcoursefinalexamination.

Thepredictivevalidityofstudentperceptionsofteachingisnotlimitedtocorrelationalevidence.Threeofthedimensionsofstudentperceptionsofteachingwiththestrongestlinkstocourseachievementincorrelationalresearch—organization/preparation(useofcourseobjectives,effectiveuseofclasstime),instructionalclarity(clearexplanations,effectiveuseofsamples),andteacherexpressiveness(eyecontact,speakingemphatically)—havebeendemonstratedwithrandomizedexperiments(Hines,Cruickshank,&Kennedy,1985;Schonwetter,Menec,&Perry,1995;Schonwetter,Perry,&Struthers,1994;Wood&Murray,1999).

Notalltheresearchonstudentperceptionsofteachingfocusesonspecificcourse-leveloutcomes.Althoughtheyconstituteamuchsmallerbodyofevidence,afewstudieshaveindicatedthatinstructionalorganization,oracombinationofinstructionalorganizationandinstructionalclarity,mayhavepositivenetimpactsonmoregeneralacademiccompetenciesandskillsnotdirectlytiedtoaspecificcourse.Researchersaffiliatedwiththe1992-95NationalStudyofStudentLearning(Pascarella,Edison,Nora,Hagedorn,&Braxton,1996)developedtwo5-itemscalestermedinstructionalorganization/preparationandinstructionalskill/clarity,thatappropriatedspecificitemsappearinginpreviousresearch(Cohen,1981;Feldman,1989,1994).Constituentitemsfortheorganization/preparationscaleincludedsuchthingsas“presentationofmaterialiswellorganized”and“classtimeisusedeffectively.”Thescalehadanalpha,internalconsistency,reliabilityof.87.Theskill/clarityscalehadconstituentitemssuchas“instructorsgiveclearpresentations”and“instructorsmakegooduseofexamplesandillustrationstoexplaindifficultpoints,”withanalphareliabilityof.86.Inaseriesofmulti-institutionalstudiesthatcontrolledforanextensivearrayofconfoundinginfluences,includingapretest,itwasfoundthatthemorestudentsreportedthattheoverallinstructiontheyreceivedincollegewashighontheorganization/preparationscale,thelargertheirgainswereonstandardizedmeasuresofcriticalthinking,readingcomprehension,andmathematics(Pascarellaetal.,1996;Edison,Doyle,&Pascarella,1998;Whitt,Pascarella,ElkinsNesheim,Marth,&Pierson,2003).Mostrecently,Bray,Pascarella,andPierson(2004)combinedthetwo5-itemscalesintoacomposite,10-itemmeasureoforganizationandclarity(alphareliability=.89).Theyfoundthat,netofextensiveconfoundinginfluences,theresultantcompositescalehadapositiveinfluenceongainsinreadingcomprehensionoverthreeyearsofcollege.

ExposuretoEffectiveInstructionandCollegePersistence

Severalscholarshavehypothesizedthatthenatureandqualityofclassroominstructionmaynotonlyinfluencestudentlearning,butmightalsoplayasignificantroleinstudentpersistenceordeparturefromaparticularpostsecondaryinstitution(Braxton,Hirschy,&McClendon,2004;Braxton&McClendon,2001-2002;Braxton&Mundy,2001-2002;Tinto,2006-2007).Asmallbodyofevidencesupportsthishypothesis(Braxton,Bray,&Berger,2000;Braxton,Milem,&Sullivan,2000;Braxton,Jones,Hirschy,&Hartley,2008;NelsonLaird,Chen,&Kuh,2008;Nora,Cabrera,Hagedorn,&Pascarella,1996;Tinto,1997).Althoughmostofthisresearchhasestimatedtheeffectsofdifferentclassroompedagogicalapproachessuchasactiveorcooperativelearning,thestudybyBraxton,Bray,andBerger(2000)mostdirectlyconsideredtheimpactsofspecificteacherbehaviors.UsingTinto’s(1975,1993)conceptualmodelasaframeworkfortheirinvestigation,Braxtonetal.hypothesizedthatstudentsexposedtofacultywhofrequentlyexhibitorganizationandclarityintheirclassroominstructionmightbemoreconfidentandrelaxedabouttheiracademicachievement.Consequently,theymightperceivethattheyhavemoretime“toinvestthepsychologicalenergynecessarytoestablishmembershipinthesocialcommunitiesoftheircollegeoruniversity”(Braxtonetal.,2000,p.216).Increasedsocialintegration,inturn,wouldenhanceinstitutionalcommitmentandintenttopersistattheinstitution.EmployingmeasuresofoverallinstructionalorganizationandclarityessentiallyidenticalwiththoseusedbyPascarellaetal.(1996),theirfindingswerequiteconsistentwiththeirhypotheses.Withimportantconfoundinginfluencescontrolledstatistically,overallexposuretoorganizedandclearinstructionenhancedbothameasureofstudentsocialintegrationandintenttoreenrollatasingleinstitutionforthesecondyearofcollege.

Itcouldbeargued,ofcourse,thatintenttore-enrollasacriterionmeasuredoesnothavethesamepredictivevalidityastheactualdecisiontore-enroll.Accordingly,arecentstudybyPascarella,Seifert,andWhitt(2008)tooktheBraxtonetal.(2000)findingstothenextlogicalstep.Analyzinglongitudinaldatafromasinglelargeresearchuniversity,andcontrollingforanextensivebatteryofconfoundinginfluences,theyfoundthatthesamemeasureofoverallexposuretoorganizedandclearinstructionemployedbyBraxtonandhiscolleagueshadasignificantpositivetotaleffectonactualreenrollmentattheinstitutionforthesecondyearofcollege.Theyfurtherfoundthatthepositiveimpactofexposuretoorganizedandclearinstructiononpersistenceintothesecondyearofcollegewaslargelymediatedthroughincreasedlevelsofstudentsatisfactionwiththeeducationtheywerereceiving.

ThefindingsofPascarellaetal.(2008)areintriguingandpotentiallyimportanttotheextentthattheysuggestthesignificantroleoflearnablefacultyinstructionalbehaviorsinstudentpersistence.However,theirstudyislimitedtoasingleinstitutionsample,andthegeneralizabilityoftheirfindingsisyettobeestablished.ThepurposeofthepresentstudywastotesttherobustnessofthePascarellaetal.findingsonamulti-institutionsampleoffirst-yearstudentsattendingresearchuniversities,regionalinstitutions,liberalartscolleges,andcommunitycolleges.Italsosoughttoextendtheirworkbydeterminingiftheeffectsonpersistenceofoverallexposuretoorganizedandclearinstructionarethesameforstudentsattendingdifferenttypesofinstitutionsorforstudentswhoenterpostsecondaryeducationwithdifferentlevelsoftestedacademicpreparation.

RESEARCHMETHODS

ConceptualModel

TheconceptualmodelguidingtheinvestigationwasbasedonanextensivebodyofresearchevidenceandisillustratedinFigures1and2.(Forasynthesisofthisbodyofevidence,seesourcessuchasBraxton,Hirschy,&McClendon,2004;Pascarella&Terenzini,1983,1991,2005;andTinto,1993.)TheseareessentiallythesameconceptualmodelsguidingthePascarellaetal.(2008)study,anddrawlargelyonresearchguidedbyTinto’s(1975,1993)theoreticalmodelofthestudentpersistence/withdrawalprocess.Tinto’smajortheoreticalcontributionwasthatheshiftedtheexplanatoryfocusofpersistence/withdrawalresearchandscholarshipfromarelianceonstudentpre-collegecharacteristics(e.g.,academicability,degreeaspirations,familybackground)toaconcernwithmeasuringastudent’slevelofintegrationintheacademicandsocialsystemsofacollegeoruniversity(Pascarella&Terenzini,2005).Whileastudententerspostsecondaryeducationwithcertainpre-collegecharacteristicsthatmayinfluenceretention(e.g.,academicability,educationalaspirations,familybackground),itislevelsofsocialandacademicintegration(e.g.,academicperformance,extracurricularinvolvement)thatarethemajordeterminantsofwhetherornotonepersistsattheinstitution.WeconceptualizedtheinfluenceofexposuretoeffectiveinstructiononpersistenceasfunctioningwithinatheoreticalmodelthatincludedmostofTinto’smajorconstructs.Wetookintoaccountnotonlyastudent’sprecollegecharacteristics,butalso,becauseofthemulti-institutionalnatureofthesample,thetypeofinstitutionattended.Inaddition,weconsiderednotonlymeasuresofacademicandsocialintegration(e.g.,collegegradesandextracurricularinvolvement),butalsofactorssuchasworkresponsibilitiesandplaceofresidenceduringcollegewhichshapesocialandacademicintegration.

Figure1modelsthehypothesizedtotaleffectofexposuretoeffectiveclassroominstruction(definedasinstructionalorganizationandclarity)onpersistenceintothesecondyearofpostsecondaryeducation(i.e.,thestudentre-enrolledforthesecondyearofpostsecondaryeducationatthesameparticipatinginstitution).Themodelassumesthatpersistenceisafunctionnotonlyofexposuretoeffectiveclassroominstruction,butalsoofstudentbackgroundcharacteristics(sex,race,testedpre-collegeacademicpreparation,pre-collegeeducationaldegreeplans,andparentaleducation),thetypeofinstitutionattended(researchuniversity,regionalinstitution,communitycollege,orliberalartscollege),andothercollegeexperiences(workresponsibilitiesduringcollege,placeofresidenceduringcollege,andinvolvementinco-curricularactivities).AccordingtotheconceptualmodelshowninFigure1,weanticipatedthatinthepresenceofstatisticalcontrolsforstudentbackgroundcharacteristics,precollegetestscores,thetypeofinstitutionattended,andothercollegeexperiences,overallexposuretoorganizedandclearinstructionduringthefirstyearofpostsecondaryeducationwouldhaveasignificantpositivetotaleffectontheprobabilityofenrollingforthesecondyearofcollegeatthatinstitution(Alwin&Hauser,1975).

Figure1abouthere

ThehypothesizeddirectandindirecteffectsofexposuretoorganizedandclearclassroominstructiononpersistencearemodeledinFigure2.Accordingtothisconceptualmodel,weanticipatedthatwhenmeasuresofcollegegradesandeducationalsatisfactionwereaddedtothetotaleffectsmodel(Figure1),twothingswouldhappen.First,netofallotherinfluences,gradesandsatisfactionwithcollegewouldhaveapositivedirectinfluenceonpersistence;andsecond,thepositiveinfluenceofexposuretoorganizedandclearinstructioninthetotaleffectsmodelwouldbecomesmallandstatisticallynonsignificant.Thiswouldindicateapositiveindirecteffectofoverallexposuretoorganizedandclearinstructiononpersistence,mediatedthroughthepositiveeffectsoforganizedandclearinstructionongradesandsatisfactionwiththeeducationbeingreceived(Alwin&Hauser,1975;Pascarella,2006).Wereasoned(asdidPascarellaetal.,2008)thatiforganizedandclearinstructionatthecourselevelimprovedcourse-levellearning,thenoverallexposuretoclearandorganizedinstructionduringthefirstyearofcollegewouldenhancecollegiateacademicachievement.AlsoconsistentwithPascarellaetal.(2008),wehypothesizedthatoverallexposuretoorganizedandclearinstructionwouldhaveanaffectivedimensionmanifestinhigherlevelsofstudentsatisfactionwiththeiroveralleducationalexperience.

Figure2abouthere

Samples

InstitutionalSample.Thesampleinthestudyconsistedofincomingfirst-yearstudentsat19four-yearandtwo-yearcollegesanduniversitieslocatedin11differentstatesfrom4generalregionsoftheUnitedStates:Northeast,Southeast,Midwest,andPacificCoast.The19institutionsdidnotincludetheresearchuniversityatwhichthePascarellaetal.(2008)studywasconducted.Institutionswereselectedfrommorethan60collegesanduniversitiesrespondingtoanationalinvitationtoparticipateintheWabashNationalStudyofLiberalArtsEducation(WNSLAE).FundedbytheCenterofInquiryintheLiberalArtsatWabashCollege,theWNSLAEisalarge,longitudinalinvestigationoftheeffectsofliberalartscollegesandliberalartsexperiencesonthecognitiveandpersonaloutcomestheoreticallyassociatedwithaliberalartseducation.Theinstitutionswereselectedtorepresentdifferencesincollegeanduniversitiesnationwideonavarietyofcharacteristicsincludinginstitutionaltypeandcontrol,size,location,andpatternsofstudentresidence.However,becausethestudywasprimarilyconcernedwiththeimpactsofliberalartscollegesandliberalartsexperiences,liberalartscollegeswerepurposefullyover-represented.

Ourselectiontechniqueproducedasamplewithawiderangeofacademicselectivity,fromsomeofthemostselectiveinstitutionsinthecountrytosomethatwereessentiallyopenadmissions.Therewasalsosubstantialvariabilityinundergraduateenrollment,frominstitutionswithenteringclassesbetween3,000and6,000,toinstitutionswithenteringclassesbetween250and500.Accordingtothe2007CarnegieClassificationofInstitutions,3oftheparticipatinginstitutionswereconsideredresearchuniversities,3wereregionaluniversitiesthatdidnotgrantthedoctorate,2weretwo-yearcommunitycolleges,and11wereliberalartscolleges.

StudentSample.Theindividualsinthesamplewerefirst-year,full-timeundergraduatestudentsparticipatingintheWNSLAEateachofthe19institutionsinthestudy.Theinitialsamplewasselectedineitheroftwoways.First,forlargerinstitutions,itwasselectedrandomlyfromtheincomingfirst-yearclassateachinstitution.Theonlyexceptiontothiswasatthelargestparticipatinginstitutioninthestudy,wherethesamplewasselectedrandomlyfromtheincomingclassintheCollegeofArtsandSciences.Second,foranumberofthesmallestinstitutionsinthestudy—allliberalartscolleges—thesamplewastheentireincomingfirst-yearclass.Thestudentsinthesamplewereinvitedtoparticipateinanationallongitudinalstudyexamininghowacollegeeducationaffectsstudents,withthegoalofimprovingtheundergraduateexperience.Theywereinformedthattheywouldreceiveamonetarystipendfortheirparticipationineachdatacollection,andwerealsoassuredinwritingthatanyinformationtheyprovidedwouldbekeptinthestrictestconfidenceandneverbecomepartoftheirinstitutionalrecords.

DataCollection

InitialDataCollection.Theinitialdatacollectionwasconductedintheearlyfallof2006with4,501studentsfromthe19institutions.Thisfirstdatacollectionlastedbetween90-100minutesandstudentswerepaidastipendof$50eachfortheirparticipation.ThedatacollectedincludedaWNSLAEprecollegesurveythatsoughtinformationonstudentdemographiccharacteristics,familybackground,highschoolexperiences,politicalorientation,educationaldegreeplans,andthelike.Studentsalsocompletedaseriesofinstrumentsthatmeasureddimensionsofintellectualandpersonaldevelopmenttheoreticallyassociatedwithaliberalartseducation.

Follow-upDataCollection.Thefollow-updatacollectionwasconductedinspring2007.Thisdatacollectiontookabouttwohoursandparticipatingstudentswerepaidanadditionalstipendof$50each.Twotypesofdatawerecollected.Thefirstwasbasedonquestionnaireinstrumentsthatcollectedextensiveinformationonstudents’experienceofcollege.Twocomplementaryinstrumentswereused:theNationalSurveyofStudentEngagement(NSSE)(Kuh,2001)andtheWNSLAEStudentExperiencesSurvey(WSES).Theseinstrumentsweredesignedtocapturestudentinvolvementinabroadvarietyofdifferentactivitiesduringcollege(e.g.,coursework,clubs,study,interactionswithotherstudents,involvementincultural/socialactivities,andthelike).Thesecondtypeofdatacollectedconsistedoffollow-up(orposttest)measuresofinstrumentsmeasuringdimensionsofintellectualandpersonaldevelopmentthatwerefirstcompletedintheinitialdatacollection.Boththeinitialandfollow-updatacollectionswereadministeredandconductedbyACT(formerlytheAmericanCollegeTestingProgram).

Oftheoriginalsampleof4,501studentswhoparticipatedinthefall2006testing,3,081participatedinthespring2007follow-updatacollection,foraresponserateof68.5%.These3,081studentsrepresented16.2%ofthetotalpopulationofincomingfirst-yearstudentsatthe19participatinginstitutions.Toprovideatleastsomeadjustmentforpotentialresponsebiasbysex,race,academicability,andinstitutioninthesampleofstudents,aweightingalgorithmwasdeveloped.Usinginformationprovidedbyeachinstitutiononsex,race,andACTscore(orappropriateSATequivalentorCOMPASSscoreequivalentforcommunitycollegestudents),follow-upparticipantswereweighteduptoeachinstitution’sfirst-yearundergraduatepopulationbysex(maleorfemale),race(Caucasian,AfricanAmerican/Black,Hispanic/Latino,Asian/PacificIslander,orother),andACT(orequivalentscore)quartile.Whileapplyingweightsinthismannerhastheeffectofmakingtheoverallsampleateachinstitutionmoresimilartothepopulationfromwhichitwasdrawn,itdoesnottotallyadjustforthepotentialresponsesofthosewhodroppedoutofthestudy(Sudman,1976).

Asecondfollowupwasconductedinthefallof2007.Atthattime,eachparticipatinginstitutionindicatedwhetherornoteachstudentwhocompletedthefirstfollowup(Spring2007)re-enrolledforthesecondyearofcollegeatthatinstitution.Thisinstitutionaldatabecamethebasisforthedependentvariableinthestudy.

Variables

DependentVariable.Thedependentvariablewaswhetherornotthestudentreenrolledataparticipatinginstitutionforthesecondyearofpostsecondaryeducation.Thevariablewascoded1=reenrolled,0=didnotreenroll.Approximately90%ofthesamplereenrolledforthesecondyearofcollegewhile10%didnot.Thedataforthisvariablecamefromtheofficialrecordsateachofthe19participatinginstitutions.

IndependentVariable:ExposuretoEffectiveClassroomInstruction.Overallexposuretoeffectiveclassroominstructionwasdefinedoperationallyasexposuretoorganizedandclearinstruction.Informationonstudentperceptionsofoverallexposuretoorganizedandclearinstructionwasgatheredbymeansofa10-itemscaleinthefirstfollow-updatacollection(Spring2007).Thequestionnairepresentedstudentswiththefollowingstem:“Belowarestatementsaboutteacherskill/clarityaswellaspreparationandorganizationinteaching.Forthemostpart,takingintoconsiderationalloftheteacherswithwhomyou’veinteractedwithat[institutionname],howoftenhaveyouexperiencedeach?”Weemployedthesame10-itemscaleofvettedreliabilityandvalidityusedbyPascarellaetal.(2008),aswellasbymanyofthestudiesdescribedpreviously(Braxton,Bray,&Berger,2000;Bray,Pascarella,&Pierson,2004;Edison,Doyle,&Pascarella,1998;Pascarellaetal.,1996;Whittetal.,2003).The10-iteminstructionalorganizationandclarityscalehasanalphareliabilityof.89.ConstituentitemsandresponseoptionsareshowninTable1.

Table1abouthere

BackgroundCharacteristicsandTestedAcademicPreparation.Studentbackgroundcharacteristicsconsistedofsex,race/ethnicity,precollegeeducationalplans,andparentaleducationallevel.Sexwascoded1=male,0=female,whileracewascoded1=white,0=personofcolor.Precollegeeducationalplanswerecoded1=graduatedegree,0=lessthanagraduatedegree.Parentaleducationwasthesumofmother’sandfather’sformaleducation.Therewereeightresponseoptionsforeachrangingfrom“l(fā)essthanhighschooldiploma”to“Doctoraldegree.”TestedprecollegeacademicpreparationwasanACTcompositescore,SATequivalent,orCOMPASSequivalentscoreforcommunitycollegestudents.Informationonsex,race,educationalplans,andparentaleducationwasgatheredontheWNSLAEprecollegequestionnaire.Testedprecollegeacademicpreparationscoreswereprovidedbyeachparticipatinginstitution.

InstitutionalType.Institutionaltypeconsistedofthreedummy(1,0)variables.Theywereresearchuniversity(vs.liberalartscollege),regionalinstitution(vs.liberalartscollege),andcommunitycollege(vs.liberalartscollege).

OtherCollegeExperiences.Othercollegeexperiencesconsistedofthreevariables:hoursofon-andoff-campusworkperweek,whetherornotonelivedoncampus,andhoursofco-curricularinvolvementperweek.Hoursofon-andoff-campusworkconsistedofthetotalnumberofhoursofremuneratedon-andoff-campusworktypicalperweek.Therewereeightresponseoptionsfrom“0hours”to“morethan30hours.”Livingoncampuswascoded:1=livedoncampus,0=didnotliveoncampus.Hoursofco-curricularinvolvementwasastudent’sreportednumberofhoursinatypicalweekinvolvedinco-curricularactivities(campusorganizations,campuspublications,studentgovernment,fraternityorsorority,intercollegiateorintramuralsports,etc.).Therewereeightresponseoptionsrangingfrom“0hours”to“morethan30hours.”Informationonworkon-oroff-campusresidence,andco-curricularinvolvementwascollectedonthefirst(Spring2007)followup.

CollegeGradesandEducationalSatisfaction.Collegegradeswerebasedonstudentself-reportstothequestion“Whathavemostofyourgradesbeenuptonowatthisinstitution?”Therewereeightresponseoptions,rangingfrom“C-orlower”to“A.”Whileitwouldhavebeenpreferabletohaveactualfirst-yeargrades,thereisevidenceindicatingsubstantialproximity(correlationsfrom.74to.96)betweenactualandreportedgrades(Baird,1976;Flowers,Osterlind,Pascarella,&Pierson,2001).Moreover,aswereportbelowreportedgradeshadarelativelystrongnetimpactonpersistence.Satisfactionwiththeoverallexperienceofcollegewasbasedonstudentresponsestothequestion“Howwouldyouevaluateyourentireeducationalexperienceatthisinstitution?”Therewerefourresponseoptions:1=“poor,”2=“fair,”3=“good,”and4=“excellent.”InformationongradesandeducationalsatisfactionwascollectedontheWNSLAEfirstfollowupinSpring2007.

DataAnalyses

Becausethisstudyfocusedonlyonthetotal,direct,andindirecteffectsofexposuretoeffectiveinstructiononpersistenceintothesecondyearofcollege,wedidnotusestructuralequationmodelingtoestimatethevalidityoftheoverallmodelshowninFigure2.SuchomnibustestsofmodelsbasedonTinto’sconstructshavealreadybeenconductedwithconsiderablefrequency(Braxton,Hirchy,&McClendon,2004;Pascarella&Terenzini,2005).Rather,welimitedouranalysestoestimatingthevariousneteffectsofexposuretoeffectiveinstructiononpersistence.

Thefirststepinthedataanalyseswastoestimatethetotaleffectofoverallexposuretoorganizedandclearinstructionduringthefirstyearofpostsecondaryeducationonpersistenceintothesecondyearofcollege.Toaccomplishthis,weusedreducedformregr